© 2006 Nature Publishing Group
The function of the immune system is to prevent
the takeover of the body by genomes other than that
encoded in the germline1. Central to this function is
the ability to kill. Given the fundamental conservation
of biochemistry, there are few chemical mechanisms for
killing a given target cell that lack the potential to kill
some other life form2. Therefore, it is no surprise that the
neutrophil, one of the body’s main cellular components
for the destruction of microorganisms, also damages cells
and tissues of the host3,4. In fact, neutrophil-mediated
tissue damage at infected sites is so common that the
host takes stock of it when judging whether to mobilize
an immune response. Indeed, tissue injury is one of the
main sources of information that launches inflammation,
which in turn launches immunity.
To view immunology as the host’s participation in
the competition between genomes helps explain what
makes the neutrophil as fascinating as it is indispen-
sable5. Surprisingly, some immunologists seem not
to share this view. Say neutrophil, and they move on,
thinking: inflammation, not immunity. They disrespect
the cell’s ‘nonspecificity’ and consider its best-studied
behaviours — crawling, eating, and disgorging pre-
packed enzymes and partially reduced molecules of
oxygen — as rudimentary. Finally, scientists who are
interested in anti-inflammatory therapeutics are dis-
couraged from targeting neutrophils because it seems
futile to try to suppress neutrophil-dependent tissue
damage without the serious side effect of increasing the
host’s risk from infection1.
The aim of this article is to dispel each of those views.
The literature on neutrophils is so vast (PubMed finds
over 76,000 papers on the subject) that few authors have
attempted to cover it comprehensively6,7. This Review
makes no such effort. On the contrary, large chunks of
neutrophil biology are neglected here, including myelo-
poiesis, chemotaxis, transendothelial migration, synthe-
sis of lipid mediators and much of signal transduction.
Instead, selected studies are marshalled to support the
following five tenets: that the neutrophil often has an
important role in launching immune responses; that
the neutrophil helps to heal tissues as well as destroy-
ing them; that the neutrophil gives instructions with
as much specificity as a lymphocyte or neuron, albeit
with specificity of a different kind; that the neutrophil
integrates information, with a circuitry of awe-inspiring
design, to tailor its responses to its spatial and temporal
context; and that the neutrophil offers potential oppor-
tunities for selective pharmacological intervention, to
both promote and restrain inflammation.
Neutrophils as decision-shapers
Neutrophils help answer one of the central questions in
immunology: what triggers an immune response? The
danger theory holds that injured host cells release alarm
signals that activate antigen-presenting cells (APCs)8. The
pattern-recognition theory posits that ‘microbial non-self’
induces an innate immune response, which in turn trig-
gers an adaptive immune response9. Both views are valid,
but they leave much unexplained. For example, surgery
injures large numbers of host cells but does not promote
immune reactions to foreign compounds, such as anti-
biotics, that are present at the time. Non-pathogenic
resident microorganisms belonging to several thousand
Department of Microbiology
and Immunology, Weill Cornell
Medical College; and
Programs in Immunology and
and Molecular Biology, Weill
Graduate School of Medical
Sciences of Cornell University,
Box 57, 1300 York Avenue,
New York 10021, USA.
17 February 2006
A theory that the trigger for
mounting an immune response
consists of an injury to host
cells, resulting in the release
of alarm signals that activate
A theory that the trigger for
mounting an immune response
consists of the recognition of
‘microbial non-self’ molecules
by receptors expressed by
innate immune cells.
Neutrophils and immunity: challenges
Abstract | Scientists who study neutrophils often have backgrounds in cell biology,
biochemistry, haematology, rheumatology or infectious disease. Paradoxically,
immunologists seem to have a harder time incorporating these host-defence cells into
the framework of their discipline. The recent literature discussed here indicates that it
is appropriate for immunologists to take as much interest in neutrophils as in their lympho-
haematopoietic cousins with smooth nuclei. Neutrophils inform and shape immune
responses, contribute to the repair of tissue as well as its breakdown, use killing mechanisms
that enrich our concepts of specificity, and offer exciting opportunities for the treatment
of neoplastic, autoinflammatory and autoimmune disorders.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | MARCH 2006 | 173
© 2006 Nature Publishing Group
Secondary lymphoid organ
Immature DC or
αβ T cell or
γδ T cell
in the bone marrow
chemokine that is generated
proteolytic activation of its
precursor. Chemerin does not
yet have a ‘chemokine ligand’
Immature dendritic cells (DCs)
with a morphology resembling
that of plasma cells.
Plasmacytoid DCs produce
type I interferons in response
to viral infection.
species express pathogen-like microbial patterns, but they
only elicit maturation of the neonatal immune system10
and tolerance, not an immune response.
A third view is that a normal immune response
results from the ongoing detection of signals that report
injury and signals that report infection1. Although many
different molecular signals might be involved, they can
be considered ‘binary’ in the sense that most of them
arise as a consequence of one of these two events, and
it generally requires at least one signal from each class
to launch a response. This binary control begins with
inflammation. Except in autoinflammatory disorders,
the triggering and continuation of an inflammatory
response generally require the simultaneous receipt of
molecular signals directly or indirectly reporting tissue
damage and the presence of a genome different from
that of the host. Signals reporting injury and infec-
tion activate epithelial cells, mast cells, macrophages,
endothelial cells, platelets and neutrophils. Binary
signalling is propagated as these cells recruit, activate
and programme APCs through further binary signals,
such as cytokines and microbial products, cytokines
and CD40 ligation, or microbial products and products
of necrotic host cells. Finally, T cells are activated and
programmed by antigen-receptor ligation together with
signals from APCs, and B cells are activated by antigen-
receptor ligation together with signals from T cells.
Therefore, inflammation imprints the immune response
with a pattern of information flow: the integration of
signals of two or more distinct classes derived directly
or indirectly from injury and infection.
As a key component of the inflammatory response,
neutrophils make important contributions to the recruit-
ment, activation and programming of APCs (FIG. 1).
Neutrophils generate chemotactic signals that attract
monocytes and dendritic cells (DCs), and influence
whether macrophages differentiate to a predominantly
pro- or anti-inflammatory state11–13. For example, neutro-
phils proteolytically activate prochemerin to generate
chemerin, one of the few chemokines that attracts both
immature DCs and plasmacytoid DCs14. Neutrophils
also produce tumour-necrosis factor (TNF) and other
cytokines that drive DC and macrophage differentia-
tion and activation12,13,15. Neutrophil activation of DCs is
fostered by cell–cell contact, in which the specific carbo-
hydrates on CD11b engage DC-specific ICAM3-grabbing
non-integrin (DC-SIGN)15. Moreover, neutrophils secrete
TNF-related ligand B-lymphocyte stimulator (BLyS)16,
which helps to drive proliferation and maturation of
B cells, and interferon-γ, which helps to drive differen-
tiation of T cells and activation of macrophages17. On
a per-cell basis, neutrophils make fewer molecules of a
given cytokine than do macrophages or lymphocytes, but
neutrophils often outnumber mononuclear leukocytes at
inflammatory sites by one to two orders of magnitude,
and they can therefore be important sources of cytokines
such as TNF at the crucial juncture at which the deci-
sion is made to mount an immune response. However,
neutrophils can also function as powerful suppressors of
T-cell activation. For example, in patients with advanced
cancer, activated neutrophils can impair T-cell receptor
(TCR) ζ-chain expression and cytokine production18.
The ability of neutrophils to augment or inhibit lym-
phocyte expansion and activation at sites of inflammation,
draining lymph nodes and in the spleen is reciprocated
by the adaptive immune system’s control of the rate of
neutrophil production in the bone marrow. This can
be appreciated by looking upstream of the cytokine
granulo cyte colony-stimulating factor (G-CSF), which
is an essential regulator of neutrophil production
through several mechanisms. Stromal-cell-derived
G-CSF triggers bone-marrow neutrophils to release
matrix metallo proteinase 9 (MMP9), which solubilizes
KIT ligand, helping to mobilize progenitor cells19. G-CSF
also acts directly on the progenitors to increase their
proliferation, while suppressing stromal-cell expression
of CXC-chemokine ligand 12 (CXCL12, also known as
Figure 1 | Neutrophils interact with monocytes, dendritic cells, T cells and
B cells in a bidirectional, multi-compartmental manner. Through cell–cell
contact and secreted products, neutrophils recruit and activate monocytes, dendritic
cells (DCs) and lymphocytes, and products of monocytes, macrophages and T cells
activate neutrophils. Tissue macrophages ingesting apoptotic neutrophils produce
less interleukin-23 (IL-23). IL-23 triggers T cells in secondary lymphoid tissues to
produce IL-17. IL-17 triggers stromal cells in the bone marrow to produce granulocyte
colony-stimulating factor (G-CSF). G-CSF promotes proliferation of neutrophil
precursors and release of neutrophils into the circulation (see text for references).
BLyS, tumour-necrosis factor-related ligand B-lymphocyte stimulator; CXCL12,
CXC-chemokine ligand 12; IFNγ, interferon-γ; TNF, tumour-necrosis factor.
174 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
A collection of liquefied tissue
containing many living or dead
neutrophils and bacteria.
A form in which activity is not
yet expressed, pending
activation by an event such as
redistribution from a particular
compartment or proteolytic
Biologically active compounds
that are primarily derived
from arachidonic acid, in part
and lipoxygenases, including
Reduction of oxygen
Donation of electrons to
molecular oxygen. Donation
of up to three electrons
gives rise to reactive oxygen
intermediates (ROIs), whereas
donation of four electrons
gives rise to water. The term
‘intermediate’ in ROI refers to
oxygen whose reduction state
is intermediate between that of
oxygen (O2) and water (H2O).
SDF1), which helps to retain neutrophils in the bone
marrow20. In turn, G-CSF production is regulated by
interleukin-17 (IL-17), which is produced by subsets
of γδ and αβ T cells (FIG. 1). T-cell production of IL-17
is governed by IL-23, which is released from extravas-
cular macrophages. The release of IL-23 is suppressed
when macrophages ingest apoptotic neutrophils21.
The homeostatic feedback loop among macrophages,
T cells and neutrophils described earlier illustrates the
fallacy of visualizing vertebrate immunity as two separate
systems — innate and adaptive — that take turns over the
course of an encounter with antigen. Instead, elements of
more and less ancient evolutionary origin work together
from before the initiation of an immune response21, into
its first hours22, and through to its resolution. Resolution
of an immune response is marked not only by the genera-
tion of immunological memory but also, in many cases,
by the innate immune system’s help in healing a wound.
Demolition and reconstruction
Wounding is a frequent initiator of immune responses
because it introduces both types of precipitants — injury
and infection — except in the unnatural instance of
surgery with aseptic technique. Although neutrophils
might delay the closure of artificial, microbiologically
sterile wounds23,24, they have three important roles in
healing natural wounds25: sterilization of microbes,
generation of signals that slow the rate of accumulation
of more neutrophils, and instigation of a macrophage-
based programme that switches the state of damaged
epithelium from pro-inflammatory and non-replicative,
to anti-inflammatory and replicative.
The principal contribution of neutrophils to wound
healing is microbial sterilization. Wounds tend to heal
poorly and lead to lethal outcomes in individuals with
insufficient neutrophils5; in individuals with neutrophils
that cannot adhere to the endothelium or extracellular
matrix and therefore fail to accumulate in infected
sites26,27; in individuals with a vasculature that is insuf-
ficient to deliver neutrophils, for example persons with
full-thickness burns; or in individuals with neutrophils
that have deficiencies in microbicidal machinery.
It is not realistic to expect neutrophils to find and eat
each bacterium in a wound before any bacteria have had
time to escape into the lymph or blood. Neutrophils that
sense tissue damage plus infection but fail to encounter a
bacterium within a short time, fire off their arsenal into
the extracellular space28. In vitro, this period of restraint
ranges from ~15 to ~45 minutes, which is thought to be
the approximate time it takes for a neutrophil to emi-
grate from the blood into the extravascular tissues. When
restraint is abandoned, what ensues is the liquefaction
of tissue — that is, the formation of pus — through neu-
trophils’ release of proteases, their activation of proteases
that are expressed in a latent form by cells resident in the
tissues, and their oxidative inactivation of anti-proteases
(proteins that specifically bind and inactivate proteases)3,4.
Tissue breakdown is usually viewed as detrimental, and
therefore neutrophils are often cast in a pejorative light.
However, early, small-scale neutrophil-mediated tissue
destruction is usually a life-saving process. It serves to
disassemble the collagen fibrils that impede neutrophil–
bacterial contact and puts pressure on surrounding tissue.
This might help to collapse lymphatics and capillaries,
thereby cutting off bacterial escape routes and trapping
the microbes in a toxic soup.
At the same time, neutrophils also generate signals
with four key actions: to retard their own accumula-
tion; to suppress their own activation; to promote their
own death; and to attract and programme macrophages
to stop the damage and orchestrate repair. Among the
main endogenously derived anti-inflammatory com-
pounds are metabolites of fatty acids. These include
neutrophil-derived lipoxins, as well as macrophage-
derived resolvins and protectins that are produced in
response to the ingestion of apoptotic neutrophils29, all
of which block neutrophil recruitment. Proteinaceous
anti-inflammatory factors such as secretory leukocyte
protease inhibitor (SLPI)30 are also important (FIG. 2),
and neutrophils, as well as macrophages30 and epi-
thelial cells, produce SLPI. Despite producing SLPI,
activated neutrophils can oxidatively inactivate this
anti-inflammatory factor31. However, SLPI protects
itself from oxidative inactivation by suppressing the
neutrophil respiratory burst32. Moreover, SLPI also
puts a brake on tissue proteolysis by inhibiting neutro-
phil elastase. Macrophages and epithelial cells also
make an epithelial-cell-growth-promoting cytokine,
proepithelin (PEPI; also known as progranulin). PEPI
synergizes with SLPI in inhibiting neutrophil activa-
tion31. Although neutrophil elastase degrades PEPI to
generate pro-inflammatory epithelins, SLPI protects
PEPI in two ways: by binding and inhibiting neutrophil
elastase, and by binding and shielding PEPI from deg-
radation by neutrophil elastase. This helps to prevent
the products of degradation of PEPI from promoting
epithelial-cell production of CXCL8 (also known as
IL-8)31, which is one of the most important neutrophil
attractants. Moreover, intact PEPI signals to epithelial
cells to proliferate and close the wound31,33.
Specificity in command
Neutrophil-derived chemokines, cytokines and
eicosanoids signal with a type of specificity that is famil-
iar to immunologists: that of ligand–receptor, protein–
protein or lipid–protein interactions. But neutrophil
specificity does not stop there. One of the main func-
tions of neutrophils is to undergo the respiratory burst:
an abrupt, non-mitochondrial reduction of oxygen to
forms that are less reduced than water and far more reac-
tive, known as reactive oxygen intermediates (ROIs).
Respiratory burst products and their derivatives include
superoxide, singlet oxygen, ozone, hydrogen peroxide
(H2O2), hypohalous acids, chloramines and hydroxyl radi-
cals34 (FIG. 3). These species can interact with an unlimited
number of macromolecules. This seems to make a travesty
of specificity, as this term is most often used in immunol-
ogy, when it characterizes the crowning achievement of
the adaptive immune system, that is, the ability of TCRs,
B-cell receptors and antibodies to make distinctions so
precise that they only permit the recognition of a single
macromolecule among tens of thousands.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | MARCH 2006 | 175
© 2006 Nature Publishing Group
A lipid whose exposure on
the outer leaflet of the plasma
membrane generally correlates
with apoptosis of the cell and
promotes its uptake by other
Of necessity, however, there are different kinds of
specificity. The infectious challenges faced by the
immune system are so diverse and dire that they can only
be met by a response in which collateral damage occurs
as a matter of course35,36. No person who lives outside
a bio-containment suite reaches maturity without their
neutrophils having formed pus on multiple occasions,
which is to say, destroying some of their tissue, and often
saving their life in the process. In the familiar sense of
the term, this is an ultimate example of nonspecificity.
In fact, the specificity of ROIs is just as exquisite as that
of antibody, but it is manifest at the atomic level, rather
than the molecular level2,37. For example, ROIs lead to
oxidation of sulphurs to sulphoxides and of sulphydryls
to disulphides or to sulphenic, sulphinic or sulphonic
acids. These are specific outcomes of the interaction of
ROIs with specific atoms. That the atomic targets of ROIs
are crucial to the function of a wide array of molecules
probably accounts for the evolutionary selection of this
means of host defence.
Indeed, there are two important innovations of innate
immunity and they both involve distinctive forms of
specificity. First, molecular specificity in the recognition
of relatively invariant and widely distributed microbial
macromolecules is imparted by Toll-like receptors, lectins
and lectin receptors, scavenger receptors, C-reactive
protein, and complement and its receptors. Second,
specific, covalent chemical reactions by ROIs and reac-
tive nitrogen intermediates with a subset of atoms that
are shared by, and crucial for, the function of diverse
microbial molecules, characterizes one of innate immu-
nity’s important effector functions. The first innovation
makes it difficult for a pathogen to evolve to a state that
escapes detection, and the second makes it difficult for
a pathogen to evolve to a state that escapes destruction38.
The price of these twin innovations in specificity is tissue
damage. But tissue damage is not wasted: it is used as
information and incorporated into the central decision
algorithm of the immune response. The immune system
cannot function and the host cannot live without these
two distinctive types of specificity.
Equally fundamental to an appreciation of neutrophil
biology is the following principle: the same products that
kill also signal. Homeostatic signalling depends on appro-
priately timed, brief exposures to low concentrations of
mediators. When present at the wrong time, for too long
and in high concentrations, the same mediators can kill.
In fact, across biology, killing is frequently the outcome
of excessive and inappropriate signalling, whether the
sender–receiver pairs are host-versus-microorganism,
or microbial-cell-versus-microbial-cell2. ROIs, for
example, are essential mediators of signalling by many
cytokine and hormone receptors, such as those for
insulin, platelet-derived growth factor, fibroblast growth
factor, nerve growth factor, TNF and angiotensin2. The
best-studied molecular mechanism for ROI-mediated
signalling is the temporary inactivation of tyrosine
phosphatases by reversible oxidation of their active-
site cysteine sulphydryls39. Moreover, neutrophil pro-
teases also regulate physiological processes19. In short,
a neutrophil-rich inflammatory site is buzzing with
information for every cell present; only at the epicentre
is the command for all the cells to die.
Cytotoxic T cells and natural killer (NK) cells can poten-
tially kill any nucleated host cell. Neutrophils match that
capacity and top it with their additional abilities to destroy
non-nucleated cells and connective tissue. In mammals,
only neutrophils are licensed to liquefy any part of the
body. Therefore, some of the main messages sent by
neutrophils in inflammatory sites can be paraphrased as
follows. To microorganisms and host cells that stand in
the way: die. To themselves: if the site still seems infected,
summon reinforcements until the crucial concentration of
neutrophils that is required to clear a given tissue volume
of bacteria is attained40; and if infection has been brought
under control, wrap yourself in a phosphatidylserine flag
and wait for macrophages to remove your apoptotic corpse
for the safe disposal of your unexploded weapons41.
Neutrophils’ decisions over their own fate are
co-ordinated through transcription factors. Hypoxia-
inducible factor 1 (HIF1) regulates both the antibacterial
activity of neutrophils42 and their apoptosis43. Forkhead
box O3A (FOXO3A) postpones neutrophil apoptosis by
Figure 2 | Neutrophils have a key role in wound healing
both by controlling microbial contamination and
by attracting monocytes and/or macrophages.
The illustration shows one pathway out of many that
are operative in wounds. Neutrophil- and macrophage-
derived secretory leukocyte protease inhibitor (SLPI)
blocks neutrophil elastase. SLPI alone, and in synergistic
combination with macrophage- and epithelial-cell-
derived proepithelin (PEPI), blocks cytokine-induced
release of proteolytic enzymes and reactive oxygen
intermediates (ROIs) by neutrophils. These actions
diminish the neutrophil-dependent proteolytic conversion
of PEPI to epithelins (EPIs), decreasing the ability of EPI
to promote epithelial-cell production of CXC-chemokine
ligand 8 (CXCL8; also known as IL-8), an important
neutrophil chemoattractant. Intact PEPI promotes
epithelial-cell proliferation, speeding closure of the
wound. Black arrows indicate processes involved in tissue
repair and regeneration. Grey arrows indicate processes
involved in host defence.
176 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
Azurophilic (also known as primary) granules:
BPI, neutrophil elastase, cathepsin G, protease 3,
Specific and tertiary granules:
Lactoferrin, lipocalin, lysozyme,
LL37, MMP8, MMP9 and MMP25
Nets that trap bacteria
and neutrophil elastase
compounds that chelate iron
and deliver it to the bacterium
through specific receptors.
Staining with azure (blue)
components of the
Romanowski-type stains used
in standard evaluations of
In neutrophils, the earliest-
formed set of granules,
which contain many antibiotic
proteins and proteases,
suppressing transcription of CD95 ligand (CD95L; also
known as FASL)44. A protein called survivin also has an
important role in delaying the death of neutrophils at
Neutrophils decide the fate of microorganisms in
two general ways: by taking away and by giving. ‘Taking
away’ means withholding essential factors that bacteria
require, for example iron. The simplest approach is for
neutrophils to secrete the iron-binding protein lacto ferrin
into the phagosome and the extracellular environment.
However, lactoferrin faces tough competition from bac-
terial iron-binding molecules, a diverse array of com-
pounds known as siderophores. The neutrophil in turn
counters this ‘counterpunch’ by secreting lipocalin-2, a
protein that embraces bacterial siderophores, preventing
them from returning to the mother ship — the bacte-
rium — with their cargo46.The dedication of neutrophils
to delivering cytotoxic messages to bacteria (that is, to
‘giving’) is evident from two perspectives: their supply of
redundant effector molecules and their use of multiple
subcellular compartments as a source. Therefore, little of
the neutrophil is wasted. As neutrophils migrate along a
concentration gradient of a stimulus, they progressively
discharge two distinct sets of granules, activate a killing
cascade at the plasma membrane, extrude their nuclear
proteins, and finally pour out their cytosol, all with anti-
microbial effects (FIG. 3). Even the nuclei of neutro phils
contribute to host defence by extruding their chromatin,
which forms extracellular nets decorated with proteases
from the azurophil granules47. Cytosol released by
necrotic neutrophils delivers large amounts of calprotec-
tin, a bacteriostatic heterodimer of MRP8 and MRP14
(REF. 48). Neutrophils also release potently bactericidal
Payloads: prepacked or made on demand. The first
granules to be discharged are peroxidase-negative, and
these include specific granules (also known as second-
ary granules) and tertiary granules (also known as
gelatinase granules), which contain overlapping sets
of proteins50. In these two types of granules are found
the gang of ‘L’s’ — lactoferrin, lipocalin, lysozyme and
LL37. LL37 is a chemotactic and antimicrobial peptide
released from its precursor, the cathelicidin CAP18
(cationic antimicrobial protein 18), by protease 3. Also
found in the peroxidase-negative granules are three ‘M’s’
— MMP8, MMP9 and MMP25. By degrading laminin,
collagen, proteoglycans and fibronectin, these MMPs
are thought to have an important role in facilitating
neutrophil recruitment and tissue breakdown50.
As the concentration of secretagogues increases, the
next granules to be emptied are peroxidase-positive
and azurophilic (also known as primary). Their bulk
contents are four α-defensins and myeloperoxidase
(MPO), the iron-containing enzyme that colours pus
green. The defensins are small, cyclic polypeptides that
trade off their relatively low molar potency against their
high abundance and broad spectrum as antimicrobial
agents51. MPO converts the relatively innocuous H2O2
into much more powerful antiseptics: hypochlorous
acid (which is the active ingredient in bleach), hypo-
bromous acid and hypoiodous acid. Hypochlorous acid
reacts with amines to give rise to longer-lived, bacteri-
cidal chloramines34. The same granules deliver a potent
protein antibiotic that is active against Gram-negative
bacteria, the lipopolysaccharide-binding bactericidal
permeability increasing protein (BPI)52, and four broad-
spectrum antibiotic proteins termed serprocidins53. The
serprocidins include three serine proteases — cathep-
sin G, neutrophil elastase and protease 3 — and a homo-
logue that lacks proteolytic activity but retains equipotent
antibacterial activity, which was independently identified
as azurocidin53 and CAP37 (REF. 54). The antimicrobial
molar potency of serprocidins is comparable to that of
formulary anti biotics, but unlike pharmaceuticals, the
enzymatically active serprocidins also degrade most of
the components of extracellular matrix. This bifunc-
tionality highlights that tissue breakdown is part of the
antimicrobial strategy of neutrophils.
Meanwhile, radical biology takes place at the plasma
membrane and its offshoot, the membrane of the nas-
cent phagosome. Studies of neutrophils have had a
crucial role in establishing the fundamental but initially
heretical principle that biochemical reactions can gener-
ate compounds with unpaired electrons (also known as
radicals) as functional products55,56. Phagocyte oxidase
(phox), which is the main ROI-producing enzyme
expressed by neutrophils, continues to lead the way in
illuminating the biochemistry of the broadly distributed
family of ROI-producing, signal-capacitating enzymes
that are termed NADPH oxidases (NOXs)57.
As a biochemical cellular component, phox is so com-
plex that its study has dominated the field of neutrophil
biology for decades. For a description of the constituents
of phox, their assembly and the diverse experimental for-
mats that investigators have used to study the activation
of phox, see Supplementary information S1 (box).
Deficiencies of phox were discovered because they
led to death from infection58. The phenotype of phox
deficiency, together with the bactericidal action of
Figure 3 | Neutrophils deliver multiple anti-microbial molecules. Microbicidal
products arise from most compartments of the neutrophil: azurophilic granules
(also known as primary granules), specific granules (also known as secondary granules)
and tertiary granules, plasma and phagosomal membranes, the nucleus and the cytosol.
BPI, bactericidal permeability increasing protein; H2O2, hydrogen peroxide; HOBr,
hypobromous acid; HOCl, hypochlorous acid; HOI, hypoiodous acid; MMP, matrix
metalloproteinase; 1O2, singlet oxygen; O2
phox, phagocyte oxidase.
–, superoxide; O3, ozone; •OH, hydroxyl radical;
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | MARCH 2006 | 177
© 2006 Nature Publishing Group
A fungal metabolite that
inhibits actin polymerization.
Cytochalasin has been widely
used in vitro to promote
activation of neutrophils
studied in suspension.
ROIs in vitro34, convinced most investigators that the
production of ROI by phox is an essential microbicidal
mechanism. ROIs can also contribute to desorption of
bactericidal neutrophil proteases from their proteoglycan
matrix in the granules59. Indeed, an inability to reproduce
the frequently reported bactericidal actions of H2O2
in vitro led one group to propose that the only role of phox
is to mobilize neutrophil proteases59. However, neutrophils
can release proteases under anaerobic conditions that pre-
clude the action of phox, or when they are congenitally
phox-deficient60, and macrophages lack the proteases
in question but can kill microbes in a phox-dependent
manner61. In addition, MPO, which does not kill by itself,
powerfully augments the killing action of ROIs, and defi-
ciency of MPO, or addition of MPO inhibitors, reduces
killing by neutrophils34. Furthermore, bacteria reacted to
phagocytosing neutrophils by trans cribing antioxidant
defense genes62, and Staphylococcus aureus deprived of an
antioxidant protein became markedly more susceptible to
phox-dependent killing by neutrophils63. In short, phox
and the azurophil granule proteases are each important
microbicidal mechanisms. They can function as mutually
redundant, synergistic or stand-alone means of defence,
depending on the experimental conditions.
Move, stop and kill: a sense of time and place
Given the beneficial and deleterious impacts of
neutro phil-mediated destruction of microorganisms
and host tissue, it is not surprising that there has
been an enormous amount of research on control
mechanisms that govern two of the most important
engines of destruction by neutrophils: the respiratory
burst and the degranulation response. The overview
of signalling mechanisms offered below is not meant
to be complete, but aims to illustrate the richness of
the topic and to set the stage for the argument that the
intricacy of neutrophil intracellular signalling affords
opportunities for selective intervention.
The cellular engineering problems solved by neutro-
phil signalling systems include the following. The cell
must remain non-sticky as it hurtles through the arte-
rial and arteriolar circulation; then it must squeeze
through capillaries smaller in diameter than itself,
without allowing collision, friction or distortion to
activate it. A fraction of the population must adhere
tightly enough to the normal endothelium of post-
capillary venules to resist being washed away in the
circulation, but loosely enough to roll while scouting
for evidence of tissue damage and microbial infection.
If such evidence is received, the cell must crawl to a
boundary between endothelial cells, penetrate the junc-
tions and the underlying basement membrane without
damaging these structures, move up a chemotactic
gradient and decide whether its original information
remains valid. If the answer is negative, the cell must
execute itself by apoptosis. If the answer is positive, the
cell must attempt to engulf and destroy microbes. If
it cannot locate microbes quickly, it must attempt to
destroy them at a distance by releasing every weapon
at its disposal.
One experimental system for studying neutrophil
activation in vitro (others are noted in Supplementary
information S1 (box)) consists of administration of
soluble, physiological stimuli to neutrophils that are
adherent to two-28 or three-dimensional40 surfaces
coated with extracellular matrix proteins. The most
striking feature of such systems is the magnitude of
responses to soluble, physiological stimuli, such as
TNF, used as single agents (that is, without cytochalasins
and without being combined with other stimuli).
In contrast to results with cells in suspension, the
responses of adherent neutrophils last over an hour
and lead to the accumulation of nearly as much
antimicrobial product as can be elicited by bacteria
or phorbol esters. For example, neutrophils that are
adherent to biological surfaces respond to soluble,
physiological stimuli by releasing about two orders of
magnitude more H2O2 than neutrophils in suspension
responding to the same stimuli28.
Under such experimental conditions, two kinds of
signals exert binary control over the respiratory burst
and degranulation: signals resulting from integrin
engagement with extracellular-matrix proteins on cellu-
lar or acellular surfaces, and signals transmitted through
cell-surface receptors for inflammatory factors, such as
formylated peptides, and cytokines and chemokines,
including TNF, G-CSF, granulocyte/macrophage colony-
stimulating factor (GM-CSF), CC-chemokine ligand 3
(CCL3; also known as MIP1α), CXCL8 and the comple-
ment component C5a28,64,65. The evidence that adhesion
in general28, and ligation of integrins in particular66,
controls the responses of cells to cytokines (FIG. 4) was
one of the first examples of a principle that is now rec-
ognized as widespread in cell biology: the joint control of
cell behaviour by integrin and non-integrin receptors67.
The control of physiological, non-phagocytic, full-scale
neutrophil activation by binary signals is an example of
the binary control that is characteristic of inflammation
Figure 4 | Neurophils spread and reorganize their cytoskeleton after activation
by dual stimuli through integrins and cytokine receptors. a | Polymerized actin is
stained with rhodamine-phalloidin. Neutrophils spread when they are allowed to
adhere to matrix-protein-coated glass and are also treated with tumour-necrosis
factor (TNF) (100 ng ml–1 for 30 min). b | Spreading is not seen when neutrophils are
allowed to adhere to matrix protein-coated glass but are not given TNF, or when they
are given TNF in suspension.
178 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
A compound that inhibits
enzymes whose activity
depends on their flavin
iodonium does so through
its structural resemblance to
a portion of the flavin molecule.
One of many small zones
that form at the surface of a
leukocyte as it adheres to
a substratum, where integrins,
integrin-binding proteins and
termini of actin microfilaments
Linked with N-acetylneuraminic
Responses of adherent neutrophils to TNF. Responses of
adherent neutrophils to one of the most important soluble
pro-inflammatory agonists, TNF, do not depend on trans-
cription or translation28. As such, these responses might
have little to do with the canonical view of TNF signalling
based on studies of other cell types. Because TNF is crucial
to the pathogenesis of certain inflammatory disorders,
such as rheumatoid arthritis, and given that neutrophils
are the most abundant cell in the joint space in rheuma-
toid arthritis, it is exciting to realize that pathways of TNF
signal transduction remain to be characterized that are
likely to be of substantial pathophysiological relevance.
After ligation of TNF by TNF receptor 1 (TNFRI)
and/or TNFRII, β2-integrins are rapidly activated in
wild-type and phox-deficient neutrophils in a manner
that can be blocked by the flavoprotein inhibitor dipheny-
lene iodonium (DPI) and mimicked using exogenous
H2O2 (REF. 68). This was interpreted as implicating the
production of very small amounts of H2O2 by an unspec-
ified flavin-dependent enzyme. The result is transient
inhibition of tyrosine phosphatase(s), whose cysteine-
dependent active sites are susceptible to reversible oxida-
tion by H2O2 (REF. 39). Such inhibition leads to increased
phosphorylation and activation of at least three sets of
tyrosine kinases: SRC-family kinases, haematopoietic
cell kinase (HCK) and FGR69; spleen tyrosine kinase
(SYK)70,71; and protein tyrosine kinase 2 (PYK2)70,72.
Also crucial at this early stage is the activation of
phosphatidy linositol 3-kinase (PI3K)70, which binds to
TNFRs73. Inhibitor studies indicate that some of the fore-
going events are mutually upstream of each other70. This
apparent paradox can be resolved by postulating iterative
interactions that constitute a feed-forward mechanism.
Onset of the respiratory burst is closely related in time
to tyrosine phosphorylation of the guanine nucleotide
exchange factor VAV74,75, which leads to the release of
RAC2 from GDP dissociation inhibitor 1 (GDI1; also
known as RHOGDI) and allows RAC2 to anchor the
assembly of phox at the membrane.
Meanwhile, the actin-based cytoskeleton, which in
suspended neutrophils is a spherical meshwork beneath
the plasma membrane, dissolves. Actin polymers
re assemble as stress fibres near the adherent surface of
the neutrophil (FIG. 4) and interconnect the podosomes.
The podosomes are sites of clustered integrins whose
intracellular domains are associated with the most
abundant of the accumulating tyrosine phospho proteins,
including paxillin76,77. This reorganization of the actin
cytoskeleton represents an extended form of what hap-
pens locally at the base of the phagocytic cup during
ingestion of particles, and serves to remove a physical
barrier that, in suspended neutrophils, might make it
difficult for granules to reach the plasma membrane.
For neutrophils to spread fully on matrix proteins,
they need to shed CD43 (also known as leukosialin), a
heavily sialylated molecule that seems to have an impor-
tant role in keeping suspended neutrophils in their non-
sticky state. Neutrophil elastase cleaves the ectodomain
of CD43 near the plasma membrane. Therefore, a small
degree of degranulation to release some neutrophil
elastase is probably a precondition to a large degree of
degranulation, which depends on cell spreading. (This
is similar to the observation that a small degree of ROI
generation from a source other than phox seems to be
necessary before TNF can trigger phox)68. Albumin
binds to CD43 and can protect CD43 from neutrophil
elastase. In the bloodstream, high concentrations of
albumin might help to prevent spontaneous neutrophil
spreading and immobilization. In inflammatory sites,
secreted neutrophil elastase is more abundant than in
blood plasma, and neutrophil elastase inhibitors and
albumin are much less abundant. Therefore, CD43 is
cleaved and neutrophils can spread78. Like neutrophil
elastase, cathepsin G can also regulate neutrophil activa-
tion79, and it will be interesting to test if the mechanism
also involves shedding of CD43.
A third process initiated in adherent neutrophils
exposed to TNF is a nearly instantaneous elevation in
the concentration of intracellular Ca2+ (REFS 80–82).
The functional impact has been understood to pertain
chiefly to actin reorganization83. However, neutrophils
were recently discovered to contain an enzyme termed
soluble adenylyl cyclase (sAC)82, which is characterized
by its lack of a transmembrane domain, independence
of G proteins and synergistic activation by Ca2+ and
bicarbonate. Activation of sAC by TNF led to activation
of RAP1A and phox82 (FIG. 5). Activation of sAC might
Figure 5 | Signalling proceeds through parallel, as well as intersecting, pathways
in adherent neutrophils responding to tumour-necrosis factor. Tumour-necrosis
factor (TNF)-triggers increased intracellular Ca2+ concentrations and activation of phos-
phatidylinositol 3-kinase (PI3K), the protein tyrosine kinases (PTKs) — spleen tyrosine
kinase, protein tyrosine kinase 2, haematopoietic-cell kinase and FGR — soluble adenylyl
cyclase (sAC) and the small G protein RAP1. These effects are all unexpected, based on
studies of TNF-induced signal transduction in other cells. The role of RAP guanine
nucleotide exchange factor 3 (RAPGEF3; also known as EPAC1) is not yet determined.
Other pathways are also activated but are not shown here. TNFR, TNF receptor.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | MARCH 2006 | 179
© 2006 Nature Publishing Group
O2–, H2O2, O3, HOCI, •OH
protease 3 and MMPs
The immediate medical
consequences of rapid
destruction of large numbers
of tumour cells, including the
release of their intracellular
(LAD). A rare hereditary
disease that is characterized
by recurrent infection and
delayed wound healing as a
consequence of defective
leukocyte adhesion. LAD type I
is caused by mutations of
β2-integrin; LAD type II is
caused by a defect in fucose
metabolism that results in a
failure to express sialyl-Lewis X,
the ligand for endothelial-cell
(E)-selectin and platelet
A genetic deficiency of
phagocyte oxidase (phox),
associated with recurrent,
and fungal infections.
therefore be another important consequence of the
TNF-induced transient increase in the level of Ca2+.
Until recently, it was thought that whatever activated
a neutrophil strongly enough to make it undergo a res-
piratory burst also forced it to degranulate. However,
two interventions were recently found to block the
TNF-induced respiratory burst without blocking
degranulation: introduction into the cell of a dominant-
negative fragment of PYK2 (REF. 84) and inhibition of
the TNF-induced Ca2+ increase by a compound termed
neucalcin82. These findings open up a new way thinking
about anti-inflammatory therapy.
Implications for therapy
The tissue-damaging power of neutrophils might
be appalling to the rheumatologist, but it should be
appealing to the oncologist. Whereas inflammation
can contribute to tumour formation and progression85,
neutro phils and/or eosinophils can have an important
role in tumour destruction in mice86–88 and, under cer-
tain circumstances, might do so in humans89. Once a
malignant tumour has become metastatic, it is fair game
to try to induce within it a neutrophil-rich inflammatory
response that is robust enough to destroy the tumour
directly89 or to damage its vasculature86. In a clinical
study, a complement-fixing, neutrophil-recruiting
tumour-specific monoclonal antibody was administered
to eight patients with end-stage metastatic melanoma90.
This was followed by a low dose of systemic TNF, in
an effort to activate intra-tumoural neutrophils. In one
patient, haemorrhagic necrosis ensued in all the tumour
deposits throughout the body, leading to the first
instance of tumour-lysis syndrome reported in a patient
with a solid tumour. Although only one of the patients
responded in this way, and he later died of a fungal infec-
tion90, neutrophil recruitment to and triggering within
tumours deserves further investigation. Neutrophils might
be especially valuable as anti-tumour effector cells when
one considers their enormous numerical preponderance
over tumour-specific cytotoxic T cells.
The usual goal of neutrophil-targeted pharmacol-
ogy is not to increase inflammation but is instead to
suppress it, for example in rheumatoid arthritis, osteo-
arthritis or chronic obstructive pulmonary disease
(FIG. 6). Unfortunately, conventional approaches to
neutrophil-based anti-inflammatory therapy have their
drawbacks. Antibodies or chemokine antagonists that
are designed to block neutrophil emigration from the
vasculature into the tissues will, at their most effective,
mimic leukocyte-adhesion deficiency, a frequently fatal set
of syndromes that impair host defence against infec-
tion26,27. Antibodies directed against adhesion molecules
expressed by lymphocytes have shown benefit in some
settings91 but have been associated with devastating
infections in others92–94. Based on the fact that life-
threatening infections generally appear much sooner in
neutropaenic patients than in lymphopaenic patients,
one might anticipate a greater risk of infection from the
administration of agents that interfere with the recruit-
ment of neutrophils to infected sites than has already
been seen after use of agents that block lymphocyte
recruitment. Efforts to hasten neutrophil apoptosis raise
the same concern.
An alternative approach has been to block individual
neutrophil enzymes. This might afford benefit, but
the approach is confounded by the redundancy of the
inflammatory pathways involved. In particular, inhibi-
tion of neutrophil elastase95 might be relatively ineffective
when one has not also inhibited cathepsin G96, protease 3
and MMPs. It has not been feasible to inhibit many pro-
teases at once because it was assumed that this would
require developing many pharmacologically acceptable
compounds and testing them together, which is unre-
alistic. Nor would one want to inhibit phox, because
this would be expected to mimic chronic granulomatous
disease, a condition whose original name began with the
The recent discovery of the Ca2+-triggered sAC path-
way in neutrophils might provide a route around these
obstacles. A screen of a chemical library was conducted to
identify compounds that block the activation of phox by
TNF or formylated peptides but not by phorbol esters82.
Compounds fulfilling these criteria did not inhibit phox
itself because triggering of phox by phorbol esters was
preserved; the compounds only inhibited activation
of phox, and even then, only its activation by soluble
inflamma tory factors. One such compound, termed neu-
calcin for its ability to block the TNF-induced elevation of
intracellular Ca2+, did not inhibit the activation of phox
by phagocytosis of bacteria, did not inhibit degranula-
tion, and as expected from the foregoing results, did
not inhibit killing of bacteria. Moreover, in the presence
of neu calcin, neutrophils migrated normally across
TNF-activated endothelial monolayers82.
Figure 6 | Interdependence of the two main classes of neutrophil tissue-damaging
products creates opportunities for anti-inflammatory strategies. Interruption
of neutrophil accumulation or shortening of neutrophil survival in the tissues risks
suppressing all neutrophil-dependent antimicrobial functions in extravascular sites.
Inhibition of an individual protease might be relatively ineffective, whereas inhibition
of the action of phagocyte oxidase (phox) risks overwhelming infection. However,
inhibition of the activation of phox by inflammatory mediators, if at the same time
preserving phox activation by bacteria, might lead to reduced oxidative damage to
proteins that would normally increase their susceptibility to proteolysis. This could
also lead to reduced oxidative activation of proteases in neutrophils and in other
cells; reduced oxidative inactivation of anti-proteases, with resulting partial
inhibition of many proteases at once; and reduced oxidative induction of pro-
inflammatory transcription factors. AP1, activator protein 1; H2O2, hydrogen peroxide;
HOCl, hypochlorous acid; MMP, matrix metalloproteinase; O2
•OH, hydroxyl radical; NF-κB; nuclear factor-κB.
–, superoxide; O3, ozone;
180 | MARCH 2006 | VOLUME 6
© 2006 Nature Publishing Group
Many questions remain unanswered, including
what is the identity of the molecular target of neu-
calcin and what actions such compounds might exert
in vivo. However, even at this early stage, the foregoing
observations raise the possibility that inhibition of the
non-bacterial, inflammatory activation of phox might
provide an indirect route to partial inhibition of pro-
teases at sites of inflammation. Inflammatory proteoly-
sis has multiple components that involve neutrophils:
first, neutrophils release their own proteases, in part
through the action of phox59; second, neutrophils
oxidatively activate some of the proteases provided
by other cells; and last, proteases are opposed by anti-
proteases. Proteolysis ensues when there is a disruption
in the protease–antiprotease balance. At least three of
the main tissue anti-proteases (SLPI, α1-antitrypsin and
α2-macroglobulin) are subject to oxidative inactivation
by neutrophils. Therefore, inhibition of the inflammatory
activation of phox might tilt the protease–anti-protease
balance against proteolysis, without impairing neutrophil
recruitment and antimicrobial activity.
It is time to set aside the view that neutrophils are destruc-
tive cells that arrive too early, lash out too blindly and live
too briefly to be of interest to immunologists. Neutrophils
are life-saving decision-makers that coach DCs, mono-
cytes and lymphocytes and help the organism decide
whether to initiate and maintain an immune response.
They are unique in their capacity both to destroy and help
heal any tissue in the body. Understanding the circuits
that confer and control such behaviour is as challeng-
ing a problem as any other in cell biology. Meeting that
challenge holds therapeutic promise in diverse settings
of intense interest to immunologists, from metastatic
tumours to ravaged joints.
Nathan, C. Inflammation: points of control. Nature
420, 846–852 (2002).
Nathan, C. Specificity of a third kind: reactive oxygen
and nitrogen intermediates in cell signalling. J. Clin.
Invest. 111, 769–778 (2003).
Henson, P. M. & Johnston, R. B. Jr. Tissue injury
in inflammation. Oxidants, proteinases, and
cationic proteins. J. Clin. Invest. 79, 669–674
Weiss, S. J. Tissue destruction by neutrophils. N. Engl.
J. Med. 320, 365–376 (1989).
Lekstrom-Himes, J. A. & Gallin, J. I. Immunodeficiency
diseases caused by defects in phagocytes. N. Engl. J.
Med. 343, 1703–1714 (2000).
Klebanoff, S. J. & Clark, R. A. The Neutrophil: Function
and Clinical Disorders (Elsevier/North-Holland,
Witko-Sarsat, V., Rieu, P., Descamps-Latscha, B.,
Lesavre, P. & Halbwachs-Mecarelli, L. Neutrophils:
molecules, functions and pathophysiological aspects.
Lab. Invest. 80, 617–653 (2000).
Matzinger, P. The danger model: a renewed sense of
self. Science 296, 301–305 (2002).
Medzhitov, R. & Janeway, C. A. Jr. Decoding the
patterns of self and nonself by the innate immune
system. Science 296, 298–300 (2002).
10. Backhed, F., Ley, R. E., Sonnenburg, J. L.,
Peterson, D. A. & Gordon, J. I. Host-bacterial
mutualism in the human intestine. Science 307,
11. Chertov, O. et al. Identification of human neutrophil-
derived cathepsin G and azurocidin/CAP37 as
chemoattractants for mononuclear cells and
neutrophils. J. Exp. Med. 186, 739–747
12. Bennouna, S., Bliss, S. K., Curiel, T. J. & Denkers, E. Y.
Cross-talk in the innate immune system: neutrophils
instruct recruitment and activation of dendritic cells
during microbial infection. J. Immunol. 171,
13. Tsuda, Y. et al. Three different neutrophil subsets
exhibited in mice with different susceptibilities to
infection by methicillin-resistant Staphylococcus
aureus. Immunity 21, 215–226 (2004).
14. Wittamer, V. et al. Neutrophil-mediated maturation of
chemerin: a link between innate and adaptive
immunity. J. Immunol. 175, 487–493 (2005).
15. van Gisbergen, K. P., Sanchez-Hernandez, M.,
Geijtenbeek, T. B. & van Kooyk, Y. Neutrophils mediate
immune modulation of dendritic cells through
glycosylation-dependent interactions between Mac-1
and DC-SIGN. J. Exp. Med. 201, 1281–1292 (2005).
16. Scapini, P. et al. Proinflammatory mediators elicit
secretion of the intracellular B-lymphocyte stimulator
pool (BLyS) that is stored in activated neutrophils:
implications for inflammatory diseases. Blood 105,
17. Ethuin, F. et al. Human neutrophils produce
interferon-γ upon stimulation by interleukin-12.
Lab. Invest. 84, 1363–1371 (2004).
18. Schmielau, J. & Finn, O. J. Activated granulocytes and
granulocyte-derived hydrogen peroxide are the
underlying mechanism of suppression of T-cell function
in advanced cancer patients. Cancer Res. 61,
19. Heissig, B. et al. Recruitment of stem and progenitor
cells from the bone marrow niche requires MMP-9
mediated release of kit-ligand. Cell 109, 625–637
20. Semerad, C. L., Liu, F., Gregory, A. D., Stumpf, K. &
Link, D. C. G-CSF is an essential regulator of
neutrophil trafficking from the bone marrow to the
blood. Immunity 17, 413–423 (2002).
21. Stark, M. A. et al. Phagocytosis of apoptotic
neutrophils regulates granulopoiesis via IL-23 and
IL-17. Immunity 22, 285–294 (2005).
22. Diefenbach, A. et al. Type 1 interferon (IFNα/β) and
type 2 nitric oxide synthase regulate the innate
immune response to a protozoan parasite. Immunity
8, 77–87 (1998).
23. Dovi, J. V., Szpaderska, A. M. & DiPietro, L. A.
Neutrophil function in the healing wound: adding
insult to injury? Thromb. Haemost. 92, 275–280
24. Martin, P. et al. Wound healing in the PU. 1 null
mouse — tissue repair is not dependent on
inflammatory cells. Curr. Biol. 13, 1122–1128 (2003).
25. Singer, A. J. & Clark, R. A. Cutaneous wound healing.
N. Engl. J. Med. 341, 738–746 (1999).
26. Roos, D. & Law, S. K. Hematologically important
mutations: leukocyte adhesion deficiency. Blood Cells
Mol. Dis. 27, 1000–1004 (2001).
27. Alon, R. & Etzioni, A. LAD-III, a novel group of
leukocyte integrin activation deficiencies. Trends
Immunol. 24, 561–566 (2003).
28. Nathan, C. F. Neutrophil activation on biological
surfaces. Massive secretion of hydrogen peroxide
in response to products of macrophages and
lymphocytes. J. Clin. Invest. 80, 1550–1560
29. Serhan, C. N. Novel ω-3-derived local mediators in
anti-inflammation and resolution. Pharmacol. Ther.
105, 7–21 (2005).
30. Jin, F. Y., Nathan, C., Radzioch, D. & Ding, A.
Secretory leukocyte protease inhibitor: a macrophage
product induced by and antagonistic to bacterial
lipopolysaccharide. Cell 88, 417–426 (1997).
31. Zhu, J. et al. Conversion of proepithelin to epithelins:
roles of SLPI and elastase in host defense and wound
repair. Cell 111, 867–878 (2002).
32. Grobmyer, S. R. et al. Secretory leukocyte protease
inhibitor, an inhibitor of neutrophil activation, is
elevated in serum in human sepsis and experimental
endotoxemia. Crit. Care Med. 28, 1276–1282
33. He, Z., Ong, C. H., Halper, J. & Bateman, A.
Progranulin is a mediator of the wound response.
Nature Med. 9, 225–229 (2003).
34. Klebanoff, S. J. Myeloperoxidase: friend and foe.
J. Leukoc. Biol. 77, 598–625 (2005).
35. Hedrick, S. M. The acquired immune system: a
vantage from beneath. Immunity 21, 607–615
36. Casanova, J. L. & Abel, L. Inborn errors of immunity
to infection: the rule rather than the exception.
J. Exp. Med. 202, 197–201 (2005).
37. Nathan, C. The moving frontier in nitric oxide-
dependent signalling. Sci. STKE 2 Nov 2004
38. Nathan, C. Nitric oxide as a secretory product of
mammalian cells. FASEB. J. 6, 3051–3064
39. Tonks, N. K. Redox redux: revisiting PTPs and the
control of cell signalling. Cell 121, 667–670 (2005).
40. Li, Y., Karlin, A., Loike, J. D. & Silverstein, S. C.
A critical concentration of neutrophils is required
for effective bacterial killing in suspension. Proc. Natl
Acad. Sci. USA 99, 8289–8294 (2002).
41. Savill, J. & Haslett, C. Granulocyte clearance by
apoptosis in the resolution of inflammation. Semin.
Cell Biol. 6, 385–393 (1995).
42. Peyssonnaux, C. et al. HIF-1α expression regulates the
bactericidal capacity of phagocytes. J. Clin. Invest.
115, 1806–1815 (2005).
43. Walmsley, S. R. et al. Hypoxia-induced neutrophil
survival is mediated by HIF-1α-dependent NF-κB
activity. J. Exp. Med. 201, 105–115 (2005).
44. Jonsson, H., Allen, P. & Peng, S. L. Inflammatory
arthritis requires Foxo3a to prevent Fas ligand-
induced neutrophil apoptosis. Nature Med. 11,
45. Altznauer, F. et al. Inflammation-associated cell cycle-
independent block of apoptosis by survivin in
terminally differentiated neutrophils. J. Exp. Med.
199, 1343–1354 (2004).
46. Flo, T. H. et al. Lipocalin 2 mediates an innate immune
response to bacterial infection by sequestrating iron.
Nature 432, 917–921 (2004).
47. Brinkmann, V. et al. Neutrophil extracellular traps kill
bacteria. Science 303, 1532–1535 (2004).
48. Striz, I. & Trebichavsky, I. Calprotectin- a pleiotropic
molecule in acute and chronic inflammation. Physiol.
Res. 53, 245–253 (2004).
49. Wright, G. W., Ooi, C. E., Weiss, J. & Elsbach, P.
Purification of a cellular (granulocyte) and an
extracellular (serum) phospholipase A2 that
participate in the destruction of Escherichia coli in a
rabbit inflammatory exudate. J. Biol. Chem. 265,
50. Faurschou, M. & Borregaard, N. Neutrophil granules
and secretory vesicles in inflammation. Microbes
Infect. 5, 1317–1327 (2003).
51. Selsted, M. E. & Ouellette, A. J. Mammalian defensins
in the antimicrobial immune response. Nature
Immunol. 6, 551–557 (2005).
52. Weiss, J., Elsbach, P., Olsson, I. & Odeberg, H.
Purification and characterization of a potent
bactericidal and membrane active protein from the
granules of human polymorphonuclear leukocytes.
J. Biol. Chem. 253, 2664–2672 (1978).
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | MARCH 2006 | 181
© 2006 Nature Publishing Group
53. Campanelli, D., Detmers, P. A., Nathan, C. F. & Gabay,
J. E. Azurocidin and a homologous serine protease
from neutrophils. Differential antimicrobial and
proteolytic properties. J. Clin. Invest. 85, 904–915
54. Pereira, H. A., Shafer, W. M., Pohl, J., Martin, L. E. &
Spitznagel, J. K. CAP37, a human neutrophil-derived
chemotactic factor with monocyte specific activity.
J. Clin. Invest. 85, 1468–1476 (1990).
55. McCord, J. M. & Fridovich, I. Superoxide dismutase.
An enzymic function for erythrocuprein (hemocuprein).
J. Biol. Chem. 244, 6049–6055 (1969).
56. Babior, B. M., Kipnes, R. S. & Curnutte, J. T. Biological
defense mechanisms. The production by leukocytes
of superoxide, a potential bactericidal agent.
J. Clin. Invest. 52, 741–744 (1973).
57. Lambeth, J. D. NOX enzymes and the biology of
reactive oxygen. Nature Rev. Immunol. 4, 181–189
58. Good, R. A. et al. Fatal (chronic) granulomatous
disease of childhood: a hereditary defect of leukocyte
function. Semin. Hematol. 5, 215–254 (1968).
59. Reeves, E. P. et al. Killing activity of neutrophils is
mediated through activation of proteases by K+ flux.
Nature 416, 291–297 (2002).
60. Roos, D., van Bruggen, R. & Meischl, C. Oxidative
killing of microbes by neutrophils. Microbes Infect. 5,
61. Shiloh, M. U. et al. Phenotype of mice and
macrophages deficient in both phagocyte oxidase and
inducible nitric oxide synthase. Immunity 10, 29–38
62. Staudinger, B. J., Oberdoerster, M. A., Lewis, P. J. &
Rosen, H. mRNA expression profiles for Escherichia
coli ingested by normal and phagocyte oxidase-
deficient human neutrophils. J. Clin. Invest. 110,
63. Liu, G. Y. et al. Staphylococcus aureus golden pigment
impairs neutrophil killing and promotes virulence
through its antioxidant activity. J. Exp. Med. 202,
64. Nathan, C. F. Respiratory burst in adherent human
neutrophils: triggering by colony-stimulating factors
CSF–GM and CSF–G. Blood 73, 301–306
65. Wolpe, S. D. et al. Macrophages secrete a novel
heparin-binding protein with inflammatory and
neutrophil chemokinetic properties. J. Exp. Med. 167,
66. Nathan, C. et al. Cytokine-induced respiratory burst
of human neutrophils: dependence on extracellular
matrix proteins and CD11/CD18 integrins.
J. Cell. Biol. 109, 1341–1349 (1989).
67. Nathan, C. & Sporn, M. Cytokines in context.
J. Cell Biol. 113, 981–986 (1991).
68. Blouin, E., Halbwachs-Mecarelli, L. & Rieu, P. Redox
regulation of β2-integrin CD11b/CD18 activation.
Eur. J. Immunol. 29, 3419–3431 (1999).
69. Mocsai, A., Ligeti, E., Lowell, C. A. & Berton, G.
Adhesion-dependent degranulation of neutrophils
requires the Src family kinases Fgr and Hck.
J. Immunol. 162, 1120–1126 (1999).
70. Fuortes, M., Melchior, M., Han, H., Lyon, G. J. &
Nathan, C. Role of the tyrosine kinase pyk2 in the
integrin-dependent activation of human neutrophils by
TNF. J. Clin. Invest. 104, 327–335 (1999).
71. Mocsai, A., Zhou, M., Meng, F., Tybulewicz, V. L. &
Lowell, C. A. Syk is required for integrin signalling in
neutrophils. Immunity 16, 547–558 (2002).
72. Yan, S. R., Huang, M. & Berton, G. Signalling by
adhesion in human neutrophils: activation of the
p72syk tyrosine kinase and formation of protein
complexes containing p72syk and Src family kinases in
neutrophils spreading over fibrinogen. J. Immunol.
158, 1902–1910 (1997).
73. Korchak, H. M. & Kilpatrick, L. E. TNFα elicits
association of PI3-kinase with the p60TNFR and
activation of PI3-kinase in adherent neutrophils.
Biochem. Biophys. Res. Commun. 281, 651–656
74. Zhao, T., Benard, V., Bohl, B. P. & Bokoch, G. M.
The molecular basis for adhesion-mediated
suppression of reactive oxygen species generation by
human neutrophils. J. Clin. Invest. 112, 1732–1740
75. Gakidis, M. A. et al. Vav GEFs are required for β2
integrin-dependent functions of neutrophils. J. Cell
Biol. 166, 273–282 (2004).
76. Fuortes, M., Jin, W. W. & Nathan, C. Adhesion-
dependent protein tyrosine phosphorylation in
neutrophils treated with tumour necrosis factor.
J. Cell Biol. 120, 777–784 (1993).
77. Fuortes, M., Jin, W. W. & Nathan, C. β2 integrin-
dependent tyrosine phosphorylation of paxillin in
human neutrophils treated with tumour necrosis
factor. J. Cell Biol. 127, 1477–1483 (1994).
78. Nathan, C., Xie, Q. W., Halbwachs-Mecarelli, L. &
Jin, W. W. Albumin inhibits neutrophil spreading and
hydrogen peroxide release by blocking the shedding of
CD43 (sialophorin, leukosialin). J. Cell Biol. 122,
79. Raptis, S. Z., Shapiro, S. D., Simmons, P. M.,
Cheng, A. M. & Pham, C. T. Serine protease cathepsin
G regulates adhesion-dependent neutrophil effector
functions by modulating integrin clustering. Immunity
22, 679–691 (2005).
80. Richter, J., Ng-Sikorski, J., Olsson, I. & Andersson, T.
Tumour necrosis factor-induced degranulation in
adherent human neutrophils is dependent on CD11b/
CD18-integrin-triggered oscillations of cytosolic free
Ca2+. Proc. Natl Acad. Sci. USA 87, 9472–9476
81. Schumann, M. A., Gardner, P. & Raffin, T. A.
Recombinant human tumour necrosis factor α induces
calcium oscillation and calcium-activated chloride
current in human neutrophils. The role of calcium/
calmodulin-dependent protein kinase. J. Biol. Chem.
268, 2134–2140 (1993).
82. Han, H. et al. Calcium-sensing soluble adenylyl
cyclase mediates TNF signal transduction in human
neutrophils. J. Exp. Med. 202, 353–361 (2005).
83. Bengtsson, T. et al. Actin dynamics in human
neutrophils during adhesion and phagocytosis is
controlled by changes in intracellular free calcium.
Eur. J. Cell. Biol. 62, 49–58 (1993).
84. Han, H., Fuortes, M. & Nathan, C. Critical role of the
carboxyl terminus of proline-rich tyrosine kinase
(Pyk2) in the activation of human neutrophils by
tumour necrosis factor: separation of signals for the
respiratory burst and degranulation. J. Exp. Med.
197, 63–75 (2003).
85. Coussens, L. M. & Werb, Z. Inflammation and cancer.
Nature 420, 860–867 (2002).
86. Rothstein, J. L. & Schreiber, H. Synergy between
tumour necrosis factor and bacterial products causes
hemorrhagic necrosis and lethal shock in normal
mice. Proc. Natl Acad. Sci. USA 85, 607–611
87. Tepper, R. I., Pattengale, P. K. & Leder, P. Murine
interleukin-4 displays potent anti-tumour activity
in vivo. Cell 57, 503–512 (1989).
88. Stoppacciaro, A. et al. Regression of an established
tumour genetically modified to release granulocyte
colony-stimulating factor requires granulocyte–T-cell
cooperation and T cell-produced interferon γ.
J. Exp. Med. 178, 151–161 (1993).
89. Soiffer, R. et al. Vaccination with irradiated autologous
melanoma cells engineered to secrete human
granulocyte-macrophage colony-stimulating factor
generates potent antitumour immunity in patients
with metastatic melanoma. Proc. Natl Acad. Sci. USA
95, 13141–13146 (1998).
90. Minasian, L. M. et al. Hemorrhagic tumour necrosis
during a pilot trial of tumour necrosis factor-α and
anti-GD3 ganglioside monoclonal antibody in patients
with metastatic melanoma. Blood 83, 56–64 (1994).
91. Feagan, B. G. et al. Treatment of ulcerative colitis
with a humanized antibody to the α4β7 integrin.
N. Engl. J. Med. 352, 2499–2507 (2005).
92. Berger, J. R. & Koralnik, I. J. Progressive multifocal
leukoencephalopathy and natalizumab — unforeseen
consequences. N. Engl. J. Med. 353, 414–416
93. Langer-Gould, A., Atlas, S. W., Green, A. J.,
Bollen, A. W. & Pelletier, D. Progressive multifocal
leukoencephalopathy in a patient treated with
natalizumab. N. Engl. J. Med. 353, 375–381 (2005).
94. Kleinschmidt-DeMasters, B. K. & Tyler, K. L.
Progressive multifocal leukoencephalopathy
complicating treatment with natalizumab and
interferon β-1a for multiple sclerosis. N. Engl. J. Med.
353, 369–374 (2005).
95. Luisetti, M. et al. MR889, a neutrophil elastase
inhibitor, in patients with chronic obstructive
pulmonary disease: a double-blind, randomized,
placebo-controlled clinical trial. Eur. Respir. J. 9,
96. de Garavilla, L. et al. A novel, potent dual inhibitor of
the leukocyte proteases cathepsin G and chymase:
molecular mechanisms and anti-inflammatory activity
in vivo. J. Biol. Chem. 280, 18001–18007 (2005).
I thank A. Ding, M. Fuortes, W. A. Muller and J. Upshaw for
critical reviews and apologize that the scope of the topic pre-
vented citation of many important studies. The Department of
Microbiology and Immunology acknowledges the support
of the William Randolph Hearst Foundation.
Competing interests statement
The author declares no competing financial interests.
The following terms in this article are linked online to:
Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
Azurocidin | cathepsin G | G-CSF | MPO | protease 3 |
neutrophil elastase | PEPI | SLPI | TNF
Carl Nathan’s laboratory: http://www.med.cornell.edu/
See online article: S1 (box)
Access to this links box is available online.
182 | MARCH 2006 | VOLUME 6