Microarrays – The Challenge of Preparing Brain Tissue
LAURA SOVERCHIA,1MASSIMO UBALDI,1FERNANDO LEONARDI-ESSMANN,2
ROBERTO CICCOCIOPPO1& GARY HARDIMAN3
1Department of Pharmacological Science and Experimental Medicine, University of Camerino, 62032 Camerino (MC), Italy,
2Department of Psychopharmacology, Central Institute of Mental Health, University of Heidelberg, 68159 Mannheim,
Germany, and3Department of Medicine, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0349,
Microarray experiments allow researchers to collect an amazing amount of gene expression data that
have the potential to provide unique information to help interpretation of the biological functions of the
central nervous system. These experiments are, however, technically demanding and present unique
difficulties when used in the context of neuroscience research, in particular. Success or failure of
microarray experiments are highly dependent on reproducible target preparations. This involves a
relatively long chain of preparation steps, such as removal of tissue from experimental animals or from
post-mortem human brains, storage, selection, and excision of brain regions. This is followed by RNA
extraction, reverse transcription, and labeling of target cDNAs or cRNAs. Additionally, it is
emphasized that the quality of microarray data largely relies on the proper handling of animals
throughout experiments and the time of the day when experiments are stopped. This article tries to
provide hints for some basic rules to be observed in preparation of samples for expression profiling
Gene expression analysis has progressed in a relatively short
time from the traditional ‘one gene at a time’ analysis to
detailed surveys of complete genomes. The increasing use and
acceptance of microarrays to study genetic and cellular
processes is clearly demonstrated by the increasing number
of citations in the literature. Two dominant methodologies
have emerged, a one-color approach that has been historically
associated with GeneChip
hybridization with cDNA based arrays (Chee et al., 1996;
Brown and Botstein, 1999). The GeneChip
many research applications and is manufactured with rigorous
quality control standards, a necessity for accurate and
sensitive genetic measurements. Rival commercial oligonu-
cleotide chip formats with similar genetic content have been
commercialized by Amersham Biosciences, (Piscataway, NJ)
the CodeLink platform and Agilent (Palo Alto, CA) (Ramak-
rishnan et al., 2002; Hughes et al., 2001; Hardiman, 2003).
Before embarking upon a microarray study, significant time
and effort is required to design an experiment that is
1oligonucleotide arrays (Affyme-
1is widely used in
biologically sound and statistically robust (Gaasterland and
Bekiranov, 2000; Kerr and Churchill, 2001). Once target
genes are identified, additional time and expense are needed
to validate the selection of relevant genes and place them in
the context of the biological process under study. However, a
prerequisite for a successful gene array experiment is a high
quality biological sample that permits high quality RNA
In the field of addiction research, pharmacological effects
on gene expression in different organs known as targets of
drugs of abuse are studied. This chapter will focus on sample
preparation and RNA extraction from brain tissue as starting
points for expression studies. There are two reasons for
choosing brain tissue: (1) the brain is the organ, which plays a
major role in drug addiction, and a growing number of
research groups are focusing their interests on gene profiling
studies involving the central nervous system and (2) the brain
has by far the most complex cytoarchitectural and functional
structure. Defining general rules for sample preparation and
RNA extraction from the brain largely covers related issues
regarding other organs.
Correspondence to: Roberto Ciccocioppo, Department of Pharmacological Science and Experimental Medicine, University of Camerino,
62032 Camerino (MC), Italy. Tel: +390737403313. E-mail: firstname.lastname@example.org.
Supported by EC-grant TARGALC QLG3-CT-2002-01048
Addiction Biology (March 2005) 10, 5–13
ISSN 1355-6215 print/ISSN 1369-1600 online
# Society for the Study of Addiction to Alcohol and Other Drugs
Taylor & Francis Group Ltd
Laboratory animal and human brain tissues
For obvious reasons, there are specific issues related to
starting an experiment with brain samples obtained from
laboratory animals or from post-mortem human specimens.
Over the years, a growing body of evidence has accumulated
showing that the manipulation of animals may result in rapid
and often dramatic changes in gene expression in the brain.
For example, studies analyzing early onset gene expression (c-
fos, c-Jun, krox) have revealed that changes in their transcript
levels may occur within a few minutes in response to animal
handling (Herdegen and Leah, 1998). Hence, always the same
experienced caretakers should be looking after the animals
during the whole experimental period and should also be
responsible for a reproducible killing procedure (Asanuma
and Ogawa, 1994). Environmental changes, resulting in
exposure to stressful events, presentation of odors or visual
cues may rapidly change the expression of those and other
genes (Herdegen and Leah, 1998; Keay and Bandler, 2001;
Broad et al., 2002). For example, when groups of animals are
transferred together from their home cages to the laboratory
where they are sacrificed, animals sacrificed later are exposed
to the new, potentially stressful, laboratory environment for a
longer period of time. Moreover, blood odor of the animals
killed earlier may cause marked behavioral arousals that likely
result in artificially induced changes in gene-expression levels.
Another important aspect is the time of the day when
experiments are stopped. Expression of most genes of an
organism is profoundly influenced by gene products of the so-
called ‘‘biological clock’’. Expression of these genes is
entrained by the dark-light cycle of the day. ‘‘Clock’’-genes
have been shown to influence and control a wide variety of
genes both in the periphery and in brain. Hence, the
expression pattern in e.g. N. accumbens or in liver of an
animal killed in the morning may be completely different from
the pattern seen in the same tissues of its littermate sacrificed
in the evening. These pitfalls may be particularly relevant in
gene array experiments where thousands of genes and
complex patterns of expression are analyzed. Researchers
should always be aware of these problems, and when brain
tissue from laboratory animals is used for gene expression
experiments the molecular aspect of the study should never be
viewed separately from behavioral analysis.
Post mortem human brain
During the last decade, concern and pessimism have prevailed
about RNA stability in post mortem brain samples. Recent
studies, however, have shown that high quality RNA can be
obtained from this tissue if samples are collected and
preserved under appropriate conditions (Bahn et al., 2001;
Cummings et al., 2001; Yasojima et al., 2001; Vonsattlel et
al., 1995). Often for legal reasons, human tissue cannot be
sampled and frozen or fixed before a certain post-mortem
interval (PMI). The longer the PMI is, the higher is the risk to
have RNA degradation, especially for those RNAs with a short
half-life (i.e., c-fos has a half-life of 15 minutes, COX-2 half-
live 30 minutes) (Shyu et al., 1989; Schramm et al., 1999). In
this regard the literature is rather confusing, but 36–48 hours
are generally considered acceptable (Johnson et al., 1986;
Barton et al., 1993). Another important factor that influences
RNA integrity is tissue pH (Kingsbury et al., 1995; Hynd et
al., 2003), intact mRNA being associated with brain pH
measurements in the range of pH 6.1–7. Tissues with low pH
(56) contain fragmented mRNAs. Death subsequent to
prolonged agony causing hypoxia may result in a drop in
tissue pH and subsequent reduction of mRNA integrity
(Harrison et al., 1991; Harrison et al., 2000). All these
variables should be taken into account before using post-
Brain tissue heterogeneity
When applying microarray technology in neuroscience, one
has to be aware of the fact that the mammalian brain is
characterized by a unique cellular and anatomical complexity.
Heterogeneous cell types (e.g., neurons, glial cells and
subtypes within) form complex cytoarchitectures that com-
partmentalize the various brain areas or are grouped within the
same nucleus into structurally distinct subregions. In addition,
these cells are organized in complex synaptic networks that
have distinct neurophysiological, neurochemical, morphologi-
organization of the brain, it is extremely difficult to obtain
homogeneous samples to use for RNA extraction and
subsequent array experiments. The problems arising in
expression profiling studies by the heterogeneous nature of
the nervous system gain particular importance when the
researcher is faced with two problems emerging from micro-
array studies (Geschwind 2000; Bonaventure et al., 2002):
Various neuronal cells react to perturbations (such as disease)
or modifications of physiological conditions with relatively
moderate changes in gene expression. Whilst in cancer
tissue, gene expression seems to change often more than
two-fold, in psychiatric diseases such as drug addiction
the relevant clinical and physiological features can often
be due to minor or moderate changes in gene expression.
For example, relatively small changes in key regulatory
gene products during early development are believed to
lead to increased susceptibility to a host of neurodeve-
lopmental conditions, including autism, attention deficit
disorder, and schizophrenia (Cowan et al., 2002). Key
gene products decisive for neuronal function, such as
neurotransmitters, receptors and their regulatory factors,
are expressed at very low levels compared to other
cellular constituents such as structural proteins. Thus in
the brain, biologically significant differences are often
due to very little changes in gene expression. The
arbitrary cut-off levels (4two-fold change) typically
employed in microarray studies often cannot be used
when applied to neurobiological tissues. This has been
documented in a number of more recent investigations
(Jiang et al., 2000; Wurmbach et al., 2002; Sokolov et
al., 2003). Therefore, discriminating real differences
from experimental noise poses a significant challenge in
microarray analysis of brain tissue.
Functional differentiation amongst neuroanatomical districts
is often due to enrichment of transcripts for only a small
Soverchia et al.
number of genes. Neuronal cells are highly differentiated
in terms of their gene expression patterns. A few gene
products enriched in a particular location could assume
a central role in shaping the functional properties of that
specific brain region or subregion. It means that this
intrinsic heterogeneity often results in the genes of
interest changing their expression levels only in a small
fraction of the overall cell population. Therefore, a
significant fold change in the expression of a particular
gene can be diluted considerably if the cell type
expressing a particular gene represents only a fraction
of the overall population being studied. Furthermore, a
relevant change such as up-regulation of a gene’s
expression in one cell population can be masked by
down-regulation of this gene’s expression in a neighbor-
ing cell population, resulting in an underestimation of
the real changes.
Signature genes for specific brain nuclei can not only be used
to delineate functional areas of the brain but also can be used
to do a post-hoc check of the precision of the sampling method
for a given brain region. This is performed by looking up the
expression levels of the site-specific cluster of genes in the
analyzed sample. This site-specific cluster of gene expression
can be constructed for several regions by searching in the
literature. In a study investigating gene expression in five
selected brain regions (including the amygdala, cerebellum,
hippocampus, olfactory bulb and periaqueductal gray), from a
total of 34,325 genes analyzed, 455 (1.3%) were enriched in
any one brain-region, relative to the other four. When more
stringent criteria for enrichment were adopted (namely the
selection of genes expressed six times higher in any given
region relative to the other four), on average 0.3% of genes
were highly enriched in each of the five regions, relative to the
others (Zirlinger et al., 2001). A similar trend was observed in
a study in which comparison of six different brain regions
(cortex, hippocampus, amygdala, entorhinal cortex, midbrain
and cerebellum) in two mouse strains revealed that out of
7,000 genes examined only 24 genes were expressed in all six
brain regions. Seventy-three genes were differentially ex-
pressed in at least one of the six brain regions between the two
strains, corresponding only to 1% of the genes examined
(Sandberg et al., 2000).
In summary, the efficacy of sophisticated genetic testing
methods is dependent upon the precision of tissue preparation
and the purity of the analyzed cell populations. The
heterogeneity of the nervous system tissues complicates this
testing as most brain nuclei contain a variety of cell populations
that are closely intermingled. Moreover, the anatomical
divisions between contiguous regions are not always clear
and the functional and anatomical areas may be very small.
Brain tissue dissection
The site-specific clustering of gene expression is one of the
reasons why there is still hope, despite the drawbacks outlined
above, to find distinct expression profiles when dissecting
specific brain areas (see also:Sandberg et al., 2000; Zirlinger
et al., 2001).
A number of techniques can be used for brain tissue
dissection. The choice depends on the level of precision the
researcher aims to achieve. A common approach for tissue
dissection, when large brain areas need to be sampled,
consists in the use of Brain Matrix Blocks. The advantage of
this method is velocity of sectioning and easiness of working in
RNAse-free conditions. When little brain areas need to be
sampled, an alternative to Brain Matrix Blocks is the use of a
cryostat. Here, brains previously stored at 7208C can be
sliced at variable thickness and at different angles (coronal,
sagittal, horizontal) and material can be punched out from
selected brain areas that are collected in RNAse-free tubes for
subsequent RNA extraction. To correctly find the regions,
one has to follow the landmarks described in brain atlases.
This method is time consuming and requires a lot of
experience. One of the advantages of this method, in addition
to the precision of isolating the correct target region, is the
integrity of the material and its low thickness, which allows
rapid and total homogenization of the sample in lysis solution
prior to RNA extraction.
These methods cannot be employed if the gene-expression
analysis is directed to the study of specific subsets of cells
located within a determined brain area. This level of accuracy
can be reached only if techniques such as Laser Capture
Microdissection (LCM) are used (Bonner et al., 1997;
Simone et al., 1998).
Less than 10 years ago, LCM technology was developed at
the National Institutes of Health (NIH). LCM was initially
developed within the cancer field (Emmert-Buck et al., 1996)
to facilitate segregation of benign and malignant cells and has
only recently been applied to the study of the CNS (Luo et al.,
1999). A commercial partnership has been established with
Arcturus Engineering (Mountain View, CA) that warranted
the successful introduction of a commercial instrument. LCM
is now well established as a tool for facilitating the enrichment
of cells under scrutiny from tissue sections, thus overcoming
the problem of tissue heterogeneity. It permits the selection
and capture of cells, cell aggregates, and discrete morpholo-
gical structures deriving from thin tissue sections. Cells are
visualized through a thermoplastic film, which is attached to
the bottom of an optically clear microfuge-tube cap. A laser
pulse, directed onto the target cells through the film, melts the
film and allows it to flow onto the targeted area where it cools
and binds with the underlying cell(s). The film including the
adhered cells or clusters is then lifted. Captured cells can then
be used for gene-expression profiling.
All the techniques described above are generally employed
to frozen brains. However, these techniques can also be used
for cutting fixed paraffin-embedded tissues (Goldsworthy et
Storage of brain tissue
High quality mRNA is relatively easy to obtain from frozen
tissue as starting material. But not always frozen tissue is
available. Sometimes, expression profiling is desired to be
performed in samples stored under a variety of other storage
conditions depending on researcher’s resources, possibilities
and interests. There may not be a low-temperature system
available for preserving brain specimens or there may only be
access to archived specimens as a source of RNA (Coombs et
al., 1999: Van Deerlin et al., 2002). In the drug addiction
field, biopsies collected from individuals with a history of drug
The Challenge of Preparing Brain Tissue Samples
dependence may be available. Similarly, fixed and paraffin-
embedded specimens often used for immunohistochemical
preparations are available from laboratory animals and in
many brains banks and pathology departments (Duyckaerts et
al., 1993, Hulette CM., 2003). The availability of these
samples, along with relative clinical histories and prognosis,
are attracting increasing attention for their use as potential
Generally, tissues can be snap-frozen with little or no risk of
RNA degradation. Brain tissue, conversely, heterogeneous as
it is, demands a different freezing procedure, aimed at
excluding alterations in the cellular morphology, which in
turn, could interfere negatively with subsequent tissue
dissection. Once the specimen has been obtained, the freezing
process begins with a few minutes of immersion in isopentane
cooled at 7408C. Subsequently, brains are stored at 7808C
Commercial RNA stabilization solutions for tissues such as
RNALater (Ambion, Inc.) are available. The RNA stabiliza-
tion solution is an aqueous reagent that permeates tissues
quickly, stabilizes and protects cellular RNA. The reagent
preserves RNA in tissues for up to 1 day at 370C, 1 week at
250C and 1 month or more at 40C (Florell et al., 2001). It
should be kept in mind that, whereas small organs such as rat
liver, kidney or spleen can be stored whole in RNA
stabilization solution, brain samples, on the other hand, are
generally too big to allow a rapid perfusion of the RNA
stabilization solution. Brain samples, therefore, should be
subdivided (e.g. 0.5 cm x 1 cm).
Fixed and paraffin-embedded tissue
Extraction of genomic DNA from formalin-fixed paraffin-
embedded tissue is well documented and it is now a routine in
diagnostic process (Mies et al., 1992) while extraction of RNA
from archived tissue is not trivial. RNA extraction from fixed
paraffin-embedded tissue strongly depends on the type of
fixative and on the experimental protocol used for fixation
(Greer et al., 1991; Foss et al., 1994). A number of different
fixatives are routinely used for histology and tissue preserva-
paraformaldehyde) function as chemical cross-linking agents
where formalin is the most commonly used fixative in routine
pathological practice, including brain tissue. By contrast,
simple organic coagulants such as ethanol, methanol or
acetone have precipitative effects. Karsten et al. (2002)
performed cDNA microarray experiments using RNA sam-
ples from frozen tissue and tissue fixed in ethanol or formalin.
The results revealed that RNA from frozen tissue yielded very
reproducible expression data while RNA obtained from fixed
tissue generated results characterized by higher variability and
lower correlation coefficients. Formalin-fixed tissue was more
severely affected than ethanol-fixed tissue in all conditions.
Furthermore, the time and the temperature of the fixation also
played an important role, especially if formalin is used as a
(such as formalinor
fixative agent (Foss et al., 1994; Van Deerlin et al., 2002). It
has been shown that with formalin, prolonged fixation time
(greater than 12–18 hours) results in a reduction of the
maximal size of amplifiable targets. Similarly, fixation under
high temperature conditions results in a reduction of RNA
integrity. In order to improve the yield of RNA from fixed or
paraffin-embedded tissue, a number of procedures have been
proposed (Koopmans et al., 1993; Krafft et al., 1997; Godfrey
et al., 2000; Specht et al., 2001). The most commonly used
method is the proteinase K digestion of the formalin fixed
paraffin embedded tissue (Masuda et al., 1999; Lewis et al.,
2001). This enzyme weakens the strong chemical bonds
formed during the fixation procedure, facilitating subsequent
RNA extraction. Today, dedicated kits for RNA extraction
from fixed tissue are commercially available (i.e., Paraffin
Block RNA Isolation Kit, Ambion).
RNA extraction is highly problematic from samples
preserved in strong cross-linking fixatives such as glutaralde-
hyde, modified formalins containing mercuric chloride, and
Bouin’s fixative (Van Deerlin et al., 2002).
Experimenters could take advantage from the existing
literature on RT-PCR application on fixed tissue. The results
of these studies can provide information on the level of RNA
degradation in fixed tissue (Greer et al., 1991; Jiang et al.,
1995; Godfrey et al., 2000).
Preparing to hybridize
Microarray analysis involves hybridizations of fluorescently
labeled targets, prepared from cellular RNA, to immobilized
DNA probes on a solid support (the chip). Several
methodologies are now routinely employed for labeling
targets and many of these systems are now supplied as
commercially available kits. In the classic experiment, targets
are prepared from RNA of a control sample and from RNA
extracted from treated cells or the tissue under investigation.
Reverse transcriptions are carried out separately for each
sample, one being labeled using cyanine 3 (Cy3) and the
other using cyanine 5 (Cy5). Both samples are then pooled
and hybridized together on the microarray. The relative
determined by a two channel detection of the fluorescence
in each spot. Traditionally, the targets have been labeled by
direct incorporation of Cy3 or Cy5-labeled deoxynucleotides
during first-strand cDNA synthesis (direct labeling). An
alternative is to label after cDNA synthesis by the amino-allyl
method (post-labeling). In this method a chemically reactive
nucleotide analog, an amino allyl-dUTP, is incorporated
during the cDNA synthesis and subsequently labeled with
monoreactive Cy3 or Cy5 dyes. An inherent disadvantage
with these methodologies is that they require relatively large
amounts of RNA. The amount of total RNA in a single cell
is estimated to be between 0.1 and 1 pg, which is difficult to
manipulate experimentally and orders of magnitude lower
than the minimum amount recommended by traditional
labeling methods. Typically 20–200 mg or greater are
required when using total RNA for monitoring gene
expression by these methods. The enrichment of 0.5 - 2 mg
poly(A)+ RNA is a prerequisite for direct labeling which
many investigators favor over indirect labeling techniques
(Zhou et al., 2002). Clearly these labeling methods are not
of each targetisthen
Soverchia et al.
readily applicable to small amounts of tissue (as with most
brain regions) or where specific cell types are investigated
and the RNA yields are poor.
Standard protocols for microarray hybridization require RNA
amounts corresponding to approximately 107cells or several
milligrams of tissue. RNA extracted from cells harvested by
needle biopsy, cell sorting, or LCM require an amplification
step. Two main approaches, signal amplification and global
poly(A)+ RNA amplification, have been developed to over-
come the hurdle of working with small tissue samples. Signal
amplification methodologies such as tyramide signal amplifi-
cation (TSA) (Karsten et al., 2002) and dendrimer technology
(Stears et al., 2000) function by increasing the fluorescence
signal emitted per labeled molecule. Global poly(A)+ RNA
amplification techniques are based either on exponential PCR
amplification (Lukyanov et al., 1997) or isothermal linear
RNA polymerase amplification (Van Gelder et al., 1990) to
increase the number of transcript equivalents and permit
sufficient labeling (Nygaard et al., 2003).
Many researchers with limited starting material employ the
classical T7 RNA Polymerase amplification method devel-
oped by Eberwine and coworkers (Van Gelder et al., 1990) or
variations of it. This method utilizes a synthetic oligo(dT)
primer annealed to a phage T7 RNA polymerase promoter to
prime synthesis of first strand cDNA by reverse transcription
of the poly(A)+ RNA pool of total RNA. Second strand
cDNA is synthesized with RNase H by degrading the RNA
strand, followed by second strand synthesis with E. coli DNA
polymerase I. Amplified antisense RNA (aRNA) is synthe-
sized via in vitro transcription of the double-stranded cDNA
(ds cDNA) template using T7 RNA polymerase. Many
commercial kits based on the Eberwine amplification techni-
que are in routine use today. Using this method, it is possible
to amplify the starting poly(A)+ RNA by up to 200-fold.
Linear RNA amplification is preferred over most of exponen-
tial amplification procedures since RNA Polymerase activity is
less influenced by template sequence or concentration than
Taq DNA polymerase. The correlation coefficient between
RNA based amplification and a non-amplified control was
found to be higher and less prone to bias than PCR based
amplification (Puskas et al., 2002). The efficacy of Eberwine
amplification protocols has been evaluated in separate studies
by Northern analysis (Eberwine et al., 1992), dot blot
differential screening (Poirier et al., 1997), comparison with
internal RNA standards, (Madison and Robinson, 1998),
hierarchical clustering analysis to compare consistency of
outlier genes upon amplification (Wang et al., 2000),
validation by quantitative RT-PCR (Puskas et al., 2002) and
comparisons of the ratio/intensity distribution of the total gene
set (Scheidl et al., 2002). Hu et al., (2002), Puska ´s et al.,
(2002) and Polacek et al., (2003) noted that additional genes
were detected in arrays hybridized with amplified material.
Absolute intensity levels were elevated on arrays hybridized
with labeled aRNA and as a consequence more genes were
noticed. Hu et al. (2002) and Polacek et al. (2003) confirmed
by other molecular techniques that these genes were in fact
expressed in the relevant cells and not the result of unspecific
binding or artefacts. This indicated that amplified RNA was
more sensitive to low abundance transcripts than the standard
method without amplification. Several studies have even
examined the degree to which the expression ratios were
preserved after 1–3 rounds of linear RNA amplification
(Xiang et al., 2003b).
RNA-based amplification although valuable require multi-
ple steps, which are labor-intensive and time-consuming.
PCR based amplification is relatively simpler but has the
stigma of biasing the abundance relationships. Nevertheless,
PCR based approaches have been employed to amplify RNA
transcripts. One approach utilizes terminal deoxynucleotide
transferase to append a homomeric tail to the 3’ end of the first
strand cDNA, followed by PCR between this primer and one
incorporated during the reverse transcription step at the 5’ end
of the cDNA. This method has the disadvantage that shorter
cDNAs are preferentially amplified and that non-specificity
arises from the use of homomeric oligo (dN) primers for PCR.
Another approach, three-prime-end amplification (TPEA),
enables global amplification of the 3’-end of all mRNAs
present in a sample (Dixon et al., 1998; Freeman et al., 1999).
PCR amplification is performed with a primer incorporated
into the first strand during reverse transcription and a second
primer used to initiate second strand synthesis. The second
strand primer is partially degenerate at the 3’-end, which
facilitates annealing every 1 Kb and results in uniformly sized
amplicons, thus promoting amplification of all mRNA species
regardless of the initial size of the transcript. Another study by
Iscove et al. (2002) noted that the exponential approach, if
through amplification as high as 3 x 1011-fold, in contrast to
common belief. The procedure involved reverse transcription
of first strand cDNA primed by oligo(dT), addition of an
oligo(dA) tail with terminal transferase and exponential
amplification with an oligo(dT) containing primer. This
procedure was independent of nucleic acid extraction or full-
length reverse transcription and the cycling and reaction
conditions were optimized to limit the extent of transcription
to only a few hundred bases of 3’ sequence. This global RT-
PCR protocol is applicable to 10 pg of RNA, or a single cell. It
is a one-tube procedure that can amplify microgram quantities
of aminoallyl-derived cDNA in less than one day. The
amplified cDNA then represents a permanent resource that
can then be reamplified as needed without compromising
abundance relationships. Furthermore, it reduces by a
million-fold the input amount of RNA needed for microarray
analysis, and yields reproducible results from the picogram
range of total RNA obtainable from single cells. (To obtain a
view of amplification methods at a glance see Table 1.)
An important aspect of RNA amplification is the degree of
reproducibility in microarray experiments. Zhao et al. (2002)
noted that when samples are amplified on the same day, the
correlations are significantly higher (0.97) compared to
samples amplified on different days (0.90). Moreover, the
amount of input total RNA affects the amplification process.
Amplified aRNA was generated from a range of total RNA
concentrations (0.25–3.0 mg). Decreasing the input RNA
within the range of 0.3–3 mg total RNA did not adversely
affect the fidelity and reproducibility of amplification. How-
The Challenge of Preparing Brain Tissue Samples
ever, when the starting total RNA was less than 300 ng, the
yield of aRNA was 53 mg and not sufficient even for one
hybridization experiment since the minimal amount of aRNA
required for microarray hybridizations is typically 10 mg (in
experiments). At starting amounts greater or
equal to 300 ng, they observed that the correlation coefficients
were approximately the same between amplified and unam-
plified samples and amongst amplified samples. Another
observation was that the fold of amplification is typically
greater with smaller starting quantities of template RNA, but
the absolute yield of aRNA is smaller.
Extended amplification times have traditionally been
employed with classical ‘Eberwine T7-Amplification’ as they
increased aRNA yield and generated greater quantities of
material for microarray hybridizations. The trend thus has
been to perform RNA synthesis for up to 20 hours. Spiess et
al. (2003) recently observed that prolonged incubation times
resulted in a decrease of aRNA quality, causing high-
background and low reproducibility on array hybridizations.
Smaller aRNA fragments were found to increase in frequency
with amplification reaction times greater than four hours. This
unexpected effect was noted irrespective of the dye which was
incorporated into the aRNA. The aRNA degradation was
directly correlated to the presence of RNA polymerase in the
amplification mixture and the degradation was observed with
de novo synthesis of aRNA, implying the presence of an
intrinsic T7 RNA Polymerase nucleolytic activity. In light of
this finding, aRNA amplification should not be carried out for
longer than four hours as the increased yield observed is at the
expense of quality. Furthermore, microarray hybridizations
should be carried out with aRNA amplified in an identical
manner and for the same amount of time in the whole set of
samples to be analyzed.
To date, the largest use of gene array technology has been in
the field of cancer research. However, it has now advanced to
the stage where it is also used in many other biomedical fields.
Particularly exciting is the possibility to use it to expand our
knowledge of the brain’s functions and to study complex
disorders of the central nervous system at the molecular level.
Considering the complexity of the brain, it is not surprising
that additional difficulties emerge in the study of such a
complex organ. In the present article, we have examined a
number of potential difficulties in using brain tissue for gene
array experiments. Amongst the major difficulties is the
preparation of good quality samples from which to obtain
the maximum amount of high quality RNA. Several factors
can interfere with this process. The cellular and structural
Table 1. Target amplification methods at a glance.
Method DescriptionAdvantage Disadvantage
Cy3 or Cy5-labeled
deoxynucleotides into first-strand
cDNA during synthesis.
Fast, cost effective. Require relatively large amounts of
cellular RNA. Potential dye Bias.
Chemically reactive nucleotide
analog, an amino allyl-dUTP, is
incorporated into the cDNA. This
is then subsequently labeled with
monoreactive Cy3 or Cy5 dyes.
Requires less input RNA than
Require relatively large amounts of
Tyramide signal amplification
method that utilizes the catalytic
activity of horseradish peroxidase
(HRP) to generate high-density
labeling of a target protein or
nucleic acid sequence in situ.
Not dependent on transcript
sequence. Increase the sensitivity
by up to 100-fold.
Potential bias introduced.
Dendrimer technology A 3DNA dendrimer is a signal
amplification molecule made
Not sequence dependent, fast.
200-fold passive enhancement of
Potential bias introduced.
Eberwine amplification Oligo(dT) primed first strand
cDNA synthesis. Second strand
cDNA synthesis and in vitro
transcription of the double-
stranded cDNA into cRNA.
Level of bias introduced into gene
expression is relatively low.
Amplified RNA is more sensitive to
low abundance transcripts than
Prolonged incubation times
resulted in a decrease in aRNA
quality. Dependent on nucleic acid
extractions, reverse transcription
and purification steps all of which
may introduce bias. Labor intensive
and time consuming. Abundance
information from smaller starting
amounts not preserved.
PCR based amplification
PCR based amplification of
oligo(dT) primed first strand
cDNA synthesis template.
Fast, no need for purification steps. Potential bias in the transcript
Soverchia et al.
heterogeneity of the brain tissue is a major concern. This has
consequences also on the amount of starting material from
which RNA can be extracted. Ideally, the experimenter should
be able to isolate and collect RNA from single cells. For this
purpose, new dissection techniques (such as LCM) and novel
reliable RNA extraction and amplification methods have been
developed. Another important consideration is the preserva-
tion of tissue. Not always frozen tissue, which is the best
material for RNA extraction, is available. Tissues of high
scientific interests archived in organ banks are often stored
fixed and/or paraffin-embedded which makes RNA extraction
more problematic. Recent studies demonstrated, however,
that under particular circumstances and using appropriate
protocols, recovery of RNA from these tissues is still possible.
Lastly, to enhance the sensitivity of the technology, several
protocols have been developed to increase the signal of targets
especially in samples with low amounts of starting material.
In summary, it can be concluded that this part of expression
profiling, using chips or other techniques, from setting up an
experiment through appropriate tissue selection and RNA
preparation until reliable labeling of targets and hybridization
encompasses a long way to go, and each step is exposed to
numerous environmental (technical) ‘‘noise’’ generating
factors that may more or less seriously obscure output data
or even negate the whole experiment.
Asanuma M, Ogawa N (1994) Pitfalls in assessment of c-fos mRNA
expression in the brain: effects of animal handling. Rev Neurosci
Bahn S, Augood SJ, Ryan M, Standaert DG, Starkey M, Emson PC
(2001) Gene expression profiling in the post-mortem human
brain – no cause for dismay. J Chem Neuroanat 22:79–94.
Barton AJ, Pearson RC, Najlerahim A, Harrison PJ (1993) Pre- and
postmortem influences on brain RNA. J Neurochem 61:1–11.
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 (2002) Nuclei and subnuclei gene expression
profiling in mammalian brain. Brain Res 943:38–47.
Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R,
Goldstein S, Liotta LA (1997) Laser capture microdissection:
molecular analysis of tissue. Science 278:1481–1483.
Broad KD, Hinton MR, Keverne EB, Kendrick KM (2002)
Involvement of the medial prefrontal cortex in mediating
behavioural responses to odour cues rather than olfactory
recognition memory. Neuroscience 114:715–729.
Brown PO, Botstein D (1999) Exploring the new world of the genome
with DNA microarrays. Nat Genet 21:33–37.
Chee M, Yang R, Hubbell E, Berno A, Huang XC, Stern D, Winkler
J, Lockhart DJ, Morris MS, Fodor SPA (1996) Accessing Genetic
Information with high-density DNA arrays. Science 274:610–
Coombs NJ, Gough AC, Primrose JN (1999) Optimisation of DNA
and RNA extraction from archival formalin-fixed tissue. Nucleic
Acids Res 27:e12.
Cowan WM, Kopnisky KL, Hyman SE (2002) The human genome
project and its impact on psychiatry. Annu Rev Neurosci 25:1–
Cummings TJ, Strum JC, Yoon LW, Szymanski MH, Hulette CM
(2001) Recovery and expression of messenger RNA from
postmortem human brain tissue. Mod Pathol 14:1157–1161.
Dixon AK, Richardson PJ, Lee K, Carter NP, Freeman TC (1998)
Expression profiling of single cells using 3 prime end amplifica-
tion (TPEA) PCR. Nucleic Acids Res 26:4426–4431.
Expression Analysis Technical Manual.
Duyckaerts C, Sazdovitch V, Seilhean D, Delaere P, Hauw JJ (1993)
A brain bank in a neuropathology laboratory (with some emphasis
on diagnostic criteria). J Neural Transm 39:107–118.
Eberwine J, Yeh H, Miyashiro K, Cao Y, Nair S, Finnell R, Zettel M,
Coleman P (1992) Analysis of gene expression in single live
neurons. Proc Natl Acad Sci USA 89:3010–3014.
Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z,
Goldstein SR, Weiss RA, Liotta LA (1996) Laser capture
microdissection. Science 274:998–1001.
Florell SR, Coffin CM, Holden JA, Zimmermann JW, Gerwels JW,
Summers BK, Jones DA, Leachman SA (2001) Preservation of
RNA for functional genomic studies: a multidisciplinary tumor
bank protocol. Mod Pathol 14:116–128.
Foss RD, Guha-Thakurta N, Conran RM, Gutman P (1994) Effects
of fixative and fixation time on the extraction and polymerase
chain reaction amplification of RNA from paraffin-embedded
tissue. Comparison of two housekeeping gene mRNA controls.
Diagn Mol Pathol 3:148–155.
Freeman TC, Lee K, Richardson PJ. (1999) Analysis of gene
expression in single cells. Curr Opin Biotechnol 10:579–582.
Gaasterland T Bekiranov S (2000) Making the most of microarray
data. Nature Genetics 24:204–206.
Geschwind DH (2000) Mice, microarrays, and the gentic diversity of
the brain. Proc Natl Acad Sci USA 97:10676–10678.
Godfrey TE, Kim SH, Chavira M, Ruff DW, Warren RS, Gray JW,
Jensen RH (2000) Quantitative mRNA expression analysis from
formalin-fixed, paraffin-embedded tissues using 5’ nuclease
quantitative reverse transcription-polymerase chain reaction. J
Mol Diagn 2(2):84–91.
Gold D, Coombes K, Medhane D, Ramaswamy A, Ju Z, Strong L,
Koo JS, Kapoor M (2004) A comparative analysis of data
generated using two different target preparation methods for
hybridization to high-density oligonucleotide microarrays. BMC
Goldsworthy SM, Stockton PS, Trempus CS, Foley JF, Maronpot RR
(1999) Effects of fixation on RNA extraction and amplification
from laser capture microdissected tissue. Mol Carcinog 25:86–
Greer CE, Lund JK, Manos MM (1991) PCR amplification from
paraffin-embedded tissues: recommendations on fixatives for
long-term storage and prospective studies. PCR Methods Appl
Hardiman G (2003) Microarray technologies 2003 - an overview.
Harrison PJ, Procter AW, Barton AJ, Lowe SL, Najlerahim A,
Bertolucci PH, Bowen DM, Pearson RC (1991) Terminal coma
affects messenger RNA detection in post mortem human
temporal cortex. Brain Res Mol Brain Res 9:161–164.
Harrison PJ, Heath PR, Eastwood SL, Burnet PW, McDonald B,
Pearson RC (1995) The relative importance of premortem
acidosis and postmortem interval for human brain gene expres-
sion studies: selective mRNA vulnerability and comparison with
their encoded proteins. Neurosci Lett 200:151–154.
Herdegen T, Leah JD (1998) Inducible and constitutive transcription
factors in the mammalian nervous system: control of gene
expression by Jun, Fos and Krox, and CREB/ATF proteins.
Brain Res Brain Res Rev 28:370–490.
Hughes TR, Mao M, Jones AR, Burchard J, Marton MJ, Shannon
KW, Lefkowitz SM, Ziman M, Schelter JM, Meyer MR,
Kobayashi S, Davis C, Dai H, He YD, Stephaniants SB, Cavet
G, Walker WL, West A, Coffey E, Shoemaker DD, Stoughton R,
Blanchard AP, Friend SH, Linsley PS (2001) Expression
profiling using microarrays fabricated by an ink-jet oligonucleo-
tide synthesizer. Nat Biotechnol 19(4):342–347.
Hu L, Wang J, Baggerly K, Wang H, Fuller GN, Hamilton SR,
Coombes KR, Zhang W (2002) Obtaining reliable information
from minute amounts of RNA using cDNA microarrays. BMC
Hulette CM (2003) Brain banking in the United States. J Neuropathol
Exp Neurol 62:715–722.
Hynd MR, Lewohl JM, Scott HL, Dodd PR (2003) Biochemical and
molecular studies using human autopsy brain tissue. J Neuro-
The Challenge of Preparing Brain Tissue Samples
Jiang YH, Davidson LA, Lupton JR, Chapkin RS (1995) A rapid RT-
PCR method for detection of intact RNA in formalin-fixed
paraffin-embedded tissues. Nucleic Acids Res 23:3071–3072.
Jiang CH, Tsien JZ, Schultz PG, Hu Y (2001) The effects of aging on
gene expression in the hypothalamus and cortex of mice. Proc
Natl Acad Sci USA 98:1930–1934.
Johnson SA, Morgan DG, Finch CE (1986) Extensive postmortem
stability of RNA from rat and human brain. J Neurosci Res
Karsten SL, Van Deerlin VM, Sabatti C, Gill LH, Geschwind DH
(2002) An evaluation of tyramide signal amplification and
archived fixed and frozen tissue in microarray gene expression
analysis. Nucleic Acids Res 30:E4.
Keay KA, Bandler R (2001) Parallel circuits mediating distinct
emotional coping reactions to different types of stress. Neurosci
Biobehav Rev 25:669–678.
Kerr M.K and Churchill G.A (2001) Experimental design for gene
expression microarrays. Biostatistics 2:183–201.
Kingsbury AE, Foster OJ, Nisbet AP, Cairns N, Bray L, Eve DJ, Lees
AJ, Marsden CD (1995) Tissue pH as an indicator of mRNA
preservation in human post-mortem brain. Brain Res Mol Brain
Koopmans M, Monroe SS, Coffield LM, Zaki SR. (1993) Optimiza-
tion of extraction and PCR amplification of RNA extracts from
paraffin-embedded tissue in different fixatives. J Virol Methods
Krafft AE, Duncan BW, Bijwaard KE, Taubenberger JK, Lichy JH
(1997) Optimization of the isolation and amplification of RNA
from formalin-fixed, paraffin-embedded tissue: The Armed
Forces Institute of Pathology Experience and Literature Review.
Mol Diagn 2:217–230.
Leonard S, Logel J, Luthman D, Casanova M, Kirch D, Freedman R
(1993) Biological stability of mRNA isolated from human
postmortem brain collections. Biol Psychiatry 33:456–466.
Lewis F, Maughan NJ, Smith V, Hillan K, Quirke P (2001) Unlocking
the archive - gene expression in paraffin-embedded tissue. J
Lukyanov K, Diatchenko L, Chenchik A, Nanisetti A, Siebert P,
Usman N, Matz M, Lukyanov S (1997) Construction of cDNA
libraries from small amounts of total RNA using the suppression
PCR effect. Biochem Biophys Res Commun 230:285–288.
Luo L, Salunga RC, Guo H, Bittner A, Joy KC, Galindo JE, Xiao H,
Rogers KE, Wan JS, Jackson MR, Erlander MG (1999) Gene
expression profiles of laser-captured adjacent neuronal subtypes.
Nat Med 5:117–122.
Madison RD, Robinson GA (1998) lambda RNA internal standards
quantify sensitivity and amplification efficiency of mammalian
gene expression profiling. Biotechniques 25:504–508.
Masuda N, Ohnishi T, Kawamoto S, Monden M, Okubo K (1999)
Analysis of chemical modification of RNA from formalin-fixed
samples and optimization of molecular biology applications for
such samples. Nucleic Acids Res 27:4436–4443.
Mies C (1992) Molecular pathology of paraffin-embedded tissue.
Current Clinical Applications. Diagn Mol Pathol 1:206–211.
Mori M, Mimori K, Yoshikawa Y, Shibuta K, Utsunomiya T,
Sadanaga N, Tanaka F, Matsuyama A, Inoue H, Sugimachi K
(2002) Analysis of the gene-expression profile regarding the
progression of human gastric carcinoma. Surgery 131:S39–47.
Miura K, Bowman ED, Simon R, Peng AC, Robles AI, Jones RT,
Katagiri T, He P, Mizukami H, Charboneau L, Kikuchi T, Liotta
LA, Nakamura Y, Harris CC (2002) Laser capture microdissec-
tion and microarray expression analysis of lung adenocarcinoma
reveals tobacco smoking- and prognosis-related molecular pro-
files. Cancer Res 62:3244–3250.
Nygaard V, Loland A, Holden M, Langaas M, Rue H, Liu F,
Myklebost O, Fodstad O, Hovig E, Smith-Sorensen B (2003)
Effects of mRNA amplification on gene expression ratios in
cDNA experiments estimated by analysis of variance. BMC
Poirier GM, Pyati J, Wan JS, Erlander MG (1997) Screening
differentially expressed cDNA clones obtained by differential
display using amplified RNA. Nucleic Acids Res 25:913–914.
Polacek DC, Passerini AG, Shi C, Francesco NM, Manduchi E,
Grant GR, Powell S, Bischof H, Winkler H, Stoeckert CJ Jr,
Davies PF (2003) Fidelity and enhanced sensitivity of differential
transcription profiles following linear amplification of nanogram
amounts of endothelial mRNA. Physiol Genomics 13:147–156.
Puska ´s LG, Zvara A, Hackler L Jr, Van Hummelen P (2002) RNA
amplification results in reproducible microarray data with slight
ratio bias. Biotechniques 32:1330–1334.
Preece P, Virley DJ, Costandi M, Coombes R, Moss SJ, Mudge AW,
Jazin E, Cairns NJ (2003) An optimistic view for quantifying
mRNA in post-mortem human brain. Brain Res Mol Brain Res
Ramakrishnan R, Dorris D, Lublinsky A, Nguyen A, Domanus M,
Prokhorova A, Gieser L, Touma E, Lockner R, Tata M, Zhu X,
Patterson M, Shippy R, Sendera TJ, Mazumder A (2002) An
assessment of Motorola CodeLink(TM) microarray performance
for gene expression profiling applications. Nucleic Acids Res
Sandberg R, Yasuda R, Pankratz DG, Carter TA, Del Rio JA,
Wodicka L, Mayford M, Lockhart DJ, Barlow C (2000) Regional
and strain-specific gene expression mapping in the adult mouse
brain. Proc Natl Acad Sci USA 97:11038–11043.
Scheidl SJ, Nilsson S, Kalen M, Hellstrom M, Takemoto M,
Hakansson J, Lindahl P (2002) mRNA expression profiling of
laser microbeam microdissected cells from slender embryonic
structures. Am J Pathol 160:801–813
Schramm M, Falkai P, Tepest R, Schneider-Axmann T, Przkora R,
Waha A, Pietsch T, Bonte W, Bayer TA (1999) Stability of RNA
transcripts in post-mortem psychiatric brains. J Neural Transm
Shyu AB, Greenberg ME, Belasco JG (1989) The c-fos transcript is
targeted for rapid decay by two distinct mRNA degradation
pathways. Genes Dev 3:60–72.
Simone NL, Bonner RF, Gillespie JW, Emmert-Buck MR, Liotta LA
(1998) Laser-capture microdissection: opening the microscopic
frontier to molecular analysis. Trends Genet 14:272–276.
Sokolov BP, Polesskaya OO, Uhl GR (2003) Mouse brain gene
expression changes after acute and chronic amphetamine. J
Sotiriou C, Khanna C, Jazaeri AA, Petersen D, Liu ET (2002) Core
biopsies can be used to distinguish differences in expression
profiling by cDNA microarrays. J Mol Diagn 1:30–36.
Spiess AN, Mueller N, Ivell R (2003) Amplified RNA degradation in
T7-amplification methods results in biased microarray hybridiza-
tions. BMC Genomics 4:44.
Specht K, Richter T, Muller U, Walch A, Werner M, Hofler H (2001)
Quantitative gene expression analysis in microdissected archival
formalin-fixed and paraffin-embedded tumor tissue. Am J Pathol
Stears RL, Getts RC, Gullans SR (2000) A novel, sensitive detection
system for high-density microarrays using dendrimer technology.
Physiol Genomics 3:93–99.
Van Deerlin VM, Gill LH, Nelson PT (2002) Optimizing gene
expression analysis in archival brain tissue. Neurochem Res
Van Gelder RN, von Zastrow ME, Yool A, Dement WC, Barchas JD,
Eberwine JH (1990) Amplified RNA synthesized from limited
quantities of heterogeneous cDNA. Proc Natl Acad Sci USA
Wang E, Miller LD, Ohnmacht GA, Liu ET, Marincola FM (2000)
High-fidelity mRNA amplification for gene profiling. Nat
Wurmbach E, Gonzales-Maeso J, Yuen T, Ebersole BJ, Mastaitis JW,
Mobbs CV, Sealfon SC (2002) Validated genomic approach to
study differentially expressed genes in complex tissues. Neuro-
chem Res 27:1027–1033.
Xiang CC, Chen M, Kozhich OA, Phan QN, Inman JM, Chen Y,
Brownstein MJ (2003a) Probe generation directly from small
numbers of cells for DNA microarray studies. Biotechniques
34(2):386–388, 390, 392–393.
Soverchia et al.
Xiang CC, Chen M, Ma L, Phan QN, Inman JM, Kozhich OA, Download full-text
Brownstein MJ (2003b) A new strategy to amplify degraded RNA
from small tissue samples for microarray studies. Nucleic Acids
Yasojima K, McGeer EG, McGeer PL. Yasojima K, McGeer EG,
McGeer PL (2001) High stability of mRNAs postmortem and
protocols for their assessment by RT-PCR. Brain Res Brain Res
Zhao H, Hastie T, Whitfield ML, Borresen-Dale AL, Jeffrey SS
(2002) Optimization and evaluation of T7 based RNA linear
amplification protocols for cDNA microarray analysis. BMC
Zirlinger M, Kreiman G, Anderson DJ (2001) Amygdala-enriched
genes identified by microarray technology are restricted to specific
amygdaloid subnuclei. Proc Natl Acad Sci USA 98:5270–5275.
The Challenge of Preparing Brain Tissue Samples