Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray gene expression analysis

An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray... © 2002 Oxford University Press Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray gene expression analysis 1 2 1 Stanislav L. Karsten, Vivianna M. D. Van Deerlin , Chiara Sabatti , Lisa H. Gill and Daniel H. Geschwind* Department of Neurology, Program in Neurogenetics, UCLA School of Medicine, 710 Westwood Plaza, Los Angeles, CA 90095-1769, USA, Department of Pathology and Laboratory Medicine, University of Pennsylvania Health System, 7.098 Founders Pavilion, 3400 Spruce Street, Philadelphia, PA 19104, USA and Department of Human Genetics and Statistics, UCLA School of Medicine, 695 Charles Young Drive South, Los Angeles, CA 90095-7088, USA Received September 5, 2001; Revised October 20, 2001; Accepted November 2, 2001 ABSTRACT specimens, such as primary cell cultures or microdissected tissue samples, one of the frequently encountered limitations of Archival formalin-fixed, paraffin-embedded and the technique can be the amount of RNA required for hybrid- ethanol-fixed tissues represent a potentially invalu- ization. In addition, while fresh frozen tissues from human able resource for gene expression analysis, as they pathological specimens of interest may not be available for are the most widely available material for studies of gene expression studies, archived tissues preserved in various human disease. Little data are available evaluating fixatives often are obtainable. whether RNA obtained from fixed (archival) tissues Various signal amplification techniques have been developed could produce reliable and reproducible microarray within the last several years (5). The MICROMAX cDNA microarray system utilizes tyramide signal amplification expression data. Here we compare the use of RNA (TSA), which requires 20–100 times less RNA than direct isolated from human archival tissues fixed in ethanol cDNA labeling. The TSA method was originally introduced to and formalin to frozen tissue in cDNA microarray improve the sensitivity of immunohistochemistry and, when experiments. Since an additional factor that can limit properly optimized to individual tissues and primary anti- the utility of archival tissue is the often small quantities bodies, has become an important tool for immunofluorescence available, we also evaluate the use of the tyramide microscopy (6,7). Although a cDNA microarray system using signal amplification method (TSA), which allows the TSA is commercially available, the accuracy and reproduci- use of small amounts of RNA. Detailed analysis indi- bility of the technique has not been extensively studied. cates that TSA provides a consistent and reproducible Since the availability of fresh disease specimens is often a signal amplification method for cDNA microarray limiting step in gene expression experiments, other RNA analysis, across both arrays and the genes tested. sources, such as formalin- and ethanol-fixed archival tissues, Analysis of this method also highlights the importance need to be considered. Fixed tissues are generally considered sub-optimal for RNA analytical techniques such as northern of performing non-linear channel normalization and blotting, but in some cases efforts using RNA isolated from dye switching. Furthermore, archived, fixed specimens formalin-fixed tissues in northern hybridization (8) and real can perform well, but not surprisingly, produce more time RT–PCR (9) have been successful. This, together with an variable results than frozen tissues. Consistent increasing need for large-scale gene expression profiling, results are more easily obtainable using ethanol- raises the issue of whether RNA from different pathological fixed tissues, whereas formalin-fixed tissue does specimens might be used in microarray-based experiments. not typically provide a useful substrate for cDNA The application of microarray-based hybridization techniques synthesis and labeling. to human pathological specimens might open new avenues in the study of diseases and subsequently facilitate the identifica- tion of new therapeutic drug targets. This is especially the case INTRODUCTION where large numbers of replicates may need to be studied to The determination of relative expression levels for a large control for human genetic diversity, compared with experi- number of genes using cDNA or oligonucleotide microarray ments using only cell lines or inbred mouse strains (3). Addi- technology is becoming a standard approach for the comparative tionally, while laser capture microdissection coupled to cDNA analysis of gene expression between different cell populations microarray analysis uses fresh frozen or ethanol-fixed specimens or tissues (1–4). Since RNA is very often extracted from rare (10), no comprehensive evaluation of its reproducibility *To whom correspondence should be addressed. Tel: +1 310 794 6570; Fax: +1 310 267 2401; Email: dhg@ucla.edu e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 2 OF 9 Table 1. Summary of 36 microarray experiments performed using the TSA method a b c d e Experiment Hybridizations Correlation Slope Hits NS (T) versus NS (T) 3 0.96 1.08 79% Frozen (T) versus frozen (T) 2 0.96 1.09 85% Frozen (P) versus frozen (P) 1 0.88 0.99 91% Frozen (P) versus frozen (T) 2 0.87 0.91 95% Ethanol (T) versus ethanol (T) 2 (4) 0.87 0.65 99% Ethanol (T) versus frozen (T) 3 (4) 0.61 0.19 99% Formalin (T) versus formalin (T) 3 (6) 0.78 1.85 47% Formalin (T) versus formalin (T) (1 h at 70°C) 2 0.66 0.86 58% Formalin (T) versus formalin (T) 1 0.48 1.05 30% Formalin (T) versus frozen (T) 8 (9) 0.55 0.3 47% Formalin (P) versus frozen (P) 2 0.72 0.16 78% All NS experiments used the mouse 9k array; the others were performed on the human array. The tissues or cell lines compared were two neurosphere cultures, NS7 and NS8, of neuronal progenitor cell lines (NS) grown in different flasks. T, total RNA; P, poly(A) RNA; 1 h at 70°C, tyramide signal amplification of RNA preheated at 70°C for 1 h prior to cDNA synthesis. Number of successful hybridizations in the experiment. The numbers in parentheses indicate the total number of hybridizations (successful and unsuccessful) that were performed. Mean correlation of signals from the two channels of successful hybridizations. Mean slope of the best fit correlation line between successful hybridizations. Number of signals above twice local background. The direct labeling method was used in the experiment. using ethanol-fixed tissues has been published. In the present was confirmed by gel electrophoresis and RT–PCR of house- study we have conducted experiments to evaluate the use of keeping genes. The extraction methods used provided optimal TSA microarray systems and made an attempt to use ethanol- RNA for RT–PCR (13). and formalin-fixed tissues as a source of RNA for gene expres- Microarrays sion experiments using custom-made cDNA microarrays. We performed 36 microarray experiments matching frozen and ethanol- and formalin-fixed samples in various combinations MATERIALS AND METHODS (Table 1). Two types of arrays were used. For experiments Tissue samples and cell cultures comparing human tissues, an array containing 95 relatively abundant human genes gridded in duplicate was used. This size To test the specificity of the TSA microarray system and variables was chosen because it was relatively inexpensive to produce of the ethanol and formalin fixation procedures, post-mortem and had the advantage of containing replicate spots on the brain tissues cut from frontal cortex were obtained from five unrelated human subjects. Post-mortem intervals ranged from same array, facilitating statistical analysis of various human 8 to 11 h. Adjacent tissue sections from each individual were tissues. A larger array containing 8700 mouse genes (http:// immediately frozen in liquid nitrogen and stored at –80°C; www.medsch.ucla.edu/som/humgen/nelsonlab) was used only others were immersed in formalin (10%, pH 7.0) or RNase-free to elucidate reproducibility of the TSA method. The mouse 70% ethanol in 10 mM Tris–HCl pH 7.6, 1 mM EDTA and array was based on Unigene clusters and contained 2700 fixed for 12–18 h at 4°C. Neural progenitor colonies (neuro- known genes for which the full-length cDNA sequences are spheres) were cultured from neonatal mouse neocortex as known and an additional 6000 ESTs. This array has the advantage described previously (11,12). that it contains genes of varying abundances and many of these genes are low abundance genes based on either previous Extraction of RNA from frozen and fixed tissues expression studies or EST sequencing data. This allowed evalua- The RNA from cultured cells and frozen tissue sections was tion of the TSA method over a large number of genes of extracted using acid phenol extraction (Trizol LS; Gibco BRL) different abundance levels. These microarrays were as recommended by the manufacturer. One milliliter of Trizol constructed in the UCLA Genetics Microarray Core using a per 50 mg tissue was used. RNA from paraffin-embedded custom arrayer and previously described methods (12). tissue was extracted using a combination of proteinase K and Briefly, clone inserts were PCR amplified using vector Trizol reagent as described previously (13), followed by primers, purified by isopropanol precipitation, resuspended in DNase treatment (37°C, 30 min). RNA was dissolved in TE high pH buffer at ∼200–800 ng/µ l concentration and gridded at buffer and stored at –80°C. The purity of RNA was checked by measuring the optical density at 260 and 280 nm. The quality high density onto poly-L-lysine-coated slides. PAGE 3 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Probe preparation and hybridization Direct RNA labeling was performed using Cy3-dCTP and Cy5- dCTP (catalog nos PA53021 and PA55021; Amersham), oligo(dT) (12–18mer) and Superscript II reverse transcriptase (catalog no. 18064-014; Life Technologies). (See http://www.tigr.org/tdb/ microarray/conciseguide.html or geschwindlab.medsch.ucla.edu for complete protocol.) The labeling reaction contained 50 µ g total RNA, 500 µ M dUTP, 500 µ M dATP, 500 µ M dTTP, 100 µ M dCTP, 1 mM Cy3–dCTP/Cy5–dCTP, 400 U Super- script II reverse transcriptase (Gibco BRL), 1 mM DTT and 1× reverse transcriptase buffer supplied by the manufacturer (NEN). The reaction mix was incubated for 3 h at 42°C and stopped with 20 mM EDTA. After degradation of the RNA by NaOH for 30 min (final concentration 25 mM) and neutralization with HCl, unincorporated fluorescent nucleotides were removed by isopropanol precipitation. The TSA probe labeling and array hybridization were performed as described in the instruction manual (MICROMAX Human cDNA Microarray System, NEN Life Science Products, Boston, MA) with minor modifications (see http://Geschwindlab.medsch.ucla.edu for detailed protocol). Biotin- and fluorescein-labeled cDNAs were generated from 0.5 and 1.5 µ g total RNA for the 92 and 8700 gene cDNA microarrays, respectively. We found that increasing the cDNA synthesis time from 1 h, as recommended by the manufacturer, to 3 h resulted in much more complete and reproducible labeling. Post-hybridization washes were performed according to the manufacturer’s recommendations (MICROMAX; NEN). The signals from specifically hybridized biotin- and fluorescein-labeled cDNAs were amplified either with streptavidin–horseradish peroxidase (HRP) and Cy5–tyramide or antifluorescein–HRP and Cy3–tyramide, respectively. After signal amplification, the cDNA microarrays were air dried and scanned in a Genetic Microsystems 418 microarray scanner. The images were analyzed using ImaGene 4.1 (Biodiscovery, Santa Monica, CA). Prior to quantitative analysis, normaliz- ation was performed using a non-linear function to control for Figure 1. Representative images of a typical TSA hybridization on the mouse 9k dye and signal intensity effects (14). Background correction microarray (A) and the 95 gene, 190 element human microarray (B). In (B), dupli- was first conducted using a smoothed background correction, cate spots are adjacent to each other and the consistency in hybridizations which resulted in a lower variance than subtracting local back- between duplicates can be observed. The image in (A) is a non-homotypic ground. All hybridizations were done at least in duplicate and hybridization, whereas (B) is homotypic hybridization of RNA extracted from human frontal cortex. repeated with the fluorophores reversed. Quantified intensity data was downloaded into GeneSpring 3.1 (Silicon Genetics) or Microsoft Excel and analyzed as described in the Results. shows representative hybridizations using TSA onto the 190 For computation and graphical display we used R, a statistical (95 genes) and 8700 element arrays. The overall images from language freely available at http://CRAN.R-Project.org. direct labeling (data not shown) and TSA appear similar and the background signals appear low. Figure 2A depicts a plot of RESULTS signal versus background using TSA amplification of RNA derived from frozen neurospheres (NS8). Eighty-three percent Sensitivity and reproducibility of the TSA microarray of the total points are 2-fold above background and are considered system ‘hits’. Figure 2B shows a plot of the log ratio of expression versus the average log intensity per spot. Two very important To assess the reproducibility and quality of the data generated technical observations follow from a careful analysis of the in frozen and ethanol- and formalin-fixed tissue samples, multiple hybridizations using the same sample, labeled with observed signals. From Figure 2A and B emerges the importance different fluorophores and co-hybridized on the same array, of using an appropriate threshold for considering what constitutes were performed. Homotypic hybridizations (frozen versus a real signal and the consideration of signal strength measures. This is most pronounced at signals <2-fold above background. frozen experiments) were used to estimate the reproducibility of the TSA method and were further used as a benchmark for A second technical issue relates to the method of normaliz- the evaluation of fixed versus fixed data quality. Figure 1 ation. Non-linear normalization has been championed in the e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 4 OF 9 Figure 2. Scatter plots of experiments with total RNA isolated from mouse neural stem cell cultures (neurospheres, NS). The same RNA sample was used for cDNA synthesis and labeled with either fluorescein-12-dCTP or biotin-11-dCTP. Both cDNAs were co-hybridized on the mouse 9k microarray (homotypic hybrid- ization) and developed according to the TSA protocol provided by the manufacturer (MICROMAX; NEN). (A) The plot (on a log scale) of Cy5 signal versus background from a representative homotypic hybridization of NS RNA on the mouse 9k array. The plots for Cy3 were similar. The points above the solid line have a signal higher than 2-fold background. The letter E indicates the signal produced by negative control spots. Notable is the clear separation between signal and background apparent in the two clusters of spots that are observed when the data are plotted in this manner. (B) A representative plot of the log ratio (Cy3/Cy5) versus the average log signal from both Cy3 and Cy5 channels for each spot in homotypic hybridization. It can be seen that the likelihood of false positive ratios decreases with signal strength. (C) Scatter plots of the signal generated from NS versus NS hybridizations from two experiments where the dyes were reversed. In this case, rather than labeling the same RNA, each NS culture, NS7 and NS8, was grown in separate flasks and processed separately. The clear non-linearity (curvature) produced by differences in dye incorporation and signal at different cDNA abundances can be observed. literature, but has not been widely adopted by most experi- spot deposition and attachment to the slide surface, slide mentalists, who typically use global normalization methods washing or local hybridization conditions, between-slide (15). Non-linearity is reflected in the curvature present in comparisons are not made with spotted cDNA arrays. In this scatter plots from globally normalized signals. This may be case, using TSA, we were curious to determine whether due to dye incorporation effects that may differ at different between-slide variability would be prohibitive to performing cDNA abundances (see for example Fig. 2C). Globally between-slide comparisons. As expected, the observed correl- normalized signals will contain errors that can be removed by ations between the same spot on different slides were less than use of a non-linear normalization to remove the curvature same spot comparisons (Fig. 3B). However, these correlations induced by dye effects. Furthermore, the effect of the dyes can were relatively high: when signals obtained on different slides be minimized by performing duplicate hybridizations with from the same spot were compared the correlation averaged reversed labels, as was done here. This is clearly an important 0.84 (range 0.77–0.87), suggesting that with a few more replicate experiments, such an approach of comparison across slides consideration when using TSA. would be technically feasible. Thus, overall the signals A high correlation between duplicate or like samples co- produced using TSA to amplify the signal from small quantities of hybridized on the 8700 element mouse array was observed. Figure 3 depicts scatter plots derived from the co-hybridization RNA were very consistent across a large number of genes assayed. of a duplicate sample labeled with different dyes. Correlations were around 0.96, even when the independent replicate Comparison of archived tissues samples from different culture dishes were co-hybridized (Fig. 3B). This demonstrates high data quality and reproduci- Quantitative gene expression analysis in matched frozen, bility of the hybridizations using TSA. Typically, because of formalin-fixed, paraffin-embedded and ethanol-fixed tissue slide-to-slide variability due to factors such as inconsistent samples were performed using the 190 human element array, PAGE 5 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Figure 3. Hybridization consistency on the 9K microarray using TSA amplification. In each graph, Cy5 signals are shown on the ordinate and Cy3 signals on the abscissa on a log scale after non-linear normalization. Each dot represents the hybridization intensity of each gene. While there was more variability in hybridizations obtained from the same sample hybridized to different arrays (B) relative to the same array (A), both showed high reproducibility. Figure 4. Comparison of RNA from frozen frontal cortex versus fixed frontal cortex. Fixed and frozen sample pairs are from an adjacent brain section from the same subject hybridized on the human array. Each dot represents the hybridization intensity of each gene. The signals for RNA from frozen tissue are shown on the ordinate in each graph on a log scale. (A) Total RNA from frozen frontal cortex versus total RNA from formalin-fixed frontal cortex. (B) Another independent sample hybridization showing the signal obtained using total RNA from frozen frontal cortex versus total RNA from formalin-fixed frontal cortex. (C) Total RNA from frozen versus ethanol fixed frontal cortex. which contains mostly high abundance genes expressed in fixed, respectively. Because of this effect, it is virtually impos- many different tissue types. To test the specificity of the sible to compare samples preserved in different fashions, as is formalin and ethanol fixation procedures, post-mortem frozen, shown by the low correlations observed when plotting signals ethanol-fixed and formalin-fixed brain specimens obtained obtained for frozen versus tissues fixed in different manners from the frontal cortex of five human subjects were used. As from the same individual brain (Fig. 4). The different behav- previously, only the spots generating signals above 2.0 times iors of frozen and fixed samples also makes it impossible to local background were considered true signals. This corre- use global normalization. For this reason, we only normalized sponded to 94% of spots for frozen tissues, 99% for ethanol- slides with homotypic hybridizations and used these compari- fixed and only 56% for formalin-fixed tissues in homotypic sons as the basis for the rest of the analysis. hybridizations (Table 1 and Fig. 4). Again, as observed on the larger array, signals obtained from Not surprisingly, the range of signals for fixed tissues was the same frozen tissue labeled with different fluorophores and significantly lower than for frozen tissues, with only 12% and co-hybridized on the array were highly reproducible (Fig. 5A). 3% of the signals above 4 times background for ethanol- and Correlations of normalized signals ranged from 0.95 to 0.99, formalin-fixed tissues, respectively (Table 1). This demon- similar to what was observed using fresh cultured neuronal strates that formalin-fixed tissues resulted in quite weak precursor cells on the 8700 array (Fig. 3A). Correlations for signals overall, which is borne out by the observed back- ethanol versus ethanol hybridizations within the same array ground-corrected, average log signal intensities, which were were lower (0.85 and 0.89), but within an acceptable range 3.34, 3.16 and 3.03 for frozen, ethanol-fixed and formalin- (Fig. 5B). Sample hybridizations were also conducted with e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 6 OF 9 Figure 5. Scatter plots of homotypic hybridization experiments using RNA isolated from the same tissue with different fixation methods. (A) Total RNA versus + + total RNA from frozen frontal cortex of a single subject. (B) Poly(A) RNA versus poly(A) RNA from frozen frontal cortex of a single subject. (C)Total RNA versus total RNA isolated from ethanol-fixed frontal cortex of the same subject shown in (A). (D) Total RNA versus total RNA from formalin-fixed frontal cortex of a different subject. poly(A) RNA as template and the correlations were no better TSA, direct labeling of the same RNA was performed. Direct than those observed using total RNA (Fig. 5C). labeling of total RNA isolated from formalin-fixed tissues also did not result in any significant improvement; only 30% of Formalin-fixed tissue posed many problems. In about half of genes produced signals above twice the local background, less the cases the experiments simply did not work; one or both dyes did not incorporate properly. Poor cDNA synthesis or than observed with TSA, showing a correlation of signals lack of dye incorporation was much less frequent and did not between duplicate samples of 0.62 and a slope of 1.3 (Table 1). happen even once in the current frozen versus frozen experi- Across-array comparisons were also performed, although ments. Thus, the failure rate using formalin-fixed tissues was the use of different human tissues would necessarily reduce the high, an important consideration given the cost of microarray correlations across arrays due to many factors, such as differ- experiments. Using formalin, an acceptable level of dye incor- ences in post-mortem interval and human genetic diversity. poration was attained in three experiments out of nine Surprisingly high correlations, similar to those observed in the attempted, yielding correlations of 0.87, 0.32 and 0.53, less tissue culture control experiments (see for example Fig. 3A), consistent than either frozen or ethanol-fixed tissues (Fig. 5D were observed between the signal intensities of the same spot and Table 1). This, coupled with the far lower dynamic range on different arrays. These inter-array correlations averaged of signal obtained using formalin, unfortunately suggests that 0.92, with a range of 0.86–0.95. Less correlation was observed its utility is very limited for current microarray analysis. between ethanol versus ethanol experiments. One of the experi- Previously, it was reported that better cDNA yield could be ments was a clear outlier, with a correlation of ∼0.2 with the obtained by simple preheating of RNA prior to reverse tran- other experiments. The other experiments showed a tighter scription (16). Preheating formalin-fixed RNA for 1 h at 70°C range of correlations, ranging from 0.62 to 0.75. This is similar in our experiments did not result in a significant improvement to what was observed in two of the formalin experiments that in cDNA synthesis or the expected increase in signal intensity showed an intra-array correlation (0.62 and 0.72). Neither of (Table 1). In order to ensure that the results obtained for these fixation methods produced signals that were as consistent formalin-fixed tissues were not due to artifacts introduced by as frozen tissue across arrays. PAGE 7 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Figure 6. Variability across arrays and across genes. (A) Box plot of the average signal per spot across five arrays hybridized with RNA from frozen human tissue on the human array. (B) Box plot of the average log ratios across arrays. (C) Quantile–quantile plots of log ratios for two representative arrays. Although this depicts only two arrays, comparisons of the others showed a virtually identical pattern (not shown). The nature of microarray noise using TSA amplification Because we performed several replicate hybridizations on the array using the same high quality frozen tissue specimen, we had the ability to look at the amount of observed variability across arrays and across genes, important issues when considering which analytical methods and assumptions to apply. Figure 6A shows box plots of the average intensity per spot in five homo- typic hybridizations using frozen tissues and TSA amplifica- tion. These distributions look very similar. Box plots of the log ratios of intensities for the same arrays are also quite similar, highlighting the global reproducibility of the labeling and hybridization (Fig. 6B). To visualize the similarity between the distributions of log ratios across arrays we considered quantile– quantile plots of log ratios across arrays (Fig. 6C provides one example). These distributions are close to the 45° line expected Figure 7. The distribution of the log ratios by gene. Ratios were generated if the distributions were identical across arrays, again demon- from homotypic hybridizations of frozen tissue on five arrays. Eighty-eight genes gave signals 2-fold above background and are represented on this plot. strating global array-to-array consistency. The line depicts the average, the box encompasses the 95% confidence interval To study variability by gene, we took the average of the log for each gene and circles lie outside the 95% confidence interval. ratios for duplicate spots on each array and considered these values across the five arrays. The distribution of the log ratios per gene is depicted in Figure 7. A conservative approach using a TSA procedure developed to allow use of a small using a standard t-test was first applied to test the null hypoth- amount of RNA in microarray experiments. As with many esis that the mean is zero. This assumes that the distribution of other RNA-based assays, the purity and quality of the starting errors is normal, which is probably not correct. However, this RNA has a significant effect on the results of microarray faulty assumption biases towards more frequent rejection of experiments. In experiments involving co-hybridizations of the the null hypothesis, which is not observed. None of the same cDNA labeled with biotin-dCTP (red pseudo-color) and 88 genes compared gave a P value <0.01. The Flinger test for fluorescein-dCTP (green pseudo-color) the TSA method homogeneity of variance was also used, since it is robust under demonstrated high sensitivity and reproducibility. We non-normality, and yielded a P value of 0.6, demonstrating no observed a similar high correlation (0.96) in experiments in difference in the variances between spots. However, for indi- which fresh frozen samples from tissue cultures labeled by vidual spots significant outliers are observed. Such outliers are TSA were hybridized on a custom 8700 murine array or human observed in most array experiments and can be easily dealt frozen tissue samples labeled by TSA were hybridized on a with analytically and removed, so as to reduce experimental smaller array with 190 elements. In both cases the reproduci- noise. bility of the hybridizations was comparable with that observed using direct labeling without any RNA or signal amplification. The level of potential false positive signals was also low, DISCUSSION demonstrating that the TSA method is of great utility to those Here we report the results from a controlled study performed in trying to use limiting amounts of RNA for microarray experi- parallel on frozen and ethanol- and formalin-fixed tissues ments. In addition, the analysis of signal and ratio variability e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 8 OF 9 over five separate arrays demonstrates that frozen post-mortem bases and it was suggested that incubation of RNA at 70°C human tissues, when handled properly, provide reproducible might significantly increase the quality of cDNA synthesis microarray results under the conditions used here. The (16). This increased variability and poor labeling using archived tissue used in this study was handled optimally; RNA formalin-fixed tissue led to more experimental failures, was extracted under RNase-free conditions after a short shelf resulting in many more attempts with this method than the storage time. This may not reflect the condition of all archived frozen and ethanol-fixed tissues, so as to be sure that this was tissues in pathology laboratories, but perhaps demonstrates the not simply a sample effect. Four different human tissues were best that is possible with this type of tissue. Clearly, the used on five arrays and only once did we obtain a reasonable optimal specimen is rapidly frozen tissue, but ethanol-fixed (>0.75) correlation between the same samples co-hybridized tissue is an alternative. Since ethanol is not as strong a preserv- on the array. ative as formalin, it is not as widely used as a fixative in We also used several simple data visualization and statistical pathology laboratories. However, given its utility for micro- methods to investigate and display the inherent noise in our array studies and gene expression analysis in general, the use array data. These methods are straightforward, standard statistical of ethanol as a fixative for limited focused storage should be tools that give considerable insight into the general quality of increasingly considered. Since RNA quality is measured data generated from microarrays (see for example Figs 6 and before expression analysis, concerns regarding potential RNA 7). Presenting such displays of data quality along with micro- degradation in tissues fixed in ethanol can be addressed. array results and gene lists would improve the interpretability Previously, it was shown that RNA and DNA isolated from of most array studies and should be encouraged to facilitate ethanol-fixed tissues are better templates for both reverse tran- data sharing and interpretation (19). scriptase and Taq polymerase compared with formalin-fixed tissues (17). Our experiments show that the RNA obtained ACKNOWLEDGEMENTS from ethanol-fixed specimens clearly results in better dye incorporation into cDNA, therefore producing more consistent We thank Zugen Chen and Stan Nelson of the UCLA and reproducible data. However, the dynamic range of signal microarray core for providing us with arrays for these studies. intensities from hybridizations with ethanol and formalin were This work was supported by grants from the NIMH, no. within a significantly narrower range compared with the data MH-60233 (D.H.G.), NIA, no. AG-16570 (D.H.G.) and generated using frozen specimens. Therefore, RNA isolated NINDS, no. NS-41408 (V.V.). from ethanol-fixed tissues can be used for microarray-based analysis only if compared with a control treated in a similar REFERENCES manner (Fig. 3). However, due to the increased variability in signal between duplicate hybridizations relative to frozen 1. Lockhart,D.J., Dong,H., Byrne,M.C., Follettie,M.T., Gallo,M.V., Chee,M.S., Mittmann,M., Wang,C., Kobayashi,M., Horton,H. et al. tissue, more replicates need to be performed to reach the same (1996) Expression monitoring by hybridization to high-density level of confidence in the results. Alternatively, follow up oligonucleotide arrays. Nat. Biotechnol., 14, 1675–1680. using another method, such as in situ hybridization or quantita- 2. DeRisi,J.L., Iyer,V.R. and Brown,P.O. (1997) Exploring the metabolic tive RT–PCR, would firmly support the key data derived from and genetic control of gene expression on a genomic scale. Science, 278, ethanol-fixed tissues (12). 680–686. 3. Geschwind,D.H. (2000) Mice, microarrays and the genetic diversity of the In spite of previously reported data on the successful use of brain. Proc. Natl Acad. Sci. USA, 97, 10676–10678. RNA from formalin-fixed tissues in real time quantitative 4. Sandberg,R., Yasuda,R., Pankratz,D.G., Carter,T.A., Del Rio,J.A., RT–PCR (9), formalin-derived RNA was not a good substrate Wodicka,L., Mayford,M., Lockhart,D.J. and Barlow,C. (2000) Regional for cDNA synthesis and clearly did not produce reliable and strain-specific gene expression mapping in the adult mouse brain. hybridizations in our microarray experiments. Generally, poor Proc. Natl Acad. Sci. USA, 97, 11038–11043. 5. Speel,E.J., Hopman,A.H. and Komminoth,P. (1999) Amplification cDNA synthesis results might be explained by prior degradation methods to increase the sensitivity of in situ hybridization: play card(s). of RNA in the tissue before or during fixation, poor yields of J. Histochem. Cytochem., 47, 281–288. RNA associated with various proteins or chemical modifica- 6. Wiedorn,K.H., Kuhl,H., Galle,J., Caselitz,J. and Vollmer,E. (1999) tion of RNA by formalin. It was shown that a high frequency Comparison of in-situ hybridization, direct and indirect in-situ PCR as of non-reproducible sequence alterations occurs in DNA or well as tyramide signal amplification for the detection of HPV. Histochem. Cell Biol., 111, 89–95. RNA isolated from formalin-fixed, paraffin-embedded tissues 7. Qian,X., Bauer,R.A., Xu,H.S. and Lloyd,R.V. (2001) In situ hybridization (18). Although the exact mechanism of RNA and DNA modi- detection of calcitonin mRNA in routinely fixed, paraffin-embedded fication in formalin-fixed specimens is not clear, it is known tissue sections: a comparison of different types of probes combined with that the rate of errors for Taq DNA polymerase is significantly tyramide signal amplification. Appl. Immunohistochem. Mol. Morphol., 9, higher (18). The high rate of misincorporations during cDNA 61–69. 8. Rupp,G.M. and Locker,J. (1988) Purification and analysis of RNA from synthesis might influence the specificity of hybridization, paraffin-embedded tissues. Biotechniques, 6, 56–60. therefore increasing the number of false positive signals on the 9. Specht,K., Richter,T., Muller,U., Walch,A., Werner,M. and Hofler,H. array. Another source of non-specific hybridization could be (2001) Quantitative gene expression analysis in microdissected archival the overall short length of cDNA molecules due to impeded formalin-fixed and paraffin-embedded tumor tissue. Am. J. Pathol., 158, cDNA synthesis. The latter can explain the more successful 419–429. 10. Luo,L., Salunga,R.C., Guo,H., Bittner,A., Joy,K.C., Galindo,J.E., use of formalin-fixed tissues in real-time PCR, as it relies on Xiao,H., Rogers,K.E., Wan,J.S., Jackson,M.R. et al. (1999) Gene much shorter cDNA amplicons and reverse transcription is expression profiles of laser-captured adjacent neuronal subtypes. Nature often performed using random hexamers, avoiding the necessity Med., 5, 117–122. of intact poly(A) tails. The impeded cDNA synthesis might be 11. Kornblum,H.I., Zurcher,S.D., Werb,Z., Derynck,R. and Seroogy,K.B. due to addition of monomethylol upon formalin attack on RNA (1999) Multiple trophic actions of heparin-binding epidermal growth PAGE 9 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 factor (HB-EGF) in the central nervous system. Eur. J. Neurosci., 11, models of variation and assessment of gene effects. Nucleic Acids Res., 3236–3246. 29, 2549–2557. 12. Geschwind,D.H., Ou,J., Easterday,M.C., Dougherty,J.D., Jackson,R.L., 16. Masuda,N., Ohnishi,T., Kawamoto,S., Monden,M. and Okubo,K. (1999) Chen,Z., Antoine,H., Terskikh,A., Weissman,I.L., Nelson,S.F. et al. Analysis of chemical modification of RNA from formalin-fixed samples (2001) A genetic analysis of neural progenitor differentiation. Neuron, 29, and optimization of molecular biology applications for such samples. 325–339. Nucleic Acids Res., 27, 4436–4443. 13. Krafft,A.E., Duncan,B.W., Bijwaard,K.E., Taubenberger,J.K. and 17. Giannella,C., Zito,F.A., Colonna,F., Paradiso,A., Marzullo,F., Alaibac,M. Lichy,J.H. (1997) Optimization of the isolation and amplification of RNA and Schittulli,F. (1997) Comparison of formalin, ethanol and Histochoice from formalin-fixed, paraffin-embedded tissue: The Armed Forces fixation on the PCR amplification from paraffin-embedded breast cancer Institute of Pathology Experience and Literature Review. Mol. Diagn., 2, tissue. Eur. J. Clin. Chem. Clin. Biochem., 35, 633–635. 217–230. 18. Williams,C., Ponten,F., Moberg,C., Soderkvist,P., Uhlen,M., Ponten,J., 14. Sabatti,C., Karsten,S.L. and Geschwind,D.H. (2002) Thresholding rules Sitbon,G. and Lundeberg,J. (1999) A high frequency of sequence for recovering a sparse signal from microarray experiments. Math. Biosci., alterations is due to formalin fixation of archival specimens. Am. J. in press. Pathol., 155, 1467–1471. 15. Tseng,G., Oh,M.-K., Rohlin,L., Liao,J. and Wong,W.H. (2001) Issues in 19. Geschwind,D.H. (2001) Sharing gene expression data: an array of options. cDNA microarray analysis: quality filtering, channel normalization, Nature Rev. Neurosci., 2, 436–438. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray gene expression analysis

Loading next page...
 
/lp/oxford-university-press/an-evaluation-of-tyramide-signal-amplification-and-archived-fixed-and-86a8WSYZGb

References (20)

Publisher
Oxford University Press
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/30.2.e4
Publisher site
See Article on Publisher Site

Abstract

© 2002 Oxford University Press Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray gene expression analysis 1 2 1 Stanislav L. Karsten, Vivianna M. D. Van Deerlin , Chiara Sabatti , Lisa H. Gill and Daniel H. Geschwind* Department of Neurology, Program in Neurogenetics, UCLA School of Medicine, 710 Westwood Plaza, Los Angeles, CA 90095-1769, USA, Department of Pathology and Laboratory Medicine, University of Pennsylvania Health System, 7.098 Founders Pavilion, 3400 Spruce Street, Philadelphia, PA 19104, USA and Department of Human Genetics and Statistics, UCLA School of Medicine, 695 Charles Young Drive South, Los Angeles, CA 90095-7088, USA Received September 5, 2001; Revised October 20, 2001; Accepted November 2, 2001 ABSTRACT specimens, such as primary cell cultures or microdissected tissue samples, one of the frequently encountered limitations of Archival formalin-fixed, paraffin-embedded and the technique can be the amount of RNA required for hybrid- ethanol-fixed tissues represent a potentially invalu- ization. In addition, while fresh frozen tissues from human able resource for gene expression analysis, as they pathological specimens of interest may not be available for are the most widely available material for studies of gene expression studies, archived tissues preserved in various human disease. Little data are available evaluating fixatives often are obtainable. whether RNA obtained from fixed (archival) tissues Various signal amplification techniques have been developed could produce reliable and reproducible microarray within the last several years (5). The MICROMAX cDNA microarray system utilizes tyramide signal amplification expression data. Here we compare the use of RNA (TSA), which requires 20–100 times less RNA than direct isolated from human archival tissues fixed in ethanol cDNA labeling. The TSA method was originally introduced to and formalin to frozen tissue in cDNA microarray improve the sensitivity of immunohistochemistry and, when experiments. Since an additional factor that can limit properly optimized to individual tissues and primary anti- the utility of archival tissue is the often small quantities bodies, has become an important tool for immunofluorescence available, we also evaluate the use of the tyramide microscopy (6,7). Although a cDNA microarray system using signal amplification method (TSA), which allows the TSA is commercially available, the accuracy and reproduci- use of small amounts of RNA. Detailed analysis indi- bility of the technique has not been extensively studied. cates that TSA provides a consistent and reproducible Since the availability of fresh disease specimens is often a signal amplification method for cDNA microarray limiting step in gene expression experiments, other RNA analysis, across both arrays and the genes tested. sources, such as formalin- and ethanol-fixed archival tissues, Analysis of this method also highlights the importance need to be considered. Fixed tissues are generally considered sub-optimal for RNA analytical techniques such as northern of performing non-linear channel normalization and blotting, but in some cases efforts using RNA isolated from dye switching. Furthermore, archived, fixed specimens formalin-fixed tissues in northern hybridization (8) and real can perform well, but not surprisingly, produce more time RT–PCR (9) have been successful. This, together with an variable results than frozen tissues. Consistent increasing need for large-scale gene expression profiling, results are more easily obtainable using ethanol- raises the issue of whether RNA from different pathological fixed tissues, whereas formalin-fixed tissue does specimens might be used in microarray-based experiments. not typically provide a useful substrate for cDNA The application of microarray-based hybridization techniques synthesis and labeling. to human pathological specimens might open new avenues in the study of diseases and subsequently facilitate the identifica- tion of new therapeutic drug targets. This is especially the case INTRODUCTION where large numbers of replicates may need to be studied to The determination of relative expression levels for a large control for human genetic diversity, compared with experi- number of genes using cDNA or oligonucleotide microarray ments using only cell lines or inbred mouse strains (3). Addi- technology is becoming a standard approach for the comparative tionally, while laser capture microdissection coupled to cDNA analysis of gene expression between different cell populations microarray analysis uses fresh frozen or ethanol-fixed specimens or tissues (1–4). Since RNA is very often extracted from rare (10), no comprehensive evaluation of its reproducibility *To whom correspondence should be addressed. Tel: +1 310 794 6570; Fax: +1 310 267 2401; Email: dhg@ucla.edu e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 2 OF 9 Table 1. Summary of 36 microarray experiments performed using the TSA method a b c d e Experiment Hybridizations Correlation Slope Hits NS (T) versus NS (T) 3 0.96 1.08 79% Frozen (T) versus frozen (T) 2 0.96 1.09 85% Frozen (P) versus frozen (P) 1 0.88 0.99 91% Frozen (P) versus frozen (T) 2 0.87 0.91 95% Ethanol (T) versus ethanol (T) 2 (4) 0.87 0.65 99% Ethanol (T) versus frozen (T) 3 (4) 0.61 0.19 99% Formalin (T) versus formalin (T) 3 (6) 0.78 1.85 47% Formalin (T) versus formalin (T) (1 h at 70°C) 2 0.66 0.86 58% Formalin (T) versus formalin (T) 1 0.48 1.05 30% Formalin (T) versus frozen (T) 8 (9) 0.55 0.3 47% Formalin (P) versus frozen (P) 2 0.72 0.16 78% All NS experiments used the mouse 9k array; the others were performed on the human array. The tissues or cell lines compared were two neurosphere cultures, NS7 and NS8, of neuronal progenitor cell lines (NS) grown in different flasks. T, total RNA; P, poly(A) RNA; 1 h at 70°C, tyramide signal amplification of RNA preheated at 70°C for 1 h prior to cDNA synthesis. Number of successful hybridizations in the experiment. The numbers in parentheses indicate the total number of hybridizations (successful and unsuccessful) that were performed. Mean correlation of signals from the two channels of successful hybridizations. Mean slope of the best fit correlation line between successful hybridizations. Number of signals above twice local background. The direct labeling method was used in the experiment. using ethanol-fixed tissues has been published. In the present was confirmed by gel electrophoresis and RT–PCR of house- study we have conducted experiments to evaluate the use of keeping genes. The extraction methods used provided optimal TSA microarray systems and made an attempt to use ethanol- RNA for RT–PCR (13). and formalin-fixed tissues as a source of RNA for gene expres- Microarrays sion experiments using custom-made cDNA microarrays. We performed 36 microarray experiments matching frozen and ethanol- and formalin-fixed samples in various combinations MATERIALS AND METHODS (Table 1). Two types of arrays were used. For experiments Tissue samples and cell cultures comparing human tissues, an array containing 95 relatively abundant human genes gridded in duplicate was used. This size To test the specificity of the TSA microarray system and variables was chosen because it was relatively inexpensive to produce of the ethanol and formalin fixation procedures, post-mortem and had the advantage of containing replicate spots on the brain tissues cut from frontal cortex were obtained from five unrelated human subjects. Post-mortem intervals ranged from same array, facilitating statistical analysis of various human 8 to 11 h. Adjacent tissue sections from each individual were tissues. A larger array containing 8700 mouse genes (http:// immediately frozen in liquid nitrogen and stored at –80°C; www.medsch.ucla.edu/som/humgen/nelsonlab) was used only others were immersed in formalin (10%, pH 7.0) or RNase-free to elucidate reproducibility of the TSA method. The mouse 70% ethanol in 10 mM Tris–HCl pH 7.6, 1 mM EDTA and array was based on Unigene clusters and contained 2700 fixed for 12–18 h at 4°C. Neural progenitor colonies (neuro- known genes for which the full-length cDNA sequences are spheres) were cultured from neonatal mouse neocortex as known and an additional 6000 ESTs. This array has the advantage described previously (11,12). that it contains genes of varying abundances and many of these genes are low abundance genes based on either previous Extraction of RNA from frozen and fixed tissues expression studies or EST sequencing data. This allowed evalua- The RNA from cultured cells and frozen tissue sections was tion of the TSA method over a large number of genes of extracted using acid phenol extraction (Trizol LS; Gibco BRL) different abundance levels. These microarrays were as recommended by the manufacturer. One milliliter of Trizol constructed in the UCLA Genetics Microarray Core using a per 50 mg tissue was used. RNA from paraffin-embedded custom arrayer and previously described methods (12). tissue was extracted using a combination of proteinase K and Briefly, clone inserts were PCR amplified using vector Trizol reagent as described previously (13), followed by primers, purified by isopropanol precipitation, resuspended in DNase treatment (37°C, 30 min). RNA was dissolved in TE high pH buffer at ∼200–800 ng/µ l concentration and gridded at buffer and stored at –80°C. The purity of RNA was checked by measuring the optical density at 260 and 280 nm. The quality high density onto poly-L-lysine-coated slides. PAGE 3 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Probe preparation and hybridization Direct RNA labeling was performed using Cy3-dCTP and Cy5- dCTP (catalog nos PA53021 and PA55021; Amersham), oligo(dT) (12–18mer) and Superscript II reverse transcriptase (catalog no. 18064-014; Life Technologies). (See http://www.tigr.org/tdb/ microarray/conciseguide.html or geschwindlab.medsch.ucla.edu for complete protocol.) The labeling reaction contained 50 µ g total RNA, 500 µ M dUTP, 500 µ M dATP, 500 µ M dTTP, 100 µ M dCTP, 1 mM Cy3–dCTP/Cy5–dCTP, 400 U Super- script II reverse transcriptase (Gibco BRL), 1 mM DTT and 1× reverse transcriptase buffer supplied by the manufacturer (NEN). The reaction mix was incubated for 3 h at 42°C and stopped with 20 mM EDTA. After degradation of the RNA by NaOH for 30 min (final concentration 25 mM) and neutralization with HCl, unincorporated fluorescent nucleotides were removed by isopropanol precipitation. The TSA probe labeling and array hybridization were performed as described in the instruction manual (MICROMAX Human cDNA Microarray System, NEN Life Science Products, Boston, MA) with minor modifications (see http://Geschwindlab.medsch.ucla.edu for detailed protocol). Biotin- and fluorescein-labeled cDNAs were generated from 0.5 and 1.5 µ g total RNA for the 92 and 8700 gene cDNA microarrays, respectively. We found that increasing the cDNA synthesis time from 1 h, as recommended by the manufacturer, to 3 h resulted in much more complete and reproducible labeling. Post-hybridization washes were performed according to the manufacturer’s recommendations (MICROMAX; NEN). The signals from specifically hybridized biotin- and fluorescein-labeled cDNAs were amplified either with streptavidin–horseradish peroxidase (HRP) and Cy5–tyramide or antifluorescein–HRP and Cy3–tyramide, respectively. After signal amplification, the cDNA microarrays were air dried and scanned in a Genetic Microsystems 418 microarray scanner. The images were analyzed using ImaGene 4.1 (Biodiscovery, Santa Monica, CA). Prior to quantitative analysis, normaliz- ation was performed using a non-linear function to control for Figure 1. Representative images of a typical TSA hybridization on the mouse 9k dye and signal intensity effects (14). Background correction microarray (A) and the 95 gene, 190 element human microarray (B). In (B), dupli- was first conducted using a smoothed background correction, cate spots are adjacent to each other and the consistency in hybridizations which resulted in a lower variance than subtracting local back- between duplicates can be observed. The image in (A) is a non-homotypic ground. All hybridizations were done at least in duplicate and hybridization, whereas (B) is homotypic hybridization of RNA extracted from human frontal cortex. repeated with the fluorophores reversed. Quantified intensity data was downloaded into GeneSpring 3.1 (Silicon Genetics) or Microsoft Excel and analyzed as described in the Results. shows representative hybridizations using TSA onto the 190 For computation and graphical display we used R, a statistical (95 genes) and 8700 element arrays. The overall images from language freely available at http://CRAN.R-Project.org. direct labeling (data not shown) and TSA appear similar and the background signals appear low. Figure 2A depicts a plot of RESULTS signal versus background using TSA amplification of RNA derived from frozen neurospheres (NS8). Eighty-three percent Sensitivity and reproducibility of the TSA microarray of the total points are 2-fold above background and are considered system ‘hits’. Figure 2B shows a plot of the log ratio of expression versus the average log intensity per spot. Two very important To assess the reproducibility and quality of the data generated technical observations follow from a careful analysis of the in frozen and ethanol- and formalin-fixed tissue samples, multiple hybridizations using the same sample, labeled with observed signals. From Figure 2A and B emerges the importance different fluorophores and co-hybridized on the same array, of using an appropriate threshold for considering what constitutes were performed. Homotypic hybridizations (frozen versus a real signal and the consideration of signal strength measures. This is most pronounced at signals <2-fold above background. frozen experiments) were used to estimate the reproducibility of the TSA method and were further used as a benchmark for A second technical issue relates to the method of normaliz- the evaluation of fixed versus fixed data quality. Figure 1 ation. Non-linear normalization has been championed in the e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 4 OF 9 Figure 2. Scatter plots of experiments with total RNA isolated from mouse neural stem cell cultures (neurospheres, NS). The same RNA sample was used for cDNA synthesis and labeled with either fluorescein-12-dCTP or biotin-11-dCTP. Both cDNAs were co-hybridized on the mouse 9k microarray (homotypic hybrid- ization) and developed according to the TSA protocol provided by the manufacturer (MICROMAX; NEN). (A) The plot (on a log scale) of Cy5 signal versus background from a representative homotypic hybridization of NS RNA on the mouse 9k array. The plots for Cy3 were similar. The points above the solid line have a signal higher than 2-fold background. The letter E indicates the signal produced by negative control spots. Notable is the clear separation between signal and background apparent in the two clusters of spots that are observed when the data are plotted in this manner. (B) A representative plot of the log ratio (Cy3/Cy5) versus the average log signal from both Cy3 and Cy5 channels for each spot in homotypic hybridization. It can be seen that the likelihood of false positive ratios decreases with signal strength. (C) Scatter plots of the signal generated from NS versus NS hybridizations from two experiments where the dyes were reversed. In this case, rather than labeling the same RNA, each NS culture, NS7 and NS8, was grown in separate flasks and processed separately. The clear non-linearity (curvature) produced by differences in dye incorporation and signal at different cDNA abundances can be observed. literature, but has not been widely adopted by most experi- spot deposition and attachment to the slide surface, slide mentalists, who typically use global normalization methods washing or local hybridization conditions, between-slide (15). Non-linearity is reflected in the curvature present in comparisons are not made with spotted cDNA arrays. In this scatter plots from globally normalized signals. This may be case, using TSA, we were curious to determine whether due to dye incorporation effects that may differ at different between-slide variability would be prohibitive to performing cDNA abundances (see for example Fig. 2C). Globally between-slide comparisons. As expected, the observed correl- normalized signals will contain errors that can be removed by ations between the same spot on different slides were less than use of a non-linear normalization to remove the curvature same spot comparisons (Fig. 3B). However, these correlations induced by dye effects. Furthermore, the effect of the dyes can were relatively high: when signals obtained on different slides be minimized by performing duplicate hybridizations with from the same spot were compared the correlation averaged reversed labels, as was done here. This is clearly an important 0.84 (range 0.77–0.87), suggesting that with a few more replicate experiments, such an approach of comparison across slides consideration when using TSA. would be technically feasible. Thus, overall the signals A high correlation between duplicate or like samples co- produced using TSA to amplify the signal from small quantities of hybridized on the 8700 element mouse array was observed. Figure 3 depicts scatter plots derived from the co-hybridization RNA were very consistent across a large number of genes assayed. of a duplicate sample labeled with different dyes. Correlations were around 0.96, even when the independent replicate Comparison of archived tissues samples from different culture dishes were co-hybridized (Fig. 3B). This demonstrates high data quality and reproduci- Quantitative gene expression analysis in matched frozen, bility of the hybridizations using TSA. Typically, because of formalin-fixed, paraffin-embedded and ethanol-fixed tissue slide-to-slide variability due to factors such as inconsistent samples were performed using the 190 human element array, PAGE 5 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Figure 3. Hybridization consistency on the 9K microarray using TSA amplification. In each graph, Cy5 signals are shown on the ordinate and Cy3 signals on the abscissa on a log scale after non-linear normalization. Each dot represents the hybridization intensity of each gene. While there was more variability in hybridizations obtained from the same sample hybridized to different arrays (B) relative to the same array (A), both showed high reproducibility. Figure 4. Comparison of RNA from frozen frontal cortex versus fixed frontal cortex. Fixed and frozen sample pairs are from an adjacent brain section from the same subject hybridized on the human array. Each dot represents the hybridization intensity of each gene. The signals for RNA from frozen tissue are shown on the ordinate in each graph on a log scale. (A) Total RNA from frozen frontal cortex versus total RNA from formalin-fixed frontal cortex. (B) Another independent sample hybridization showing the signal obtained using total RNA from frozen frontal cortex versus total RNA from formalin-fixed frontal cortex. (C) Total RNA from frozen versus ethanol fixed frontal cortex. which contains mostly high abundance genes expressed in fixed, respectively. Because of this effect, it is virtually impos- many different tissue types. To test the specificity of the sible to compare samples preserved in different fashions, as is formalin and ethanol fixation procedures, post-mortem frozen, shown by the low correlations observed when plotting signals ethanol-fixed and formalin-fixed brain specimens obtained obtained for frozen versus tissues fixed in different manners from the frontal cortex of five human subjects were used. As from the same individual brain (Fig. 4). The different behav- previously, only the spots generating signals above 2.0 times iors of frozen and fixed samples also makes it impossible to local background were considered true signals. This corre- use global normalization. For this reason, we only normalized sponded to 94% of spots for frozen tissues, 99% for ethanol- slides with homotypic hybridizations and used these compari- fixed and only 56% for formalin-fixed tissues in homotypic sons as the basis for the rest of the analysis. hybridizations (Table 1 and Fig. 4). Again, as observed on the larger array, signals obtained from Not surprisingly, the range of signals for fixed tissues was the same frozen tissue labeled with different fluorophores and significantly lower than for frozen tissues, with only 12% and co-hybridized on the array were highly reproducible (Fig. 5A). 3% of the signals above 4 times background for ethanol- and Correlations of normalized signals ranged from 0.95 to 0.99, formalin-fixed tissues, respectively (Table 1). This demon- similar to what was observed using fresh cultured neuronal strates that formalin-fixed tissues resulted in quite weak precursor cells on the 8700 array (Fig. 3A). Correlations for signals overall, which is borne out by the observed back- ethanol versus ethanol hybridizations within the same array ground-corrected, average log signal intensities, which were were lower (0.85 and 0.89), but within an acceptable range 3.34, 3.16 and 3.03 for frozen, ethanol-fixed and formalin- (Fig. 5B). Sample hybridizations were also conducted with e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 6 OF 9 Figure 5. Scatter plots of homotypic hybridization experiments using RNA isolated from the same tissue with different fixation methods. (A) Total RNA versus + + total RNA from frozen frontal cortex of a single subject. (B) Poly(A) RNA versus poly(A) RNA from frozen frontal cortex of a single subject. (C)Total RNA versus total RNA isolated from ethanol-fixed frontal cortex of the same subject shown in (A). (D) Total RNA versus total RNA from formalin-fixed frontal cortex of a different subject. poly(A) RNA as template and the correlations were no better TSA, direct labeling of the same RNA was performed. Direct than those observed using total RNA (Fig. 5C). labeling of total RNA isolated from formalin-fixed tissues also did not result in any significant improvement; only 30% of Formalin-fixed tissue posed many problems. In about half of genes produced signals above twice the local background, less the cases the experiments simply did not work; one or both dyes did not incorporate properly. Poor cDNA synthesis or than observed with TSA, showing a correlation of signals lack of dye incorporation was much less frequent and did not between duplicate samples of 0.62 and a slope of 1.3 (Table 1). happen even once in the current frozen versus frozen experi- Across-array comparisons were also performed, although ments. Thus, the failure rate using formalin-fixed tissues was the use of different human tissues would necessarily reduce the high, an important consideration given the cost of microarray correlations across arrays due to many factors, such as differ- experiments. Using formalin, an acceptable level of dye incor- ences in post-mortem interval and human genetic diversity. poration was attained in three experiments out of nine Surprisingly high correlations, similar to those observed in the attempted, yielding correlations of 0.87, 0.32 and 0.53, less tissue culture control experiments (see for example Fig. 3A), consistent than either frozen or ethanol-fixed tissues (Fig. 5D were observed between the signal intensities of the same spot and Table 1). This, coupled with the far lower dynamic range on different arrays. These inter-array correlations averaged of signal obtained using formalin, unfortunately suggests that 0.92, with a range of 0.86–0.95. Less correlation was observed its utility is very limited for current microarray analysis. between ethanol versus ethanol experiments. One of the experi- Previously, it was reported that better cDNA yield could be ments was a clear outlier, with a correlation of ∼0.2 with the obtained by simple preheating of RNA prior to reverse tran- other experiments. The other experiments showed a tighter scription (16). Preheating formalin-fixed RNA for 1 h at 70°C range of correlations, ranging from 0.62 to 0.75. This is similar in our experiments did not result in a significant improvement to what was observed in two of the formalin experiments that in cDNA synthesis or the expected increase in signal intensity showed an intra-array correlation (0.62 and 0.72). Neither of (Table 1). In order to ensure that the results obtained for these fixation methods produced signals that were as consistent formalin-fixed tissues were not due to artifacts introduced by as frozen tissue across arrays. PAGE 7 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 Figure 6. Variability across arrays and across genes. (A) Box plot of the average signal per spot across five arrays hybridized with RNA from frozen human tissue on the human array. (B) Box plot of the average log ratios across arrays. (C) Quantile–quantile plots of log ratios for two representative arrays. Although this depicts only two arrays, comparisons of the others showed a virtually identical pattern (not shown). The nature of microarray noise using TSA amplification Because we performed several replicate hybridizations on the array using the same high quality frozen tissue specimen, we had the ability to look at the amount of observed variability across arrays and across genes, important issues when considering which analytical methods and assumptions to apply. Figure 6A shows box plots of the average intensity per spot in five homo- typic hybridizations using frozen tissues and TSA amplifica- tion. These distributions look very similar. Box plots of the log ratios of intensities for the same arrays are also quite similar, highlighting the global reproducibility of the labeling and hybridization (Fig. 6B). To visualize the similarity between the distributions of log ratios across arrays we considered quantile– quantile plots of log ratios across arrays (Fig. 6C provides one example). These distributions are close to the 45° line expected Figure 7. The distribution of the log ratios by gene. Ratios were generated if the distributions were identical across arrays, again demon- from homotypic hybridizations of frozen tissue on five arrays. Eighty-eight genes gave signals 2-fold above background and are represented on this plot. strating global array-to-array consistency. The line depicts the average, the box encompasses the 95% confidence interval To study variability by gene, we took the average of the log for each gene and circles lie outside the 95% confidence interval. ratios for duplicate spots on each array and considered these values across the five arrays. The distribution of the log ratios per gene is depicted in Figure 7. A conservative approach using a TSA procedure developed to allow use of a small using a standard t-test was first applied to test the null hypoth- amount of RNA in microarray experiments. As with many esis that the mean is zero. This assumes that the distribution of other RNA-based assays, the purity and quality of the starting errors is normal, which is probably not correct. However, this RNA has a significant effect on the results of microarray faulty assumption biases towards more frequent rejection of experiments. In experiments involving co-hybridizations of the the null hypothesis, which is not observed. None of the same cDNA labeled with biotin-dCTP (red pseudo-color) and 88 genes compared gave a P value <0.01. The Flinger test for fluorescein-dCTP (green pseudo-color) the TSA method homogeneity of variance was also used, since it is robust under demonstrated high sensitivity and reproducibility. We non-normality, and yielded a P value of 0.6, demonstrating no observed a similar high correlation (0.96) in experiments in difference in the variances between spots. However, for indi- which fresh frozen samples from tissue cultures labeled by vidual spots significant outliers are observed. Such outliers are TSA were hybridized on a custom 8700 murine array or human observed in most array experiments and can be easily dealt frozen tissue samples labeled by TSA were hybridized on a with analytically and removed, so as to reduce experimental smaller array with 190 elements. In both cases the reproduci- noise. bility of the hybridizations was comparable with that observed using direct labeling without any RNA or signal amplification. The level of potential false positive signals was also low, DISCUSSION demonstrating that the TSA method is of great utility to those Here we report the results from a controlled study performed in trying to use limiting amounts of RNA for microarray experi- parallel on frozen and ethanol- and formalin-fixed tissues ments. In addition, the analysis of signal and ratio variability e4 Nucleic Acids Research, 2002, Vol. 30, No. 2 PAGE 8 OF 9 over five separate arrays demonstrates that frozen post-mortem bases and it was suggested that incubation of RNA at 70°C human tissues, when handled properly, provide reproducible might significantly increase the quality of cDNA synthesis microarray results under the conditions used here. The (16). This increased variability and poor labeling using archived tissue used in this study was handled optimally; RNA formalin-fixed tissue led to more experimental failures, was extracted under RNase-free conditions after a short shelf resulting in many more attempts with this method than the storage time. This may not reflect the condition of all archived frozen and ethanol-fixed tissues, so as to be sure that this was tissues in pathology laboratories, but perhaps demonstrates the not simply a sample effect. Four different human tissues were best that is possible with this type of tissue. Clearly, the used on five arrays and only once did we obtain a reasonable optimal specimen is rapidly frozen tissue, but ethanol-fixed (>0.75) correlation between the same samples co-hybridized tissue is an alternative. Since ethanol is not as strong a preserv- on the array. ative as formalin, it is not as widely used as a fixative in We also used several simple data visualization and statistical pathology laboratories. However, given its utility for micro- methods to investigate and display the inherent noise in our array studies and gene expression analysis in general, the use array data. These methods are straightforward, standard statistical of ethanol as a fixative for limited focused storage should be tools that give considerable insight into the general quality of increasingly considered. Since RNA quality is measured data generated from microarrays (see for example Figs 6 and before expression analysis, concerns regarding potential RNA 7). Presenting such displays of data quality along with micro- degradation in tissues fixed in ethanol can be addressed. array results and gene lists would improve the interpretability Previously, it was shown that RNA and DNA isolated from of most array studies and should be encouraged to facilitate ethanol-fixed tissues are better templates for both reverse tran- data sharing and interpretation (19). scriptase and Taq polymerase compared with formalin-fixed tissues (17). Our experiments show that the RNA obtained ACKNOWLEDGEMENTS from ethanol-fixed specimens clearly results in better dye incorporation into cDNA, therefore producing more consistent We thank Zugen Chen and Stan Nelson of the UCLA and reproducible data. However, the dynamic range of signal microarray core for providing us with arrays for these studies. intensities from hybridizations with ethanol and formalin were This work was supported by grants from the NIMH, no. within a significantly narrower range compared with the data MH-60233 (D.H.G.), NIA, no. AG-16570 (D.H.G.) and generated using frozen specimens. Therefore, RNA isolated NINDS, no. NS-41408 (V.V.). from ethanol-fixed tissues can be used for microarray-based analysis only if compared with a control treated in a similar REFERENCES manner (Fig. 3). However, due to the increased variability in signal between duplicate hybridizations relative to frozen 1. Lockhart,D.J., Dong,H., Byrne,M.C., Follettie,M.T., Gallo,M.V., Chee,M.S., Mittmann,M., Wang,C., Kobayashi,M., Horton,H. et al. tissue, more replicates need to be performed to reach the same (1996) Expression monitoring by hybridization to high-density level of confidence in the results. Alternatively, follow up oligonucleotide arrays. Nat. Biotechnol., 14, 1675–1680. using another method, such as in situ hybridization or quantita- 2. DeRisi,J.L., Iyer,V.R. and Brown,P.O. (1997) Exploring the metabolic tive RT–PCR, would firmly support the key data derived from and genetic control of gene expression on a genomic scale. Science, 278, ethanol-fixed tissues (12). 680–686. 3. Geschwind,D.H. (2000) Mice, microarrays and the genetic diversity of the In spite of previously reported data on the successful use of brain. Proc. Natl Acad. Sci. USA, 97, 10676–10678. RNA from formalin-fixed tissues in real time quantitative 4. Sandberg,R., Yasuda,R., Pankratz,D.G., Carter,T.A., Del Rio,J.A., RT–PCR (9), formalin-derived RNA was not a good substrate Wodicka,L., Mayford,M., Lockhart,D.J. and Barlow,C. (2000) Regional for cDNA synthesis and clearly did not produce reliable and strain-specific gene expression mapping in the adult mouse brain. hybridizations in our microarray experiments. Generally, poor Proc. Natl Acad. Sci. USA, 97, 11038–11043. 5. Speel,E.J., Hopman,A.H. and Komminoth,P. (1999) Amplification cDNA synthesis results might be explained by prior degradation methods to increase the sensitivity of in situ hybridization: play card(s). of RNA in the tissue before or during fixation, poor yields of J. Histochem. Cytochem., 47, 281–288. RNA associated with various proteins or chemical modifica- 6. Wiedorn,K.H., Kuhl,H., Galle,J., Caselitz,J. and Vollmer,E. (1999) tion of RNA by formalin. It was shown that a high frequency Comparison of in-situ hybridization, direct and indirect in-situ PCR as of non-reproducible sequence alterations occurs in DNA or well as tyramide signal amplification for the detection of HPV. Histochem. Cell Biol., 111, 89–95. RNA isolated from formalin-fixed, paraffin-embedded tissues 7. Qian,X., Bauer,R.A., Xu,H.S. and Lloyd,R.V. (2001) In situ hybridization (18). Although the exact mechanism of RNA and DNA modi- detection of calcitonin mRNA in routinely fixed, paraffin-embedded fication in formalin-fixed specimens is not clear, it is known tissue sections: a comparison of different types of probes combined with that the rate of errors for Taq DNA polymerase is significantly tyramide signal amplification. Appl. Immunohistochem. Mol. Morphol., 9, higher (18). The high rate of misincorporations during cDNA 61–69. 8. Rupp,G.M. and Locker,J. (1988) Purification and analysis of RNA from synthesis might influence the specificity of hybridization, paraffin-embedded tissues. Biotechniques, 6, 56–60. therefore increasing the number of false positive signals on the 9. Specht,K., Richter,T., Muller,U., Walch,A., Werner,M. and Hofler,H. array. Another source of non-specific hybridization could be (2001) Quantitative gene expression analysis in microdissected archival the overall short length of cDNA molecules due to impeded formalin-fixed and paraffin-embedded tumor tissue. Am. J. Pathol., 158, cDNA synthesis. The latter can explain the more successful 419–429. 10. Luo,L., Salunga,R.C., Guo,H., Bittner,A., Joy,K.C., Galindo,J.E., use of formalin-fixed tissues in real-time PCR, as it relies on Xiao,H., Rogers,K.E., Wan,J.S., Jackson,M.R. et al. (1999) Gene much shorter cDNA amplicons and reverse transcription is expression profiles of laser-captured adjacent neuronal subtypes. Nature often performed using random hexamers, avoiding the necessity Med., 5, 117–122. of intact poly(A) tails. The impeded cDNA synthesis might be 11. Kornblum,H.I., Zurcher,S.D., Werb,Z., Derynck,R. and Seroogy,K.B. due to addition of monomethylol upon formalin attack on RNA (1999) Multiple trophic actions of heparin-binding epidermal growth PAGE 9 OF 9 Nucleic Acids Research, 2002, Vol. 30, No. 2 e4 factor (HB-EGF) in the central nervous system. Eur. J. Neurosci., 11, models of variation and assessment of gene effects. Nucleic Acids Res., 3236–3246. 29, 2549–2557. 12. Geschwind,D.H., Ou,J., Easterday,M.C., Dougherty,J.D., Jackson,R.L., 16. Masuda,N., Ohnishi,T., Kawamoto,S., Monden,M. and Okubo,K. (1999) Chen,Z., Antoine,H., Terskikh,A., Weissman,I.L., Nelson,S.F. et al. Analysis of chemical modification of RNA from formalin-fixed samples (2001) A genetic analysis of neural progenitor differentiation. Neuron, 29, and optimization of molecular biology applications for such samples. 325–339. Nucleic Acids Res., 27, 4436–4443. 13. Krafft,A.E., Duncan,B.W., Bijwaard,K.E., Taubenberger,J.K. and 17. Giannella,C., Zito,F.A., Colonna,F., Paradiso,A., Marzullo,F., Alaibac,M. Lichy,J.H. (1997) Optimization of the isolation and amplification of RNA and Schittulli,F. (1997) Comparison of formalin, ethanol and Histochoice from formalin-fixed, paraffin-embedded tissue: The Armed Forces fixation on the PCR amplification from paraffin-embedded breast cancer Institute of Pathology Experience and Literature Review. Mol. Diagn., 2, tissue. Eur. J. Clin. Chem. Clin. Biochem., 35, 633–635. 217–230. 18. Williams,C., Ponten,F., Moberg,C., Soderkvist,P., Uhlen,M., Ponten,J., 14. Sabatti,C., Karsten,S.L. and Geschwind,D.H. (2002) Thresholding rules Sitbon,G. and Lundeberg,J. (1999) A high frequency of sequence for recovering a sparse signal from microarray experiments. Math. Biosci., alterations is due to formalin fixation of archival specimens. Am. J. in press. Pathol., 155, 1467–1471. 15. Tseng,G., Oh,M.-K., Rohlin,L., Liao,J. and Wong,W.H. (2001) Issues in 19. Geschwind,D.H. (2001) Sharing gene expression data: an array of options. cDNA microarray analysis: quality filtering, channel normalization, Nature Rev. Neurosci., 2, 436–438.

Journal

Nucleic Acids ResearchOxford University Press

Published: Jan 15, 2002

There are no references for this article.