Microarray Analysis
Gene Representation in Amplified vs. Unamplified RNA

Microarray experiments typically require microgram amounts of total RNA (5-20 µg) or poly(A) mRNA (0.2-5 µg) per slide. Obtaining these amounts of RNA can be problematic, especially when working with small biopsy or Laser Capture Microdissection (LCM) samples. To circumvent this problem, RNA from limited samples can be amplified using a multistep procedure developed by Van Gelder and Eberwine (Van Gelder et al., 1990). Ambion has incorporated this technology into the MessageAmp™ aRNA Kit, which includes all the reagents necessary to carry out the complete Van Gelder and Eberwine procedure (Figure 1). In brief, RNA samples are reverse transcribed to synthesize first strand cDNA. This is followed by synthesis of second strand cDNA using a DNA polymerase mix. In vitro transcription of the resulting double stranded DNA template using T7 RNA polymerase results in linear amplification of the original RNA sample.

Figure 1. A schematic of the MessageAmp™ Procedure.

Variation in RNA mass and quality, and the probe labeling systems used, can affect amplification of the RNA and thus the outcome of gene expression analysis. How can Ambion's MessageAmp Kit be used with different microarray platforms, with different sample types, and under different probe labeling conditions? This is the first in a series of articles in which we present data obtained in collaboration with various laboratories using our MessageAmp Kit under varied conditions.

Does Amplification Affect Gene Representation?
Nalini Raghavachari and coworkers in the Biological Technologies Division of Corning, Inc. compared the expression profile of RNA amplified from HepG2 cells using MessageAmp to that of un-amplified total RNA, using Corning Cancer Glass Microarrays (Raghavachari et al., 2002). If significant bias is introduced during the amplification procedure, a change in the expression profile will be observed when the data from the two arrays is compared. To determine the extent of variation caused by the amplification process, total RNA was isolated from 1000, 3000, 10,000 and 50,000 HepG2 cells. All RNA samples had an A₂₆₀/A₂₈₀ ratio of 1.8 or higher. The yields of total RNA and aRNA obtained from these cells are tabulated in Figure 2. Greater than 1000-fold amplification was observed for each RNA sample.

Labeled cDNA probes generated from one half of the RNA amplified from the 3000 cells and total RNA isolated from 50,000 cells were prepared by reverse transcription with Cy3-dCTP, according to the protocol described in Bao et al. (2002). A Universal Human reference RNA (5 µg; Stratagene) was labeled with Cy5-dCTP and used to normalize the signal intensity across the various arrays that were tested. Efficient incorporation of dye molecules with 35-38 Cy3-dCTP per 1000 nucleotides and 48-51 Cy5-dCTP per 1000 nucleotides was obtained. Corning cancer arrays are printed on Corning® CMT-GAPS coated slides and contain 1607 cancer related genes (PCR products). Hybridization of the CyDye labeled probes to the array at 42°C, and high stringency posthybridization washes were carried out according to the procedure developed at Corning (Raghavachari et al., 2002). Microarrays were scanned on a GenePix® 4000A (Axon Instruments) fluorescence scanner at a PMT setting of 750-950 volts. Images were analyzed using GenePix® Pro 3.0 analysis software (Axon Instruments).

Figure 3. Gene Expression Profiling with Amplified vs. Unamplified RNA. A. Images of microarrays hybridized with DNA made from amplified RNA (1500 cells) or unamplified total RNA (50,000 cells). B. A scatter plot analysis for the array data compares expression profiles obtained from total RNA vs. amplified RNA. Cy3 labeled samples (unamplified total RNA) and Cy5 labeled reference RNA were hybridized to Corning Cancer arrays. Fluorescence intensities (signal intensity minus background) were normalized to median array densities to obtain normalized measures for each gene across all of the samples. The log Cy3/Cy5 ratio for each gene was calculated relative to human reference RNA.

The gene expression profiles obtained with probes made from amplified and unamplified total RNA from HepG2 cells on the Corning arrays are depicted in Figure 3. Scatter plot analysis shows that the differential expression ratios derived from amplified and unamplified RNA were highly comparable (R² = 0.867). This suggests that the amplification process did not introduce significant bias in the vast majority of mRNA levels. In addition, there was no reduction in signal intensity or dynamic range with aRNA obtained from as few as 1000 cells when compared to total RNA from 100,000 cells (data not shown). Based on this simple but informative example, it is clear that the MessageAmp Kit is very useful for amplifying RNA from as few as 1000 cells and the resulting data is representative to the original (unamplified) mRNA profile.

We would like to thank Nalini Raghavachari, Paul Bao, Yulong Hong, and Uwe Muller from Biochemical Technologies, Corning Inc. for sharing their data with us.

Cy is a tradmark of Amersham Biosciences.

REFERENCES

Related Articles

Array Analysis Step by Step: Ambion's Array Analysis Products

Array Analysis: The Basics

Selected References on Microarray Analysis