See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/13394869

False positive staining in the TUNEL assay to detect apoptosis in liver and
intestine is caused by endogenous nucleases and inhibited by diethyl
pyrocarbonate

Article  in  Molecular Pathology · September 1998

DOI: 10.1136/mp.51.4.204 · Source: PubMed

CITATIONS READS

113 612

5 authors, including:

Ulrich Marti Marc Solioz

MCL Medizinische Laboratorien Tomsk State University
53 PUBLICATIONS   1,901 CITATIONS    169 PUBLICATIONS   9,289 CITATIONS   

SEE PROFILE SEE PROFILE

Juerg Reichen
Universität Bern
352 PUBLICATIONS   12,284 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Copper toxicity mechanisms View project

History of copper in biology View project

All content following this page was uploaded by Juerg Reichen on 01 June 2014.

The user has requested enhancement of the downloaded file.

204 J Clin Pathol: Mol Pathol 1998;51:204–208

Papers

False positive staining in the TUNEL assay to

detect apoptosis in liver and intestine is caused by
endogenous nucleases and inhibited by diethyl
pyrocarbonate
B J Stähelin, U Marti, M Solioz, H Zimmermann, J Reichen

Abstract Apoptosis is an important mechanism in the

Background—The terminal transferase regulation of tissue homeostasis and, in the
uridyl nick end labelling (TUNEL) assay liver, it is an important feature of diVerent
allows the easy demonstration of cell pathological conditions.1 An important
death as a result of apoptosis. However, method for visualising apoptotic cells is the
when this assay is applied to liver tissue, terminal transferase uridyl nick end labelling
the number of TUNEL positive cells is (TUNEL) method, described by Gavrieli et al,2
dependent on the time of incubation with although this method fails to discriminate
proteinase K. between apoptosis and necrosis.3
Aim—To test whether false positive results Bile duct ligation induces marked prolifera-
are the result of the release of endogenous tion of ductular epithelial cells4 5 and these
endonucleases by proteinase K and can be hyperplastic ductules regress rapidly after
abolished by pretreatment with diethyl biliary decompression by Roux-en-Y biliary
pyrocarbonate (DEPC). anastomosis.6 Morphologically, this can be
Methods—Involution of hyperplastic duc- seen to be the result of apoptosis.7 We were
tules in bile duct ligated rats after biliary interested in applying the TUNEL assay to
decompression by Roux-en-Y anastomo- measure apoptosis after biliary decompression
sis and acute CCl4 intoxication were stud- but found an undue dependence of the number
ied as models of apoptosis and necrosis, of apoptotic cells on the time of incubation
respectively. A standard TUNEL assay with proteinase K, a step required to render the
was applied to formalin fixed tissue sec- nuclear DNA fragments accessible to the
tions mounted with cement. To inhibit terminal transferase. We hypothesised that this
putative endogenous endonucleases, tis- phenomenon was the result of DNA nicking by
sue slides were pre-incubated with DEPC. endogenous endonucleases set free by the pro-
Results—In the standard TUNEL assay, teinase K treatment.
the number of positive nuclei was highly
dependent upon the length of time that Materials and methods
sections were incubated with proteinase ANIMAL TREATMENT
K. After pretreatment with DEPC, only Male Sprague-Dawley rats were obtained from
cells that also exhibited morphological Süddeutsche Versuchstierfarm Hartmut and
features of apoptosis stained positive. Voss (Tuttlingen, Germany) and were kept
DEPC pretreatment abolished false posi- under a 12 hour light–dark cycle with free
tive staining in CCl4 induced hepatocyte access to standard rat feed and tap water.
necrosis and blocked interference by en- Biliary cirrhosis was induced by bile duct liga-
dogenous alkaline phosphatase in intes- tion and dissection according to Kontouras et
Department of Clinical tine. The method of gluing the tissue al,4 as described previously by our
Pharmacology, laboratories.5 Biliary decompression was per-
University of Berne, section to the glass slide was found to be of
Murtenstrasse 35, 3010 utmost importance because the eVect of formed by a Roux-en-Y biliodigestive anasto-
Berne, Switzerland DEPC was abolished on silanised slides. mosis as described previously.6 To induce acute
B J Stähelin Conclusions—False positive staining in hepatocyte necrosis, CCl4 was given by gavage
U Marti the TUNEL assay in the liver is caused by (CCl4 in corn oil (1/4 vol/vol) 2 ml/kg of body
M Solioz the release of endogenous endonucleases weight); studies were performed 24 hours after
H Zimmermann intoxication.3 All animal experiments had been
J Reichen as a result of proteinase treatment. This
can be abolished by pretreatment of tissue approved by the State Board on Ethics in Ani-
Correspondence to: slides with DEPC. mal Experimentation following strict inter-
Professor Reichen. (J Clin Pathol: Mol Pathol 1998;51:204–208) national guidelines.
email: reichen@ikp.unibe.ch Organs (liver, thymus, and intestine) were
Accepted for publication Keywords: immunohistochemistry; in situ nick end obtained under phenobarbital anaesthesia
21 April 1998 labelling; apoptosis; liver (50 mg/kg intraperitoneally) and were immer-
False positive staining in the TUNEL assay 205

Figure 1 Results of the TUNEL assay in diVerent tissues; all slides are counterstained with haematoxylin and shown at a
final magnification of ×250. (A) Standard TUNEL assay in liver after 45 minutes of proteinase K pretreatment showing
an obvious false positive reaction in up to 75% of all nuclei. (B) After DEPC pretreatment only cells meeting the
morphological criteria of apoptosis stained positive. (C) Detection of DNA fragmentation by the standard TUNEL assay
in liver tissue slides 24 hours after CCl4 intoxication demonstrating false positive staining in ballooned hepatocytes and in
apparently normal hepatocytes. (D) In the modified TUNEL assay, this false positive staining is abolished by DEPC
pretreatment. (E) TUNEL assay in intestine without DEPC pretreatment. Endogenous alkaline phosphatase activity is
seen along the apical section of the epithelial cell lining; furthermore, the nuclei also show some false positive staining. (F)
After DEPC pretreatment, the endogenous alkaline phosphatase is inhibited completely and no false positive staining of
nuclei occurs.

sion fixed immediately in buVered formalde- Thereafter, the sections were incubated for 90
hyde (4% vol/vol). Liver tissue was subjected to minutes at 37°C with terminal deoxyribonucle-
systematic random sampling8 for stereological otidyl transferase (75 U/ml) and digoxigenin-
analysis, as described previously.5 After dehy- 11-dUTP (5 nmol/ml) in potassium cacodylate
dration, tissues were embedded in paraYn wax buVer (200 mmol/l; pH 8.0) containing serum
and cut into 4–6 µm sections. bovine albumin (50 µg/ml) and CoCl2
(2.5 mmol/l). After 90 minutes, the slides were
TUNEL ASSAY washed with SSC buVer (150 mmol/l NaCl,
The assay was performed as described by 15 mmol/l sodium citrate, pH 7.0), followed by
Gavrieli et al.2 The paraYn sections were stuck Tris/HCl (10 mmol/l, pH 8.2) in 150 mmol/l
on to glass slides with cement (Cementit weiss, NaCl. Non-specific binding was blocked with
5% in distilled water; Merz and Benteli SA, the blocking reagent from Boehringer Man-
Niederwangen, Switzerland), dewaxed with nheim (for nucleic acid hybridisation and detec-
xylene, and rehydrated through a series of tion) for 30 minutes at room temperature.
decreasing concentrations of ethanol. Next, Labelled nick ends of DNA strands were visual-
the slides were partially digested with protein- ised with the alkaline phosphatase reaction,
ase K (10 µg/ml; Boehringer Mannheim, using Fab fragments against digoxigenin linked
Mannheim, Germany) in Tris/HCl buVer to alkaline phosphatase, and fast red chromogen
(20 mmol/l, pH 8.1) containing EDTA as a substrate (Boehringer Mannheim). The
(5 mmol/l) at 37°C for up to 60 minutes; reaction was stopped after 15–30 minutes by
digestion was stopped with H2O and the slides washing with H2O. Slides were counterstained
were washed four times in distilled water. lightly with haematoxylin.
206 Stähelin, Marti, Solioz, et al

100

75
Positive cells (%)

50

25

0
15 30 45 60
Time (min)
Figure 2 The eVect of proteinase K on the outcome of the TUNEL assay without
pretreatment (closed circle) and after DEPC pretreatment (open triangle) of liver tissue
slides. Positive nuclei were quantified stereologically; mean (SD) values are given.

To inhibit interfering enzyme activities, after

dewaxing, the slides were incubated in an eth-
anolic solution of diethyl pyrocarbonate
(DEPC; 4% vol/vol) for 30 minutes at 4°C,
after which the procedure described above was Figure 3 DNA digestion assay. Lane 1, genomic human
DNA without an incubation step; lane 2, DNA incubated for
carried out. In another set of experiments, 12 hours with proteinase K at 37°C; lane 3, DNA incubated
digestion was carried out with pepsin (0.5% on liver tissue slides for 12 hours without proteinase K; lane 4,
wt/vol) in HCl (0.1 mmol/l) and NaCl DNA after DEPC pretreatment with proteinase K; lane 5,
DNA without DEPC pretreatment in the presence of
(0.1 mmoles/l) instead of proteinase K. proteinase K; lane M, molecular weight markers.
As a positive control, DNA nicks were
induced by incubation of the slides with DATA AND STATISTICAL ANALYSIS
DNAase I (Boehringer Mannheim), as pro- The volume fraction of TUNEL positive cells
posed by Wijsman et al.9 After digestion with was quantified by a point counting procedure11
proteinase K, the slides were washed twice with as described previously.5 TUNEL positive cells
Tris/HCl (10 mmol/l, pH 8.1) in 150 mmol/l were counted similarly and related to the
NaCl, and then incubated with DNAase I volume fraction of the cell population under
(0.2 µg/ml) in Tris/HCl (10 mmol/l, pH 7.4) consideration (bile duct epithelial cells). Re-
containing NaCl (10 mmol/l), MgCl2 (5 mmol/ sults are reported as mean (standard deviation
l), CaCl2 (0.1 mmol/l), and KCl (25 mmol/l) (SD)). Mean values from diVerent groups were
for 15 minutes at 37°C. Further processing was compared by analysis of variance.12
as described above.
Results
In control liver, less than one in 1000 cells
ENDOGENOUS NUCLEASE ACTIVITY exhibited features of apoptosis in haematoxylin
DNA digestion by proteinase K and endog- and eosin stained sections using the criteria of
enous endonucleases was assessed as follows: Kerr.13 In contrast, using the standard TUNEL
untreated tissue slides and slides pretreated assay, up to 20% of cells stained positive, most
with DEPC were incubated for 24 hours at of them in the absence of related morphologi-
37°C with Tris/EDTA buVer containing ge- cal features of apoptosis, in particular conden-
nomic DNA (1 µg/ml) and proteinase K sation of chromatin at the nuclear membrane
(10 µg/ml). Control slides were incubated (fig 1A).
under the same conditions with DNA alone. To DiVerent factors that could be responsible
assess whether proteinase K had DNAse activ- for this false positive staining were investigated.
ity, it was incubated with DNA under similar Replacing proteinase K with pepsin as pro-
conditions but without exposure to tissue. posed by Wijsman and colleagues9 gave identi-
After incubation, 200 µl of the incubation cal results. Performing a time course of
solution was used for extraction of DNA. DNA incubation with proteinase K showed no posi-
extraction was performed with phenol/ tive cells with an incubation of up to 30
chloroform and precipitation was performed minutes; thereafter, there was a linear increase
with ethanol.10 DNA was then dissolved in in cell positivity, with all cells becoming
Tris/EDTA buVer (10 mM Tris-HCl, 1 mM positive after 45 minutes. However, the per-
EDTA, pH 7.5). Aliquots (1 µg) of DNA were centage of positive cells varied greatly from one
characterised by electrophoresis in a 0.8% aga- batch to another, making standardisation of
rose gel containing Tris/borate/EDTA buVer proteinase K treatment impossible (fig 2).
and ethidium bromide. Gels were run at 100 V This non-specific DNA degradation was not
for 45 minutes and were then photographed caused by proteinase K because incubation of
under UV light. DNA with proteinase K did not induce DNA
False positive staining in the TUNEL assay 207

Evolution of TUNEL positive bile duct epithelial cells after city of the method is obvious because terminal
biliary decompression in rats with ligation/dissection of the transferase (the enzyme used to perform the
common bile duct
nick end labelling) will recognise any 3' nick
Time after surgery (days) TUNEL positive cells (%) ends, regardless of their origin. This is because
unlike all other DNA polymerases, terminal
0 <0.1
1 5.4 (1.5)* deoxyribonucleotidyl transferase does not re-
2 1.3 (1.1)* quire template instruction for polymerisation.20
3 3.5 (4.4)* This is exemplified by the study of Grasl-
7 <0.1
Kraupp et al, who were unable to diVerentiate
Results (mean (SD)) were analysed by analysis of variance between necrotic, autolytic, or apoptotic cell
compared to day 0. death using the TUNEL assay.3
*Significant diVerence (p < 0.05).
In attempting to quantify apoptotic bodies
degradation (fig 3). In contrast, there was some with the TUNEL assay, we found that protein-
spontaneous DNA degradation when DNA ase K digestion—introduced into the assay to
was incubated with tissue sections pretreated make cleaved DNA accessible to the terminal
or not with DEPC (fig 3). When the complete transferase—induced a rapid increase in posi-
procedure including proteinase K treatment tive nuclei and that this step could not be con-
was performed on pretreated and untreated trolled reproducibly. This has been noted pre-
tissue sections, it became evident that DEPC viously by another group of investigators.9
almost completely inhibited DNA degradation According to their suggestion, we replaced
by these putative endogenous endonucleases proteinase K with pepsin; this did not abolish
(fig 3). false positive staining, suggesting that, at least
Therefore, the TUNEL procedure was in liver, there is a general eVect of proteases on
modified by inclusion of an inactivation step the outcome of the TUNEL assay, perhaps
with DEPC as described in the methods related to the release of endogenous endonu-
section. Mounting the sections with cement cleases. It is known that the secondary
proved to be crucial because silanised slides structure of proteins remains unchanged after
(which are often used in in situ hybridisation) formaldehyde fixation21 and alkaline phos-
prevented the inactivating eVect of DEPC phatase remains active in formalin fixed tissue
(data not shown). Pre-incubation of the slides (fig 1E). Therefore, we attempted to inhibit
with DEPC abolished the false positive staining such putative endonucleases by pretreatment
completely: only cells with clear features of of the histology slides with DEPC, a non-
apoptosis stained positive in the TUNEL assay specific inhibitor of DNAases and RNAases
(fig 1B). Accordingly, stereological analysis of commonly used in molecular biology. DEPC
the involution of bile duct epithelial cells was pretreatment abolished the false positive
parallelled by a wave of apoptosis (table 1). staining—positive cells all exhibited the mor-
Even more rewarding is the fact that DEPC phological features of apoptosis. Moreover,
pretreatment eliminated false positive staining DEPC pretreatment also abolished the false
in necrosis (fig 1C) induced by acute CCl4 positive staining in necrosis induced by CCl4,
intoxication (fig 1D), in contrast to the suggesting that the proposed modification
conventional TUNEL assay reported by Grasl- might be able to overcome the lack of
Kraupp et al.3 specificity of the TUNEL assay.
In addition, in the intestine, there were many It is of interest that not all tissues were
false positive cells and a strong reaction caused vulnerable to proteinase digestion. In the intes-
by the endogenous alkaline phosphatase of the tine there was false positive nuclear staining
brush border (fig 1E), which was eliminated by and a strong reaction of endogenous alkaline
pretreatment with DEPC (figure 1F). How- phosphatase with the chromogen; both reac-
ever, thymus and mammary gland were found tions were abolished by DEPC treatment. In
to be insensitive to DEPC pretreatment (data contrast, neither thymus nor involuting mam-
not shown). mary gland showed a false positive TUNEL
reaction, which might be related to varying
Discussion amounts of the putative endogenous endonu-
Apoptosis is a common mode of cell death in cleases in diVerent organs.
diVerent liver diseases.14 DNA laddering is con- DEPC did not aVect the accessibility of
sidered to be the hallmark of apoptosis15 and it is nuclear DNA to the terminal transferase as
detected most reliably by gel electrophoresis, shown by positive staining of all nuclei in tissue
where a characteristic laddering pattern is seen, sections treated with DNAse, which should be
which results from cleavage of DNA into strands included as a positive control.9 Finally, it
of about 200 kilo bases in length.16 Even though should be noted that the method of gluing the
it occurs late in the process, it is considered to be tissue section to the glass slide is of utmost
one, if not the most, valuable feature for the importance and that in particular the eVect of
detection of cell death by apoptosis14 17 18; DEPC is abolished on silanised slides.
however, it cannot be demonstrated in all Bile duct ligation in the rat induces marked
experimental settings.19 Furthermore, the ductular proliferation,4 5 which (in addition to
method cannot be applied to archival, paraYn the associated fibrosis) is reversible after biliary
wax embedded material. Therefore, the decompression.6 The hyperplastic bile duct
TUNEL assay developed by Gavrieli and epithelial cells are removed by apoptosis, as
colleagues2 is widely used because it permits in demonstrated morphologically by Bhathal et
situ visualisation of DNA cleavage by inserting al.7 The data in table 1 confirm and quantitate
a marker at the 3' nick end. The lack of specifi- this finding, showing that there is a rapid wave
 208 Stähelin, Marti, Solioz, et al

of TUNEL positive cells leading to removal of 7 Bhathal PS, Gall JAM. Deletion of hyperplastic biliary epi-
thelial cells by apoptosis following removal of the prolifera-
most bile duct epithelial cells within a week. tive stimulus. Liver 1985;5:311–25.
We conclude that the inclusion of an 8 Cruz-Orive LM, Weibel ER. Sampling design for stereology.
J Microsc 1981;122:235–7.
incubation step with DEPC inactivates endog- 9 Wijsman JH, Jonker RR, Keijzer R, et al. A new method to
enous endonucleases that can give false detect apoptosis in paraYn sections: in situ end labeling for
fragmented DNA. J Histochem Cytochem 1993;41:7–12.
positive results in the TUNEL assay in liver 10 Anonymous. Short protocols in molecular biology. New York:
and intestine. This step also should allow a John Wiley and Sons, 1992.
better diVerentiation between death by necro- 11 Weibel ER. Stereological methods. London: Academic Press,
1979.
sis and apoptosis. 12 Snedecor GW, Cochran WG. Statistical methods. Iowa: Iowa
State University Press, 1967.
13 Kerr JFR. Shrinkage necrosis: a distinct mode of cellular
Supported by a grant from the Swiss National Foundation for death. J Pathol 1971;105:13–20.
Scientific Research (No. 32.45349.95) to JR. 14 Patel T, Gores GJ. Apoptosis and hepatobiliary disease.
Hepatology 1995;21:1725–41.
1 Que FG, Gores GJ. Cell death by apoptosis: basic concepts 15 Alles A, Alley K, Barrett JC, et al. Apoptosis: a general com-
and disease relevance for the gastroenterologist. Gastroen- ment. FASEB J 1991;5:2127–8.
terology 1996;110:1238–43. 16 Wyllie AH. Glucocorticoid induced thymocyte apoptosis is
2 Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of associated with endogeneous endonuclease activation.
programmed cell death in situ via specific labelling of Nature 1980;284:555–6.
nuclear DNA fragmentation. J Cell Biol 1992;119:493– 17 RaV MC. Social control on cell survival and cell death.
501. Nature 1992;356:397–9.
3 Grasl-Kraupp B, Ruttkay-Nedecky B, Koudelka H, et al. In 18 Williams GT, Smith C. Molecular regulation of apoptosis:
situ detection of fragmented DNA (TUNEL assay) fails to genetic control of cell death. Cell 1993;74:777–9.
discriminate among apoptosis, necrosis, and autolytic cell 19 Oberhammer F, Bursch W, Tiefenbacher R, et al. Apoptosis
death: a cautionary note. Hepatology 1995;21:1465–8. is induced by transforming growth factor â1 within 5 hours
4 Kontouras J, Billing BH, Scheuer PJ. Prolonged bile duct in regressing liver without significant fragmentation of the
obstruction: a new experimental model for cirrhosis in the DNA. Hepatology 1993;18:1238–46.
rat. Br J Exp Pathol 1984;65:305–11. 20 Gorczyca W, Gong J, Darzynkiewicz Z. Detection of DNA
5 Gross JB, Reichen J, Zeltner T, et al. The evolution of strand breaks in individual apoptotic cells by in situ termi-
changes in quantitative liver function tests in a rat model of nal deoxyribonucleotidyl transferase and nick translation
cirrhosis: correlation with morphometric measurement of assay. Cancer Res 1993;53:1945–51.
hepatocyte mass. Hepatology 1987;7:457–63. 21 Mason JT, O’Leary TJ. EVects of formaldehyde fixation on
6 Zimmermann H, Reichen J, Zimmermann A, et al. Revers- protein secondary structure: a calorimetric and infrared
ibility of secondary biliary fibrosis by biliodigestive anasto- spectroscopic investigation. J Histochem Cytochem 1991;39:
mosis in the rat. Gastroenterology 1992;103:579–89. 225–9.

View publication stats