Review Article
Improved microRNA quantification in total RNA from clinical samples
Ditte Andreasen 1
, Jacob U. Fog 1
, William Biggs, Jesper Salomon, Ina K. Dahslveen *, Adam Baker,
Peter Mouritzen
Exiqon A/S, Skelstedet 16, DK-2950 Vedbaek, Denmark
a r t i c l e i n f o
Keywords:
microRNA
qPCR
Locked Nucleic Acid
Biomarker
Plasma
FFPE
a b s t r a c t
microRNAs are small regulatory RNAs that are currently emerging as new biomarkers for cancer and
other diseases. In order for biomarkers to be useful in clinical settings, they should be accurately and reliably
detected in clinical samples such as formalin fixed paraffin embedded (FFPE) sections and blood
serum or plasma. These types of samples represent a challenge in terms of microRNA quantification. A
newly developed method for microRNA qPCR using Locked Nucleic Acid (LNA™)-enhanced primers
enables accurate and reproducible quantification of microRNAs in scarce clinical samples. Here we show
that LNA™-based microRNA qPCR enables biomarker screening using very low amounts of total RNA
from FFPE samples and the results are compared to microarray analysis data. We also present evidence
that the addition of a small carrier RNA prior to total RNA extraction, improves microRNA quantification
in blood plasma and laser capture microdissected (LCM) sections of FFPE samples.
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
1.1. microRNAs are promising biomarkers
microRNAs comprise a family of highly conserved small noncoding
RNAs ($22 nt). As regulators of post-transcriptional gene
expression, microRNAs play an essential role in a large number
of biological and pathological processes. It is now established that
altered microRNA expression profiles are associated with a number
of different diseases including heart disease, neurological disorders
and human cancers [1–3]. This has suggested the use of microRNAs
as biomarkers for disease diagnosis and prognosis. A number of recent
reports have confirmed the usefulness of microRNAs as biomarkers
in cancer [4,5].
As microRNAs are more stable than mRNAs, they are good candidates
for use as biomarkers [6]. However, due to their short
length and the high sequence similarity within microRNA families,
reliable and accurate quantification is still a challenge. A valuable
biomarker requires robust and reproducible assays that work in
clinically available samples as well as archived material. With
the advent of more complete screening protocols where precious
clinical samples are being analyzed for a variety of biomarkers,
the importance of reduced sample requirements is increasing, generating
a need for extremely sensitive assays.
The most commonly available archived material often used for
biomarker discovery and validation is FFPE tissue [7]. However, total
RNA isolated from FFPE material represents a challenge due the
fact that formalin fixation causes cross links and degradation that
result in isolation of short degraded RNA fragments. Formalin also
causes chemical modifications that negatively affect qPCR efficiency
[8]. This results in poor RNA yields and inferior qPCR performance.
In addition, variation between separate RNA extractions
from FFPE samples is significant.
Blood serum and plasma also represent important sample types
for investigating microRNAs as biomarkers [9,10]. The nucleic acids
present in carefully prepared serum and plasma are presumed to
be extracellular. The level of microRNAs in such samples is very
low and efficient and reproducible recovery of this RNA is problematic.
In addition, RNA purified from plasma can also contain
inhibitors that affect qPCR efficiency. Taken together, these issues
generate a great need for reproducible RNA isolation and miRNA
quantification methods that are extremely sensitive and robust.
1.2. A unique method for microRNA quantification
Several different methods for real-time quantification of
microRNAs have been developed attempting to achieve the optimal
combination of sensitivity and specificity [11]. Gene specific
reverse transcription (RT) can add sensitivity and specificity, but
leads to increased sample requirement and complicated experimental
procedures. A universal RT step allows hundreds of targets
to be quantified from the same cDNA synthesis, leading to dramatically
decreased sample requirements and increased ease of use.
However, microRNA amplification using DNA-based primers leads
to restricted sensitivity, especially for AT-rich targets as well as difficulties
with single nucleotide mismatches.
1046-2023/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.ymeth.2010.01.006
* Corresponding author. Fax: +45 45661888.
E-mail address: ina@exiqon.com (I.K. Dahslveen).
1
These authors contributed equally.
Methods 50 (2010) S6–S9
Contents lists available at ScienceDirect
Methods
journal homepage: www.elsevier.com/locate/ymeth
We have developed a new method for microRNA qPCR which
starts with a cDNA synthesis reaction incorporating poly-adenylation
and reverse transcription in a single step followed by microRNA
amplification using two microRNA-specific, LNA™-enhanced
primers. The inclusion of LNA™ in the PCR primers allows the design
of a short, yet highly specific forward primer covering most
of the microRNA sequence. The reverse primer, also enhanced with
LNA™, covers the 30
end of the microRNA sequence. This unique
combination of two microRNA-specific primers means that the system
retains specificity at the same time as being extremely sensitive
due to the combination of Universal RT and LNA™-enhanced
PCR.
Due to the extreme sensitivity and robustness, the new system
for microRNA qPCR is especially useful for quantification of
microRNAs in challenging samples. Here, we report how LNA™-enhanced
microRNA qPCR can be used to reliably detect microRNAs
in difficult samples. Ready-to-use PCR panels of 730 human
microRNAs enables comprehensive screening for biomarkers in
substantially lower amounts of FFPE derived sample than can be
achieved using microarray analysis. In addition, we report how
microRNA detection can be improved by addition of a small RNA
carrier prior to total RNA extraction.
2. Experimental methods
2.1. Preparation of total RNA
Total RNA from plasma was purified using the Qiagen miRNAeasy
mini kit with minor modifications according to Wang et al.
[9]. Carrier RNA (MS2 RNA, Roche) was added at 1 lg per 750 ll
Qiazol reagent (before mixing with plasma samples). Mock purifications
without plasma (but including MS2) were also performed.
Total RNA from FFPE sections including laser capture microdissected
(LCM) samples were purified using the Qiagen miRNAeasy
FFPE kit (50) (PN 217404, Qiagen) according to the manufacturer’s
instructions. One microgram of MS2 RNA (Roche) was added to
each sample when indicated. A blank purification (with MS2)
was included in every batch.
2.2. Microarray analysis
Microarray analysis was performed using miRCURY LNA™
microRNA Arrays according to the manufacturer’s instructions
(Exiqon). Duplicate microarrays were run using 500 ng total RNA
from tumor and normal samples labeled with Hy3, against a universal
reference labeled with Hy5. Signals were normalized using
global Lowess and results were converted to relative values
(log 2 ratios) between tumor and normal sample.
2.3. microRNA qPCR
Real-time PCR for microRNAs was performed using the miRCURY
LNA™ Universal RT microRNA PCR system (Exiqon) according to
the manufacturer’s instructions with minor modifications (see below).
For biomarker screening, Ready-to-use microRNA PCR Human
Panel I and II were used (Product number 203602)
according to the instruction manual (Exiqon). For each RT reaction
23 ng total RNA was used. Three RT replicates per sample were
used for real-time amplification on ready-to-use PCR plates run
on a Roche LightCyclerÒ
480 real-time PCR system. Average Cq values
were normalized to three stably expressed reference genes
using the Exiqon GenEx software. Results were converted to relative
values (log 2 ratios) between tumor and normal sample.
For individual microRNA assays, primer sets with the following
product numbers were used: 204099, 204326, 204063, and 204230
(Exiqon). For qPCR on total RNA from plasma, 2 ll of eluted RNA
(equivalent to 8 ll original plasma sample) was used in a 10 ll
RT reaction. The resulting cDNA was diluted 50Â and 5 ll used
in 10 ll PCR amplification reactions run on a Roche LightCycler
480 real-time PCR system. For qPCR on total RNA from LCM samples,
2 ll of the 30 ll purification was used in 10 ll RT reactions
as described above. Negative controls included mock purifications
and no template controls.
2.4. Data analysis
Comparison between microarray and qPCR data was performed
as described by the MAQC Consortium [12].
3. Results and discussion
3.1. microRNA biomarker screening in FFPE samples using qPCR
compared to microarray
The first step in biomarker discovery is to screen for differentially
expressed microRNAs in normal and disease tissue (or tissue
from different stages of disease). When sufficient sample is available,
initial screening using microarrays provides a comprehensive,
fast and affordable method for biomarker discovery. However,
microarrays are limited to a larger minimum sample amount compared
to more sensitive techniques such as real-time PCR. We have
compared the use of miRCURY LNA™ microRNA Arrays and LNA™enhanced
qPCR for detection of differences between tumor and
normal tissue samples.
For expression analysis using microarrays, 500 ng total RNA
from normal and tumor FFPE tissue was used to screen a total of
about 1300 human microRNAs. On the arrays, 16 microRNAs were
shown to be more than 2-fold differentially expressed between tumor
and normal samples. LNA enhanced microRNA qPCR enables
similar expression analysis of 730 human microRNAs in two
ready-to-use PCR plates from only 23 ng total RNA from normal
and tumor FFPE tissue. Out of the 730 microRNAs, 68 were shown
to be more than 2-fold differentially expressed between tumor and
normal samples.
A comparison of 220 microRNAs with robust detection (signals
below 36 Cq values) in qPCR and corresponding probes on the
microarray shows a very good correlation between the two platforms
(Fig. 1). Compared to a perfect fit between array and qPCR
data, the correlation shows that qPCR enables larger differences
to be detected due to a larger dynamic range. The results show that
the new LNA™-enhanced microRNA qPCR method allows biomarker
screening in clinical samples using 20Â less starting material
and providing more extensive results compared to microarray
screening.
3.2. Improved quantification of microRNAs in plasma and LCM sections
The small size of microRNAs makes reproducible recovery from
very small amounts of challenging samples difficult. In order to improve
microRNA recovery from clinical samples, we implemented
the addition of a small carrier RNA prior to total RNA extraction.
To test the effects of the carrier in blood plasma we quantified
three microRNAs known to be expressed at different levels, hsamiR-192,
hsa-let-7a and hsa-miR-103. The addition of carrier
means that the RNA in the samples cannot be accurately quantified,
however, the same amount of total RNA equivalent to 8 ll
of plasma was used in each RT reaction. This was considered a reasonable
approach as all samples were handled in the same way
from sampling of blood to further processing and RNA preparation.
The improvement in microRNA quantification after addition of
D. Andreasen et al. / Methods 50 (2010) S6–S9 S7
RNA carrier to plasma samples is shown by an average decrease of
1–2 Cq values compared to total RNA isolated without carrier
(Fig. 2A). The positive effect of the carrier was slightly higher for
microRNAs with lower expression levels. In addition to lower Cq
values, the addition of carrier also improved reproducibility between
separate RNA extractions from an average standard deviation
of 1.18 Cq values between different extractions without
carrier RNA to an average standard deviation of 0.21 Cq values between
different extractions with carrier RNA. The improvement in
microRNA detection is likely due to improved microRNA recovery
in the presence of carrier, but additional beneficial effects of the
presence of carrier cannot be ruled out. A mock extraction using
the same amount of carrier RNA did not generate any signal for
any of the three microRNAs tested (data not shown).
The effects of addition of carrier during RNA extraction was
also tested on LCM sections from FFPE samples. Areas ranging
from 0.04 mm2
to 0.64 mm2
(equivalent to approximately 25–
400 cells) were microdissected from two FFPE sections, and in
each case, two independent LCM dissections were made. Total
RNA was isolated with and without carrier RNA. Three microRNAs
(hsa-miR-21, hsa-miR-103, has-let-7a) were quantified using
LNA™-enhanced qPCR in each sample. In order to investigate
the correlation between the area dissected and microRNA detection,
the Cq values obtained from the various areas were compared
to the Cq values from the lowest amount of input
material (0.04 mm2
). The average delta Cqs with and without
the addition of carrier RNA are shown in Fig. 2B. The results show
that the addition of carrier RNA prior to RNA extraction improves
microRNA quantification from very small amounts of LCM starting
material. The variation in microRNA quantification between separate
RNA isolations is large in the absence of carrier. The addition
of carrier leads to reduced variation between separate extractions.
A mock isolation using the equivalent amounts of carrier, but no
sample gave no signal for any of the microRNAs tested (data not
shown).
The new microRNA qPCR method using Universal RT and
LNA™-enhanced primers allows microRNA quantification in
minute amounts of challenging clinical samples such as blood plasma
and LCM sections from FFPE material. Our results suggest that
the addition of an RNA carrier prior to total RNA extraction improves
the quantification of microRNAs in challenging samples.
Inclusion of carrier probably improves recovery of microRNAs from
the samples, but might also have additional beneficial effects on
microRNA quantification, such as improved stability or availability
in extracted RNA. The fact that the addition of carrier also seems to
improve reproducibility between separate extractions is particularly
interesting when considering biomarker screening and validation
in large numbers of samples collected and isolated at
different time points. Furthermore, the low yield of RNA in such
samples may prevent RNA quantification in which case robustness
in the extraction method is prerequisite.
Array
qPCR
3
2
1
0
-1
-2
-3
-4
-5
-5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00
Fig. 1. Correlation between microRNA qPCR and microarray expression analysis.
The same total RNA extracted from FFPE sections from tumor and normal samples
was analyzed using either miRCURY LNA arrays or LNA microRNA qPCR. The
amount of total RNA used in the microarray analysis was 500 ng per sample whilst
the amount of total RNA used for qPCR was 23 ng per sample. A total of 220
microRNAs detected by qPCR and with probes available on the array are included.
Log 2 ratios from both arrays and qPCR between tumor and normal samples were
compared according to reference [12] and gives a Pearson correlation of 0.84. The
linear regression (black line) compared to a perfect fit (stippled line) of the two data
sets indicates that qPCR can be used to detect larger differences between samples.
0.04
Area dissecected from FFPE section in mm 2
0.16 0.32 0.64
7
6
5
4
3
2
1
0
-1
DeltaCqtolowestinputamount
No carrier
With Carrier
25
27
29
31
33
35
37
39
miR-192 let-7a miR-103
MeanCq
Plasma I, no carrier
Plasma II, no carrier
Plasma I, with carrier
Plasma II, with carrier
A
B
Fig. 2. The addition of carrier improves microRNA quantification. Quantification of
three microRNAs in plasma shows that the use of an RNA carrier improves both
mean Cq values and reproducibility of parallel extractions from the same sample
(A). For each microRNA, two separate RNA isolations were performed from plasma I
and plasma II. Error bars represent the standard deviation between parallel RNA
isolations from the same sample. microRNA recovery and quantification from FFPE
samples is also improved by the addition of carrier (B). Areas of varying size were
isolated from FFPE sections using LCM. Duplicate RNA isolations were performed on
duplicate LCM samples and three microRNAs quantified by qPCR (hsa-miR-21, hsamiR-103,
has-let-7a). The average delta Cq values compared to the lowest input
amount is shown. Error bars represent variation between duplicate isolations. The
inclusion of carrier RNA leads to improved recovery and quantification of microRNA
as well as reduced variation.
S8 D. Andreasen et al. / Methods 50 (2010) S6–S9
Acknowledgments
We thank Jannie R. Christensen and Tina Sommer Bisgaard for
excellent technical assistance.
References
[1] N. Meola, V.A. Gennarino, S. Banfi, Pathogenetics 2 (1) (2009) 7.
[2] R. Visone, C.M. Croce, Am. J. Pathol. 174 (4) (2009) 1131–1138.
[3] E.R. Hutchison, E. Okun, M.P. Mattson, Neuromol. Med. 11 (3) (2009) 153–161.
[4] M.A. Saunders, L.P. Lim, RNA Biol. 6 (3) (2009) 324–328.
[5] S.M. Metias, E. Lianidou, G.M. Yousef, J. Clin. Pathol. 62 (9) (2009) 771–776.
[6] A. Liu, M.T. Tetzlaff, P. Vanbelle, D. Elder, M. Feldman, J.W. Tobias, A.R.
Sepulveda, X. Xu, Int. J. Clin. Exp. Pathol. 2 (6) (2009) 519–527.
[7] F. Lewis, N.J. Maughan, V. Smith, K. Hillan, P. Quirke, J. Pathol. 195 (1) (2001)
66–71.
[8] N. Masuda, T. Ohnishi, S. Kawamoto, M. Monden, K. Okubo, Nucleic Acids Res.
27 (22) (1999) 4436–4443.
[9] Kai Wang, Shile Zhang, Bruz Marzolf, Pamela Troisch, Amy Brightman, Zhiyuan
Hu, Leroy E. Hood, David J. Galas, PNAS 106 (11) (2009) 4402–4407.
[10] P.S. Mitchell, R.K. Parkin, E.M. Kroh, B.R. Fritz, S.K. Wyman, E.L. PogosovaAgadjanyan,
A. Peterson, J. Noteboom, K.C. O’Briant, A. Allen, D.W. Lin, N.
Urban, C.W. Drescher, B.S. Knudsen, D.L. Stirewalt, R. Gentleman, R.L. Vessella,
P.S. Nelson, D.B. Martin, M. Tewari, Proc. Natl. Acad. Sci. USA 105 (30) (2008)
10513–10518.
[11] V. Benes, M. Castoldi, Methods 50 (2010) 244–249.
[12] MAQC Consortium, Nat. Biotechnol. 24 (2006) 1151–1161.
D. Andreasen et al. / Methods 50 (2010) S6–S9 S9