Experimental and Molecular Pathology 94 (2013) 188–194

Contents lists available at SciVerse ScienceDirect

Experimental and Molecular Pathology

journal homepage: www.elsevier.com/locate/yexmp

Non-formalin fixative versus formalin-fixed tissue: A comparison of histology and

RNA quality
Daniel Groelz a,⁎, Leslie Sobin d, Philip Branton c, Carolyn Compton c, e, Ralf Wyrich a, Lynne Rainen b
a
Qiagen GmbH, Research and Development, Hilden, Germany
b
PreAnalytiX GmbH, Research and Development, Franklin Lakes, USA
c
Office of Biorepositories and Biospecimen Research, National Cancer Institute, National Institutes of Health, Bethesda MD, USA
d
Frederick National Laboratory for Cancer Research, National Cancer Institute, Rockville, MD, USA
e
Critical Path Institute (C-Path), Tucson, AZ, USA

a r t i c l e i n f o a b s t r a c t

Article history: Preanalytical handling of tissue samples can influence bioanalyte quality and ultimately outcome of analytical
Received 22 June 2012 results. The aim of this study was to compare RNA quality, performance in real time RT PCR and histology of
Available online 17 July 2012 formalin-fixed tissue to that of tissue fixed and stabilized with a formalin-free fixative, the PAXgene Tissue
System (PAXgene), in an animal model under highly controlled preanalytical conditions. Samples of rat
Keywords:
liver, kidney, spleen, intestine, lung, heart muscle, brain, and stomach tissue were either fixed in formalin
PAXgene Tissue
RNA
or fixed in PAXgene or fresh frozen in liquid nitrogen. RNA was extracted from all samples, examined for in-
RNA integrity score tegrity in microcapillary electrophoresis, and used in a series of quantitative RT PCR assays with increasing
H&E histology amplicon length. Histology of paraffin-embedded samples was determined by staining with hematoxylin
Quantitative RT PCR and eosin.
Histology of all formalin-fixed and PAXgene fixed samples was comparable. RNA with acceptable integrity
scores could be isolated from all embedded tissues, 4.0 to 7.2 for formalin and 6.4 to 7.7 for PAXgene, as com-
pared to 8.0 to 9.2 for fresh frozen samples. While RNA with acceptable RINs (RNA integrity number) could be
isolated from formalin-fixed samples, in microcapillary electrophoresis this RNA separated with a slower mi-
gration rate and displayed diffuse, less focused peaks for ribosomal RNA as compared to RNA from frozen or
PAXgene fixed samples. Furthermore, RNA from formalin-fixed tissues exhibited inhibition in quantitative RT
PCR assays which increased with increasing amplicon length, while RNA from PAXgene fixed samples did not
show such inhibition.
In conclusion, our results demonstrate that excluding other preanalytical factors, PAXgene Tissue System
preserves histology similarly to formalin, but unlike formalin, does not chemically modify RNA. RNA purified
from PAXgene fixed tissues is of high integrity and performs as well as RNA from fresh frozen tissue in RT PCR
regardless of amplicon length.
© 2012 Elsevier Inc. Open access under CC BY-NC-ND license.

Introduction medicine, it is often necessary to analyze biomarkers from a clinical

In recent years, new molecular methods have enabled testing for
global gene, protein, and protein pathway activation expression pro-
files and/or somatic mutations in cancer from individual patients.
These methods have enabled identification of biomarkers which
better define prognosis, and treatment options and lead to the devel-
opment of new drugs which interfere with the molecular mechanisms
of malignancy in an individual patient. For this so called personalized
Abbreviations: PFPE, PAXgene fixed, paraffin-embedded.
⁎ Corresponding author at: QIAGEN GmbH, QIAGEN Straße 1, 40724 Hilden, Germany.
Fax: +49 2103 29 26213.
E-mail addresses: Daniel.groelz@qiagen.com (D. Groelz), leslie.sobin@nih.gov
(L. Sobin), philip.branton@nih.gov (P. Branton), CCompton@c-path.org (C. Compton),
Ralf.Wyrich@qiagen.com (R. Wyrich), Lynne_Rainen@bd.com (L. Rainen).
sample, most often a primary tumor or biopsy from a suspected malig-
nancy. Many preanalytical factors have an influence on the quality of
the biomarker to be analyzed and, ultimately, on the outcome of ana-
lytical results. These factors include both warm and cold ischemia
times, type of fixative, fixation time, tissue sample size, transport con-
ditions, tissue processing protocol, and, most important for biobanks,
sample storage conditions (Chung et al., 2008; Engel and Moore,
2011; Hatzis et al., 2011; Hewitt et al., 2008). The lack of access to
well characterized, highly preserved tissue samples was recognized
as a major bottleneck for identification of new biomarkers and the
main reason so few ostensible biomarkers have been validated for
routine clinical use (Poste, 2011).
Formalin, a 4% aqueous formaldehyde solution, is the most widely
used fixative for preservation of tissue morphology and examination
by histological or immunohistochemical techniques. Formalin has

0014-4800 © 2012 Elsevier Inc. Open access under CC BY-NC-ND license.

doi:10.1016/j.yexmp.2012.07.002
 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194
been used for this purpose with only a few minor changes since its
introduction in the late 19th century (Blum, 1894). Although its mech-
anism of action is still not fully understood, formalin fixes tissue by
chemical modification of biomolecules, forming crosslinks between
proteins and nucleic acids (Masuda et al., 1999). Large archives
of formalin-fixed, paraffin-embedded (FFPE) blocks of tissue exist
throughout the world, and huge efforts have been made to improve
nucleic acid purification from these stored tissues in order to make
these archives accessible for discovery of new biomarkers. This effort
has been largely unsuccessful. Due to the harsh conditions necessary
to break the protein-nucleic acid crosslinks and reverse or partially
reduce chemical modifications, biomolecules isolated from FFPE tissue
are of limited use for molecular analysis (McSherry et al., 2007;
Tournier et al., 2012; von Smolinski et al., 2005; Williams et al.,
1999). Snap-freezing of tissues in liquid nitrogen and storage at low
temperatures, widely accepted to be the best way to preserve proteins
and nucleic acids in tissues, can be inappropriate for a clinical environ-
ment because of intracellular ice formation and consequent morpho-
logical changes in the tissue (Desciak and Maloney, 2000), and
restrictions on the use of liquid nitrogen in clinics.
In the last few years, a number of fixatives have been proposed as
alternatives to formalin (Delfour et al., 2006; Gillespie et al., 2002;
Mueller et al., 2011; Olert et al., 2001; Stanta et al., 2006; Vincek et
al., 2003), but none of these reagents has gained broader acceptance.
Recently, a new tissue fixation method, the PAXgene® Tissue System
(PAXgene) was developed by PreAnalytiX GmbH (Hombrechtikon,
Switzerland). The two-reagent system consists of non-crosslinking
fixative (PAXgene® Tissue Fix) and stabilizer (Paxgene® Tissue Stabi-
lizer) and is being evaluated as a part of the EU FP7 Project SPIDIA
(Standardization and improvement of generic preanalytical tools
and procedures for in vitro diagnostics), a consortium of European
universities and biotechnology companies. The aim of the SPIDIA con-
sortium is to standardize and improve generic preanalytical tools and
procedures for in vitro molecular diagnostics. Studies conducted by
SPIDIA consortium members have demonstrated that tissue samples
fixed and stabilized with PAXgene can be used with conventional his-
tological stains such as hematoxylin and eosin (H&E) or immunohis-
tochemical staining (Kap et al., 2011). Other SPIDIA studies have
shown that PAXgene preserves proteins (Ergin et al., 2010), RNA,
miRNA and DNA (Viertler et al., 2012) in clinical samples.
Since many preanalytical variables extant in clinical settings are
known to have a profound influence on the quality and analytical
results of tissue biomolecules, we used an animal model (Rattus
norvegicus) and tightly controlled specimen collection and handling
conditions to minimize preanalytical variables and directly compare
histology and RNA in PAXgene fixed tissue to that of formalin-fixed
tissues. For our studies, warm and cold ischemia times, fixative quality,
fixation time, tissue processing and RNA extraction protocols were stan-
dardized and optimized to yield optimal histological results as well as
the highest possible quality RNA from all fixed samples. RNA from
fresh frozen tissues was always used as control RNA representing the
highest quality. As a corollary to our investigation of the effects of fixa-
tives on RNA quality, we also investigated the effects of tissue process-
ing and embedding in paraffin on RNA from formalin or PAXgene
fixed tissues.
Materials and methods
Tissue collection and preservation
Liver, kidney, spleen, intestine, lung, stomach, heart muscle and
brain were obtained from R. norvegicus. Rats were maintained and
sacrificed at the Universitätsklinikum Düsseldorf Tierversuchsanlage
(Düsseldorf, Germany) in accordance with the German protection of
animals act. Rats were raised to a weight of approximately 500 g
after which time the animals were sacrificed and organs removed
189
within 10 min of sacrifice. Three adjacent, equally sized samples no
larger than 15 x 15× 4 mm were grossed from each organ. One sample
was fresh frozen (FF) immediately in liquid nitrogen, transported on
dry ice, and stored at −80 °C. For tissue fixation, the remaining two
samples were placed into standard tissue cassettes (Simport Plastic
Ltd., Beloeil, Canada) and completely submerged in a container filled
with either 4% neutral buffered formaldehyde (NBF) (Sigma-Aldrich,
Steinheim, Germany) or PAXgene Tissue Fix (PreAnalytiX GmbH,
Hombrechtikon, CH) with fixation solutions in a ratio of at least 20
parts fixative to one part of tissue (v/v). Fixation was performed at
room temperature with formalin for 24 h or with the PAXgene Tissue
Fix for 4 h followed by a transfer into the PAXgene Tissue Stabilizer
(PreAnalytiX Gmbh, Hombrechtikon, CH) and stored for 72 h, to simu-
late transport or storage times in clinical settings. Effects of reversing
the recommended PAXgene protocol were investigated by incubating
tissue samples in the PAXgene Tissue Stabilizer reagent for 4 h before
fixation with PAXgene Tissue Fix for 72 h. To investigate the effects of
tissue processing and paraffin embedding on the quality of RNA, RNA
from duplicate paired organ samples (two samples from each organ)
were either extracted immediately after fixation in formalin or fixation
and stabilization in PAXgene or after fixation, stabilization processing,
and embedding in paraffin as described below.
Tissue processing
Samples fixed with formalin or PAXgene were processed in separate
runs on an automated tissue processor (TP1020, Leica-microsystems,
Wetzlar, Germany). After fixation, formalin-fixed samples were washed
briefly with deionized water and transferred to 70% ethanol. PAXgene
fixed samples were directly transferred into 70% ethanol. Samples
remained in 70% ethanol for up to 8 h prior to initiation of processing.
The processing protocol for all samples, regardless of fixation method
was as follows: incubation at 80%, 90%, 99% ethanol (2×), followed by
isopropanol (2×), xylene (2×) for no longer than 1 h at each position.
Tissue processing reagents were replaced after five runs to avoid re-
agent dilution and incomplete dehydration of the tissue samples. Ex-
treme care was taken to prevent the exposure to even trace amounts
of formalin of tissue fixed and stabilized in PAXgene during the tissue
processing step by using separate batches of reagents for PAXgene and
formalin-fixed tissues.
In order to prevent RNA degradation due to extreme temperatures, a
low-melting point paraffin (Surgipath Paraplast-XTRA, Carl Roth GmbH,
Karlsruhe, Germany) was used for infiltration and embedding of
samples regardless of method of fixation. For infiltration of tissue with
paraffin, samples were incubated (3× 1 h) under vacuum at 56 °C.
Within 30 min after infiltration, samples were, embedded. All blocks
of paraffin-embedded tissue were kept dry at 4 °C until use, and RNA
was isolated from the paraffin blocks within eight weeks from freshly
cut sections only.
Hematoxylin and eosin (H&E) staining
Sections of 4 μm thickness from FFPE and PAXgene fixed,
paraffin-embedded (PFPE) tissue blocks were cut on a microtome
(RM2245, Leica-microsystems, Wetzlar, Germany) and stretched on a
water bath filled with fresh, deionized water heated to 40 °C for PFPE
and to 45 °C for FFPE tissue. Paraffin sections were removed from the
water bath within 1 min and placed on slides which were air-dried
overnight at room temperature. Staining was performed manually in
staining dishes starting with a de-waxing step in xylene, then rehydra-
tion with successive incubations in 96%, 80%, 70%, 60% ethanol, and
finally tap water. Hematoxylin (Merck KGaA, Darmstadt, Germany)
was applied for 1 min followed by a 3 min wash with tap water then
followed by staining for 3 min with eosin Y (Merck KGaA). After wash-
ing with tap water and dehydration with successive washes of 80%,
96%, 100% ethanol, and xylene, slides were mounted with Entellan
190 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194
(Merck KGaA, Darmstadt, Germany) and air-dried prior to microscopic
examination.
RNA preparation
Total RNA was purified from 10 mg of fresh frozen (FF), formalin or
PAXgene fixed tissue or from three sections, each 10 μm thick of FFPE
and PFPE tissue. For RNA isolation, the miRNeasy Mini kit (QIAGEN
GmbH, Hilden, Germany) was used for FF tissue, the miRNeasy FFPE
kit (QIAGEN) for formalin-fixed and FFPE tissue, and the PAXgene Tis-
sue miRNA kit (PreAnalytiX GmbH) for PAXgene fixed and PFPE tissue.
All preparations were done in triplicate with tissue samples from two
different animals for each fixation method (FF, formalin-fixed, PAXgene
fixed, PFPE and FFPE). RNA yield was determined by spectrophotomet-
ric absorbance (Nanodrop ND-1000 spectrophotometer, Nanodrop
Technologies, Wilmington, USA) at 260 nm (A260). RNA integrity
was assessed by microcapillary electrophoresis on an Agilent 2100
Bioanalyzer and analyzed with the Agilent 2100 expert software
(Agilent Technologies, Waldbronn, Germany) to give RNA integrity
scores (RINs).
PCR amplifications
Using one forward primer and ten different reverse primers
(Table 1), PCR primers specific to the rat beta-actin gene (NC_005111)
were used to design individual quantitative, real-time RT PCR assays
which amplified ten different transcript sequences with lengths of
between 109 and 610 nucleotides (nt). One-step, RT PCR assays were
performed with 10 ng of total RNA on a Rotor-Gene Q (QIAGEN)
using the Rotor-Gene SYBR Green RT PCR Kit (QIAGEN). Thermal cycling
conditions started with reverse transcription for 30 min at 60 °C
followed by an activation step of 15 min at 95 °C and 40 cycles of 15 s
at 94 °C, 30 s at 60 °C and 30 s at 70 °C. PCR reactions were run in
duplicate for each of the triplicate RNA extractions of tissue samples
from two different animals, resulting in 12 single reactions per assay
for each sample type (FF, PFPE or FFPE), 360 reactions per tissue type
or 2,880 PCR reactions in total. The acceptance criteria for single reac-
tions were that the cycle threshold (CT) of the reaction must be b40
and the difference in CT between duplicate reactions must be ≤2. The
validity of each individual PCR reaction was confirmed by melting
point analysis. Melting curves had to be free of extraneous peaks or
peaks which indicated non-specific amplification to be accepted as
valid. All data not meeting these acceptance criteria were excluded
from data analysis.
Results
Eight different rat tissue types, including liver, kidney, spleen,
intestine, lung, stomach, heart muscle and brain were used in a compar-
ison of morphology and RNA preservation in tissue samples fixed either
with formalin or PAXgene. For optimal preservation of morphology and
RNA, all preanalytical steps of tissue procurement, fixation, and
Table 1
Oligonucleotide primer sequences for RT PCR.
PCR primer Primer sequence (5′ → 3′)
Rn_ACTB rev109 ACGCTCGGTCAGGATCTTCATG
Rn_ACTB rev158 TAATGTCACGCACGATTTCCC
Rn_ACTB rev189 AAGTCTAGGGCAACATAGCAC
Rn_ACTB rev236 TCTTCTCCAGGGAGGAAGAGGATG
Rn_ACTB rev287 GGAACCGCTCATTGCCGATAG
Rn_ACTB rev331 TTCCATACCCAGGAAGGAAGG
Rn_ACTB rev438 TACATGGTGGTGCCACCAGAC
Rn_ACTB rev465 TTCTGCATCCTGTCAGCAATG
Rn_ACTB rev569 ACAGTGAGGCCAGGATAGAG
Rn_ACTB rev610 TACTCCTGCTTGCTGATCCAC
processing were performed according to recommendation of Hewitt
et al. (2008). RNA was purified from formalin-fixed or PAXgene fixed
tissue samples, before and after embedding into paraffin using the
QIAGEN miRNeasy FFPE Kit and the PAXgene Tissue miRNA Kit respec-
tively. As a reference for RNA quality and performance in RT PCR, repli-
cate samples from each tissue type were also fresh frozen in liquid
nitrogen and stored at -80 °C prior to RNA isolation.
Tissue histology
Fig. 1 depicts H&E stained tissue sections from all tissue types
obtained for this study. As can be seen in Fig. 1, both PAXgene and for-
malin fixation gave generally comparable results that were adequate for
histologic identification and structural status for multiple tissues. There
were differences, however, in that there was more cytoplasmic eosino-
philia evident in PAXgene fixed tissues than in formalin-fixed tissues.
This characteristic was especially evident in liver, heart, and brain
(Fig. 1a,m,o), but the increased eosinophilic staining seen in PAXgene
fixed tissue seemed to enhance cytoplasmic detail as compared to
formalin-fixed tissue. Nuclear features were good in both PAXgene
fixed and formalin-fixed tissues although somewhat more intense in tis-
sues treated with PAXgene. Red cell lysis appeared to be characteristic of
PAXgene fixation which tends to limit the depiction of capillaries (Fig. 1c,
d kidney) and the assessment of vascular congestion (Fig. 1e,f spleen) as
compared to formalin-fixed tissues. Lung structure in formalin-fixed
tissue is likewise easier to discern than in PAXgene fixed tissues be-
cause the capillaries retain red cells (Fig. 1i,j); however, the distinc-
tion between the basophilic parietal and eosinophilic chief cells in
stomach tissue (Fig. 1k,l) is better in PAXgene than in formalin-fixed
tissue.
The recommended procedure for fixing and stabilizing tissue with
the PAXgene Tissue System is first to immerse the tissue in the PAXgene
Tissue Fix for a minimum of 2 h followed by immersion in the PAXgene
Tissue Stabilizer. The reagents are currently presented to the user in a
two-chambered container with the chambers clearly marked. While
reversing the recommended procedure (immersion in PAXgene Tissue
Stabilizer followed by transfer to the PAXgene Tissue Fix) is a re-
mote, but nevertheless possible mistake, we investigated whether
this improper use of the system affected histology of the mishandled tis-
sue. We found that immersion of tissue into PAXgene Tissue Stabilizer
before fixation followed by the correct protocol could easily be detected
histologically by the appearance of a characteristic eosinophilic blurred
tissue border (Fig. 2). The broadness of the border depended on the
time the sample remained in the stabilizer before fixation. After 4 h in
the stabilization reagent and 72 h in the fixative (the reverse of the
recommended procedure) this zone measured between 0.25 and
0.33 mm. Tissues placed in stabilizer first for shorter periods displayed
narrower eosinophilic borders. The border was more prominent in cer-
tain tissues like liver and kidney compared to lung (data not shown).
For all tissues for which the recommended procedure was followed,
the resulting morphology as revealed by H&E staining was histologically
comparable to tissue fixed with formalin.
RNA quality
In our study, RNA with moderately high or adequate RINs could be
Size (nt) isolated from all samples (Table 2) regardless of the method used to
109 preserve the tissue. RNA with the highest RIN scores (>8.0) was iso-
158 lated from FF tissue. RINs for RNA from formalin-fixed tissue as well
189 as for RNA from PAXgene fixed tissue were predominantly in the
236
range of 7 to 9, nearly as high as those from FF tissue even with an ad-
287
331
ditional 72 h room temperature storage time after the initial 4 h fix-
438 ation period in the case of the PAXgene fixed tissue samples. Tissue
465 processing and embedding in paraffin using our methods had only a
569 slight negative effect on RINs for PAXgene fixed tissue samples with
610
an average decrease from 7.9 to 7.1. This decrease in RINs was higher
 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194 191

Fig. 1. Morphology of H&E stained rat liver (a, b), kidney (c, d), spleen (e, f), small intestine (g, h), lung (i, j), stomach (k, l), heart (m, n) and brain (o, p) tissue, from PAXgene fixed
and stabilized, paraffin-embedded (PFPE: a, c, e, g, i, k, m, o) or formalin-fixed, paraffin-embedded (FFPE: b, d, f, h, j, l, n, p) samples; original magnifications for spleen, stomach and
brain ×200, liver, kidney, small intestine, lung and heart ×400.

for formalin-fixed tissue samples with an average decrease from 7.4
to 5.3. In addition the RIN scores of RNA extracted from FFPE tissue,
showed greater variability depending on tissue type. For liver and
kidney tissues, RIN scores, consistently higher than 6 were obtained,
and for liver, RIN scores as high as 8 could be achieved. For other tis-
sue types e.g. brain or heart muscle, RIN scores were in the range of 3
to 6.
Despite adequate or even high RIN scores for RNA from formalin-
fixed tissue, consistent differences in the actual electropherograms
between RNA from FFPE and PFPE or FF tissues could be observed.
As depicted in Fig. 3, the electropherogram of RNA from formalin-
fixed tissues showed that this RNA separated at a slower migration
rate as compared to RNA from PFPE or FF tissue, resulting in 18S
and 28S rRNA peaks which were slightly shifted toward higher
192 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194

(Arzt et al., 2011; Chung et al., 2006; von Ahlfen et al., 2007). Because
our study focused specifically on the quality of RNA extracted from tis-
sues fixed in either formalin or in a non-crosslinking fixative system,
we chose to use an animal model in which sample handling, fixation,
processing conditions and RNA purification were tightly controlled and
favored the preservation of RNA. Our goal was to compare RNA quality
and performance in RT PCR assays between RNA extracted from
formalin-fixed tissues and tissues fixed with the PAXgene Tissue System
with all other preanalytical factors equal and favorable to recovery of in-
tact RNA.
Fig. 2. Morphology of PAXgene fixed, H&E stained rat liver. (a) Recommended procedure Using this animal model, it was possible to minimize pre- and
with fixation stopped after 4 h by transfer into PAXgene Tissue Stabilizer and (b) improp- post-excision times and standardize sample size, fixative quality, fix-
er use with reversed procedure, e.g. 4 h in PAXgene Tissue Stabilizer and transfer into ation times, and tissue processing reagents and conditions. PAXgene
PAXgene Tissue Fix reagent resulting in an eosinophilic blurred border (arrow); original fixed tissue was subjected to an additional three days of storage in
magnification ×100.
the stabilizer at room temperature to simulate conditions found in
clinical settings. We compared the morphology of PFPE tissue to that
of FFPE tissue using H&E staining as a common staining method. We
molecular weights. These peaks were not only shifted, but they were also able to investigate the influence of the new PAXgene Tissue
appeared more diffuse and less focused when compared to the 18S fixation method as well as the effect of processing conditions on RNA
and 28S peaks of ribosomal RNA from FF and PFPE samples. These dif- quality by isolating RNA after fixation/stabilization and after fixation/
ferences, however, were not reflected in the RIN scores as calculated stabilization and processing including paraffin embedding. RNA from
by the Agilent 2100 expert software. fresh frozen replicate samples obtained immediately after excision
was always used as reference for RNA of the highest possible quality.
While there were some differences in staining intensities between
RNA performance in RT PCR
PFPE and FFPE tissues PAXgene Tissue was essentially comparable to
formalin-fixed tissue for histology, except that red blood cells are
While RIN scores for all samples would indicate adequate to good
lysed in PAXgene fixed tissue. Overall, the differences in histologic
quality RNA had been extracted from tissue samples regardless of
features between PFPE and FFPE tissues do not appear to be diagnos-
preservation method, discrepancies in performance in RT PCR were
tically significant.
observed for RNA from different sources. Again, although RIN scores
RIN scores obtained from microcapillary electrophoresis are very
for RNA from FFPE and PFPE tissues were, in some tissue types, com-
often used to indicate overall RNA quality. The RIN is an algorithm
parable, real time RT PCR assays revealed RT PCR inhibition of RNA from
for assigning integrity values to purified RNA and is largely based on
all FFPE tissues. Furthermore, inhibition increased with increasing
the run times and peak heights of 18S and 28S ribosomal RNA which
amplicon length (Fig. 4). Using the same amount of RNA for each
are widely accepted as markers for total RNA, including transcript
assay, the assay with the shortest amplicon (109 nt) demonstrated
RNA, quality (Schroeder et al., 2006). The algorithm assigns a 1 to 10
the lowest difference in CT values between RNA from FF and FFPE tissue
RIN score where level 10 RNA is completely intact. RIN scores of b10
samples with ΔCT values (ΔCT = CT[FFPE] − CT[FF]) of 3 to 5, depending on
but >5 indicate various levels of partial RNA degradation while RIN
the tissue type. These differences increased linearly to 14 in the case of
scores of b 5 indicate higher levels of degradation such that RNA with
the assay with the largest amplicon (610 nt). In contrast, using the
RIN scores b3 are generally considered to be of very poor quality.
same amount of RNA from PFPE and FF tissue samples for RT PCR, the
Whether or not this poor quality RNA is useful for gene expression anal-
ΔCT values (ΔCT = CT[PFPE] − CT[FF]) were consistently in the range of
ysis depends in large part on the size of the (degraded) transcript to be
−1 to 2 for all tissue types and did not vary with amplicon length.
amplified, and RT PCR assays are often designed to amplify short tran-
script sequences to accommodate RNA of poor quality (Antonov et al.,
Discussion 2005). This approach, unfortunately limits the amount and quality of
gene expression data researchers can obtain from degraded tissue
Preanalytical variables such as ischemia time, sample size, volume samples.
and type of fixative, fixation time, and processing conditions can have In our study, the average RIN scores for six replicates of eight
a profound influence on the quality of tissue and biomolecule preser- different rat tissues from fresh frozen samples was high at 8.7
vation (Chung et al., 2008; Hewitt et al., 2008; Verderio, 2012). Most (Table 2). The fixation process itself, whether by formalin or by PAXgene,
studies involving RNA from clinical, formalin-fixed samples report with all other preanalytical conditions kept under tight control, slightly
extensive RNA degradation with typical fragment sizes of 100 to 200 nt affected the RNA integrity as measured by RIN scores. Average RIN
scores of unprocessed formalin-fixed and PAXgene fixed tissues de-
creased to 7.4 and 7.9 respectively. The tissue processing protocol did
not greatly affect RIN scores from PAXgene fixed tissue with a decrease
on average from 7.9 to 7.1. For formalin-fixed tissue, this decrease due
Table 2
Average RNA integrity values from fixed rat tissue, 6 replicates each.
to tissue processing was greater with the average RIN score dropping
from 7.4 to 5.3. Therefore, FFPE and PFPE samples showed an average de-
Fresh frozen PAXgene Formalin PFPE FFPE crease in RIN scores of 2.1 and 0.8 respectively when compared to
Liver 9.0 ± 0.4 7.9 ± 0.2 7.8 ± 0.2 7.2 ± 0.3 7.2 ± 0.7 paired, unprocessed samples. This indicates that in our model system,
Kidney 8.0 ± 0.1 7.0 ± 0.1 7.6 ± 0.6 6.4 ± 0.1 6.4 ± 0.8 processing, which included embedding in low melting point paraffin,
Spleen 8.6 ± 0.9 8.5 ± 0.5 8.2 ± 0.7 7.5 ± 0.3 5.5 ± 1.1
had only a minor effect on RNA integrity in PAXgene fixed tissue but a
Intestine 8.3 ± 0.4 8.7 ± 0.3 8.0 ± 0.8 7.7 ± 0.2 4.6 ± 0.5
Lung 9.2 ± 0.6 8.5 ± 0.2 7.9 ± 0.7 6.9 ± 0.3 4.9 ± 2.7 moderate effect on RNA integrity in formalin-fixed tissue.
Heart 9.1 ± 0.3 8.0 ± 0.3 6.0 ± 1.5 7.1 ± 0.6 4.0 ± 0.8 Insofar as the quality of RNA extracted from fixed tissue is
Brain 9.1 ± 0.3 7.2 ± 0.3 6.8 ± 1.0 7.2 ± 0.3 4.1 ± 1.7 concerned, an often neglected factor in obtaining high quality RNA
Stomach 8.4 ± 0.5 7.4 ± 0.3 7.2 ± 1.5 7.0 ± 0.1 5.7 ± 1.6 from tissue is the method used for RNA extraction. In recent studies
Overall 8.7 ± 0.6 7.9 ± 0.7 7.4 ± 1.1 7.1 ± 0.5 5.3 ± 1.7
of non-crosslinking fixatives, some investigators were only able to
 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194 193

Fig. 3. Analysis of total RNA isolated from rat liver fresh frozen (FF), PAXgene fixed and stabilized paraffin-embedded (PFPE) or formalin-fixed, paraffin-embedded (FFPE) by
microcapillary electrophoresis. Slower migration rate of RNA from FFPE is visible in (a) the electropherogram and (b) on the gel.

extract RNA with RIN scores in the range of 1.8 to 2.5 from surgically
obtained specimens (Arzt et al., 2011; Moelans et al., 2011) While
some might conclude that these findings point to the lack of an
advantage in preserving RNA for non-crosslinking fixatives, low RIN
scores can be partially explained by longer ischemia times, varied
preanalytical conditions seen in surgical pathology settings, or the
RNA extraction methods employed. To explain the last of these
three possibilities, it is well known that commercially available kits
for RNA extraction from FFPE tissue include optimized proteinase K
digestion steps often followed by heating steps in order to break the
methylene bridges that form crosslinks in formalin-fixed tissue
(Masuda et al., 1999). Under these conditions, RNA is released from
a network of crosslinked proteins and nucleic acids. These methods,
however, should be applied only when necessary, i.e. when crosslinks
found in formalin-fixed tissue must be removed from extracted RNA,
an unnecessary step in non-crosslinked tissue. This was illustrated by
the fact that in our study, RNA extracted from PFPE samples using kits
for RNA extraction from FFPE (miRNeasy FFPE kit, QIAGEN GmbH),
RIN scores of the resulting RNA were reduced by ≥3.0 as compared to
the RNA extracted with the PAXgene Tissue miRNA kit (data not
shown). This finding is due directly to the less harsh conditions
employed in the PAXgene Tissue miRNA kit. These RNA-friendly condi-
tions include the elimination of the 80 °C heating step after proteinase K
digestion.
Nevertheless, in our study although we were able to isolate high
molecular weight RNA with relatively high RIN scores from FFPE tis-
sue, electropherograms of this RNA showed a shift of the ribosomal
RNA bands toward higher molecular weights, and the bands were
broader and less focused. This most likely indicates the presence of
chemical modifications of the RNA such as covalently linked residual
amino acids. This is consistent with the finding that, even with a com-
bination of optimized digestion and heating steps, it is not possible to
completely remove all chemical modifications from nucleic acids in
formalin-fixed tissue (Masuda et al., 1999). When RNA is highly
degraded, as this is often the case in surgical samples fixed with formalin,
such modifications are no longer detectable because 18S and 28S RNA
peaks are no longer visible in the electropherogram. Since our methods
yielded FFPE tissue RNA with RIN scores as high as 8.0 (liver), it would
seem that the algorithm for the calculation of RIN score does not reflect
slight changes in molecular weight due to chemical modification of the
RNA.
The results in RT PCR also indicate that the RNA from FFPE samples
most likely contained residual methyl groups, amino acids or other
chemical modifications which led to a premature stop of the reverse
transcription step. Furthermore, our results showed that, for formalin-
fixed tissue RNA, the longer the transcript length, the greater the likeli-
hood for the premature stop of reverse transcription. Unlike RNA from
formalin-fixed tissue, the RNA from PAXgene fixed samples appeared

Fig. 4. Reverse transcription and amplification using 10 ng RNA each from rat tissue fresh frozen (FF), PAXgene fixed paraffin-embedded (PFPE) or formalin-fixed,
paraffin-embedded (FFPE), in ten different SYBR-Green real time one step RT PCR assays. Amplicons of the rat beta-actin gene ranged from 109 to 610 nucleotide (nt). Average
delta-CT values (ΔCT = CT[FFPE] − CT[FF] or ΔCT = CT[PFPE] − CT[FF]) are shown for triplicate extractions from two different tissue samples amplified in duplicates for each type of tissue
and fixation method.
194 D. Groelz et al. / Experimental and Molecular Pathology 94 (2013) 188–194
not to be burdened with chemical modifications as revealed both by
microcapillary electrophoresis and RT PCR results. For RNA from PFPE
tissue, reverse transcription was not inhibited, and RNA purified from
PFPE samples performed similarly to RNA isolated from fresh frozen tis-
sue regardless of amplicon length. Since in our model system the RIN
scores obtained for RNA from FFPE and PFPE tissues were in some tissue
types nearly equal, we conclude that RIN scores cannot be used as a pre-
dictor for performance in downstream assays of RNA from formalin-fixed
tissue while they do seem to indicate performance of RNA from PAXgene
fixed tissue, the performance of which is comparable to RNA from fresh
frozen tissue. Based on these results, caution should be taken when com-
paring RNA from FFPE tissues to RNA from fresh frozen or PAXgene fixed
samples RT PCR assays.
In conclusion, tissue fixed with the PAXgene Tissue System is mor-
phologically comparable to formalin-fixed tissue but yields RNA
which performs as well as RNA from fresh frozen tissue in RT PCR as-
says. While RNA with acceptable RIN scores could be isolated from
both FFPE and PFPE tissue, only RNA from PFPE tissue was comparable
to RNA from fresh frozen tissue in real time RT PCR assays and for all
amplifications of transcript sequences from 109 to 610 nt. For studies
involving prospective tissue collection for which the downstream anal-
ysis either requires RNA or is unspecified at the time of collection, the
PAXgene Tissue System might be useful, since histology of PAXgene
fixed tissue is comparable to that of formalin-fixed tissue, but the per-
formance of PAXgene Tissue RNA, unlike formalin-fixed RNA, is identi-
cal to FF tissue RNA in RT PCR assays.
Conflict of interest statement
The authors disclose the following: Daniel Groelz and Ralf Wyrich
are employed by QIAGEN and have stock options or bond holdings in
the company's pension plan. Lynne Rainen is employed by BD. The
data presented in this manuscript refer to the PAXgene Tissue System
that has been developed by PreAnalytiX, a joint venture between BD
and QIAGEN, and evaluated within the EU FP7 SPIDIA consortium.
All other authors have no conflict of interest.
Acknowledgments
The authors wish to thank Nadine Dettmann, Isabell Blassnig and
Evelyn Traenert for their technical assistance. The research leading to
these results has received funding from the European Union Seventh
Framework Programme [FP7/2007–2013] under grant agreement no.
222916.
References
Antonov, J., Goldstein, D.R., Oberli, A., Baltzer, A., Pirotta, M., Fleischmann, A., Altermatt,
H.J., Jaggi, R., 2005. Reliable gene expression measurements from degraded RNA by
quantitative real-time PCR depend on short amplicons and a proper normalization.
Laboratory Investigation 85, 1040–1050.
Arzt, L., Kothmaier, H., Quehenberger, F., Halbwedl, I., Wagner, K., Maierhofer, T.,
Popper, H.H., 2011. Evaluation of fromalin-free tissue fixation for RNA and
microRNA studies. Experimental and Molecular Pathology 91, 490–495.
Blum, F., 1894. Notiz ueber die Anwendung des Formaldehyds (Formol) als Hartungs- und
Konservierungsmittel. Anatomischer Anzeiger 9, 224.
Chung, J.Y., Braunschweig, T., Hewitt, S.M., 2006. Optimization of recovery of RNA from
formalin-fixed, paraffin-embedded tissue. Diagnostic Molecular Pathology 15,
229–236.
Chung, J.Y., Braunschweig, T., Williams, R., Guerrero, N., Hoffmann, K.M., Kwon, M.,
Song, Y.K., Libutti, S.K., Hewitt, S.M., 2008. Factors in tissue handling and process-
ing that impact RNA obtained from formalin-fixed, paraffin-embedded tissue.
JHC 56, 1033–1042.
Delfour, C., Roger, P., Bre, C., Berthe, M.L., Rachiax, P., Kalfa, N., Raynaud, P., Bibeau, F.,
Maudelonde, T., Boulle, N., 2006. RCL2, a new fixative, preserves morphology and
nucleic acid integrity in paraffin-embedded breast carcinoma and microdissected
breast tumor cells. JMD 8, 157–169.
Desciak, E.B., Maloney, M.E., 2000. Artifacts in frozen section preparation. Dermatologic
Surgery 26, 500–504.
Engel, K.B., Moore, H.M., 2011. Effects of preanalytical variables on the detection of
proteins by immunohistochemistry in formalin-fixed, paraffin-embedded tissue.
Archives of Pathology & Laboratory Medicine 135, 537–543.
Ergin, B., Meding, S., Langer, R., Kap, M., Viertler, C., Schott, C., Ferch, U., Riegmann, P.,
Zatloukal, K., Walch, A., Becker, K.F., 2010. Proteomic analysis of PAXgene-fixed
tissues. Journal of Proteome Research 9, 5188–5196.
Gillespie, J.W., Best, C.J., Bichsel, V.E., Cole, K.A., Greenhut, S.F., Hewitt, S.M., Ahram, M.,
Gathright, Y.B., Merino, M.J., Strausberg, R.L., Epstein, J.I., Hamilton, S.R., Gannot, G.,
Baibakova, G.V., Calvert, V.S., Flaig, M.J., Chuaqui, R.F., Herring, J.C., Pfeifer, J.,
Petricoin, E.F., Linehan, W.M., Duray, P.H., Bova, G.S., Emmert-Buck, M.R., 2002.
Evaluation of non-formalin tissue fixation for molecular profiling studies. American
Journal of Pathology 160, 449–457.
Hatzis, C., Sun, H., Yao, H., Hubbard, R.E., Meric-Bernstam, F., Babiera, G.V., Wu, Y.,
Pusztai, L., Symmans, W.F., 2011. Effects of tissue handling on RNA integrity and
microarray measurements from resected breast cancers. National Cancer Institute
103, 1871–1883.
Hewitt, S.M., Lewis, F.A., Cao, Y., Conrad, R.C., Cronin, M., Danenberg, K.D., Goralski, T.J.,
Langmore, J.P., Raja, R.G., Williams, M., Palma, J.F., Warrington, J.A., 2008. Tissue
handling and specimen preparation in surgical pathology. Archives of Pathology
& Laboratory Medicine 132, 1929–1935.
Kap, M., Smedts, F., Oosterhuis, W., Winther, R., Christensen, N., Reischauer, B., Viertler,
C., Groelz, D., Becker, K.F., Zatloukal, K., Langer, R., Slotta-Huspenina, J., Bodo, K., de
Jong, B., Oelmuller, U., Riegman, P., 2011. Histological assessment of PAXgene
tissue fixation and stabilization reagents. PLoS One 6, e27704.
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 Research 27, 4436–4443.
McSherry, E.A., McGoldrick, A., Kay, E.W., Hopkins, A.M., Gallagher, W.M., Dervan, P.A.,
2007. Formalin-fixed paraffin-embedded clinical tissues show spurious copy
number changes in array-CGH profiles. Clinical Genetics 72, 441–447.
Moelans, C.B., Oostenrijk, D., Moons, M.J., von Diest, P.J., 2011. Formaldehyde substitute
fixatives: effects on nucleic acid preservation. Journal of Clinical Pathology 64,
960–967.
Mueller, C., Edmiston, K.H., Carpenter, C., Gaffney, E., Ryan, C., Ward, R., White, S.,
Memeo, L., Colarossi, C., Petricoin, E.F., Liotta, L.A., Espina, V., 2011. One-step
preservation of phosphoproteins and tissue morphology at room temperature for
diagnostic and research specimens. PLoS One 6, e23780.
Olert, J., Wiedorn, K.H., Goldman, T., Kkühl, H., Mehraein, Y., Scherthan, H., Niketeghad,
F., Vollmer, E., Müller, A.M., Müller-Navia, J., 2001. HOPE fixation: a novel fixing
method and paraffin-embedding technique for human soft tissues. Pathology,
Research and Practice 197, 823–826.
Poste, G., 2011. Bring on the biomarkers. Nature 469, 156–157.
Schroeder, A., Mueller, O., Stocker, S., Salowsky, R., Leiber, M., Gassmann, M., Lightfoot,
S., Menzel, W., Granzow, M., Ragg, T., 2006. The RIN: an RNA integrity number for
assigning integrity values to RNA measurements. BMC Molecular Biology 7, 3.
Stanta, G., Mucelli, S.P., Petrera, F., Bonin, S., Bussolati, G., 2006. A novel fixative
improves opportunities of nucleic acids and proteomic analysis in human archive's
tissues. Diagnostic Molecular Pathology 15, 115–123.
Tournier, B., Chapusot, C., Courcet, E., Martin, L., Lepage, C., Faivre, J., Piard, F., 2012.
Why do results conflict regarding the prognostic value of the methylation status
in colon cancers? The role of the preservation method. BMC Cancer 12, 12.
Verderio, P., 2012. Assessing the clinical relevance of oncogenic pathways in neoadjuvant
breast cancer. Journal of Clinical Oncology 30, 1912–1915.
Viertler, C., Groelz, D., Gündisch, S., Kashofer, K., Reischauer, B., Riegman, P., Winther,
R., Wyrich, R., Becker, K.F., Olmüller, U., Zatloukal, K., 2012. A New Technology
for Stabilization of Biomolecules in Tissues for Combined Histological and Molecu-
lar Analyses. Journal of Molecular Diagnostics Jun 28 [Epub ahead of print].
Vincek, V., Nassiri, M., Nadji, M., Morales, A.R., 2003. A tissue fixative that protects mac-
romolecules (DNA, RNA, and protein) and histomorphology in clinical samples.
Laboratory Investigation 83, 1427–1435.
von Ahlfen, S., Missel, A., Bendrat, K., Schlumpberger, M., 2007. Determinants of RNA
quality from FFPE samples. PLoS One 2, e1261.
von Smolinski, D., Leverkoehne, I., von Samson-Himmelstjerna, G., Gruber, A.D., 2005.
Impact of formalin-fixation and paraffin-embedding on the ratio between mRNA
copy numbers of differently expressed genes. Histochemistry and Cell Biology
124, 177–188.
Williams, C., Ponten, F., Moberg, C., Söderkvist, P., Uhlen, M., Ponten, J., Sitbon, G.,
Lundeberg, J., 1999. A high frequency of sequence alterations is due to formalin
fixation of archival specimens. American Journal of Pathology 155, 1467–1471.