Tyramide Signal Amplification Method in Multiple-
Label Immunofluorescence Confocal Microscopy
Guoji Wang, Cristian L. Achim, Ronald L. Hamilton, Clayton A. Wiley,
and Virawudh Soontornniyomkij
Department of Pathology (Neuropathology), University of Pittsburgh School of Medicine,
Pittsburgh, Pennsylvania 15213
The tyramide signal amplification (TSA) method has recently
been introduced to improve the detection sensitivity of immuno-
histochemistry. We present three examples of applying this
method to immunofluorescence confocal laser microscopy: (1)
single labeling for CD54 in frozen mouse brain tissue; (2) double
labeling with two unconjugated primary antibodies raised in the
same host species (human immunodeficiency virus type 1 p24
and CD68) in paraffin-biopsied human lymphoid tissue; and (3)
triple labeling for brain-derived neurotrophic factor, glial fibrillary
acidic protein, and HLA-DR in paraffin-autopsied human brain
tissue. The TSA method, when properly optimized to individual
tissues and primary antibodies, is an important tool for immuno-
fluorescence microscopy. Furthermore, the TSA method and en-
zyme pretreatment can be complementary to achieve a high
detection sensitivity, particularly in formalin-fixed paraffin-
embedded archival tissues. Using multiple-label immunofluores-
cence confocal microscopy to characterize the cellular localiza-
tion of antigens, the TSA method can be critical for double
labeling with unconjugated primary antibodies raised in the same
© 1999 Academic Press
Detection sensitivity may be an important limitation
of immunohistochemistry. Formalin fixation and par-
affin embedding, while preserving excellent histology,
can adversely affect immunohistochemical detection of
antigens. A variety of antigen retrieval procedures
have been employed to unmask antigenic epitopes
within tissues, e.g., enzyme pretreatment (trypsin,
pepsin, pronase, proteinase K), alkaline hydrolysis,
formic acid treatment, and high-temperature heating
(1, 2). Recently, the tyramide signal amplification
(TSA) method, originally described by Bobrow et al. as
“catalyzed reporter deposition,” has been used to im-
prove detection sensitivity in a variety of techniques
including enzyme immunoassays, Western blotting, in
situ hybridization, and immunohistochemistry (3–9).
This method has also been successfully applied to im-
embedded autopsied brain tissue for detecting human
immunodeficiency virus type 1 (HIV-1) p24 (10) and
brain-derived neurotrophic factor (BDNF) (11). The
amplification is achieved by biotin- or fluorophore-
conjugated tyramide which acts as substrate for horse-
radish peroxidase (HRP). The HRP reacts with hydro-
gen peroxide and the phenolic portion of tyramide to
produce highly reactive tyramide radicals that then
covalently bind toelectron-rich moieties (e.g., tyrosine)
within tissues (3). Owing to the extremely short half-
life of tyramide radicals, only tyrosine residues in close
vicinity of the HRP will bind tyramide (5). The biotin-
conjugated tyramide can
fluorophore- or enzyme-conjugated streptavidin. Pri-
mary antibodies can be applied for short incubation
times or at low concentrations, resulting in decreased
background. In general, both thebiotin-conjugated and
fluorophore-conjugated tyramides can significantly im-
proveimmunohistochemical detection sensitivity (6–8,
Multiple immunofluorescence (IF) labeling can usu-
ally be achieved with conventional methods using dif-
ferent fluorophore-conjugated secondary antibodies, if
theprimary antibodies used areraised in different host
species. In detecting two unconjugated primary anti-
bodies raised in the same species, recognition of both
primary antibodies by each of the two different
fluorophore-conjugated secondary antibodies is a major
concern. With application of the TSA method, at a very
low concentration of the primary antibody, the antigen
cannot be detected by a conventional fluorophore-
conjugated secondary antibody, but is detectable after
then be visualized by
METHODS 18, 459–464 (1999)
Article ID meth.1999.0813, available online at http://www.idealibrary.com on
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
F IG. 1.
method (B), modified with permission from the specification sheet of TSA-Direct (NEN Life Science Products, Boston, MA). (C, D)
Immunofluorescence labeling of CD54 in frozen sections of ts-1 virus-infected mouse brain tissue. At the same concentration of biotinylated
anti-CD54 antibody, the CD54 signal on vascular endothelia is barely visible with the conventional immunofluorescence method (C), but is
distinctly visible with the tyramide signal amplification method (D). Bar ? 20 ?m.
(A, B) Schematic representations of the conventional immunofluorescence method (A) and of the tyramide signal amplification
WANG ET AL.
TSA. Therefore, another primary antibody raised in
the same host species can be applied and visualized
with a different fluorophore in subsequent conven-
tional IF labeling on the same tissue section (13, 14).
According to Hunyady et al., successful double immu-
nolabeling can be presumed when the first antigen can
be clearly detected by the TSA method at a concentra-
tion of primary antibody at least 10-fold lower than
that used in the conventional IF method (14).
In our laboratory, we have successfully applied the
TSA method to immunohistochemistry of neural and
lymphoid tissues. Using this method, we are able to
demonstrate the cellular localization of several anti-
gens by immunofluorescence confocal microscopy in
both frozen and paraffin tissues. Schematic represen-
tations of the conventional IF method and the TSA
method are shown in Figs. 1A and 1B, respectively.
Further, we present three examples of using the TSA
method: (1) single IF labeling in frozen mouse brain
tissue (Figs. 1C and 1D); (2) double IF labeling with
unconjugated primary antibodies raised in the same
host species in paraffin-biopsied human lymphoid tis-
sue (Fig. 2); and (3) triple IF labeling in paraffin-
autopsied human brain tissue (Fig. 3).
DESCRIPTION OF METHOD
Tissue sections were mounted on Fisherbrand Su-
perfrost Plus glass slides (Fisher Scientific, Pitts-
burgh, PA) to ensure adequate tissue adherence dur-
ing immunohistochemical procedures. Specifications
and sources of the primary and secondary antibodies
and critical reagents used are listed in Tables 1 and
2. The commercially available kit Tyramide Signal
Amplification (TSA-Direct, NE N Life Science Prod-
ucts, Boston, MA) was used in the experiments de-
scribed below. TNB buffer (0.1 M Tris–HCl, pH 7.5,
0.15 M NaCl, 0.5% Blocking Reagent) and TNT
buffer (0.1 M Tris–HCl, pH 7.5, 0.15 M NaCl, 0.05%
Tween 20) were prepared according to the manufac-
turer’s protocols (NE N Life Science Products). The
negative reagent controls were generated by replac-
ing primary antibodies with normal sera from the
same host species with equivalent protein concentra-
rinsed in phosphate-buffered saline (PBS, 3 ? 5 min)
and then mounted with Gelvatol. This water-soluble
mounting medium can withstand the heat generated
during laser-scanning microscopy. The “in-house”
recipe for Gelvatol was modified from Current Proto-
cols in Molecular Biology (15). Briefly, polyvinyl al-
cohol (20 g) was added slowly to PBS (100 ml) while
constantly stirring [room temperature (RT), 4–6 h].
The solution was covered and under continuous stir-
ring (4°C, overnight). Sodium azide (0.03 g) and glyc-
erol (50 ml) were added and mixed thoroughly. The
solution was then centrifuged at 10,000 rpm (4°C, 20
min), aliquoted into 10-ml syringes, and stored at
4°C until use.
F IG. 2.
conventional method, green). Note that both unconjugated primary antibodies were raised in mice. Multinucleated giant cells in tonsillar
lymphoid tissue from an HIV-1-infected individual, labeled with a macrophage marker CD68, express HIV-1 p24. Bar ? 20 ?m.
F IG. 3. Triple immunofluorescence labeling of BDNF (with the tyramide signal amplification method, green), GFAP (with the conventional
method, red), and HLA-DR (with the conventional method, blue). A senile plaque in brain tissue from an individual with Alzheimer’s disease
contains an HLA-DR-labeled microglial cluster and BDNF-labeled dystrophic neurites (arrowheads) within its core. GFAP-labeled astrocytes
are located at the periphery of the plaque and extend their processes (arrows) into the plaque core. Bar ? 20 ?m.
Double immunofluorescence labeling of HIV-1 p24 (with the tyramide signal amplification method, red) and CD68 (with the
TYRAMIDE SIGNAL AMPLIFICATION IN CONFOCAL MICROSCOPY
Single Immunofluorescence Labeling for CD54
in Frozen Mouse Brain Tissue
Conventional Method (Fig. 1C)
Hemispheres of ts-1 virus (a temperature-sensitive
mutant of Moloney murine leukemia virus)-infected
mouse brains were fixed in 2% paraformaldehyde (2–4
h) and then cryoprotected by immersing in 30% sucrose
(4°C, 2 days). The tissues were frozen on dry ice and
20-?m-thick tissue sections were taken. After being
rinsed in PBS (6 ? 5 min), the sections were incubated
with biotinylated hamster anti-mouse CD54 antibody
(1:1000 dilution in PBS with 0.01% Tween 20, (4°C,
overnight). After being rinsed in PBS (3 ? 5 min), the
sections were incubated with FITC-conjugated strepta-
vidin (1:200 in PBS with 0.01% Tween 20) (RT, 1 h).
Tyramide Signal Amplification (Fig. 1D)
The frozen sections were rinsed in PBS (6 ? 5 min)
and then incubated with biotinylated hamster anti-
mouse CD54 antibody (1:1000 in PBS with 0.01%
Tween 20, (4°C, overnight). After being rinsed in PBS
(3 ? 5 min), the sections were incubated with TNB
buffer (RT, 30 min), excess TNB buffer was blotted,
and then the sections were incubated with HRP-
conjugated streptavidin (1:500 in TNB Buffer), (RT, 30
min). The sections were rinsed in PBS (3 ? 5 min) and
then incubated with FITC-conjugated tyramide (1:100
in 1? Amplification Diluent), (RT, 10 min).
Double-Immunofluorescence Labeling for HIV-1 p24 and
CD68 in Paraffin-Biopsied Tonsillar Lymphoid Tissue
Using Unconjugated Primary Antibodies Raised in the
Same Species (Fig. 2)
Formalin-fixed paraffin-embedded sections were de-
waxed (3 ? 5 min), rehydrated in serial ethanol,
treated with 3% hydrogen peroxide in methanol (30
min), rinsed in water, treated with 0.1% Pronase
(37°C, 10 min), rinsed in water, rinsed in PBS (5 min),
and incubated with 10% normal goat serum (RT, 20
min), and excess serum was drained. The sections were
incubated with mouse anti-HIV-1 p24 antibody (1:100,
(4°C, overnight). At this concentration of anti-HIV-1
p24 antibody, the HIV-1 p24 antigen could not be de-
tected by Cy2-conjugated goat anti-mouse IgG serum
(1:200, (RT, 1 h) in the conventional IF method (in
which the optimal dilution of anti-HIV-1 p24 antibody
was 1:5), but was still detectable after TSA.
The sections were rinsed in PBS with agitation (3 ?
5 min) and then incubated with biotinylated goat anti-
mouse IgG serum (1:500, (RT, 30 min). After being
rinsed in TNT with agitation (3 ? 5 min), the sections
were incubated with TNB buffer (RT, 30 min), excess
TNB buffer was blotted, and then the sections were
incubated with HRP-conjugated streptavidin (1:500 in
TNB Buffer, (RT, 30 min). The sections were rinsed in
TNT with agitation (3 ? 5 min) and then incubated
with TRITC-conjugated tyramide (1:100 in 1? Ampli-
fication Diluent, RT, 10 min). After being rinsed in
TNT with agitation (3 ? 5 min) and in PBS (5 min), the
sections were incubated with mouse anti-human CD68
antibody (1:100, RT, 2 h). The sections were rinsed in
PBS (3 ? 5 min) and then incubated with Cy2-
conjugated goat anti-mouseIgG serum (1:200, RT, 1 h).
Triple Immunofluorescence Labeling for BDNF, GFAP, and
HLA-DR in Paraffin-Autopsied Alzheimer’s Disease Brain
Tissue (Fig. 3)
treated with thesameprotocol as described abovefor the
lymphoid tissue, except that instead of being incubated
with Pronase, the sections were immersed in preheated
Target Retrieval Solution (95–99°C, 1 h). The sections
were incubated with a mixture of rabbit anti-human
BDNF antibody (1:400) and mouse anti-human HLA-DR
antibody (neat, 4°C, overnight). At this concentration of
anti-BDNF antibody, the BDNF antigen could not be
detected by Cy3-conjugated goat anti-rabbit IgG serum
(1:100, RT, 1 h) in the conventional IF method (in which
the optimal dilution of anti-BDNF antibody was 1:20),
but was still detectable after TSA.
The sections were rinsed in PBS with agitation (3 ? 5
min) and then incubated with biotinylated goat anti-
rabbit IgG serum (1:200, RT, 30 min). After being rinsed
in TNT with agitation (3 ? 5 min), the sections were
incubated with TNB buffer (RT, 30 min), excess TNB
buffer was drained, and then thesections wereincubated
with HRP-conjugated streptavidin (1:400 in TNB buffer,
RT, 30 min). The sections were rinsed in TNT with agi-
tation (3 ? 5 min) and then incubated with FITC-
T ABL E 1
Antibody toClone (isotype)Manufacturer
Kal-1 (mouse IgG1)
San Diego, CA
Santa Cruz, CA
PG-M1 (mouse IgG3)
N-20 (rabbit IgG)
LN-3 (mouse IgG2b)
aICAM-1, intercellular adhesion molecule-1; HIV-1, human im-
munodeficiency virus type 1; BDNF, brain-derived neurotrophic fac-
tor; GFAP, glial fibrillary acidic protein.
WANG ET AL.
conjugated tyramide (1:100 in 1? Amplification Diluent,
RT, 10 min). After being rinsed in TNT with agitation
(3 ? 5 min) and in PBS (5 min), the sections were incu-
bated with rabbit anti-bovineGFAP antibody (1:100, RT,
2 h). The sections were rinsed in PBS (3 ? 5 min) and
then incubated with a mixture of Cy5-conjugated goat
anti-mouse IgG serum (1:100) and Cy3-conjugated goat
anti-rabbit IgG serum (1:100, RT, 1 h).
The IF sections were analyzed with a Molecular Dy-
namics confocal laserscanning microscope (Sunnyvale,
CA). This instrument is equipped with a Nikon in-
verted microscope with Plan-Apo 20? 0.75-N.A. (air),
40? 1.00-N.A. (oil), and 60? 1.40-N.A. (oil) objective
lenses. The illumination is provided by the argon/
krypton laser with 488-, 568-, and 647-nm primary
emission lines. Each image was scanned along the z
axis for a total of 10 sectional planes with a 0.5-?m step
(512 ? 512 pixels per sectional plane, 0.34 ? 0.34 ?m
per pixel). Images were collected on a Silicon Graphics
Inc. computer (Operating System Release 5.3, Farm-
ington, MI) and analyzed using the Image Space soft-
ware (Version 3.2, Molecular Dynamics). All multiple-
label IF images are 10-section projections.
For triple labeling with FITC, Cy3, and Cy5 as an
example, the specimen was first scanned for FITC and
Cy5 signals. Using a 488/647 dual-bandpass laser
wavelength filter, FITC and Cy5 were excited by the
488-nm line and the 647-nm line, respectively. Fluo-
rescent light emitted by FITC and Cy5 passed through
a 488/647 primary dual dichroic beamsplitter, and then
was separated by a 650-nm secondary dichroic beam-
splitter. Emitted light from FITC and Cy5 passed
through a 530DF30 (between 515 and 545 nm) band-
pass filter and a 660-nm long-pass filter, respectively.
Subsequently, the specimen was rescanned for Cy3
signal with the same starting plane and ending plane
along the z axis. Using a 568-nm bandpass laser wave-
length filter, Cy3 was excited by the 568-nm line. Flu-
orescent light emitted by Cy3 passed through a 488/
568 primary dual dichroic beamsplitter, and then
passed through a 600DF40 (between 580 and 620 nm)
bandpass filter. No secondary dichroic beamsplitter
was used at this step.
The use of fluorophore-conjugated TSA can signif-
icantly improve the detection sensitivity in IF mi-
croscopy. Nevertheless, the variable reactivity of dif-
ferent markers with the TSA method underlines the
necessity for individual testing of every primary an-
tibody candidate to determine the optimal protocol
for each type of tissue and fixation. With this
method, both specific signal and nonspecific back-
ground are amplified. Important staining variables
that need to be optimized include concentrations of
primary antibodies, biotinylated secondary antibod-
ies, HRP-conjugated streptavidin, and fluorophore-
conjugated tyramide, incubation times, and temper-
ature. In general, nonspecific binding of primary
antibodies is diminished by the use of concentrations
lower than those necessary to achieve comparable
signal intensity with conventional IF methods (6).
Consistent with the report of Adam et al. (6), we
found that biotinylated secondary antibodies, al-
though affinity-purified, can contribute to significant
background, probably due to their nonspecific bind-
T ABL E 2
Secondary Antibodies and Reagents
Secondary antibodies and reagentsa
Biotinylated goat anti-mouse IgG (H ? L)
Biotinylated goat anti-rabbit IgG (H ? L)
Cy2-conjugated goat anti-mouse IgG (H ? L) (excitation: 490 nm, emission: 508 nm)
Cy3-conjugated goat anti-rabbit IgG (H ? L) (excitation: 550 nm, emission: 570 nm)
Cy5-conjugated goat anti-mouse IgG (H ? L) (excitation: 650 nm, emission: 670 nm)
FITC-conjugated streptavidin (excitation: 494 nm, emission: 517 nm)
TRITC-conjugated tyramide (excitation: 550 nm, emission: 570 nm)
Target Retrieval Solution
Caltag Labs, San Francisco, CA
J ackson ImmunoResearch Labs, West Grove, PA
J ackson ImmunoResearch Labs
J ackson ImmunoResearch Labs
J ackson ImmunoResearch Labs
NEN Life Science Products, Boston, MA
NEN Life Science Products
NEN Life Science Products
NEN Life Science Products
Dako Corp., Carpinteria, CA
aAll secondary antibodies wereaffinity-purified. Noteapproximatepeak wavelengths of excitation and emission for fluorophore-conjugated
secondary antibodies. Cy3, indocarbocyanine; Cy5, indodicarbocyanine; FITC, fluorescein isothiocyanate; TRITC, tetramethylrhodamine
isothiocyanate; HRP, horseradish peroxidase.
TYRAMIDE SIGNAL AMPLIFICATION IN CONFOCAL MICROSCOPY
ing to the tissue. We overcome this issue by optimiz- Download full-text
ing the concentration of secondary antibodies, as
well as by incubating the tissue sections (prior to the
incubation with primary antibodies) with normal
sera from the same species as secondary antibodies.
To achieve a high detection sensitivity, particularly
in formalin-fixed paraffin-embedded archival tis-
sues, some antigen retrieval procedures may be
needed to complement the TSA method (9–11).
The TSA method is a promising tool when conven-
tional IF methods fail to detect specific signals. Confo-
cal microscopy is an essential method in characterizing
the cellular localization of antigens using multiple IF
labeling in the same tissue section. Since in many
circumstances the availability of primary antibodies is
limited, the TSA method can be critical for double IF
labeling with unconjugated primary antibodies raised
in the same host species.
This work was supported by the NIH Grants NS35731 and
MH46790 to C.L.A. and C.A.W.
1. McQuaid, S., McConnell, R., McMahon, J ., and Herron, B. (1995)
J . Pathol. 176, 207–216.
2. Shi, S. R., Cote, R. J ., and Taylor, C. R. (1997) J . Histochem.
Cytochem. 45, 327–343.
3. Bobrow, M. N., Harris, T. D., Shaughnessy, K. J ., and Litt, G. J .
(1989) J . Immunol. Methods 125, 279–285.
4. Bobrow, M. N., Shaughnessy, K. J ., and Litt, G. J . (1991) J . Im-
munol. Methods 137, 103–112.
5. Bobrow, M. N., Litt, G. J ., Shaughnessy, K. J ., Mayer, P. C., and
Conlon, J . (1992) J . Immunol. Methods 150, 145–149.
6. Adams, J . C. (1992) J . Histochem. Cytochem. 40, 1457–1463.
7. Berghorn, K. A., Bonnett, J . H., and Hoffman, G. E. (1994)
J . Histochem. Cytochem. 42, 1635–1642.
8. van Gijlswijk, R. P., Zijlmans, H. J ., Wiegant, J ., Bobrow, M. N.,
Erickson, T. J ., Adler, K. E., Tanke, H. J ., and Raap, A. K. (1997)
J . Histochem. Cytochem. 45, 375–382.
9. Soontornniyomkij, V., Wang, G., K apadia, S. B., Achim, C. L.,
and Wiley, C. A. (1998) Arch. Pathol. Lab. Med. 122, 534–
10. Strappe, P. M., Wang, T. H., McK enzie, C. A., Lowrie, S.,
Simmonds, P., and Bell, J . E . (1997) J . Virol. Methods 67,
11. Soontornniyomkij, V., Wang, G., Pittman, C. A., Wiley, C. A., and
Achim, C. L. (1998) Neuropathol. Appl. Neurobiol., in press.
12. Von Wasielewski, R., Mengel, M., Gignac, S., Wilkens, L.,
Werner, M., and Georgii, A. (1997) J . Histochem. Cytochem. 45,
13. Shindler, K. S., and Roth, K. A. (1996) J . Histochem. Cytochem.
14. Hunyady, B., Krempels, K., Harta, G., and Mezey, E. (1996)
J . Histochem. Cytochem. 44, 1353–1362.
15. Zellor, R., and Rogers, M. (1996) in Current Protocols in Molec-
ular Biology (Ausubel, F. M., Brent, R., Kingston, R. E., Moore,
D. D., Seidman, J . G., Smith, J . A., and Struhl, K., Eds.), p.
14.5.4, Wiley, New York.
WANG ET AL.