
* School of Pathology, University of New South Wales, Sydney, Australia; and
Department of Anatomical Pathology, Prince of Wales Hospital, Randwick, Australia
Correspondence: R. K. Kumar, School of Pathology, University of New South Wales, Sydney, Australia 2052. E-mail: R.Kumar{at}unsw.edu.au
|
|
|---|
Key Words: acute inflammation neutrophil activation immuno-histochemistry S100 proteins
|
|
|---|
High levels of extracellular A8/A9 complex are suggested markers of disease activity in inflammatory disorders such as arthritis and ulcerative colitis [9 , 10 ]. Whether the complex is released by dying cells [11 ] or its presence is a result of active secretion remains unclear. The proteins lack signal sequences associated with classical secretory pathways, although a tubulin-dependent mechanism of secretion by monocytes is suggested [12 ]. There is no information available about mechanisms of secretion of A8 and/or A9 by neutrophils.
Our laboratory has a long-standing interest in a murine S100 protein, originally designated CP-10 [13 ], the nearest human counterpart to which is S100A8 [14 ]. Unlike human S100A8, from which it exhibits substantial sequence divergence, CP-10 is a potent chemoattractant for phagocytic cells [15 ]. However, like the human protein, it is expressed in high amounts by neutrophils and by activated macrophages [16 , 17 ] so that it has come to be known as murine S100A8 (mA8) [18 ]. Previously, we demonstrated that in murine experimental-acute alveolitis induced by bleomycin, large quantities of this protein are released into the extracellular fluid in a biologically active, monomeric form, apparently by recruited neutrophils [19 ]. In the present study, we have used antibodies to human and murine S100A8 to examine the expression of the A8 protein by human neutrophils, in vivo and in vitro. We show for the first time that homodimeric S100A8 is present in the cytoplasm of human neutrophils and that following phagocytosis, neutrophils exhibit rapid loss of cytoplasmic immunoreactivity for homodimeric S100A8, which may serve as a marker of localized neutrophil activation in tissues.
|
|
|---|
Immunoblotting
Recombinant S100A8, S100A9, or S100A12 (2 µg) purified from C4
reverse-phase HPLC [21
] was lyophilized, resuspended in
50 mM Tris-HCl, pH 7.5, and oxidized in 2 mM copper sulfate at room
temperature for 30 min. The formation of homo- or heterodimers was
confirmed by polyacrylamide gel electrophoresis (PAGE) and silver
staining. Samples were mixed with an equal volume of 2x loading
buffer, heated at 100°C for 3 min, and analyzed by 10% tricine
sodium dodecyl sulfate (SDS)-PAGE [22
]. Gels run in
parallel were silver stained or transferred to polyvinylidene
difluoride membranes (Millipore, Bedford, MA). Antigenic reactivity of
the S100 protein monomers and dimers to the rat mAb CP10.1 or the 85-C2
mouse mAb (BMA Biomedicals AG, Augst, Switzerland) was detected using
horseradish peroxidase (HRP)-conjugated, goat anti-mouse or anti-rat
IgG (Bio-Rad, Hercules, CA), respectively, and visualized by
chemiluminescence [enhanced chemiluminescence (ECL) system, Amersham
Australia, Sydney, Australia].
Immunostaining
Immunostaining was performed using a capillary-action staining
system as described previously [19
]. Cells expressing
cytoplasmic immunoreactivity for S100 A8 were identified using CP10.1
or 85-C2. For sections of archival, formalin-fixed blocks of tissue
from tuberculous lymph nodes and acute appendicitis, antigen retrieval
was performed by boiling in 0.01 M citrate buffer, pH 6.0, for 10 min.
Sections were washed with phosphate-buffered saline (PBS) containing
0.1% Triton X-100 and blocked with 50% normal rabbit serum for 10 min
at 37°C. They were then incubated for 30 min at 37°C with an
appropriate concentration of the primary antibody. Sections were washed
four times with PBS-Triton after this and each subsequent incubation.
Bound antibody was detected by sequential, 30-min incubations with
biotinylated rabbit anti-rat or anti-mouse IgG (Dako, Sydney,
Australia), followed by a complex of streptavidin (Boehringer Mannheim,
Sydney, Australia) and biotinylated HRP in PBS-1% bovine serum albumin
(BSA). Diaminobenzidine was used as the substrate, and sections were
counterstained lightly with hematoxylin. For cytospin preparations, a
similar procedure was used, but antigen retrieval was not required, and
nuclei were counterstained with methyl green.
Enzyme immunoassay
Wells of a microtitre plate (Nunc, Roskilde, Denmark) were
coated with 500 ng recombinant human or murine S100A8 (prepared
according the protocols described in ref. [16
] and
comprising a mixture of monomeric and dimeric forms) in
carbonate-bicarbonate buffer, pH 9.6, for 2 h at 37°C, washed
four times with Tris-buffered saline (TBS; pH 7.4) containing 0.01%
Tween-20, blocked with TBS-Tween containing 0.1% ovalbumin for 60 min
at 37°C, and washed again. To compare the binding of CP10.1 with the
two forms of S100A8, serial dilutions of mAb supernatant were added to
the wells in triplicate and incubated overnight at 4°C. To assess
competitive inhibition, a 1:25 dilution of CP10.1 antibody was added to
each well of a plate coated with human S100A8 in the presence of
0.050.5 µg/ml recombinant murine or human S100A8. The wells were
then washed four times, and bound antibody was detected by incubation
for 60 min at room temperature with biotinylated rabbit, anti-rat IgG,
followed by additional washes and incubation with streptavidin-HRP
complex (Amersham Australia). Enzyme activity was revealed by
incubation with ABTS substrate (Kirkegaard & Perry,
Gaithersburg, MD), and the reaction product was read using a Titertek
Multiskan photometer with a 405 nm filter.
Stimulation of neutrophils in vitro
Freshly collected blood samples were layered onto Mono-Poly
Resolving medium (Ficoll-Hypaque, ICN Biomedicals, Aurora, OH),
according to the manufacturers instructions and centrifuged at 600
g for 1 h. Neutrophils removed from the interface were
washed twice and resuspended in RPMI 1640 at 5 x 106
cells/ml.
Neutrophils were exposed to N-formyl-Met-Leu-Phe (fMLP; 10-7 M; Sigma) or to opsonized zymosan for 15 min or 2 h. Opsonized zymosan was prepared according to the method of Markert et al. [23 ] and used at dilutions ranging from 1:11:20. After incubation, aliquots of the neutrophil suspension were removed for assessment of viability using trypan blue exclusion and for preparation of cytospin slides. The latter were air-dried, fixed in 2% paraformaldehyde for 30 min, rinsed in PBS for at least 5 min, and stored at 4°C.
|
|
|---|
![]() View larger version (29K): [in a new window] |
Figure 1. Immunoblotting of monomeric and dimeric forms of S100A8, S100A9, and
S100A12 using CP10.1 and 85-C2 mAbs reacted against monomeric S100A8
(A8m), homodimeric S100A8 (A8d), monomeric S100A9 (A9m), homodimeric
S100A9 (A9d), a 1:1 mixture of oxidized S100A8 and S100A9 (A8/9), or
S100A12 (A12). (A) 85-C2 reacts with monomeric and dimeric forms of
S100A8 but not with S100A9 or S100A12. (B) CP10.1 reacts only with
homodimeric S100A8 and exhibits no cross-reactivity with S100A9 or
S100A12. (C) The corresponding silver-stained gel shows the migration
patterns of the S100 proteins. Note that A8 + 9 contains
homodimeric A8 (20 kDa), A8/A9 heterodimeric complex (24 kDa), and
homodimeric A9 (28 kDa).
|
![]() View larger version (17K): [in a new window] |
Figure 2. Binding serial dilutions of CP10.1 antibody to wells coated with 500 ng
recombinant human or murine S100A8 or to control wells. Values are
optical density at 405 nm (mean±SE of triplicate
samples).
|
![]() View larger version (143K): [in a new window] |
Figure 3. Immunolabeling leukocytes by antibodies to S100A8 in sections of human
lymph node involved by tuberculous lymphadenitis. (A) Using 85-C2
antibody, reactivity of intravascular neutrophils and tissue
macrophages in subcapsular tissue is apparent, whereas in B, only
neutrophils are labeled by CP10.1 antibody. (C) 85-C2 antibody labels
epitheloid cells in a tuberculous granuloma, as well as neutrophils in
the caseous necrotic center, and in D, only neutrophils exhibit
immunoreactivity with CP10.1 antibody.
|
Immunolabeling neutrophils in acute appendicitis by CP10.1
Using a polyclonal antibody to S100A8, we showed that in an
experimental inflammatory lesion in mice, neutrophils were apparently
activated to release S100A8, with loss of immunoreactivity in tissue
sections [19
]. To investigate whether a similar process
occurred in human inflammatory lesions, sections of acute appendicitis
were immunostained for S100A8 using CP10.1 antibody. An obvious
gradation of intensity of immunostaining of neutrophils was observed
from the serosa to the lumen. Intense staining of cells was noted in
the serosa, where the neutrophils are recruited as part of an exudative
inflammatory response (Fig. 4A
). Staining intensity diminished through the muscularis and the
mucosa, which are sites of bacterial invasion in established
appendicitis (Fig. 4B and 4C)
. In the lumen, which contained large
numbers of neutrophils participating in the suppurative inflammatory
response, virtually none of the cells exhibited positive staining (Fig. 4D) . Identical results with respect to immunostaining of neutrophils
were obtained with the 85-C2 antibody (unpublished results).
![]() View larger version (164K): [in a new window] |
Figure 4. Immunolabeling neutrophils for S100A8 in various layers of the appendix
in acute appendicitis, using CP10.1 antibody. (A) Intense reactivity of
neutrophils in the serosa is obvious, whereas in B, the staining of
neutrophils in between the smooth-muscle fibersis less
marked. Neutrophils infiltrating the mucosal layer in C are still less
strongly reactive, and the cells in the (D) lumen exhibit no
immunostaining for S100A8.
|
![]() View larger version (73K): [in a new window] |
Figure 5. Isolated human neutrophils immunolabeled for S100A8 using CP10.1
antibody. (A) Neutrophils exposed to 10-7 M
fMLP remained strongly immunoreactive. (B) Many of the neutrophils that
had phagocytosed opsonized zymosan exhibited markedly diminished
reactivity (e.g., vertical arrow) with some cells completely failing to
stain for S100A8 (horizontal arrow).
|
|
View this table: [in a new window] |
Table 1. Viability of Isolated Human Neutrophils after Stimulation In
vitro
|
|
|
|---|
Demonstration of changes in cytoplasmic immunoreactivity of neutrophils was facilitated by a novel mAb, which we designated CP10.1, raised against murine S100A8 and cross-reactive with human S100A8. This exhibited specific staining of the protein in human neutrophils. In earlier studies, there has been conflicting information about patterns of immunostaining for S100A8 in different inflammatory lesions (e.g., refs. [24 , 25 ]). However, most such antibodies were poorly characterized in terms of epitopes recognized. Variability in staining patterns may be a consequence of cross-reactivity with other S100 proteins, which is likely in view of their high degree of sequence homology, particularly in the calcium-binding and N-terminal domains [26 ]. At least some mAbs to S100A8 cross-react with S100A9 [27 ], and others recognize the A8/A9 complex preferentially [28 ]. Among antibodies specific for S100A8, different commercially available antibodies apparently recognize different epitopes on S100A8, because some permit demonstration of immunoreactivity in formalin-fixed, paraffin-embedded tissue (with or without antigen retrieval), whereas others do not.
However, differential staining of neutrophils and monocytes by an antibody to S100A8 has not been shown previously. The CP10.1 antibody specifically recognizes an epitope on homodimeric S100A8, which is evidently present on human and murine S100A8, is distinct from the epitopes recognized by other antibodies, and is expressed only by the protein in neutrophils. Of particular value for immunocytochemistry is the preservation of the neutrophil-specific epitope in paraffin-embedded tissues. Thus, the antibody to murine S100A8 provides a unique reagent with which to investigate the expression of this protein by human neutrophils.
In acute appendicitis, the inflammatory response is triggered by bacterial proliferation within the lumen, with subsequent spread of organisms through the wall of the appendix leading to transmural inflammation. Therefore, neutrophils in the lumen are virtually certain to have encountered and phagocytosed bacteria, whereas those in the serosal layers are less likely to have participated in a phagocytic response. In this context, the differential intensity of staining for S100A8 of neutrophils in the various layers of the appendicular wall could be explained in terms of phagocytic activation following interaction with the bacteria. Whether this represents extracellular release of the protein or alteration of the intracellular protein leading to loss of immunoreactivity is unclear. Release of neutrophil S100A8 in response to bacterial infection has not been demonstrated in human disease, although previously, we have shown high levels of S100A8 in murine abscess fluid [29 ].
To ascertain whether the loss of immunoreactivity was more likely to have been elicited by exposure to bacterial products or by the phagocytic response, we examined neutrophils activated by the potent bacterial chemoattractant fMLP, compared with cells that had phagocytosed particles of opsonized zymosan. There was no alteration in immunostaining following exposure to an optimal chemoattractant concentration of fMLP, whereas reactivity was reduced rapidly and strikingly after phagocytosis. This suggests that phagocytosis is an important trigger for the diminution of cytoplasmic immunoreactivity for S100A8, although it is unlikely to be the only one, because previously, we have demonstrated loss of neutrophil immunoreactivity for S100A8 in a model of alveolitis in which no particulate irritant was present [19 ]. The reduction in immunostaining was not merely a consequence of leakage of the protein from dead or dying cells, because there was no evidence of any selective decrease in neutrophil viability after engulfment of particles. Thus, it is possible that the loss of immunoreactivity might be a consequence of release of S100A8 by active secretion. However, we were unable to demonstrate convincingly increased levels of S100A8 in cellular supernatants commensurate with the loss of immunoreactivity by immunoblotting or enzyme immunoassay, although clearly, A8 and A9 were present in the supernatants (unpublished results).
Earlier studies demonstrate Ca2+-dependent translocation of the S100A8/A9 complex to the membrane of neutrophils activated with fMLP or activated zymosan [3 , 30 ], and these proteins are postulated to regulate Ca2+-dependent processes important in neutrophil migration, phagocytosis, and activation. Guignard et al. [31 ] show that maximal A8/A9 translocation was provoked by zymosan, whereas that by fMLP was relatively weak. The mechanism by which cytoplasmic immunoreactivity for A8 is lost is unclear. One possibility is that it might be related to alteration of protein conformation or redox state, e.g., by hypochlorite generated by activated neutrophils, so that the epitope recognized by the CP10.1 antibody on homodimeric S100A8 is no longer available. However, immunostaining using the 85-C2 antibody was diminished in parallel, in vitro and in vivo, suggestive of extracellular release of the protein. In any case, loss of immunoreactivity for S100A8 appears to be a marker of phagocytically activated neutrophils.
If neutrophil S100A8 is released, it would be of interest to establish whether this occurs as the free protein (monomeric or oligomeric) or in association with S100A9. In parallel experiments using a polyclonal antibody to human S100A9, we noted comparable loss of immunostaining for A9 (unpublished results), consistent with release of the complex. An important and as yet unresolved issue is the function of S100A8 following release. In the light of studies of anti-microbial activity of the A8/A9 complex and of zinc sequestration by A9 in the complex [5 ], release in response to phagocytosis of bacteria could be important in microbial defense. Of particular interest, however, is that S100A8 can be oxidized rapidly by hypochlorite to yield novel oxidation products ([32 ] and unpublished results), and an intriguing possibility is that it may minimize nonspecific tissue damage by neutrophil-derived, reactive oxygen species. Clearly, further investigation of these issues is needed.
Received June 19, 2000; revised February 21, 2001; accepted February 22, 2001.
|
|
|---|
This article has been cited by other articles:
![]() |
C. Ryckman, C. Gilbert, R. de Medicis, A. Lussier, K. Vandal, and P. A. Tessier Monosodium urate monohydrate crystals induce the release of the proinflammatory protein S100A8/A9 from neutrophils J. Leukoc. Biol., August 1, 2004; 76(2): 433 - 440. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Berthier, M.-H. Paclet, S. Lerouge, F. Roux, S. Vergnaud, A. W. Coleman, and F. Morel Changing the Conformation State of Cytochrome b558 Initiates NADPH Oxidase Activation: MRP8/MRP14 REGULATION J. Biol. Chem., July 3, 2003; 278(28): 25499 - 25508. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. J. Raftery, Z. Yang, S. M. Valenzuela, and C. L. Geczy Novel Intra- and Inter-molecular Sulfinamide Bonds in S100A8 Produced by Hypochlorite Oxidation J. Biol. Chem., August 31, 2001; 276(36): 33393 - 33401. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||