(Journal of Leukocyte Biology. 2000;68:593-602.)
© 2000
by Society for Leukocyte Biology
Cellular signaling in macrophage migration and chemotaxis
Gareth E. Jones
The Randall Centre for Molecular Mechanisms of Cell Function, King’s College London, United Kingdom
Correspondence: Gareth E. Jones, The Randall Centre for Molecular Mechanisms of Cell Function, King’s College London, New Hunt’s House, Guy’s Campus, London SE1 1UL, UK. E-mail:
gareth.jones{at}kcl.ac.uk
 |
ABSTRACT
|
|---|
Whereas most cells in adult tissues are fixed in place by cell
junctions, leukocytes are motile and able to migrate actively through
the walls of blood vessels into surrounding tissues. The actin
cytoskeleton of these cells plays a central role in locomotion,
phagocytosis, and the regulation of cell shape that are crucial
elements of neutrophil and monocyte/macrophage function. This review
will concentrate on how macrophages in particular control the actin
cytoskeleton to generate cell movement and the shape changes required
for chemotaxis. It has recently become evident that a complex of seven
proteins known as the Arp2/3 complex regulates the assembly of new
actin filament networks at the leading front of moving cells. Proteins
of the Wiskott-Aldrich Syndrome Protein (WASP) family bind directly to
the Arp2/3 complex and stimulate its ability to promote the nucleation
of new actin filaments. Upstream of the WASP family proteins, receptor
tyrosine kinases, G-protein-coupled receptors, phosphoinositide-3-OH
kinase (PI 3-kinase), and the Rho family of GTPases receive and
transduce the signals that lead to actin nucleation through WASP-Arp2/3
action. Although many gaps remain in our understanding, we are now in a
position to consider completing signaling pathways that are initiated
from outside the cell to the actin rearrangements that drive cell
motility and chemotaxis.
Key Words: phosphoinositide 3-kinases • GTPases • signal transduction • actin cytoskeleton • cell polarity • cell locomotion • review
|
INTRODUCTION
|
|---|
The chemoattraction of macrophages and neutrophils into tissues is
an essential step in the host response to infection. It is commonly
held that this so-called innate or nonspecific immune response protects
us from at least 98% of the pathogens we encounter, so there is little
doubt that the migratory capacity of these leukocytes is of fundamental
importance to our survival. Although the dynamics of cell migration are
most frequently analyzed in tissue-cultured fibroblasts
[1
, 2
], there are fundamental similarities
in the locomotory steps within all vertebrate cells that suggest that
they share a common motile machinery. This concept is best demonstrated
by the observation that isolated cultured macrophages display short
surface protrusions, termed filopodia or microspikes, which are
extensions of about 0.1–0.2 µm in diameter and up to 20 µm in
length, supported by a core of bundled actin filaments
(microfilaments). In both macrophages and fibroblasts filopodia support
thin veils or sheets of membrane-enclosed cytoplasm, termed
lamellipodia, containing a meshwork of myosin II-associated
microfilaments. In macrophages, as in other cell types, the actin
meshwork within the lamellipodia, in association with numerous
structural and regulatory proteins, constitutes the molecular motor for
cell locomotion. Interested readers are referred to the excellent core
reviews cited above plus more recent material [3
].
The locomotory apparatus of most cells works against cell-to-substratum
adhesions usually referred to as focal contacts or focal adhesions.
These structures link adhesive extracellular matrix proteins to myosin
II-containing bundles of cytoplasmic microfilaments (stress fibers) via
members of the integrin family of proteins [4
]. While
integrin-mediated contacts to the substratum also exist in macrophages,
these take two forms: focal complexes that are structurally related to
focal contacts but lack stress fibers [5
], and
podosomes, distinct circular structures that seem to be largely
restricted to cells of the myeloid lineage [6
,
7
].
The forward movement of macrophages can be divided into steps:
protrusion of filopodia and lamellipodia at the leading front, adhesion
of the protruding edge to the substratum via focal complexes,
contraction of the cytoplasmic actomyosin, and finally release from
contact sites at the tail of the cell [8
,
9
]. A number of molecular events need to be integrated in
order to allow a cell to move across a substratum, and it appears that
this coordination is largely mediated by the actin microfilaments
within the cytoplasm. In addition, the actin cytoskeleton is a key
mediator of cell polarization and the directed migration of macrophages
toward chemoattractant (chemotaxis). This review will examine the
evidence that supports these conclusions, dealing with aspects of
chemoattractant signals, regulatory intracellular signaling pathways,
and cytoskeletal reorganization as appropriate.
|
CHEMOATTRACTANT SIGNALS AND RECEPTORS
|
|---|
Leukocytes sense a wide variety of chemoattractants that can in
principle guide their migratory path within tissues [10
,
11
]. Cells are responsive to bacterial components,
leukotrienes, complement factors, and chemokines, an ever-expanding
family of attractants controlling leukocyte migration
[12
]. All these attractants in both neutrophils and
monocytes/macrophages interact with specific serpentine (heptahelical)
receptors [13
], embedded in the plasma membrane, that
transduce ligand-induced signals by coupling to heterotrimeric G
proteins (Fig. 1
). The G protein complex dissociates into 
and ß
subunits,
which in turn bind and activate target enzymes such as phospholipase C,
phosphoinositide 3-kinase (PI 3-kinase), or adenyl cyclase. These
enzymes generate intracellular messengers that initiate a cascade of
events that culminate in the cytoskeletal and chemoattraction response
to ligand-induced receptor activation. Similar systems are also found
in primitive organisms such as the free-living amoeba
Dictyostelium discoideum [14
], demonstrating
the near-universality of this signal transduction system. Five recent
articles [15
16
17
18
19
] together provide convincing evidence
of a role for PI 3-kinase in serpentine receptor signaling and
chemotaxis. At least four Class I PI 3-kinase isoforms exist in
mammalian cells [20
], but only one form, a single Class
IB variant containing the p110 catalytic subunit
complexed with a 101-kDa regulatory protein, is thought to interact
with G-proteins in leukocytes. Because leukocyte chemokine responses
are sensitive to pertussis toxin, it has been assumed that chemokine
receptors are coupled to a Gi, and a large number of
observations support this view [13
]. However, this
assumption requires modification because there are many data in the
literature implicating not only other families of G-protein
subunits in signaling [21
], but also Gß subunits
[22
]. Whereas some have reported that Class
IB PI 3-kinase is responsive to activation by G subunits
[23
], others have shown that p110 becomes activated
through an interaction with the ß subunits [24
].
Whatever the case, all appear to converge on a common pathway where the
outcome will lead to the phosphorylation of
phosphatidylinositol-4,5-bisphosphate (PIP2) by activated
PI 3-kinase. As a result, phosphatidylinositol-3,4,5-triphosphate
(PIP3) is generated (Fig. 1)
. The work cited earlier
[15
] on mice lacking the p110 catalytic subunit of PI
3-kinase convincingly show that leukocytes lacking this
G-protein-specific isoform are unable to produce PIP3 and
that this has a striking effect on the ability of macrophages and
neutrophils to migrate. Chemotaxis mediated via serpentine receptors is
severely reduced and there was a general failure to clear bacteria from
the peritoneal cavity [17
].
Although it is commonly assumed that the chemotactic behavior induced
in leukocytes will act through serpentine receptors
[25
], monocytes, and macrophages at least are also
capable of responding to chemoattractants that signal through surface
receptor tyrosine kinases [26
]. This type of receptor
(Fig. 2
) is linked to the signal transduction pathways that regulate
mitogenesis and cell differentiation, and indeed monocytes and
macrophages posses a canonical receptor tyrosine kinase that regulates
just these parameters: colony-stimulating factor-1 receptor (CSF-1R)
[27
, 28
]. However, it was shown some time
ago [29
, 30
] that the actin cytoskeleton
was also a target for CSF-1. Re-addition of CSF-1 to quiescent BAC1.2F5
macrophages stimulates rapid cytoskeletal reorganization and cell
motility within a few minutes, followed by chemotactic migration up a
gradient of diffusing cytokine [31
]. Given the fact that
major sources of endogenous CSF-1 include activated endothelia and
tissue fibroblasts as well as macrophage-recruiting mammary gland
carcinomas [32
], the chemotactic properties of this
cytokine for macrophages and monocytes is likely to be physiologically
significant. Unlike serpentine receptors, receptor tyrosine kinases
such as CSF-1R directly interact with a host of substrates after
autophosphorylation induced by ligand binding, although in
CSF-1-treated mouse macrophages, PI 3-kinase is the major protein
associated with the activated receptor [33
]. A full
review of the biology of CSF-1 signal transduction is available
[27
], but it is worth emphasizing that three Class
IA isoforms of PI 3-kinase are involved. The p110 subunits
in these PI 3-kinases exist in complex with a p85 protein that has two
Src-homology-2 (SH2) domains (Fig. 2) . The latter bind to
phosphorylated tyrosine residues found on activated CSF-1R, thus
allowing translocation to the plasma membrane where their lipid (such
as PIP2) and other substrates are found
[34
]. All mammalian cells so far examined express at
least one Class IA PI 3-kinase isoform: p110 and p110ß
are widely distributed in tissues, but p110
is normally restricted
to leukocytes. It was shown earlier that CSF-1R induces direct
interaction of PI 3-kinase (via its p85 subunit) with the SH2/SH3
adaptor protein Grb2 [35
]. More recently, we found that
although all three Class IA p110 isoforms were equally
recruited to activated CSF-1R, subsequent signaling to the actin
cytoskeleton of macrophages was differentially regulated. Antibody
against p110 blocked CSF-1-induced DNA synthesis but did not affect
CSF-1-induced actin rearrangements or cell migration. However,
antibodies against p110ß and p110 had the converse effect, with
the latter isoform completely abrogating cell migration
[36
]. As was described for serpentine receptors, the net
result of PI 3-kinase activation through receptor tyrosine kinases is
the generation of PIP3 (Fig. 2)
.
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Early signal transduction events mediated through receptor tyrosine
kinases. Homodimers of cytokine (such as CSF-1) induce a rapid sequence
of changes. Ligand-induced covalent dimerization of receptor leads to
autophosphorylation at a number of tyrosine residues on the cytoplasmic
tail, which initiates signaling events that precede the rapid
internalization and subsequent degradation of receptor-ligand
complexes. Multiple signaling events occur on activated receptors,
including recruitment of Class 1A PI 3-kinase isoforms via
binding of the p85 subunit to SH2 binding sites created on the receptor
by autophosphorylation. Once again, recruitment to the plasma membrane
of any of the p110 catalytic subunits results in the generation of
PIP3.
|
|
PIP3 is a target for many pleckstrin-homology (PH)
domain-containing proteins [37
], which activate kinases
and small GTPases. PIP2- and PIP3-binding PH
domains are found in a diverse array of proteins, including
nucleotide-exchange factors (e.g., Vav, GRP1, ARNO, Tiam-1, Sos1),
GTPase-activating factors (e.g., GAP1m), phospholipases
(e.g., PLC2), protein kinases (e.g., PKB, Btk, PDK1), and adaptor
proteins [38
]. As will be described later, these are
important components in the regulation of the actin cytoskeleton
leading to the migration response of cells.
|
RHO FAMILY PROTEINS
|
|---|
Small GTPases of the Rho family are pivotal regulators of
signaling networks that are activated by chemokine and cytokine
receptors as well as other receptor types [39
]. An
introduction to these proteins is provided below, but for those who
wish to see more detail, many excellent reviews exist and the reader is
referred to recent examples [40
].
The Rho family is part of the Ras superfamily of small (around 21 kDa)
GTP-binding proteins. To date, 15 mammalian members of the Rho family
have been identified: Rho (A, B, C, D, E, G), Rac (1, 2, 3), Cdc42 (two
alternatively spliced variants of the same gene with different
carboxy-terminal sequences), Rnd1/Rnd6, Rnd2/Rho7, TC10, and TTF. Of
the mammalian proteins, the best characterized for their ability to
regulate actin organization are RhoA, Rac1, and Cdc42. Rho was the
first member of this family to be cloned in 1985, followed a few years
later by Rac and Cdc42. The most frequently used tool for studying Rho
function is C3 transferase, an exoenzyme from Clostridium
botulinum, which ADP-ribosylates and inactivates Rho
[41
]. Treatment of many cell types with C3 transferase
induces loss of stress fibers, and this was the first indication that
Rho influences the actin cytoskeleton. Subsequently, Rac was also shown
to regulate actin organization, and at the same time was independently
purified as an essential cofactor for the NADPH oxidase in phagocytic
cells [42
].
|
RHO FAMILY PROTEINS AND THE ACTIN CYTOSKELETON
|
|---|
The roles of Rho, Rac, and Cdc42 in regulating actin organization
were first characterized in detail in Swiss 3T3 fibroblasts. These
cells have proven to be a good model system for analyzing rapid changes
in the actin cytoskeleton, because when confluent and serum-starved
they lose practically all of the two most prominent actin
filament-containing structures found in fibroblasts: stress fibers and
lamellipodia. Stress fibers are bundles of actin filaments associated
with myosin II filaments and other proteins, forming contractile
fibers. They terminate at the plasma membrane in focal adhesions, where
transmembrane integrins are clustered and associate both with
extracellular matrix proteins outside the cell and with a large number
of proteins inside the cell [43
]. Lamellipodia are
broad, highly dynamic membrane protrusions that extend and retract
through a combination of actin polymerization at the plasma membrane,
depolymerization within the cytoplasm, and myosin-mediated rearward
movement of the actin fibers [2
].
Constitutively active mutants of Rho and Rac induce the formation of
stress fibers and lamellipodia, respectively, when microinjected into
quiescent Swiss 3T3 cells. Conversely, microinjection of C3 transferase
to inhibit Rho or of a dominant-negative Rac mutant to inhibit Rac
inhibits growth factor-induced formation of these structures
[41
]. Activated Cdc42 protein induces the extension of
filopodia [44
], finger-like plasma membrane protrusions
containing actin filament bundles, which actively protrude and retract.
Under appropriate conditions, Cdc42, Rac, and Rho can activate each
other sequentially in a cascade: Cdc42 can induce Rac-mediated
lamellipodium formation, and Rac can induce Rho-mediated stress fiber
formation [45
].
Rho, Rac, and Cdc42 also regulate the assembly of adhesion sites
to the extracellular matrix in fibroblasts. Rho mediates the formation
of focal adhesions, whereas Rac and Cdc42 induce the formation of
smaller adhesion sites (focal complexes) to the extracellular matrix,
located in lamellipodia and at the bases of filopodia
[41
]. To determine how Rho, Rac, and Cdc42 act in cell
types other than fibroblasts, we used a mouse macrophage cell line,
BAC1.2F5, that resembles primary macrophages in being dependent upon
CSF-1 for survival and proliferation as well as exhibiting many of the
markers of normal activated macrophages [29
].
Constitutively active and dominant-negative mutants of Rho, Rac, and
Cdc42 were injected into BAC1 macrophages and assessed for their
effects on the actin cytoskeleton and on adhesion sites
[5
]. As in fibroblasts, Rac induces the formation of
lamellipodia and membrane ruffles and is required for the formation of
these structures in response to CSF-1. Cdc42 induces rapid formation of
filopodia and again is required for CSF-1-induced filopodium extension.
These cells do not possess stress fibers, but have very fine actin
cables within the cytoplasm, running parallel to the plasma membrane
and around the nucleus. These cables are not detectable in
cytokine-starved cells, but re-appear upon stimulation with CSF-1 after
15–30 min. Rho is required for this response and is activated
downstream of Rac. The ability of Rac to activate Rho is thus conserved
between fibroblasts and macrophages. Activated Rho also stimulates the
formation of these cables in cytokine-starved cells, and induces cell
contraction (Fig. 3
). It is interesting that BAC1 macrophages have focal complexes
that are regulated by Cdc42 acting upstream of Rac [5
].
Again, the link between Cdc42 and Rac is present in these cells, as in
fibroblasts. These focal complexes contain proteins normally associated
with fibroblast focal adhesions, including ß1-integrin,
vinculin, a focal adhesion kinase, and paxillin. In BAC1 cells, Rho
does not regulate the formation of focal complexes, suggesting that it
does not directly modulate cell adhesion, at least via
integrin-containing complexes. Unfortunately, these macrophages only
rarely develop podosomes in culture so it has not been possible to
examine the regulation of these structures by Rho proteins through the
use of this model.
View larger version (32K):
[in this window]
[in a new window]
|
Figure 3. Model of the actions of Cdc42, Rac, and Rho in BAC1.2F5 macrophages.
The cytokine CSF-1 independently activates Cdc42 and Rac, leading to
the formation of filopodia and lamellipodia, respectively. Cdc42 also
acts upstream of Rac to regulate the assembly of focal complexes. Rac
activation leads to the Rho-mediated assembly of actin cables. It is
suggested that filopodia regulate cell polarization toward a source of
chemoattractant, whereas Rac and Rho together regulate cell
migration.
|
|
|
PROTEINS CONTROLLING RHO GTPASE ACTIVITY
|
|---|
All Rho family proteins bind GTP, and the majority have been shown
to act as GTPases and cycle between an active, GTP-bound form and an
inactive, GDP-bound form. Three different types of protein have been
found to regulate the cycling of Rho family proteins.
Nucleotide exchange factors (GEFs) stimulate the release of nucleotide
allowing GTP, which is at a higher concentration than GDP in cells, to
bind and thereby activate the protein. Well over 30 potential exchange
factors for Rho family proteins have been identified
[46
]; all contain a homologous Dbl domain that is
sufficient to stimulate exchange, adjacent to a PH domain. Several
mechanisms have been suggested for how GEF activity might be regulated
in cells, one favored hypothesis being that PIP3 binds to
their PH domains (Fig. 4
). In the case of Vav, a GEF expressed only in hematopoietic cells,
there is good evidence for this hypothesis [47
] because
Vav is activated by tyrosine phosphorylation in response to
extracellular signals, and binding of PIP3 enhances this
phosphorylation [48
].
View larger version (26K):
[in this window]
[in a new window]
|
Figure 4. Early signaling events in chemotaxis. On binding a chemoattractant, the
activated receptor, be it a serpentine or tyrosine kinase receptor,
signals through specific PI 3-kinase subunits to generate lipid signals
such as PIP3. Proteins with pleckstrin homology (PH)
domains, such as Vav GEF, bind the lipid signaling elements. GEFs are
then activated, leading to activation of the Rho family of GTPases.
|
|
The intrinsic GTPase activity of the Rho proteins is enhanced by
GTPase-activating proteins (GAPs). GAPs for Rho family proteins all
share a related 140-amino-acid domain, the RhoGAP domain, which is
sufficient to confer GAP activity. So far, little is known about how
GAPs are regulated in cells, although changes in subcellular
localization will probably be important. Finally, Rho, Rac, and Cdc42
have all been shown to complex with proteins known as GDIs (guanine
nucleotide dissociation inhibitors), which prevent their interaction
with other regulatory proteins and keep them sequestered in the
cytoplasm [49
, 50
]. Once an appropriate
stimulus induces dissociation of the complex, binding of the Rho
protein to the membrane is possible, consistent with GDIs acting as
negative regulators of Rho signaling. Of the three mammalian GDIs
specific for the Rho family, only RhoGDI has a high affinity for RhoA,
Rac1, and Cdc42 [51
]. In addition to keeping Rho family
proteins in an inactive complex in the cytoplasm, RhoGDI interacts with
ERM (ezrin/radixin/moesin) proteins, which in turn interact with
transmembrane proteins such as CD44 and ICAM-1, and also with actin
[52
]. Because ERM proteins can adopt either a closed or
open conformation, stimuli that induce unfolding of ERM proteins could
lead to release of Rho proteins from RhoGDI and their subsequent
availability for activation via GEFs. ERM proteins can also bind to the
Rho GEF, Dbl [53
], suggesting that they may coordinate
the release of Rho proteins from GDIs and enhance exchange of GDP for
GTP.
|
ACTIVATION OF CDC42, RAC, AND RHO
|
|---|
It is generally assumed that the major form of activation of Rho
GTPases is via activation of GEFs, although this assumption must be
considered in light of the complex pattern of controls outlined above.
The GTPases are relatively small molecules acting as switches, whereas
GEFs are usually large in size and contain many domains that allow them
to interact with different regulatory cellular components and thus
integrate various signals. Rac is activated by a wide variety of
tyrosine kinase receptors including the CSF-1R [5
] and
serpentine receptors [54
]. Data from a number of studies
suggest that PI 3-kinase is a crucial component in this pathway but the
details are not yet clear. Several studies have demonstrated that
lamellipodia formation and membrane ruffling induced by activated Rac
is not inhibited by pharmacological inhibitors of PI 3-kinase, and the
ruffling induced by PI 3-kinase is blocked by dominant-negative
versions of Rac [55
]. These data all suggest that PI
3-kinase is acting upstream of Rac. For signaling via serpentine
receptor, the Gß heterodimer is likely to be the conduit to the
p110 isoform of PI 3-kinase that leads to Rac activation via
PIP3 production [56
]. Ma and colleagues
reported that a dominant-negative variant of Cdc42 failed to block PI
3-kinase-dependent effects, thus implying that the pathway between PI
3-kinase and the actin cytoskeleton is dependent on Rac but not on
Cdc42. On the other hand, Benard and co-workers [54
]
demonstrated a clear activation of Cdc42 in addition to Rac. Our own
work on CSF-1R in macrophages supports a similar pathway for signaling
via receptor tyrosine kinases [36
], although evidence
for a specific Rac and/or Cdc42-mediated pathway has not yet been
presented.
PIP3 binds to the PH domain of at least three GEFs, Tiam1,
Sos, and the leukocyte-specific Vav [57
]. In the latter
study on Vav, PIP3 binding was found to enhance GDP-to-GTP
exchange on Rac and Cdc42 as well as Rho, lending support to earlier
findings that Vav activates Rho, Rac, and Cdc42, and not just Rac
[58
]. An attractive hypothesis is one that predicts that
PIP3 binding to the PH domain of Vav elicits a
conformational change, which then activates this GEF (Fig. 4)
. It
should not be forgotten that many other possibilities exist. It has
been shown that N-formyl-methionyl-leucyl-phenylalanine
(fMLP) receptor stimulation will also cause activation of the
src-related kinase, Lyn [59
], directly
indicating the involvement of tyrosine kinase activity leading to Rac
and Cdc42 activation through serpentine receptor activation. Activation
of Rho itself can also be stimulated through G-proteins, with both
G12 and G13 being implicated
[60
]. However, the exact mechanism by which G-protein
subunits lead to activation of Rho is yet to be fully elucidated,
although a candidate GEF in this pathway has been identified as
p115RhoGEF [61
]. Although G subunits can activate Rho
through this GEF, additional signaling events must also be involved,
the complexities of which are discussed in a recent review
[62
].
|
REGULATION OF THE CYTOSKELETON
|
|---|
Once activated, Rho family proteins can interact with downstream
target (effector) proteins, stimulating signaling pathways that lead to
the observed cellular responses. For example, many observations on Rho
function are consistent with its ability to regulate cell contractility
[63
]. This is seen clearly in macrophages, where
microinjection of activated Rho protein rapidly leads to cell
contraction [5
]. As well as stimulating actomyosin-based
contractility, Rho can also stimulate actin polymerization in a more
cell type-restricted manner [64
]. Many targets for Rho
family proteins have recently been identified, including protein
kinases, phosphoinositide kinases, and adaptor proteins, which have no
enzymatic function but have the ability to interact with one or more
other proteins [65
]. A number of targets have the
potential to link Rho, Rac, or Cdc42 directly with the actin
cytoskeleton. For example, IQGAP1, which is abundant in lamellipodia,
binds to actin filaments and also to Rac and Cdc42 [66
].
Cdc42 directly stimulates actin polymerization in leukocyte extracts in
a manner that seems independent of either PIP2 or
PIP3 [67
]. A Cdc42-interacting protein,
CIP4, shows sequence homology to a small region of ERM proteins and may
act as a transducer to the actin cytoskeleton [68
]. In
addition, Rho-kinase/ROK can induce the phosphorylation of myosin II
light chain kinase [69
], whereas the Rac/Cdc42-specific
PAK family of kinases can phosphorylate myosin I heavy chain, although
it is debatable whether this occurs in mammalian cells
[70
].
Indirect mechanisms linking Rho proteins to the actin cytoskeleton are
more common. One member of the PAK family, the Cdc42-specific PAK4,
induces the formation of filopodia [71
]. PAKs also
promotes turnover of focal complexes under the influence of Cdc42 or
Rac [72
], suggesting a role in the breakdown of
attachments to the substratum that may be critical for macrophage
locomotion. Finally, the Rho target p140mDia can bind to profilin
[73
], an actin-binding protein with the potential to
enhance actin polymerization at the leading edge of migrating cells.
The precise mechanisms whereby these targets, and undoubtedly others,
act to regulate actin reorganization, have yet to be fully elucidated
[69
].
|
WAS AND WASP
|
|---|
For leukocytes in general and macrophages in particular, interest
has recently focused on Wiskott-Aldrich Syndrome protein (WASP). The
Wiskott-Aldrich Syndrome (WAS) is a rare inherited X-linked recessive
disease characterized by immune disregulation and
microthrombocytophenia [74
, 75
]. The
clinical phenotype of the immune disorder includes susceptibility to
pyrogenic, viral, and opportunistic infections [76
].
Several lines of evidence suggest a role for this hematopoietic
cell-restricted protein as a mediator of Cdc42, and to a much lesser
extent Rac, effects on the cytoskeleton. Three groups have demonstrated
that WASP interacts directly with Cdc42 in a GTP-dependent manner
[77
78
79
]. Overexpression of WASP induced the formation
of actin filament clusters in several cell types and this clustering
could be inhibited with dominant-negative Cdc42 [79
].
More recently, inducible targeting of WASP to the plasma membrane has
been shown to induce filopodia formation in the presence of activated
Cdc42 [80
]. This same group also demonstrated that WASP
acts downstream of Cdc42 in membrane protrusion formation.
In recent months, a fuller picture of WASP and related family proteins
(N-WASP and SCAR proteins) has emerged [81
]. It was
found that members of the WASP family all share a conserved
amino-terminal portion called the EVH1 domain, which shares some
similarities with PH domains. As well as binding phosphoinositides, it
seems that EVH1 domains perform a unique function, interacting with
proline-rich target sequences [82
]. For WASP, the most
likely candidate is the WASP-interacting protein WIP
[83
]. Just carboxy-terminal to the EVH1 domain of WASP
is the CRIB motif, which confers interaction with Cdc42, and more
centrally there are proline-rich sequences that can interact with
SH3-containing proteins such as the adaptor proteins Nck and Grb2,
protein tyrosine kinases of the c-Src family such as Fyn, and the
actin-binding protein profilin [84
]. The carboxy
terminus of WASP family proteins is made up of two regions. One is an
actin-binding motif [85
] known as the WH2 motif. The
second is the A motif, which includes a cluster of acidic residues that
mediate binding to the Arp2/3 complex [85
]. Interest in
the Arp2/3 complex has been intense because it proved essential for the
reconstitution of actin-based motility and polymerization on the
surface of the intracellular parasite Listeria monocytogenes
[86
]. The complex, which was first discovered in 1994,
localizes to the leading edge of motile cells [87
]
including macrophages [G. Jones, unpublished data], where it is also
found in WASP-enriched podosomes. The Arp2/3 complex caps the
slow-growing (pointed) ends of actin filaments [88
], and
also binds to the sides of filaments, allowing cross-linking and
branching of the actin cytoskeleton [89
].
|
CELL MIGRATION
|
|---|
The ability of Rho, Rac, and Cdc42 to regulate cell adhesion and
actin organization suggests that they could be involved in cell
migration and chemotactic responses. We have tested this hypothesis on
BAC1 macrophages. Migration of microinjected cells can be directly
measured in a Dunn chemotaxis chamber by time-lapse microscopy followed
by computer-assisted analysis [31
]. This allows the
immediate effects of microinjecting proteins to be determined, avoiding
the less direct approach of establishing cell lines that show altered
migratory responses as a result of long-term changes in gene expression
induced by the exogenously expressed GTPases.
We have observed that activated forms of Rho, Rac, and Cdc42 inhibit
the migration of macrophages in response to CSF-1 [8
].
This most likely reflects the fact that expression of these proteins
dramatically reorganizes the cytoskeleton such that a cell can no
longer be polarized sufficiently to allow migration [5
].
Inhibition of Rho and Rac, by microinjection of C3 transferase and of
the dominant-negative N17Rac1 protein, respectively, also prevents the
migration of cells. The effect of N17Rac1 is expected, because
lamellipodia are universally observed at the leading edge of migrating
cells, and are required for forward protrusion of the membrane. The
effect of inhibiting Rho suggests that the contractile actin cables
regulated by Rho play an important role in mediating cell migration,
probably pulling the body of the cell forward. In the absence of Rho,
cells gradually extend dendritic processes, terminating in lamellipodia
[5
]. This suggests that the cell is still able to extend
its leading edge, but the cell body does not follow and net
translocation is not achieved.
Unlike inhibiting Rho and Rac, which completely abrogates cell
migration, inhibiting Cdc42, and thus the formation of filopodia, does
not prevent cells moving in response to CSF-1, but actually enhances
their migration rate compared to control injected cells
[8
]. However, the dominant-negative N17Cdc42 protein
does prevent cells recognizing a chemotactic gradient of CSF-1, and the
cells migrate in random directions. This effect of N17Cdc42 resembles
the response of BAC1 cells to tumor necrosis factor (TNF-),
which abolishes their ability to detect a gradient of CSF-1 without
altering their speed of locomotion [31
]. Consistent with
a role for Cdc42 in responding to the concentration gradient, TNF-
also inhibits CSF-1-induced filopodium formation without affecting
lamellipodium formation or membrane ruffling [90
].
|
A ROLE FOR INTEGRINS
|
|---|
One explanation for why macrophages move faster when Cdc42 is
inhibited is that the adhesions to the substratum are weaker, because
the integrin-containing focal complexes are dispersed
[5
]. It has been shown that cell migration is maximal at
a critical strength of interaction between the cell and the
extracellular matrix, dependent on ligand concentration, integrin
expression, and integrin-ligand affinity [91
]. Possibly,
by decreasing the clustering of integrins at focal complexes, the cells
can move faster [1
].
Using mouse knockouts for a series of Src-family tyrosine kinases,
Lowell’s group have defined a signaling cascade leading from integrins
involving the adapter protein c-Cbl and PI 3-kinase that is required
for macrophage spreading upon a fibronectin substratum
[92
]. Thus it seems likely that PIP3 is once
again implicated in signaling, this time from activated integrins, to
the Rho proteins Cdc42 and Rac that regulate macrophage spreading via
focal complexes. A similar conclusion can be drawn from the observation
[16
, 17
] that p110 knockout mice retain
a greater number of leukocytes in the circulation, coupled with a
marked inability to migrate into tissues. This observation may suggest
that p110 is an important component of the signaling pathways that
regulate leukocyte selectin and integrin molecules. A description of
integrins is beyond the scope of this review, but a full discussion of
integrin signal transduction in myeloid cells has recently been
published [93
], as have descriptions of the role of Rho
proteins in cell adhesion [69
].
|
A ROLE FOR WASP
|
|---|
The generation of regulated immune responses is dependent on the
ability of cells to migrate in response to integrated chemical signals
[94
]. In addition to the influx of neutrophils and
macrophages to sites of inflammation, these processes are responsible
for the relocation of dendritic cells from their surveillance positions
in non-lymphoid tissue to the secondary lymphoid organs, and migration
of lymphocytes into their specialized B and T cell zones
[95
]. The cellular mechanisms that allow these
trafficking processes to occur are, therefore, intimately related to
the regulation of the cytoskeletal apparatus. Cells from WAS patients
can thus provide natural models with which to answer some questions
about cytoskeletal rearrangements and the migration of leukocytes
[96
].
To investigate the role of WASP in generating cell polarization induced
by chemoattractants, we analyzed the chemotactic responses of WAS
macrophages to linear concentration gradients of CSF-1. Chemotaxis of
WAS macrophages to CSF-1 was found to be abolished, whereas normal
human macrophages show a strong chemotactic response to this cytokine
[97
]. However, the speed of cell motility was
indistinguishable from normal human cells. This result is reminiscent
of the findings for BAC1 cells exposed to gradient of CSF-1 in the
presence of TNF- [31
] where the loss of filopodia is
linked to a failure to sense gradients of chemoattractant. Furthermore,
polymerization of actin on the ventral surfaces of WAS macrophages and
extension of filopodia and lamellipodia were severely compromised,
adding further support for the role of WASP as a physiological effector
for Cdc42 in hematopoietic cells [97
]. Similar
disturbances of WAS cell motility and cell polarization have been
reported from other laboratories [98
]. Because WASP is
expressed in all hematopoietic cells, similar abnormalities of cell
polarization and chemotaxis-driven trafficking may be apparent in other
hematopoietic lineages such as T lymphocytes [99
] and
dendritic cells [100
].
|
A POSSIBLE MODEL FOR MIGRATION AND CHEMOTAXIS
|
|---|
These data all suggest a pathway through which multiple signaling
cascades activated by both serpentine and tyrosine kinase receptors
could activate the actin polymerization required for cell migration and
chemotaxis. From model systems [18
] it seems that
proteins with phosphoinositide-binding PH domains are selectively
recruited to the leading edge of chemotaxing cells. This intracellular
asymmetric distribution generates a much sharper gradient than the
causative gradient of chemoattractant. The mechanism of recruitment of
PH-containing proteins involves the production and degradation of
phosphoinositides, especially PIP3. Additional support for
this view comes from mice deficient in SHIP (SH2-containing
inositol-5-phosphatase), an enzyme that hydrolyzes PIP3.
These mice suffer from a lethal infiltration of the lungs by
macrophages and neutrophils, suggesting that abnormal persistence of
the PIP3 signal leads to excessive leukocyte recruitment
and inflammation [101
].
Polarized aggregation of serpentine chemotactic receptors to the
leading edge of migrating cells was previously thought to be the
mechanism by which the directional signaling to cells was maintained
[102
], but the application of more advanced imaging
techniques have discounted this otherwise attractive hypothesis
[103
] and it is now accepted that chemokine receptors
remain diffusely distributed over the cell surface
[104
]. These findings show that the continued sensing of
a gradient does not require an asymmetric localization or
redistribution of serpentine receptors, or indeed of the G-proteins
linked to these receptors [19
], but does involve
localized recruitment of PH domain-containing proteins
[18
].
How do macrophages respond to a gradient of CSF-1? Previous work on
BAC1 cells has shown that uniformly distributed CSF-1R is rapidly
endocytosed after stimulation with CSF-1 [30
] and that
no detectable CSF-1R re-appears on the surface for at least 20–30 min
[105
]. This makes it most unlikely that the cells are
continuously sensing and responding to the gradient, as would be the
case with serpentine receptors. This hypothesis is supported by the
observation that a cell which initially lies directly downgradient to
another macrophage, thus sensing a distorted diffusion gradient,
appears to move persistently away from a source of CSF-1 even after
entering an area in which the normal gradient has been retained
[31
]. This observation suggests that initial receptor
activation is sufficient to generate cell polarization, which is
retained in the absence of further receptor availability. A possible
model for how macrophages polarize in response to gradients of CSF-1 is
considered here (Fig. 5
). The earliest response of rounded, cytokine-starved macrophages
to CSF-1 is rapid extension of filopodia and lamellipodia all around
the free margin of the cell, a response that is seen whether the CSF-1
is presented as an isotropic [30
] or gradient
[31
] stimulus. This initial response is followed by cell
spreading, reaching a maximum spread cell area within 5 min of addition
of CSF-1. It is likely that the gradient is detected during this
initial stage, and that the extension of filopodia is necessary for
this. The filopodia may act as sensing devices because they have a high
surface area with the potential to carry large numbers of CSF-1
receptors. When they later retract back into the cell as the spreading
process ensues, they may create a concentrated "hot spot" of
activated receptor that is subsequently endocytosed but still actively
signaling [106
]. The cell then senses the difference in
signal intensity between the up- and down-gradient ends of the cell,
and polarizes its actin cytoskeleton accordingly. Possibly, the sites
of filopodium retraction act as centers to drive the polarization,
recruiting further PH-domain containing proteins and actin-associated
proteins to the up-gradient hot spot. This asymmetric recruitment will
favor the formation of a dominant lamellipodia on the up-gradient face
of the spreading cell, which subsequently results in directed movement
up the cytokine gradient [107
]. Cdc42, in stimulating
the formation of filopodia, thereby initiates the process of gradient
detection. In the absence of Cdc42, the cells still migrate, but not in
response to the gradient of chemoattractant. The extent of cell
polarization is far less and the cells remain quite rounded
[8
]. It is likely that this latter behavior is due to a
stochastic process of lamellipodial extension [1
].
View larger version (35K):
[in this window]
[in a new window]
|
Figure 5. A model for the roles of Rho GTPases in macrophage migration and
chemotaxis. In the absence of CSF-1, BAC1 cells have a rounded
morphology. Upon addition of CSF-1, macrophages extend filopodia and
subsequently spread via the extension of lamellipodia, forming new
focal complex adhesions to the substratum. In a gradient of CSF-1, the
cell polarizes its actin cytoskeleton such that lamellipodium extension
becomes restricted to the up-gradient face of the cell. Macrophage
polarization requires Cdc42 activity, and cell migration is mediated
through the actions of Rac, which extends the lamellipodia, and Rho,
which causes retraction of the cell body so as to detach the rear of
the cell.
|
|
|
FUTURE DIRECTIONS
|
|---|
Clearly there is still a long way to go before we have a
comprehensive description of cell polarization, motility, migration,
and chemotaxis. Although the identification of Rho family proteins as
key regulators of the actin cytoskeleton was a stunning breakthrough,
the precise means by which Rho proteins control the formation and
breakdown of filopodia, lamellipodia, and cell adhesion to the
substratum is far from well understood. Similarly, although many
examples in the literature support a generalized role for Cdc42 in
generating cell polarization [108
], precisely how this
is achieved is currently only a subject of active research rather than
detailed modeling. As far as chemotaxis is concerned, the recent
evidence pointing to a critical role for lipid signaling and PH
domain-containing proteins in creating the asymmetric response after
ligand binding to both serpentine and tyrosine kinase surface receptors
serves as the foundation for work to identify the cellular targets that
actually generate asymmetric cellular extensions. Finally, it would be
facile to assume that the signal transduction pathways discussed in
this review act in isolation: any suggestion that separate linear
pathways lead to the biological response under investigation is an
oversimplification. Instead, complex interdependent signaling networks
must be involved and the untangling of these networks must be the focus
of future work. The genetic tools that allow us to visualize the
chemotaxis-induced redistribution of PH domain-containing proteins and
cytoskeleton-associated proteins in living cells are now being
developed. Together with the use of mutant cells with defects in
chemotaxis this approach will undoubtedly contribute dramatically to
our future understanding of cell polarization and chemotaxis.
|
ACKNOWLEDGEMENTS
|
|---|
This work was supported by grants from the UK Medical Research
Council, The Wellcome Trust, and the Arthritis Research Campaign. I
thank Graham Dunn (Randall Centre, London), Anne Ridley (Ludwig
Institute, London), and Adrian Thrasher (Institute of Child Health,
London) for numerous stimulating discussions.
|
REFERENCES
|
|---|
-
Lauffenburger, D. A., Horwitz, A. F. (1996) Cell migration: a physically integrated molecular process Cell 84,359-369[Medline]
-
Mitchison, T. J., Cramer, L. P. (1996) Actin-based cell motility and cell locomotion Cell 84,371-379[Medline]
-
Puius, Y. A., Mahoney, N. M., Almo, S. C. (1998) The modular structure of actin-regulatory proteins Curr. Opin. Cell Biol. 10,23-34[Medline]
-
Critchley, D. R., Holt, M. R., Barry, S. T., Priddle, H., Hemmings, L., Norman, J. (1999) Integrin-mediated cell adhesion: the cytoskeletal connection Biochem. Soc. Symp. 65,79-99[Medline]
-
Allen, W. E., Jones, G. E., Pollard, J. W., Ridley, A. J. (1997) Rho, Rac and Cdc42 regulate actin organization and cell adhesion in macrophages J.Cell Sci. 110,707-720[Abstract]
-
DeFife, K. M., Jenney, C. R., Colton, E., Anderson, J. M. (1999) Cytoskeletal and adhesive structural polarizations accompany IL-13-induced human macrophage fusion J. Histochem. Cytochem. 47,65-74[Abstract/Free Full Text]
-
Correia, I., Chu, D., Chou, Y. H., Goldman, R. D., Matsudaira, P. (1999) Integrating the actin and vimentin cytoskeletons. adhesion-dependent formation of fimbrin-vimentin complexes in macrophages J. Cell Biol. 146,831-842[Abstract/Free Full Text]
-
Allen, W. E., Zicha, D., Ridley, A. J., Jones, G. E. (1998) A role for Cdc42 in macrophage chemotaxis J. Cell Biol. 141,1147-1157[Abstract/Free Full Text]
-
Sheetz, M. P., Felsenfeld, D., Galbraith, C. G., Choquet, D. (1999) Cell migration as a five-step cycle Biochem. Soc. Symp. 65,233-243[Medline]
-
Foxman, E. F., Campbell, J. J., Butcher, E. C. (1997) Multistep navigation and the combinatorial control of leukocyte chemotaxis J. Cell Biol. 139,1349-1360[Abstract/Free Full Text]
-
Foxman, E. F., Kunkel, E. J., Butcher, E. C. (1999) Integrating conflicting chemotactic signals. The role of memory in leukocyte navigation J. Cell Biol. 147,577-588[Abstract/Free Full Text]
-
Turner, L., Scotton, C., Negus, R., Balkwill, F. (1999) Hypoxia inhibits macrophage migration Eur. J. Immunol. 29,2280-2287[Medline]
-
Boulay, F., Naik, N., Giannini, E., Tardif, M., Brouchon, L. (1997) Phagocyte chemoattractant receptors Ann. NY Acad. Sci. 832,69-84[Abstract/Free Full Text]
-
Parent, C. A., Blacklock, B. J., Froehlich, W. M., Murphy, D. B., Devreotes, P. N. (1998) G protein signaling events are activated at the leading edge of chemotactic cells Cell 95,81-91[Medline]
-
Li, Z., Jiang, H., Xie, W., Zhang, Z., Smrcka, A. V., Wu, D. (2000) Roles of PLC-ß2 and -ß3 and PI3K in chemoattractant-mediated signal transduction Science 287,1046-1049[Abstract/Free Full Text]
-
Sasaki, T., Irie-Sasaki, J., Jones, R. G., Oliveira-dos-Santos, A. J., Stanford, W. L., Bolon, B., Wakeham, A., Itie, A., Bouchard, D., Kozieradzki, I., Joza, N., Mak, T. W., Ohashi, P. S., Suzuki, A., Penninger, J. M. (2000) Function of PI3Kg in thymocyte development, T cell activation, and neutrophil migration Science 287,1040-1046[Abstract/Free Full Text]
-
Hirsch, E., Katanaev, V. L., Garlanda, C., Azzolino, O., Pirola, L., Silengo, L., Sozzani, S., Mantovani, A., Altruda, F., Wymann, M. P. (2000) Central role for G protein-coupled phosphoinositide 3-kinase in inflammation Science 287,1049-1053[Abstract/Free Full Text]
-
Servant, G., Weiner, O. D., Herzmark, P., Balla, T., Sedat, J. W., Bourne, H. R. (2000) Polarization of chemoattractant receptor signaling during neutrophil chemotaxis Science 287,1037-1040[Abstract/Free Full Text]
-
Jin, T., Zhang, N., Long, Y., Parent, C. A., Devreotes, P. N. (2000) Localization of the G protein bg complex in living cells during chemotaxis Science 287,1034-1036[Abstract/Free Full Text]
-
Vanhaesebroeck, B., Waterfield, M. D. (1999) Signaling by distinct classes of phosphoinositide 3-kinase Exp. Cell Res. 253,239-254[Medline]
-
Amatruda, T. T., Gerard, N. P., Gerard, C., Simon, M. I. (1993) Specific interactions of chemoattractant factor receptors with G-proteins J. Biol. Chem. 268,10139-10144[Abstract/Free Full Text]
-
Clapham, D. E., Neer, E. J. (1993) New roles for G-protein beta gamma-dimers in transmembrane signaling Nature 365,403-406[Medline]
-
Murga, C., Laguinge, L., Wetzker, R., Cuadrado, A., Gutkind, J. S. (1998) Activation of Akt/protein kinase B by G protein-coupled receptors. A role for alpha and beta gamma subunits of heterotrimeric G proteins acting through phosphatidylinositol-3-OH kinasegamma J. Biol. Chem. 273,19080-19085[Abstract/Free Full Text]
-
Neptune, E. R., Bourne, H. R. (1997) Receptors induce chemotaxis by releasing the betagamma subunit of Gi, not by activating Gq or Gs Proc. Natl. Acad. Sci. USA 94,14489-14494[Abstract/Free Full Text]
-
Parent, C. A., Devreotes, P. N. (1999) A cell’s sense of direction Science 284,765-770[Abstract/Free Full Text]
-
Weiss, F. U., Daub, H., Ullrich, A. (1997) Novel mechanisms of RTK signal generation Curr. Opin. Genet. Dev. 7,80-86[Medline]
-
Hamilton, J. A. (1997) CSF-1 signal transduction J. Leukoc. Biol. 62,145-155[Abstract]
-
Stanley, E. R., Berg, K. L., Einstein, D. B., Lee, P. S., Pixley, F. J., Wang, Y., Yeung, Y. G. (1997) Biology and action of colony-stimulating factor-1 Mol. Reprod. Dev. 46,4-10[Medline]
-
Morgan, C., Pollard, J. W., Stanley, E. R. (1987) Isolation and characterization of a cloned growth factor dependent macrophage cell line, BAC1.2F5 J. Cell. Physiol. 130,420-427[Medline]
-
Boocock, C. A., Jones, G. E., Stanley, E. R., Pollard, J. W. (1989) Colony-stimulating factor-1 induces rapid behavioural responses in the mouse macrophage cell line, BAC1.2FS J. Cell Sci. 93,447-456[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
