|
|
||||||||

* Instituto de Bioquímica (Centro Mixto CSIC-UCM) and
Centro de Citometría de Flujo y Microscopía Confocal, Facultad de Farmacia. Universidad Complutense, 28040 Madrid, Spain
1Correspondence: Instituto de Bioquímica, Facultad de Farmacia, 28040 Madrid. Spain. E-mail: boscal{at}eucmax.sim.ucm.es
| ABSTRACT |
|---|
|
|
|---|

m) with specific fluorescent
probes showed a rapid and persistent increase of 
m, a
potential that usually decreases in cells undergoing apoptosis through
mitochondrial-dependent mechanisms. Using confocal microscopy, the
release of cytochrome c from the mitochondria to the
cytosol was characterized as an early event preceding the rise of

m. The cytochrome c from cells treated
with nitric oxide donors was modified chemically, probably through the
formation of nitrotyrosine residues, suggesting the synthesis of
peroxynitrite in the mitochondria. These results indicate that nitric
oxide-dependent apoptosis in macrophages occurs in the presence of a
sustained increase of 
m, and that the chemical
modification and release of cytochrome c from the
mitochondria precede the changes of 
m.Hortelano,
S., Alvarez, A. M., Boscá, L. Nitric oxide induces tyrosine
nitration and release of cytochrome c preceding an
increase of mitochondrial transmembrane potential in macrophages.
Key Words: apoptosis NO nitric oxide synthase mitochondria
| INTRODUCTION |
|---|
|
|
|---|
NOS-2 synthesizes for large periods of time important amounts of NO; NO
alone, when released at these high concentrations, has been recognized
as an inducer of apoptotic cell death for a wide range of cells
including macrophages, thymocytes, chondrocytes, smooth muscle cells,
and diverse cells of the neural system (11
12
13
14
15)
. Regarding
the macrophage, cell activation is usually followed by apoptotic death
as part of the physiological response (11
, 14
15
16)
. In the
case of defense against pathogens, this mechanism ensures the
elimination of infected cells and therefore precludes the development
of intracellular parasitic strategies, as well as extension of the
infection by potential pathogens that are exposed to humoral responses
(14
, 15
, 17
18
19)
.
Mitochondrial function is very sensitive to the presence of NO
(20
, 21)
, and the mechanism by which NO induces apoptosis
in thymocytes and other cell types appears to include changes in the

m. Indeed, incubation of isolated
mitochondria with NO releases molecules that promote nuclear DNA
fragmentation when assayed in ex vivo reconstituted systems
(22
23
24)
, suggesting a prominent role for mitochondria in
NO-dependent apoptosis. In recent years, our knowledge of the
mitochondrial contribution to apoptotic death in response to several
stresses has been improved, and release of cytochrome c and
apoptosis-inducing factor from the mitochondria to the cytosol has been
recognized as a key event for commitment to apoptosis (3
, 4
, 22
, 25)
. One of the early events involved in mitochondrial-dependent
apoptosis consists of a fall in the 
m
(22)
. From biochemical and pharmacological data, this
dissipation of 
m appears to be the initial
change leading to apoptosis in thymocytes, fibroblasts, and monocytes
(among other cells) triggered by several proapoptotic stimuli
(22
, 23
, 26)
. Moreover, this collapse of the

m seems to establish an irreversibility of
the apoptotic process (22)
. After the irreversible fall of

m, cytochrome c and other
proapoptotic factors are released from the mitochondrial intermembrane
space to the cytosol. However, there is a certain ambiguity regarding
the precise sequence of events involving the changes of

m and the release of cytochrome
c (22
, 27
, 28)
. Interaction between cytochrome
c and Apaf1, a mammalian homologue of CED-4, favors the
formation of a complex, including caspase-9, that after activation
initiates the proteolytic processing of caspase-3
(29
30
31)
, starting the executioner step of apoptosis
(3
, 22)
.
In the present work we investigated the mitochondrial changes involved
in the apoptotic response in macrophages treated with NO concentrations
in the range of those prevailing in activated cells. As our data show,
challenge of macrophages with NO donors promotes an early and
irreversible increase of 
m that persists
for a long time. Moreover, NO triggers a rapid modification of
cytochrome c in the mitochondria, probably via tyrosine
nitration, followed by the release of this molecule to the cytosol.
Detection of cytochrome c in the cytosol preceded the
changes of 
m, suggesting that NO-induced
apoptosis is initiated in macrophages by a NO-dependent modification of
cytochrome c.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Cells and in vitro culture conditions
Balb/c (10-wk-old) mice were maintained free of pathogens and 4
days prior to use were intraperitoneal injected with 1 ml of sterile
10% thioglycollate broth. Peritoneal macrophages were prepared as
described previously (15)
and cells were seeded at 1 x 106/cm2 in phenol red
free RPMI 1640 medium containing 10% fetal calf serum. Nonadherent
cells were removed 2 h after seeding by extensive washing with
medium. Thymocytes were obtained from 6-wk-old Balb/c mice as described
previously (23)
. The murine monocyte/macrophage cell line
RAW 264.7 and Jurkat T cells were maintained in RPMI 1640 medium
supplemented with 10% fetal calf serum, L-glutamine, HEPES, and
antibiotics.
Flow cytometric analysis of apoptosis by plasma membrane
permeability to propidium iodide (PI)
In vivo PI staining was performed after incubation of
the cells with the indicated stimuli in the presence of 0.005% PI
(32
, 33)
. Cells were carefully resuspended and run in a
FACScan cytometer (Becton & Dickinson, San Jose, Calif.) equipped with
a 25 mW argon laser. The analysis of apoptotic cells was realized using
a dot plot of the forward scatter against the PI fluorescence
(32
, 34)
.
Flow cytometric analysis of mitochondrial transmembrane potential
To measure the 
m, cells were
incubated at 37°C for 15 min in the presence of 30 nM chloromethyl
X-rosamine (CMXRos), 40 nM of 3,3'-dihexyloxacarbocyanine iodide
or 10 ng/ml of Rh123
(23)
, followed by immediate analysis of fluorochrome
incorporation in a FACScan flow cytometer. Incubation with 10 µM of
the uncoupling agent m-chlorophenylhydrazone carbonylcyanide (mClCCP)
was used as a control to decrease 
m in
apoptotic and nonapoptotic cells and, therefore, the fluorochrome
fluorescence (23
, 32)
. 
m was
calculated as the percentage of change in fluorochrome fluorescence
with respect to control cells.
Confocal microscopy
RAW cells were incubated with S-nitrosoglutathione (GSNO) and
peroxynitrite for the indicated periods of time, labeled with CMXRos
(15 min at 37°C; 100 nM), and fixed with methanol at -20°C for 2
min. The cells were blocked with 3% BSA for 30 min at room
temperature, followed by incubation for 30 min with 1:100
anti-cytochrome c monoclonal antibody (mAb) (PharMingen,
clone 6H2.B4). After two washes with ice-cold phosphate-buffered
saline, the cells were revealed by using a secondary Ab (1:300) against
mouse immunoglobulin conjugated with Oregon green (Molecular Probes,
Eugene, Oreg.). Plates were visualized using an MRC-1024 confocal
microscope (Bio-Rad). Both fluorescences were acquired at the same time
and electronically evaluated. Laser sharp software (Bio-Rad) was used
to determine the percentage of colocalization, the relative intensity
of the fluorescence per pixel, and the intensity of both fluorescences
per time.
Immunoprecipitation and Western blot analysis
The cell layers were washed twice with ice-cold buffer A (10 mM
HEPES, pH 7.9; 1 mM EDTA, 1 mM EGTA, 10 mM KCl, 1 mM DTT, 0.5 mM PMSF,
2 µg/ml aprotinin, 10 µg/ml leupeptin, 2 µg/ml TLCK, 5 mM NaF, 1
mM NaV04, 10 mM
Na2MoO4) containing 120 mM
NaCl. Lysis of the cells was performed at 4°C with 0.8 ml of buffer A
supplemented with 0.5% Nonidet P-40 and under continuous shaking.
After centrifugation in an Eppendorf centrifuge, equal amounts of
supernatant protein (50 µg) were immunoprecipitated with either
anti-NO-Y Ab (a gift from Dr. J. S. Beckman, University of Alabama
at Birmingham) or anti-cytochrome c mAb (PharMingen, clone
7H8.2C12) following the instructions of the Ab suppliers. After
overnight rocking and extensive washing, the proteins were
size-separated in 12% sodium dodecyl sulfate-polyacrylamide gel
electrophoresis, transferred to a PVDF membrane, and incubated with the
indicated Ab. The bands recognized were visualized by the ECL
technique. Purified cytochrome c (from rat heart; Sigma) was
incubated with peroxynitrite and treated as described for the cell
extracts.
Synthesis of peroxynitrite
ONOO- was synthesized by the reaction of
NaNO2 (1.8 M) with
H2O2 (2.1 M), as described
(35
, 36)
, and its actual concentration was determined
spectrophotometrically (
302=1670
M-1 cm-1).
Caspase assay
The activity of caspase was measured using as substrate
N-acetyl-DEVD-7-amino-4-methylcoumarin (a good substrate for caspase-3
and -7) and following the instructions of the supplier (PharMingen).
The corresponding peptide aldehyde and Z-VAD.fmk were used to inhibit
the caspase activity in vitro and in vivo,
respectively, and to ensure the specificity of the reaction. The
linearity of the caspase assay was determined over a 30 min reaction
period.
Statistical analysis
The data shown are the means ± SE of three or
four experiments. Statistical significance was estimated with
Students t test for unpaired observations. A P
< 0.05 was considered significant. In studies of Western blot
analysis, linear correlation between increasing amounts of input
protein and signal intensity were observed (correlation coefficients
higher than 0.84)
| RESULTS |
|---|
|
|
|---|
(IFN-
) induced
apoptotic death as deduced by the increase of the in vivo
staining with propidium iodide when analyzed by flow cytometry
(Fig. 1A
m has
been described in mitochondrial-mediated apoptosis (22
m was determined by the changes of the
fluorescence of CMXRos, and the potential increased notably both in
peritoneal macrophages and RAW 264.7 cells treated with GSNO or LPS
plus IFN-
(Fig. 1B
m was observed in thymocytes and Jurkat
cells undergoing apoptosis, data that were in agreement with previous
work (23
but did not affect the apoptosis triggered by GSNO. Under these
conditions the 
m remained unchanged in the
absence of NO synthesis, but the increase persisted in cells treated
with GSNO.
|
The increase of 
m in macrophages treated
with GSNO was associated with a shift in the distribution of the
fluorescence, which involved a cell population comparable to that
exhibiting a decrease in thymocytes incubated under identical
conditions (Fig. 2
, left). The intracellular fluorescence distribution of CMXRos in RAW
cells was analyzed by confocal microscopy; this probe was detected
mainly (85% of the cellular fluorescence) in the cytosolic compartment
(Fig. 2
, central and right panels), colocalizing with the mitochondria
(not shown). Moreover, when the time course of

m change was followed both in RAW cells and
thymocytes treated with GSNO, an opposite behavior was observed
depending on the cell type: the changes in potential were clearly
established after 12 h of incubation with the NO donor and persisted
for at least 6 h (Fig. 3A
). Indeed, the uncoupling agent mClCCP reduced the CMXRos
fluorescence to similar levels in RAW 264.7 cells and thymocytes,
irrespective of the treatment with GSNO, indicating that the changes in
fluorescence were not due to an alteration in the capacity of the
mitochondria to release the fluorochrome (Fig. 3B
). When
apoptosis was triggered in RAW 264.7 cells with other NO donors that
exhibit different kinetics of NO release, a consistent increase of

m was measured, although with quantitative
differences among them (Fig. 3C
). Also, the increase of

m observed in RAW cells was independent of
the reagent used to evaluate the transmembrane potential, since similar
results were obtained with Rh123 and DiOC6 (3) as
probes (Fig. 3D
).
|
|
NO induces cytochrome c modification preceding the
changes of 
m
To further study the mechanisms by which NO induces alterations in
the mitochondrial integrity, experiments were performed to establish
the temporal correlation between the changes of CMXRos fluorescence and
the subcellular localization of cytochrome c. As Fig. 4A
shows, control cells only exhibited fluorescence of CMXRos
(red), with a minimal contribution of cytochrome c-dependent
fluorescence (green). This was probably because the anti-cytochrome
c mAb used either poorly recognized the native structure of
this protein or had a restricted accessibility to the protein in the
mitochondria (vide infra). However, when cells were treated
for 1 h with GSNO, the fluorescence due to cytochrome c
increased homogeneously in the cytosol, reflecting the release of this
protein from the mitochondria. Moreover, whereas cytochrome
c colocalized with CMXRos in control cells (>78%), the
colocalization decreased (<19%) when cells were treated with GSNO
(Fig. 4A
, insets). The time-dependent changes of the
fluorescence of cytochrome c and CMXRos were analyzed
simultaneously by confocal microscopy and the results are shown in Fig. 4B
. The fluorescence due to cytochrome c
increased after 10 min of incubation with GSNO without noticeable
changes in the fluorescence of CMXRos. This shift in the changes of
fluorescence indicates that the release of cytochrome c from
the mitochondria to the cytosol precedes the increase of

m. Moreover, in cells treated with GSNO and
analyzed for short periods of time (10 min or less), an important
increase in the mitochondrial-associated fluorescence of cytochrome
c was observed prior to the release of this protein from the
organelle. Similar results were obtained when RAW 264.7 cells were
treated with 100 µM peroxynitrite (20 min incubation). These data
suggest the occurrence of a NO-dependent structural modification of
cytochrome c in mitochondria, probably nitrosylation, that
facilitates the interaction with the Ab.
|
To investigate the mechanisms by which NO promotes changes in the
fluorescence of cytochrome c, we explored the possibility of
a NO-dependent modification of the protein as the most likely
mechanism. Figure 5A
shows that treatment of macrophages for 20 min with GSNO
promoted an important nitration/nitrosylation of cytochrome
c as deduced by the immunoprecipitation and detection with
distinct combinations of anti-cytochrome c and anti-NO-Y
Abs. Moreover, incubation of purified cytochrome c with 50
µM peroxynitrite resulted in a modified protein that behaved like
cytochrome c from GSNO-treated cells in terms of
immunoprecipitation and detection. Activation of caspases that cleave a
florigenic DEVD substrate was measured after 3 h of incubation
with GSNO, reflecting the initiation of a proteolytic process. Caspase
activity increased notably under these conditions (Fig. 5B
),
and this process was blocked when cells were incubated with the caspase
inhibitor Z-Val-Ala-DL-Asp-fluoromethylketone (Z-VAD).
|
| DISCUSSION |
|---|
|
|
|---|

m, which remained elevated for a long time
(at least 6 h), a period when DEVD-specific caspases were fully
active (3 h). This behavior of macrophages appears to be unique among
the cells analyzed by us and others, since dissipation of

m has been observed in apoptotic cells
(22
m
measurements are supported by the use of different potential-sensitive
probes (23
m in thymocytes and T cells treated not
only with NO donors, but also with dexamethasone and other drugs
inducing apoptosis. In RAW 264.7 cells, the release of cytochrome
c from the mitochondria clearly precedes the rise of

m, indicating that at least in this
particular model, the delivery of cytochrome c is
accomplished without relevant changes of 
m,
as deduced by the use of specific fluorescent probes and visualization
by confocal microscopy. This sequence of events at the mitochondrial
level, although atypical, is not unique to this model, and this issue
is under debate (27
m that occurs more than 4 or 5 h
after the death-inducing signal (39)
m are not involved in the process as
observed using cell-free system assays, including isolated
mitochondria, after the release of apoptotic factors on
mitochondria-depleted Xenopus eggs (41)
As already mentioned, NO disrupts several mitochondrial functions
because of its high reactivity with heme and catalytically relevant
thiolic groups from the proteins (8
, 9
, 43
, 44)
. In
particular, mitochondrial respiration is efficiently inhibited by NO;
complex IV is reversibly inhibited whereas the inhibition of complex I
progresses irreversibly in time (21
, 45
, 46)
. The capacity
of the cell to maintain a threshold of ATP levels favors the
development of an apoptotic response instead of a necrotic process
(4
, 22
, 47)
This reactivity of NO with mitochondrial
protein targets can be influenced by the presence of other NO-reactive
molecules such as intramitochondrial glutathione or the synthesis of
ROI. Oxygen superoxide can accumulate after mitochondrial dysfunction
and when it reacts with NO forms peroxynitrite, a very reactive species
that could be responsible for the effects of NO in these cells
(21
, 36
, 48)
. RAW cells release under basal conditions
significant amounts of ROI as deduced by the high staining with
hydroethidine, a fluorescent probe specific for oxygen radicals.
Moreover, when these cells were treated with a NO donor, the
fluorescence of hydroethidine decreased, suggesting formation of the
corresponding oxygen and nitrogen intermediates (not shown). For this
reason, the formation of NO-dependent cytochrome c
modificationstyrosine nitration in particularwas investigated.
Indeed, murine cytochrome c contains four tyrosine residues
in its sequence (49)
; presumably, some of them can be
affected by NO or by the formation of peroxynitrite.
The mechanism by which NO leads to cytochrome c release in
the absence of changes of the mitochondrial transmembrane potential
remains unexplained. It is possible that NO, in addition to the effects
on the cytochrome c molecule, might act directly on proteins
of the Bcl-2 family (Bax, for example), the localization and function
of which can be altered, as shown for Bax in other systems (26
, 50)
. In RAW cells incubated with GSNO, the fluorescence
associated with cytochrome c increased and colocalized with
CMXRos in the early moments after treatment and was followed by a
release to the cytosol. These results suggest that the accessibility of
cytochrome c to the mAb used for immunodetection is notably
altered after treatment with NO, possibly due to a chemical
modification of its structure. Indeed, when the nitration of cytochrome
c was measured at very early times (10 min) by
immunoprecipitation with anti-NO-Y Ab, an important increase in the
amount of immunoprecipitated cytochrome c was recovered. To
confirm that cytochrome c can be nitrated in tyrosine (in
addition to other NO-dependent modifications) (8
, 10
, 51)
,
we performed experiments with peroxynitrite; an intense band
corresponding to nitrated cytochrome c was detected,
suggesting the occurrence of tyrosine nitration.
Our results show a rapid apoptotic effect in response to GSNO challenge of macrophages. However, it should be mentioned that under in vivo conditions, NO is synthesized several hours after initiation of the activation process; therefore, it is proposed that the apoptotic effects of NO are exerted when the activation process is almost complete. Also, simultaneous analysis of mitochondrial functions (i.e., respiration), together with the changes of cytochrome c and CMXRos fluorescence in cells treated with more stable NO donors (for example DETA-NO), might allow us to evaluate the contribution of a sustained supply of NO to alterations of these parameters and their relative role in the onset of apoptotic vs. necrotic death.
Finally, from a mechanistic point of view, our results show that
NO-dependent apoptosis in macrophages involves a chemical modification
of cytochrome c that alters its structure and facilitates
release from the mitochondria, regardless of the changes of

m. Unraveling the biological relevance of
this modification of cytochrome c in other cells in which NO
triggers a rapid fall of 
m might provide
additional clues to understanding the relative contribution of the
different mitochondrial changes observed in the course of
apoptosis.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. C. Carreras and J. J. Poderoso Mitochondrial nitric oxide in the signaling of cell integrated responses Am J Physiol Cell Physiol, May 1, 2007; 292(5): C1569 - C1580. [Abstract] [Full Text] [PDF] |
||||
![]() |
H.-J. An, O. Maeng, K.-H. Kang, J.-O. Lee, Y.-S. Kim, S.-G. Paik, and H. Lee Activation of Ras Up-regulates Pro-apoptotic BNIP3 in Nitric Oxide-induced Cell Death J. Biol. Chem., November 10, 2006; 281(45): 33939 - 33948. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Zeini, P. G. Traves, R. Lopez-Fontal, C. Pantoja, A. Matheu, M. Serrano, L. Bosca, and S. Hortelano Specific Contribution of p19ARF to Nitric Oxide-Dependent Apoptosis. J. Immunol., September 1, 2006; 177(5): 3327 - 3336. [Abstract] [Full Text] [PDF] |
||||
![]() |
W.-P. HO, T.-L. CHEN, W.-T. CHIU, Y.-T. TAI, and R.-M. CHEN Nitric Oxide Induces Osteoblast Apoptosis through a Mitochondria-Dependent Pathway Ann. N.Y. Acad. Sci., May 1, 2005; 1042(1): 460 - 470. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. J. Oursler, E. W. Bradley, S. L. Elfering, and C. Giulivi Native, not nitrated, cytochrome c and mitochondria-derived hydrogen peroxide drive osteoclast apoptosis Am J Physiol Cell Physiol, January 1, 2005; 288(1): C156 - C168. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Ranjan, P. Shrivastava, S. M. Singh, A. Sodhi, and N. H. Heintz Baculovirus P35 inhibits NO-induced apoptosis in activated macrophages by inhibiting cytochrome c release J. Cell Sci., June 15, 2004; 117(14): 3031 - 3039. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Hortelano, P. G. Traves, M. Zeini, A. M. Alvarez, and L. Bosca Sustained Nitric Oxide Delivery Delays Nitric Oxide-Dependent Apoptosis in Macrophages: Contribution to the Physiological Function of Activated Macrophages J. Immunol., December 1, 2003; 171(11): 6059 - 6064. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. PARK, Y. AHN, C.-K. PARK, H.-Y. CHUNG, and Y. PARK Interleukin-6 Protects MIN6 {beta} Cells from Cytokine-Induced Apoptosis Ann. N.Y. Acad. Sci., November 1, 2003; 1005(1): 242 - 249. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. T. Grey, C. Longo, T. Shukri, V. I. Patel, E. Csizmadia, S. Daniel, M. B. Arvelo, V. Tchipashvili, and C. Ferran Genetic Engineering of a Suboptimal Islet Graft with A20 Preserves {beta} Cell Mass and Function J. Immunol., June 15, 2003; 170(12): 6250 - 6256. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Castrillo, P. G. Traves, P. Martin-Sanz, S. Parkinson, P. J. Parker, and L. Bosca Potentiation of Protein Kinase C {zeta} Activity by 15-Deoxy-{Delta}12,14-Prostaglandin J2 Induces an Imbalance between Mitogen-Activated Protein Kinases and NF-{kappa}B That Promotes Apoptosis in Macrophages Mol. Cell. Biol., February 15, 2003; 23(4): 1196 - 1208. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y.-C. T. Huang, J. Soukup, S. Harder, and S. Becker Mitochondrial oxidant production by a pollutant dust and NO-mediated apoptosis in human alveolar macrophage Am J Physiol Cell Physiol, January 1, 2003; 284(1): C24 - C32. [Abstract] [Full Text] [PDF] |
||||
![]() |
C.-J. Tsi, Y. Chao, C.-W. Chen, and W. W. Lin Aurintricarboxylic Acid Protects against Cell Death Caused by Lipopolysaccharide in Macrophages by Decreasing Inducible Nitric-Oxide Synthase Induction via Ikappa B Kinase, Extracellular Signal-Regulated Kinase, and p38 Mitogen-Activated Protein Kinase Inhibition Mol. Pharmacol., July 1, 2002; 62(1): 90 - 101. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Beltran, M. Quintero, E. Garcia-Zaragoza, E. O'Connor, J. V. Esplugues, and S. Moncada Inhibition of mitochondrial respiration by endogenous nitric oxide: A critical step in Fas signaling PNAS, June 25, 2002; 99(13): 8892 - 8897. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Almeida, J. Almeida, J. P. Bolanos, and S. Moncada Different responses of astrocytes and neurons to nitric oxide: The role of glycolytically generated ATP in astrocyte protection PNAS, December 6, 2001; (2001) 261560998. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Castrillo, D. J. Pennington, F. Otto, P. J. Parker, M. J. Owen, and L. Bosca Protein Kinase C{epsilon} Is Required for Macrophage Activation and Defense Against Bacterial Infection J. Exp. Med., October 29, 2001; 194(9): 1231 - 1242. [Abstract] [Full Text] [PDF] |
||||