Biomedical Engineering Reference
In-Depth Information
Sensitivity increases can be gained via one of two basic approaches: amplification of the
amount of signal generated on the array or amplification of the amount of starting material
(Table 5.2).
Signal amplification methodologies work by bringing additional detection moieties to the
hybridization site, either through the use of branched dendrimeric molecules or through enzy-
matic deposition of fluorophores (Tyramide Signal Amplification (TSA); MICROMAX
™
TSA
, PE) [13]). The advantages of these techniques lie in the fact that the original RNA
sample does not need to be unduly manipulated, and as such have the potential for less bias
due to enzymatic manipulation. In addition, signal amplification techniques tend to be fairly
rapid, taking on the order of 1-3 h total. Despite these advantages, signal amplification
technologies still are limited by the amount of RNA present in the initial sample (e.g. the
dendrimers can only bind to the cDNA generated from the RNA, and so the fewer cDNA
molecules there are, then so too the fewer dendrimers). As such, their ability to lower sample
requirements tends to be limited to around one order of magnitude.
The greatest improvements in sample reduction are realized through sample (RNA) ampli-
fication techniques. RNA amplification methodologies (whether they produce RNA or DNA
copies of the original mRNA) tend to fall into three categories: (i) linear amplification
methodologies that use T7-mediated transcription (Figure 5.1), (ii) NuGen RiboSPIA
™
™
amplification processes (Figure 5.2) and (iii) exponential amplification processes that incor-
porate polymerase chain reaction (PCR) methodologies (Figure 5.3) [14-16].
One of the first methodologies applied to the amplification of RNA for microarrays
was the T7-mediated approaches (also known as IVT) first described by Eberwine and
co-workers and others [7, 9, 17]. This technique has been used extensively in microarray
research and is the
de facto
standard for most of the Affymetrix, Agilent and Illumina gene
expression arrays. T7-based amplification has several inherent advantages, particularly in
that it is has been shown to be reproducible [17-21], is theoretically linear in nature and is
not affected by template sequence [16]. There have also been several modifications of the
IVT protocol that aim to improve sensitivity, reliability or functionality of the technique [22,
23]. The major drawbacks of the IVT method are that it is time consuming (a single round of
amplification takes 1.5-2 days) and there is a limit to its sensitivity. An additional concern
is that T7/IVT produces RNA-based products for profiling which are inherently less stable.
Although the initial descriptions of the T7 amplification technique used single neurons
to assay gene expression, this was not a global analysis [7]. Several groups have indicated
that a single round of T7-based amplification will provide between 1000- and 2000-fold
increases in RNA levels [15, 16, 24]. Multiple rounds of amplification can be performed,
but usually each subsequent round leads to increased bias and noise [17, 24, 25]. Many
groups have reported that, when working with less than 10 ng of total RNA, the results
of microarray profiling tend to be less reliable [17, 26, 27]. As such, microarray profiling
of single cells using T7-based amplification strategies has not been extensively conducted;
however, one report showed that it was possible with four rounds of amplification [28].
While T7/IVT amplification strategies have undoubtedly been the most popular to date for
microarray-based gene expression profiling, scientists looking to profile single-cell expres-
sion patterns have more often turned toward PCR-based amplification strategies. It has been
demonstrated that global cDNA libraries and gene expression profiles could be obtained
using PCR-based approaches [29-33]. It has also been shown that these methods can
profile RNA samples down to the low picogram range [34]. Since these initial reports,
several single-cell microarray profiles have been obtained using these or other PCR-based
approaches [6, 35, 36]. PCR-based methodologies afford many advantages: they are quick,
