Gene Expression Analysis Using Microdissected Mouse Brain Regions: Technology Evaluation

Ivonne Moon, Marianna Goldrick, Bob Setterquist, Sharmili Moturi, Charmaine San Jose

Abstract

Laser Capture Microdissection (LCM) is a technique that enables the selection of specific cell types from a tissue section. LCM combined with microarray technology shows promise for gene expression profiling of individual cell populations within heterogeneous tissues. The specific visualization of desired cell types prior to LCM is essential for achieving selective capture. A mouse brain model of neurons micro dissected from the granular cell layer of the hippocampus and from the granular cell layer of the cerebellum was studied using LCM. We have developed staining reagents, RNA isolation, and amplification protocols for reproducible gene expression analysis. The staining reagents and protocols for the two dyes allow a clear visualization of target cells for LCM, while preserving RNA integrity. The isolation of total RNA from these limited samples, and its subsequent T7 RNA amplification can be challenging and inconsistent. We report the integration of several technologies (a new Staining Kit, LCM, RNAqueous®-Micro RNA Isolation Kit, MessageAmp™ II aRNA Kit) with Affymetrix® Mouse 430A GeneChip® and real-time RT-PCR analysis. Ten genes were selected for validation of the original total RNA using real-time RT PCR. The results demonstrate that reproducible mRNA expression profiles can be obtained from mouse brain LCM selected cells.

Introduction

To test the feasibility of Ambion's new LCM Staining Kit reagents, we selected two distinct mouse brain structures: hippocampus and cerebellum. Using the Arcturus Pixel® IIe LCM system we micro dissected a dense layer of tiny neurons denoted as the granular cell layer, leaving behind the molecular layer, which is nearly cell-free and is occupied mostly by axons and dendrites (Figure 1).





Figure1. Successful Staining and Microdissection Using Ambion's LCM Staining Kit. (A) Hippocampus stained with Acridine Orange (top) and Cresyl Violet (bottom); (B) Cerebellum stained with Acridine Orange (top) and Cresyl Violet (bottom). In each panel, the image on the left side shows the stained section prior to LCM, the middle image shows the remaining section after LCM, and the image on the right shows the micro dissected tissue adhered to the LCM cap.

RNA from both mouse brain structures was isolated using Ambion's RNAqueous®-Micro Kit. RNA quality was assessed by the Agilent® 2100 bioanalyzer (Agilent Technologies) after separation by an RNA LabChip® Kit. RNA yields and intactness are compared between RNA isolated from an unstained section and RNA extracted from the remaining tissue section on the slide before and after several hours of LCM (Figure 2). Tissues prepared for LCM using the LCM Staining Kit yield high quality RNA suitable for qRT-PCR and microarray analysis. Note that the starting material (unstained tissue) was compromised in quality even before processing (staining and LCM).







Figure 2. Total RNA: Quality. Electropherograms of RNA from serial sections of unstained tissue or tissue stained (LCM Staining Kit, Ambion) with Cresyl Violet or Acridine Orange before and after several hours of LCM.

Approximately 20 ng of total RNA from each microdissected region was linearly amplified (two rounds) using Ambion's MessageAmp™ II Kit. The quality of the aRNA was examined on an RNA LabChip (Agilent Technologies).








Figure 3. Successful aRNA Two Round Amplification Using MessageAmp™ II Kit. Agilent bioanalyzer results showing great yields (mean ± SD, 142 ± 12 µg) and size distribution (median, 1100 nt) of the aRNA after two rounds of amplification using either Acridine Orange (AO) or Cresyl Violet (CV) stain.

Figure 4. Gene Expression Profiles Comparing Both Mouse Brain Regions Were Examined on an Affymetrix® GeneChip® Array. (A) Graph summarizing Percent Present Call results for each sample. (B) Scatter plot summarizing the overall results shows correlation between results from the two brain samples using the GCOS software (Affymetrix).

We chose 10 genes that showed either similar or differential expression between Hippocampus and Cerebellum based on the GeneChip results. Three genes were overexpressed in Hippocampus, three were overexpressed in Cerebellum, and four showed no change between the two tissues. qRT-PCR was used to validate the GeneChip results (see Figure 5). When the Log2 expression ratios were compared between the two assay methods an R2 value of 0.918 was observed (see Figure 5, inset). This correlation demonstrates excellent concordance between the two methods. These results demonstrate the utility of using our LCM methods for gene expression analysis with microarrays. Even starting with as little as 20 ng LCM RNA, the MessageAmp II aRNA kit (with 2 rounds of amplification) was able to accurately detect expression differences.







Figure 5. Excellent Correlation Between the Two Detection Methods. Log2 ratios (Hippocampus/Cerebellum) of gene expression were compared for GeneChip or qRT-PCR experiments. GeneChip ratios were determined by averaging signals from triplicate biological replicates. qRT-PCR ratios were performed in duplicate. Notice that the Ptpn8 RNA was called absent in hippocampus on GeneChips and was not detected in that brain structure by qRT-PCR.

• The reproducible Percent Present Call rates (Figure 4A) indicated excellent, reproducible sensitivity when tissue sections were stained with either Acridine Orange or Cresyl Violet (LCM Staining Kit, Ambion). Both stains, Acridine Orange and Cresyl Violet, proved to be equally effective for cell selection.

• RNA quality and quantity obtained from the LCM samples using Ambion's RNAqueous-Micro Kit demonstrated that this kit is suitable for processing of LCM samples

• The high percentage of genes detected on the Affymetrix® Mouse 430A GeneChip® showed the ability of Ambion's MessageAmp™ II aRNA Kit to successfully amplify minute amounts of RNA from microdissected samples (approximately 40,000-fold from the starting mRNA amount using two rounds of linear amplification).

• The excellent correlation between the two detection methods, microarray and qRT-PCR, in the expression profiling of cerebellar and hippocampal LCM samples demonstrated the ability of MessageAmp™ II aRNA Kit to linearly amplify (with two rounds) minute samples with high reproducibility.

Samples

For each brain structure, we used two different stains for cell identification. Samples stained with Cresyl Violet were microdissected in triplicate. RNA extraction and amplification was done independently for each sample. Whole sections from stained and unstained tissue (from the same set of slides used for LCM) were processed as control RNA. This RNA was used to quantify RNA from LCM samples via qRT-PCR [used for Standard Curve with mouse fatty acid synthase (Fasn) as target], to assess RNA quality after LCM, and as control for aRNA amplification efficiency.

Approximately 20 ng from each LCM replicate and from each whole section sample were used in the linear amplification reactions.

Tissue Preparation and Laser Capture Microdissection

The brain from an 8-10 week old female Swiss Webster mouse was treated in 30% sucrose for 24 hours, embedded in OCT, and slowly frozen on dry ice. Transversal sections (10 um) were cut in a –24ºC cryostat and mounted onto pre-cleaned glass slides (VWR Scientific Products, West Chester, PA). Slides were stored on dry ice until sectioning was finished and immediately transferred to a –80ºC freezer for long-term storage. The slides were stored for 2 months before the LCM.

Slides were taken out of storage and stained with Cresyl Violet and Acridine Orange following Ambion's LCM Staining Kit protocols. After staining we immediately proceed to LCM. Stained slides were stored in a closed box with desiccant until microdissection. The Pixel® IIe LCM System from Arcturus Engineering, Inc. (Mountain View, CA) was used to lift the desired tissue.

Three to four hippocampal structures from 3 different slides stained with Cresyl Violet were microdissected using three independent caps. Four to five hippocampal structures were microdissected from one slide stained with Acridine Orange for stain comparison. One complete structure from cerebellum was microdissected from each of the slides stained with either Cresyl Violet or Acridine Orange (a total of three caps for Cresyl Violet and one cap for Acridine Orange).

RNA Extraction

After the LCM was completed, RNA from both mouse brain structures was independently isolated from all three caps using RNAqueous Micro Kit (Ambion, Inc., Austin, TX). The thermoplastic films were removed from the caps and added to 0.5 ml tubes containing 100 ul of lysis solution from the kit. The tissue was incubated for 30 minutes at 42° C before RNA extraction. The RNAqueous Micro Kit is a phenol-free silica filter based isolation method. A novel reagent is included in the kit to facilitate recovery of the minute amounts of RNA recovered from microdissected samples. The small size of the silica filter allows the RNA to be eluted in less than 20 ul of Elution Solution. A DNase treatment was performed on the eluted RNA using 2 units of DNAse I for 20 minutes. The DNase I was subsequently removed using the novel DNase Inactivation Reagent from Ambion's DNA-free system. All reagents for DNase treatment and inactivation are included in the RNAqueous Micro Kit.

RNA yields and intactness were compared between RNA isolated from an unstained section, from a stained section before LCM, and the RNA extracted from the remaining tissue section on the slide after several hours of LCM. To assess RNA quality, 1 ul of the RNA obtained from the whole-section samples was directly analyzed on an RNA LabChip (Agilent Technologies, Palo Alto, CA) following the manufacture's instructions.

RNA yields from LCM samples were assessed by qRT-PCR using a standard curve derived from total RNA extracted from the whole-section samples. RNA yields from the whole-section samples were determined by Abs 260 nm [triplicate readings on Nanodrop® ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE)]. Primers and TaqMan® probe targeting mouse Fasn mRNA were used. The average total RNA yields from the Hippocampal and Cerebellar LCM was 34 ng and 53 ng, respectively.

aRNA Amplification and Labeling

Approximately 20 ng of total RNA from each sample was used to generate biotin-labeled aRNA for hybridization to GeneChips Mouse Genome 430A 2.0 microarrays (Affymetrix, Santa Clara, CA). The LCM total RNA and the whole section samples were processed in parallel for two rounds of T7-based RNA amplification using MessageAmp™ II aRNA Kit. (Ambion, Inc.) The total RNA was converted to first strand cDNA during a reverse transcription step, then the cDNA was made double stranded during a second strand synthesis step using a DNA polymerase enzyme mix. The filter-purified double stranded cDNA with a T7 promoter now makes an efficient template for T7 polymerase-based in vitro transcription. During the in vitro transcription, Biotin 16-UTP (PerkinElmer, Boston, MA) was incorporated so as to yield biotinylated aRNA.

The quality of the aRNA after two rounds of amplification was examined on an RNA LabChip (Agilent Technologies). The average size of the amplified RNA was grater than 1000 nt for all samples. The two rounds of amplifications yielded on average ~400,000 times the starting mRNA material (~150 ug of aRNA.)

GeneChip® Hybridization and Data Analysis

The biotinylated aRNA was fragmented using the Affymetrix fragmentation buffer and hybridized to the Affymetrix® GeneChip array consisting of 22690 probe sets. Following a 16-hour hybridization, the arrays were stained and washed according to the Affymetrix recommended protocol. The arrays were scanned using the Affymetrix GeneChip Scanner 3000. All arrays passed the Affymetrix QC specifications and were further analyzed for correlations within replicate arrays and between different conditions. Over 55% of the genes were called present in all samples and scaling factors were similar in all arrays.

qRT-PCR Validation

The following 10 genes were selected from the array results and validated by qRT-PCR: Fasn GB:NM_007988 was used to normalize the data and primers/probe were designed in-house. Primers were ordered from Integrated DNA Technologies (Coralville, IA): 5'-AGA GAT CCC GAG ACG CTT CTG –3' and 5'-AGCCGG TTG GCC ATC ATT –3', and the modified internal probe from Synthegen (Houston, TX): 5'- CGC TGG CAG CCC ACC ATG CT –3'. Gabra6 GB:NM_008068, Mmp24 GB:NM_010808, Ptpn8 GB:NM_008979, Bcat1 GB:NM_007532, Pde1a GB:NM_016744, Igfbp4 GB:NM_010517, G3bp GB:NM_013716, Brca2 GB:NM_009765, Accn1 GB:NM_007384. Primers/probe sets were obtained through TaqMan® Gene Expression assays (Applied Biosystems, Foster City, CA.)

The targets included genes that were unchanged, up-regulated, and down-regulated between the two brain regions. A two-step qRT-PCR was performed on the total RNA used for aRNA amplification. A standard curve made from total mouse brain using an equally expressed gene (Fasn) was used to ensure equal amounts of RNA were used from both cerebellar and hippocampal samples. The RT reaction was setup using SuperScript™ II Reverse Transcriptase (Invitrogen, Carlsbad CA) with a reaction volume of 30 ul using random decamers and following the manufacturer's instructions. The temperature profile was 10 minutes at 25ºC, 50 minutes at 42ºC and an RT inactivation step of 15 minutes at 70ºC. The qPCR step was setup using Ambion's SuperTaq™ reagents. The qPCR reaction volume was 25 ul including 5 ul of cDNA input from the RT reaction. The final RNA input amount was 1 ng for each sample. The temperature profile for the qPCR was 2 minutes at 50ºC, 10 minutes at 95ºC followed by 40 cycles of 15 seconds at 95ºC and 1 minute at 60ºC on the MX3000P Real-Time instrument (Stratagene, La Jolla, CA.)

1. Bonaventure P, Guo H, Tian B, Liu X, Bittner A, Roland B, Salunga R, Ma XJ, Kamme F, Meurers B, Bakker M, Jurzak M, Leysen JE, Erlander MG, Nuclei and subnuclei gene expression profiling in mammalian brain. Brain Res. 2002 Jul 5;943(1):38–47.


2. Torres-Munoz JE, Van Waveren C, Keegan MG, Bookman RJ, Petito CK. Gene expression profiles in microdissected neurons from human hippocampal subregions. Brain Res Mol Brain Res. 2004 Aug 23;127(1-2):105–14.


3. Luzzi V, Holtschlag V, Watson MA. Expression profiling of ductal carcinoma in situ by laser capture microdissection and high-density oligonucleotide arrays. Am J Pathol. 2001 Jun;158(6):2005–10.


4. Shimamura M, Garcia JM, Prough DS, Hellmich HL. Laser capture microdissection and analysis of amplified antisense RNA from distinct cell populations of the young and aged rat brain: effect of traumatic brain injury on hippocampal gene expression. Brain Res Mol Brain Res. 2004 Mar 17;122(1):47–61


5. Ginsberg SD, Che S. Combined histochemical staining, RNA amplification, regional, and single cell cDNA analysis within the hippocampus. Lab Invest. 2004 Aug;84(8):952–62.