From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
THE E.DONNALLTHOMASLECTURE,
1993
The Machinery of Blood Cell Movements
By Thomas P. Stossel
I
T IS A SPECIAL HONOR to receive an award named
after E. Donna11 Thomas, whose research has directly
benefitted so many patients with blooddisorders. My journey
to this award started over 20 years ago with studies of neutrophil function; meandered through reductionist biochemistry
of cellular proteins, biophysics of polymers, and molecular
biology; and most recently has been searching for clinical
applications. This odyssey would not have been possible
without many wonderful students and independent colleagues, and it would nothave even started without the initial
training and support I received from Martha Vaughan at
the National Institutes of Health or the encouragement and
independence offered by David Nathan at Boston Children’s
Hospital.
The excitement and challenge of studying cell motion
derives from the fact that it encompasses many aspects of
cell biology, hematology, and human physiology. My colleagues and I have focussed on the intracellular machinery
that mediates cell crawling but have tried to relate this machinery to the larger picture of how it relates to instructions
that activate crawling and to find ways to make these relationships useful to cell biologists and clinicians in a variety
of disciplines. In this review, I provide a description of cell
crawling movements, propose a minimal molecular mechanism for them, and suggest some directions where this research is going that might be of clinical importance.
DESCRIPTION OF CELL CRAWLING MOVEMENTS
Overview
Neutrophils, which I studied in the early 1970s, were an
excellent starting point for a career analyzing how cells
crawl, because neutrophils are the “crawliest” cells in the
human body. Each day, an average-sized person produces
over 100 billion mature neutrophils.’ These cells crawl out
of the bone marrow,2 get a brief ride in the circulation, and
then crawl again out of the vasculature for a few millimeters
in search of microorganisms or necrotic tissue. One hundred
billion objects moving that distance in the aggregate is equivalent to over twice around the earth. Neutrophils are also
the fastest cells in the human body. They move at speeds
up to 30 pm per minute, an order of magnitude faster than
crawling rates of cells required for development of the embryo, wound healing, and tumor invasion (Fig 1). It is the
inexorability, not the speed, of tumor cell migration that is
lethal. Blood platelets do not exactly crawl in the sense of
locomotion, but they undergo speedy crawling movements
as they change shape from flat discs to spheres covered
with multiple spines or to flattened configurations resembling
fried eggs with a few spines to plug leaks in injured vessels
(or occlude atherosclerotic coronary arteries). Similarly, neutrophils and mononuclear phagocytes exhibit surface crawling behavior as they engulf particulate objects.
Stimulation of Cell Crawling Movements
Cells initiate crawling movements in response to a wide
variety of external stimuli. Particularly numerous are moleBlood, Vol 84, No 2 (July 15). 1994: pp 367-379
cules that elicit neutrophil locomotion. These chemoattractants arise when inflammatory agents initiate proteolytic
cascades in extracellular fluids and stimulate cells, predominantly macrophages, to secrete bioactive lipids and polypeptide cytokines. A partial list of these agents includes C5a,
kinins, leukotriene B4,interleukin-8, and N-formyl oligopeptides3 C3b and C3bi and the Fc domain of IgG bound to
particles induce phagocytosis by neutrophils and mononuclear phagocytes. A number of activated blood clotting factors, especially thrombin, and ADP elicit platelet shape
changes. Thrombin; mitogens such as platelet-derived
growth factor, epidermal growth factor, insulin-like growth
factor-l, insulin, tumor growth factor-pl, and others; and
extracellular matrix fragments induce crawling movements
in fibroblasts, epithelial cells, and tumor cells! All of these
stimuli operate by binding specific receptors that, when so
engaged, turn on one or more intracellular metabolic cascades collectively designated as signal transduction reactions.
While the molecules mentioned above bind to several
types of specific receptors on cell surfaces, relatively nonspecific chemical and physical perturbations can also induce
crawling movements. An example of particular importance
to hematology is cold-induced platelet shape change. Chilling of platelets to temperatures less than about 15°C causes
platelets to alter their shape in a manner resembling agonistinduced activation. This phenomenon is paradoxical, because platelet activation is normally dependent on intermediary metabolism and the signal transduction reactions that
should be depressed at lowered temperatures. Cold-induced
platelet activation precludes keeping platelets produced for
transfusion at refrigeration temperature, a fact that limits
platelet storage to 5 days.5
Cortical Protrusion
Over two centuries of observations with the light microscope has provided important inferences about the mechanism of cell crawling (Figs l and 2). First, crawling movements originate from protrusive activity of a special
organelle-excluding region surrounding the cell periphery
known as the cortex (the dark shading in Fig 1). One or
more protrusions of this cortex are consistent features of
crawling movements. These protrusions may be in the form
offlat sheets, inwhich case they are called lamellae or
From theExperimental Medicine and Hematology/Oncology Divisions, Brigham and Women’s Hospital, Harvard Medical School,
Boston, MA 021 15.
Supported by US Public Health Services (National Institutes of
Health) Grants No. HL19429 and HLO7680,by an American Cancer
Society Professorship, and by the E.S. Webster Foundation.
Address reprint requests to Thomas P. Stossel, MD, Director,
Division of Experimental Medicine, Brigham and Women’S Hospital,
221 Longwood Ave, LMRC 3, Boston, MA 02115.
0 1994 by The American Society of Hematology.
0006-4971/94/8402-0139$3.00/0
367
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
THOMAS P. STOSSEL
368
Fig 1. Types and speeds of
blood cell crawling movements.
lamellipodia; when they are more bulbous, they are designated pseudopodia or lobopodia.” These protrusions
can originate fromanywherearound
thecellsurface,
but in cells
undergoing locomotion: such as neutrophils, fibroblasts, and
tumor cells. they are invariably localized at the front end of
the moving cell (Fig I ) . In contrast, platelets and other cells
that spread in preference t o crawling extend protrusions circumferentially. Another important type of cell surface protrusion is filopodia, hair-like projections that provide redundant surface area to accommodate cell shape changes and
are useful for pulling on extracellular structures, eg, the retraction of fibrin strands by platelets (Fig I ) .
are so coordinated that cells appear to glide. A large number
of adhesion molecules important for cell crawling have been
identified. The mostextensivelystudiedaretheintegrins.‘
The first recognized clinical example of a blood cell crawling
defect involving impaired adhesion was a case of genetically
determined deficiency in expression of a phagocyte-specific
integrin, CD I 1 b/CD 18.’ Inappropriate adhesion of neutrophils has also been shown to mediate catastrophic inflammatory reactions, and efforts have been directed at inhibiting
cell adhesionasastrategy
to abrogateinappropriate
inflammatory reactions.x
Revtrsihlc, Adhesion
Sol-Gel Trnnsformations and Gel Contructiotl
After protrusion of lamellae. part of these structures must
become adherent to the underlying substrate if the cell is to
crawl (orretractextracellular objects centripetally). Adhesion provides a traction force that permits thc cell body to
be pulled forward.Thereafter, propagation of locomotion
requires that the adhesion release and be re-established distally after another round of protrusion(Fig 2). Sequential
bouts of protrusion, adhesion, and forward contractionof the
cell body can sometimes be observed. but usually the steps
One of theearliestmechanisticinferencesabout
cell
crawlingmade by watching cells in the light tnicroscope
was that these movements were derived from changes of the
cell substance betweenliquid-likeandsolid-likestates.”
A
generalconceptualthreadappearing
in discussions of cell
crawling behavior was that the body of the unstilnulated cell
is a liquid (a “sol”) andthat, upon stimulation, the extended
protrusionsrepresentsolidification
of this liquid material
(into a “gel”) that is subject to compression (contractility).“’
“Sol-gel transformations” and gel contraction are, I believe,
not mere quaint adumbrations but rather bear importantly on
the mechanism of cell crawling, have been very helpful i n
characterizing molecular interactions of cell crawling, and
may turn out to be clinically useful a s well.
Sols are liquids that deform irreversibly whcn subjcctcd
to stress. Liquids with dense concentrations of asymmetrical
solutes can have a very high viscosity, ie, it may take a l o t
of forcetodeform them, because of friction between the
solute and the solvent. The internal contents of cells, which
includeconcentrated proteinsolutions,membranes,anda
variety of solidobjects, are very viscous, on theorder of
100,000 poise.” Gels are also liquids, but because they are
permeated with some kind of cohcrent network of solute,
they are elastic and; when transiently deformed short
of a
Chemoattractants
Growth Factors
Other stimuli
I
Induction
II
Protrusion &
Adhesion
111
Contraction &
Detachment
IV
Propagation
Fig 2.
+ Protrusion, etc.
Steps in cell crawling.
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
BLOOD CELL MOVEMENTS
369
in these cells in the millimolar range. Other cytoskeletal
elements, the microtubules and intermediate filaments, were
considered less likely to be directly involved in cell crawling,
because, first, in contrast to actin, they were notconcentrated
in the cell periphery. Second, cytochalasins, implicated as
specific inhibitors of actin assembly, blockedmany cell
crawling movements, whereas inhibitors of microtubule assembly did not have this effect."
Gel
,
Myosins
Fig 3. A general mechanism for cell crawling movements. Remodeling of a cortical gel under contractile tension.
breaking point, return to their original position." Another
important property of gels is their ability to impede water
flow, like a ~ p o n g e . Relatively
'~
small changes in the configuration of polymers within a gel can also result in large
volume alterations, namely contraction and expan~ion.'~
A Mechanism for Cell Crawling
With this general information we can propose a mechanical explanation for cell crawling behavior (Fig 3). A resting
cell is a liquid encased in a gel. Upon stimulation, the gel
undergoes contraction; because liquids are incompressible,
no shape change takes place until a focal weakening in the
cortical gel, perhaps at the site of initial or strongest stimulation, permits the membrane to push outwards. This protrusion is not usually spherical like an aneurysm, because the
cortical network modulates fluid flow so that the expansion
is smooth. The protrusion is flattened into a lamella by formation and contraction of gel contents from subunits introduced by the flow of sol into the extending protrusion, by
adhesion of the membrane to the substrate, and by contraction of the gel. Continued cycles of contraction, weakening,
flow, and gel reconstruction propagate crawling. Adhesion
receptors make and break connections to the extracellular
matrix and to the underlying lamellar gel. This model, in
general, is not new,I5and it incorporates others that variably
emphasized contraction at the front, at the tail, or all around
the
The challenge has been to flesh such a model
out with a minimal degree of molecular definition.
THEMOLECULAR
BASIS OF CELLCRAWLING
The Polymers of the Cell Periphery-Focus
on Actin
When we began to work on cell crawling in the 1970s,
the major polymers of the cell cortex were known (Fig 4).
The model of the plasma membrane as a fluid bilayer proposed by Singer and Nicholson was well accepted. Marchesi,
Branton, Steck, and others had described the spectrin-dominated two-dimensional protein network laminating the lipid
bilayer of erythrocytes now known to be related to numerous
hemolytic di~orders,'~.'~
although there was no evidence at
the time for such a structure in nonerythroid cells. Most
compelling to those studying cell locomotion was the probable role of the muscle proteins actin and myosin, identified
during the 1960s in nonmuscle cells by Hatano and Oosawa
and by Adelman and Taylor. Actin was shown to represent
from 5% to 20%of cell protein, representing a concentration
Much of the initial interest in the actin system in cell
crawling focussed on myosin, which had also been identified
in nonmuscle cells in general and from platelets and leukocytes, respectively, by Zucker-Franklin andby Senda and
colleagues. Some of this interest arose from the fact that the
first nonmuscle myosins identified resembled muscle myosin
in forming bipolar filaments capable of exerting tension by
sliding along actin filaments. Like muscle actin-myosin complexes, aggregates of actin and myosin from nonmuscle cells,
including blood cells, demonstrated a characteristic polarity
identifiable by the angle at which globular myosin head domains extending from filament-forming helical tails bound
to the double helical needle-like actin filaments (Fig 5). The
angle that the bound myosin head groups appeared to have
in electron micrographs conferred a characteristic arrowhead
appearance, and sliding actin filaments moved, concomitant
with ATP hydrolysis, from the barbed to the pointed direction." Immunofluorescent micrographs of cultured fibroblasts and epithelial cells stained with anti-actin antibodies
showed beautiful brightly staining actin filament bundles,
resembling muscle sarcomeres, suggesting that forces for
movement might arise from actin-myosin contractile interactions as in muscle. Fascination with myosin also arose from
the fact that Ebashi had shown that the link between striated
muscle contraction and signal transduction was calcium. Calcium ions act on troponins and tropomyosin to activate ATP
hydrolysis and contractile activity in the myosin head domain.
Therefore, our first workon the biochemistry of leukocyte
crawling was to confirm that neutrophils and macrophages
contained myosin molecules. Adelstein and colleagues
showed that smooth muscle and blood cell myosins are also
regulated by calcium, but by a different mechanism than in
skeletal muscle. Instead of directly regulating control proteins like troponins in muscle, calcium indirectly activates
platelet and macrophage myosins for actin-activated ATPase
activity and contractility by inducing phosphorylation of serine residues in the myosin head domain.*' Bloodcells probably also contain one or more varieties of a different kind of
myosin, known as "unconventional" myosin or myosin I,
first discovered in an ameoba species by Pollard and Kom.
These myosins do not form bipolar filaments but rather use
their helical tails to bind membranes."
ABP and the Cortical Actin Gel
Our focus on blood cell myosin was diverted by the discovery in 1974 that, when macrophages were disrupted in
solutions containing micromolar free calcium concentrations, the cell-free extacts, which contained about 1 to 2 mg
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
370
THOMAS P. STOSSEL
Receptors
/
Lipid
membrane
bilayer
\.
\
rm
Protein membrane3c,
skeleton
Fig 4. Polymers of the cell periphery.
of actin per milliliter, remained liquid, but if the free calcium
was lowered by chelation, the extracts solidified. The gelled
extracts then underwent a contraction provided that ATP
was present. These observations resonated with the old ideas
about sol-gel transformations and contractility of gels and
also with the fact that highly motile cells such as neutrophils
do not form actin filament bundles but rather show diffuse
actin staining in the leading lamella. We therefore began to
try and understand the nature of the gelation process and
determine its relevance to cell crawling movements. As a
physical state transformation, the actin gelation phenomenon
is complex and differs conceptually from the usual enzymatic, transport or binding reactions ordinarily studied by
biologists. A variety of approaches were mobilized to address this problem.
Biochemistry. We connected the cell extract gelation observation to our earlier finding that stirred macrophage extracts precipitated large amounts of actin withfew other
proteins save a high molecular weight protein. Purification
of this protein and adding it back to purified actin filaments
efficiently precipitated the actin.22This purified protein also
appeared to cause actin solutions to solidify when added
without stirring.23We named this protein-thefirst
new
nonmuscle actin binding protein-actin-binding protein or
ABP. An isoform of this protein, named filamin, was subsequently isolated from chicken gizzard. Both nameshave been
used interchangeably, although it is now evident thatthe
muscle and nonmuscle forms are structurally as well as functionally d i s t i n ~ t . ~ ~ " , ~ ~ ~ . ' ~
Physics. More stringent evidence for the importance of
ABP in the gelation of cell actin came from considerations
of the physics of gelation. The coherence of a gel is achievable at low solute concentration if the solute is in the form
of highly extended polymers. Because actin filaments can
be many micrometers long, they can easily overlap and therefore produce the solute coherence required for gelation at
very low (microgram) concentrations. Long overlapping rods
cannot easily rotate and therefore can have gel-like properties. Because a few very long actin filaments will gel due to
overlaps, operationally defined as cross-links, and because
the length to diameter ratio of such actin filaments is very
high, the individual filaments are flexible so that such a gel
is notvery rigid. At the concentrations of actin filaments
found in cells (7 to 15 mg/mL,), the filaments have a marked
tendency because of entropy to align in parallel (anisotropic)
bundle^.^^^^^ Such bundling compromises their ability to fill
optimally a three-dimensional space that can generate a uniform and strong gel. Moreover, for purposes of impeding
water flow, a disadvantage of a gel containing aligned filament bundles is that the pore size is not uniform.
Actin cross-linking proteins fall into two general classes,
one that stabilizes the tendency of actin filaments to form
"Barbedend"
49
OO
Actin
subunits
Fig 5. Actin-myosin interaction:fundamentals.
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
31 1
BLOOD CELL MOVEMENTS
Fig 6. Fighting entropy: organization of actin filaments into
bundles and gels by actin-crosslinking proteins.
,l
L
Uniform Gel
bundles and another, of which ABP is the prototype, that
counteract this bundling by causing the actin filaments to
branch at high angles (Fig 6). When very dilute actin polymerizes in vitro, the filaments interpenetrate as they assemble, and as the filaments are trapped in a relatively isotropic
configuration, it is possible for nearly any cross-linking protein to find overlaps and cement them in place.26This experimental artifice is probably not relevant to the situation in vivo
in which more concentrated actin filaments would bundle if
not prevented from doing so first by the actin cross-linking
proteins specifically designed for this function.
All of the actin filament cross-linking proteins are bivalent
in that they have two actin filament binding sites, and they
bind actin filaments with approximately equivalent affinities
in the micromolar range.27*27a
What distinguishes them is the
structure conferred by the amino acid sequences separating
the two actin binding sites. When the distance between these
sites is small or the intervening structure is rodlike, the crosslinking protein is better designed to stabilize bundles; when
the spacing between the sites is large and when it contains
limited flexibility, the protein is capable of promoting isotropic branching of the actin filaments. ABP, for example,
is a large dimer with a short self-association site at one end,
the carboxy terminus, separated by 160 nm from two actin
binding sites. The spacer arms of the ABP subunits contain
23 beta sheet repeats interrupted in two places by random
coil sequences designated hinges. The beta sheet repeats
presumably confer rigidity on the subunit arms while the
hinges provide for strategic flexibility that permits the ABP
dimer to function as a leaf spring to hold actin filaments at
near perpendicular angles to one a n ~ t h e r . ~ ' , ~ ~
The sol-gel transformation of linear polymers represents
the formation of a single giant coherent molecule from dispersed subunits and occurs at a critical degree of polymer
~ross-linking.~'
As a critically determined transition, gel formation is analogous to the boiling or freezing of liquids and
is thus a particularly adept mechanism for regulating cell
structure-small changes in actin composition or configuration can have profound effects on cell consistency. The transition from sol to gel can be monitored by an abrupt change
Bundles
in the consistency of the actin solution as ABP is added;
at the same critical ABP concentration, an actin solution
effectively inhibits solvent
The orthogonal branching
of actin filaments by ABP was inferred from the fact that
ABP induces an abrupt sol to gel transformation of an actin
solution at the lowest concentration compared with other
actin filament cross-linking
This concentration
can be as low as nearly one ABP dimer per actin filament,
equivalent to a molar ratio of one ABP per thousands of
actin subunits.
Ultrastructure. Electron micrographs prepared by a variety of techniques of actin filaments cross-linked by ABP
support the conclusion derived from physical studies that
ABP promotes the high-angle branching of actinfilam e n t ~ .The
~ ~extended
.~~
lamellae of macrophages, moreover,
contain orthogonal actin filament networks consisting of 1pm-long filaments branching every 100 nm at close to right
angles. A similar actin network is visible in the spread lamellae of activated blood platelets. These networks resemble
actin filament gels assembled with low ABP concentrations
in
and ABP molecules have been localized histochemically at the branch points between actin filaments.36
Actin Gelation, Contraction, and the Mechanism of Cell
Crawling
Returning to the general model proposed above for cell
crawling movements, we can now envisage the cortical gel
as some kind of network of filamentous actin (Fig 7). Cell
activation induces contraction of the network, presumably by
inducing the phosphorylation of myosin molecules through
transient increases in cytosolic calcium3' (although see below). The hydrostatic force imposed by network contraction
forces the membrane bilayer outwards at a region where
it becomes detached from the underlying protein network.
Simultaneously at the site of membrane expansion, actin
filaments assemble from monomeric subunits into relatively
short (1 pm) linear filament^.^' These filaments are crosslinked by ABP and other cross-linking proteins into a threedimensional uniform network. This network of actin filaments is dynami~,~'
but nonetheless it is coherent, and it
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
372
THOMAS P. STOSSEL
Myosin
l
l
ABP
Fig 7. Actin-based
mechanism for cell crawling movements: vectorial actin assembly,
gelation, and contraction.
modulates the flow of sol, provides strength to the forming
protrusion, and serves as an anchor for adhesion molecules
welding the protrusion to the underlying substrate. In addition, the cross-linking of initially freely diffusing actin filaments into an immobilized state imparts contractility on the
solvent included within the gel, and the squeezing action on
the solvent can provide additional force to protrude membrane forward at the leading edge. When the membrane is
tightly affixed to the underlying extracellular matrix, and is
thereby prevented from advancing, the contractile force can
produce waves on the dorsal surface of the leading lamella,
and these waves move back to cell body, a phenomenon
known as “cortical flow.”’6
Some of the best evidence for the validity of this model
has come from the opportunity to askthe question, what
happens if actin filaments cannot be efficiently cross-linked
into a uniform network? The opportunity arose when we
were provided with permanent cell lines cultivated from malignant melanomas excised from seven unrelated patients4”
The cells of four of these lines exhibit typical features of
crawling cells in culture. When stimulated by factors present
in serum, the cells acquire a polarized morphology with one
or more protruded lamellae, and they are capable of directed
migration in response to a gradient of these factors. All of
these cells express ABP at levels about 1 dimer per 500
actin molecules. In contrast, cells from the other three lines
express barely detectable quantities of ABPprotein and
&A.
The gene for ABP, which is on the X chromohowever, is present in these cells. These ABPminus cells respond to serum factors by extending spherical
projections, designated blebs, from around the entire cell
circumference. The blebs expand rapidly to about 10 pm in
diameter, stabilize, and then gradually retract. These cells
move poorly if at all toward a chemoattractant gradient.
Restoration of stable ABP expression to one of the ABPnull by transfection of the full-length ABP cDNA causes the
cells to take on the phenotype of wild-type ABP-expressing
cells, to cease constitutive blebbing on stimulation, and to
acquire the ability to crawl directionally in proportion to
the expression level of ABP (Fig 8). At the ratio of ABP to
actin found in wild-type cells, the formerly ABP null cell
lines have rates of chemotaxis indistinguishable from wildtype cells. Interestingly, cells expressing supranormal ABP
amounts have somewhat slower chemotactic
The findings with these melanoma cell lines demonstrate
unequivocally the importance of ABP for efficient stabilization of cell cortical actin. In the ABP null cells, stimulation
induces cortical contraction and focal weakening as in the
ABP-expressing cell. However, the former lack the ability
to modulate solvent flow, and the membrane blows out in
an aneurysmal fashion. Actin assembles at the leading edge
of the protruded membrane but is not efficiently linked into
a uniform network. Eventually sufficient actin assembles to
stabilize the structure by filament interpenetration, and contractile forces squeeze the bleb back into the cell body as
blowouts elsewhere reduce the hydrostatic pressure inthe
initial blebs.
Stimulus-Response Coupling, Actin Assembly and
Disassembly, and Cell Crawling
As mentioned above, cells respond to myriad stimuli with
crawling movements. A simple and reproducible assay capitalizing on the specific binding of fluorescent phalloidin to
actin filaments has permitted investigators to document that
net actinpolymerization sometimes followed by depolymerization accompanies cell ~timulation.4~
Many signal transduction mechanisms, eg, calcium transients, pH changes, protein
phosphorylation reactions, phospholipid turnover, lipoxygenase activity, GTP-binding proteins, and others, have been
proposed to mediate these changes. The large number of
candidates and the curious problem of nonspecific stimuli
such as cold-inducing actin assembly make the problem of
signal-response coupling confusing at face value. To address
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
BLOOD CELL MOVEMENTS
373
Cells
without
Stimulation
ABP
+
4
Transfect with
ABP cDNA
Fig 8. Effect of ABP expression on the surface behavior of
human melanoma cells.
this difficulty, it is helpful to consider twopeculiarities about
the assembly of actin (Fig 9).
Actin Jilmment u.s.rembly from subunits. The first peculiarity is thefact that the spontaneous assembly of actin
subunits into filaments requires a nucleation
step in which
two or three monomers aggregate. This nucleation reaction
ishighlyunfavorable:
the criticalconcentrationatwhich
such nucleation is likely to occur is i n the near millimolar
range. The second is based on the “barbed” and “pointed”
end polarity of actinfilaments first recognized by myosin
binding. The barbed ends of actin filaments serve as nuclei
onto which actin subunits can rapidly add when present at
concentrations more than a few micromolar. Cytochalasins
compete for actin subunits for binding to the barbed actin
filament ends, explainingin part the inhibitory effect of these
compoundsoncellularactinassembly.
A nearly10-fold
higher subunit concentration isrequired to promote addition
of subunits to the pointed ends. The lower critical concentration for addition to the barbed end depends on the presence
of ATP bound to the actin subunits that hydrolyzes to ADP
and orthophosphate during assembly.44
Cells capitalize on these characteristics of actin assembly
Fast
0on
Barbed
W
t Slow
4
Q*
Pointed
Elongation
Fig 9.
Spontaneous actin polymerization
0
by using two generaltypes of controlproteins(Fig
IOA).
One of these, of which /?-thymosin and, to a lesser extent,
profilin are the principal representatives,
bind:, unpolymerized actin subunits to inhibit their spontaneous assembly.45
The affinity of thymosin for actin is in the micromolar range,
however, such that, should free barbed ends be present, they
can compete effectivelyforactinsubunits;
actin assembly
might be inhibited somewhat, but nevertheless would occur
efficiently. Because about half of the total actin in resting,
unstimulated cells is polymeric, plenty of barbed ends theoretically would be present to cause most of the actin to be
polymerized. This contradiction is solved by the second type
of control protein that blocks subunit exchange on the barbed
ends of actin filaments. Two general types of these so-called
capping proteins have been identified: the gelsolinprotein
superfamily (Table I ) and a highly conserved capping protein, also named capZ. Uncapping of actin filament barbed
ends by removal of these proteins can determine where and
when actin filament elongation occurs.“
Actin jilument disussembly. Capping the barbed ends of
elongating filaments can stop further filament growth. But,
because exchange of subunits with the pointed filament end
isslow(the
off-rate constant is 1 persecond orslower),
capping is not an efficient mechanism for disassembly. To
achieve disassembly rapidly, cells use proteins that sever or
nibble actin filaments, which simultaneously shortcns them
and increases their number, providing more ends from which
depolymerization can take place (Fig I OB). Some members
of the gelsolin family bind to the sides of actin filaments,
sever the noncovalent bonds holding subunits together in the
filaments, and remain firmly affixed to the severed barbed
ends, thereby preventing reannealing of the broken filaments.
A second class of actin disassembling proteins, examples of
which are cofilin, actin-depolymerizing factor (ADF), dcpactin, and actophorin, nibbles actin monomers o u t of filaments.
possibly at places where there are defects or
kinks, which
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
374
THOMAS P. STOSSEL
Actin Awemblv in Cells
l. Actin subunit binding by ( m) 8-thymosin inhibits the
spontaneous nucleation of actin subunits.
Q
Q
11. Actin subunit binding proteins have lower
affinities for actin than free barbed filament ends.
Q
O
O
Q
Ill. Thus, together with reversible barbed
end capping, they regulate actin assembly.
Actin Disassemblv in Cells
--
1. Barbed end capping terminates assembly free
subunit-binding protein concentratlon increased
11. Actin filameht severing and nibbling
+ barbed
end capping creates many pointed ends for
deDolvmerization.
Table 1. The Gelsolin Family
Functions
Protein
Gelsolin
Distribution
Wide; abundant in plasma,
platelets, and leukocytes
Villin
Brush border epithelial
cells
Fragmin
Physarum polycephalum
Severin
Dictyostelium discoideum;
rat Lewis carcinoma
cells
Cap-G
Macrophages, fibroblasts
Protovillin
Dictyostelium
discoideum
Adseverin Adrenal and platelets
Nighfless Drosophila, humans
Cap
Sever
Nucleate
V
v
V
V
V
V
V
v
v
v
V
V
W
V
v
CrossLink
v
Fig 10. Control of intracellular actin assembly and
disnuembly. Not shown is that profilin
promotes actin assembly
by catalyzingadeninenuclsotid.
exchange on actin subunita,w. and
that some capping proteins also
work by nucleatingactinsubunits and then remaining bound
to the barbed end of the oligomers formed.
also results in increased numbers of shortened filaments.
These proteins are much less efficient than the gelsolinrelated proteins in shortening actin filaments, but they are
abundant proteins and, together with capping proteins, could
play an important role in filament shortening.
Control of actin assembly by signal transduction. The
actin filament severing protein gelsolin has been characterized in considerable detail. Its original discovery came from
our efforts to purify the factor responsible for calcium's
prevention of cytoplasmic extract actin gelation described
above?' Gelsolin's cDNA was cloned in 1986,48and part of
its three-dimensional crystal structure in a complexwith
actin solved last year.49 This last study has shed important
insights into the severing mechanism. Research on gelsolin
has also connected cell signal transduction to actin assembly.
Micromolar calcium and more recently proton concentrations have been shown to activate gelsolin for actin filament
binding, severing, and capping.50
However, removal of calcium or protons is insufficient to
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
375
BLOOD CELL MOVEMENTS
K%p
\
Polyphosphoinositides
l
Uncapping
Fig 11. Regulation of gelsolin
bv
ions
and
membrane
PO~Vphosphoinositides.
Phosphatidylinositol 4 8t 4 3 phosphate
cause gelsolin to dissociate from actin filament barbed ends.
The only nondenaturating condition shown to effect such
dissociation is the binding of gelsolin-actin complexes to a
class of membrane lipids, the polyphosphoinositides implicated in signal transduction5' (Fig 11). The action of these
phospholipids, which also dissociate actin subunits from profilin:2 requires that they have two acyl chains, a glycerol
backbone, and phosphates in the 4 or 4 and 5 positions of
an inositol ring. The physical chemistry of presentation of
the lipid with the protein is also important, and indirect
evidence suggests that phospholipid clusters are involved. It
is now appreciated that nearly all of the known actin filament
barbed end capping proteins are inhibited by polyphosphoinositides."
This information provides for a unifying picture linking
cell surface stimulation to actin assembly and disassembly
(Fig 12). During receptor-mediated cell stimulation, reaction
EGF
PDGF
Aggregated
'gG
Thrombin
N-formyl
peptides
Cold
Leukotrienes
Interleukind
cascades lead to the breakdown or synthesis of polyphosphoinositides. The former are coupled to release of calcium (and
possibly protons) from intracellular compartments, and the
resulting effect of these ions on gelsolin and related proteins
predictably results in actin filament disassembly. Conversely, reactions leading to the local synthesis or clustering
of polyphosphoinositides would promote actin assembly by
inducing the uncapping of actin filament fragments capped
by gelsolin and other capping proteins. Actin assembly and
disassembly occur in different parts of the cell at different
times, and it is therefore reasonable for the cell to use different signals for the two processes. The intermediary of phospholipids in promoting actin assembly may explain the basis
of the apparently nonspecific stimuli inducing actin assembly. Physical and mechanical perturbations may cause perturbations in the configuration of membrane lipids, promoting
clustering of membrane polyphosphoinositides.
Glass
Electrical
fields
'
E
1ACTIN
ASSEMBLY
Phospholipase C y
4
-
Phospholipase C-p /DISASSEMBLY
ACTIN
-
Inositol -3
phosphate
CALCIUM
Protons
Fig 12. Integrated mechanism for control of actin assembly and disassembly by cell signalling processes.
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
THOMAS P. STOSSEL
376
It has been difficult to obtain unambiguous evidence in
favor of the circuitry shown in Fig 12 for controlling actin
assembly in cells on the basis of correlating changes in signal
transduction intermediates with net changes in cell actin assembly. The relative proportions of assembled and disassembled actin in human neutrophils has been shown to be
roughly inversely proportional to the free calcium concentrat i ~ nalthough
, ~ ~ neutrophil locomotion can take place on certain surfaces with intracellular free calcium levels clamped
at nanomolar concentration^.^^ Because of the probable redundancy of signalling mechanisms (eg, both calcium and
protons activate gelsolin), compartmentation of signalling
reactions, and ambiguities in the experimental approaches,
the failure to achieve clean correlations is not surprising.
The best evidence in support of this proposed mechanism
has come from studies of platelet activation. First, recent
work building on earlier studies by White, Bainton, Boyles,
Fox, Nachmias, and others has established that the resting
blood platelet has a spectrin-based submembrane lamina resembling the red blood cell c y t o s k e l e t ~ n(Fig
~ ~ ~13).
~ ~ The
platelet skeleton is not only attached to transmembrane proteins as in erythrocytes but also to an internal gel of actin
cross-linked by ABP. The actin filaments of the internal
skeleton are linked to the membrane directly to spectrin
and indirectly to the glycoprotein IbfiX complex by ABP.
Activation of platelets by thrombin or glass leads to rapid
severing of the actin filaments at the periphery of the internal
skeleton that mobilizes the membrane and its underlying
skeleton that expands. The large area of the open canalicular
system of the platelet provides excess membrane to accommodate this expansion. Thereafter, two types of actin filament assembly occur. One leads to the formation of one or
more actin filament bundles that invest long filopodia. The
other consists of growth of actin filaments onto the severed
fragments in the platelet periphery, and this assembly leads
to the expansion of a circumferential lamella.s7
The initial fragmentation of actin filaments during platelet
activation is absolutely dependent on a transient increase in
intracellular free calcium into the micromolar range, strongly
implicating involvement of a gelsolin-related actin filament
severing mechanism. In contrast, the subsequent actin assem-
Membrane
bilayer and
skeleton
Calcium-mediated
severing of the
peripheralactingel
bly events are independent of changes in intracellular free
calcium. In fact, if platelets activate withtheir calcium
clamped to nanomolar concentrations with an intracellular
chelator, they become progressively distorted by the internal
growth of enormously long actin filaments that coil around
the cell periphery. Allof these actinassembly events are
inhibited by cytochalasin B, indicating that they represent
barbed end growth. Gelsolin molecules can be localized at
the barbed ends of actin filaments in resting platelets, and
new filament growth off of the barbed ends of filaments in
activated platelets has beendemonstrated by electron microscopy.57Allof this evidence strongly implicates uncapping
of capped barbed actin filaments, some of which had been
severed, as the intermediary steps of platelet activation.
However, additional work is required to prove unequivocally
that polyphosphoinositides are the final signalling step leading to actin filament barbed end uncapping in vivo.
An important aspect of control of cell crawling by the actin
system is the formation and breakage of specific connections
between membrane proteins and actin filaments. The mechanism for disconnecting the membrane from the underlying
actin skeleton involving severing of actin filaments described
in the previous section concerning initial platelet activation
is relatively crude. More complex and precise interactions,
involving the formation and breakdown of multicomponent
complexes between the extracellular surface and the intracellular actin gel, however, probably mediate many connections
between the actin network and the extracellular matrix, and
these are subject to regulation by protein phosphorylation
and other modifications.6358~sy".5'a
CLINICAL IMPLICATIONS OF THE MACHINERY OF CELL
CRAWLING MOTIONS
The usual reasons given to justify research on cell crawling is that if we could understand the mechanisms of it better,
we could modulate crawling for physiologically useful purposes-speed it up in, for example, wound healing, and slow
it down in the setting of inflammation or tumor metastasis.
Proof of the concept that such modification might bepossible was provided when we showed that murine fibroblasts
made to overexpress modest quantities of human gelsolin by
Polyphosphoinositidemediated
of
peripheral actln filaments
Open
gel
canalicular
system
membrane skeleton
and of filopodia1
actin -----> clot retraction
Fig 13. Actin-madiated control of platelet shape change.
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
BLOOD CELL MOVEMENTS
stable transfection had increased chemotactic responsiveness
in proportion to the amount of gelsolin overexpressed. Interestingly, it was not possible to increase gelsolin expression
by more than twofold in these cells, except by transient
transfection methodology. In the latter case, the cells
rounded up as their actinfilament architecture disaggregated.60Thus, as described above for ABP transfection experiments, there appears to be an optimal ratio of regulatory
proteins to actin with a relatively limited window for altering
cell function by changing this ratio.
Recently, we have engineered mice by homologous recombination that express no gelsolin!’ Although apparently
healthy and able to breed normally in captivity, these mice
have a long bleeding time and a delayed inflammatory response to intraperitoneal irritation. These physiologically altered responses correspond to slow activation of platelets
and blunted chemotactic responsiveness of neutrophils in
vitro. The findings imply that gelsolin is required for rapid
crawling responses but that other members of the gelsolin
family and other capping proteins compensate against major
impairment in cell function. Platelets, for example, contain
at least one other member of the gelsolin gene family, adseverim6’ Why is there such redundancy? The explanation lies
in the analogy that an automobile engine with multiple cylinders runs much more smoothly than a lawn mover with
one cylinder. The different capping proteins also provide
for specific responses to specific stimuli. The existence of
superficially healthy gelsolin-null mice with blunted inflammatory and thrombotic responses also points to gelsolin
and other components of the cell crawling machinery as
potential pharmacologic targets. Such targets are at the final
steps of cell shape change and motility, distal to stimuli,
receptors, and signal transduction processes, and therefore
should be broadly effective sites of intervention.
Because no specific gelsolin inhibitors are currently
known, a more immediate clinical direction has come from
another angle. In the 1980s, it was determined that extracellular fluids of humans contain actin depolymerizing activity.
Subsequent work showed that this activity was the result of
two proteins. One of these, GC globulin or the vitamin Dbinding protein, is present in blood at a concentration of 8
pmol/L andbinds specifically to actin subunits withhigh
affinity. The other is a gelsolin isoform, synthesized by the
same gene as cellular gelsolin on chromosome 9 and processed for export by alternative mRNA splicing. Blood
plasma contains about 3 pmol/L gelsolin. A mutant form of
this plasma protein is the cause of an uncommon autosomal
dominant genetic disorder, Finnish-type (Meretoja) amyloidosis, characterized by late onset cranial neuropathy and
corneal lattice dystrophy caused by deposition of gelsolin
fragments in the affected tissues.63Collectively, these two
proteins bind actin released into the extracellular space by
injury or cell death and represent what has been called the
extracellular actin scavenger
Extracellular actin complexed to GCprotein or to gelsolin
has been detected along with partial depletion of the scavenger proteins in pregnancy, in fulminant hepatic failure, in
hemolytic falciparum malaria, in adult respiratory distress
syndrome, in septic shock, and in bum wound
In
addition to preventing fluid stasis due to pileup of long actin
377
filaments, the scavenger system may ameliorate other toxic
effects of actin. These include activation of blood platelets
by ADP bound to actin subunits@and interference with fibrin
formation and fibrinoly~is.6~*~*
It has been proposed that repletion of actin scavenger proteins might be clinically beneficial.Evidence in support of this contention has come from an
unexpected source. Cystic fibrosis, the most frequent genetic
disorder of Caucasians, is caused by a defect in a chloride
transport regulator. This defect somehow results among
other effects in the early inflammation of airways and colonization by bacteria that exacerbates the inflammation. The
inflammatory response leads to thick airway mucus that is
difficult to expectorate, and the obstruction furthers the inflammatory insult. Intense neutrophilic infiltration is the hallmark of this inflammation that ultimately destroys airways,
and irreversibly damages lung function.
The high consistency of cystic fibrosis sputum is not
unique. Sputum of high viscosity is generally associated with
purulence-intense neutrophil infiltration. Sputum is a complex material comprised of mucus glycoproteins, serum proteins, and, when purulent, the contents of degenerating inflammatory cells. Because DNA was first discovered in pus,
it has been widely assumed that the high viscosity of purulent
sputum is due to DNA polymers released from the nuclei of
deteriorating neutrophils. On the basis of this belief, bovine
pancreatic deoxyribonuclease I (DNAse I) wasused as a
mucolytic in the 1950s until adverse reactions led to its
abandonment. Recently, DNAse I has been revived as a
therapeutic agent in the form of human recombinant protein6’
and was recently approved by the Food and Drug Administration for use in cystic fibrosis.
DNAse I, in addition to being a DNA-hydrolyzing enzyme, is also an actin-subunit binding protein7’ that depolymerizes actin by preventing subunits dissociating from actin filament ends from polymerizing back onto filament.7’
This fact suggested that some of DNAse 1’s mucolytic activity might result from its actin depolymerizing activity, We
documented the presence of filamentous actin in cystic fibrosis
Because the viscosity of actin filaments is
proportional to the fifth power of actin filament length:3
shortening of actin filaments ought to have a marked effect
on cystic fibrosis sputum viscosity if actin filaments are important contributors to this physical property.
In a direct test of this hypothesis, we showed that human
plasma gelsolin diminished the viscosity of cystic fibrosis
sputum at relatively low (< 100 nmol/L) concentrations in
vitro. DNAse I also decreased the sputum viscosity, but
higher (>500 nmoVL) amounts were required. These results
were consistent with DNAse 1’s effect on the sputum being
mediated by slow actin depolymerization by subunit sequestration that would require large amounts of DNAse I compared with gelsolin, which instantaneously severs actin filaments7* (Fig 14). Because gelsolin is a normalhuman
extracellular protein that can be produced by recombinant
technology, it is a candidate for mucolytic therapy in cystic
fibrosis and other diseases characterized by purulent sputum.
Presumably, gelsolin normally enters inflamed airways but
in insufficient quantities to disaggregate the very large actin
burden present.
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
378
THOMAS P. STOSSEL
GC globulin
DNAse I
Subunit
sequestration
I
Severing
t
Fig 14. Comparison of actin filament severing and actin subunit
sequestration mechanisms for shortening actin filament length.
If gelsolin is shown to be effective by appropriate clinical
investigation, this research will have come full circle: from
neutrophils back to neutrophils (albeit from live neutrophils
to dead neutrophils) after a digression into reductionist studies of actin gelation and its regulation. There are many reasons to hope that gelsolin therapy will be useful in cystic
fibrosis and other airway disorders, but in particular it would
support the Donna11 Thomas tradition of hematologic research eventuating in benefit for patients.
REFERENCES
1. Cartwright GE, Athens JW, Wintrobe MM: The kinetics of
granulopoiesis in normal man. Blood 24:780, 1964
2. Lichtman MA, Chamberlain JK, Weed RI, Pincus A, Santillo
PA: The regulation of the release of granulocytes from normal marrow, In Greenwalt T, Jamieson G (eds): The Granulocyte, Function
and Clinical Utilization. New York, NY, Liss, 1977, p 53
3. Springer TA: Traffic signals for lymphocyte recirculation and
leukocyte emigration: The multistep paradigm. Cell 76:301, 1994
4. Stetler-Stevenson WG, Aznavoorian S, Liotta LA: Tumor cell
interactions with the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol 9541, 1993
5. Murphy S, Gardner FH: Effect of storage temperature on maintenance of platelet viability-Deleterious effect of refrigerated storage. N Engl J Med 280: 1094, 1969
6. Condeelis J: Life at the leading edge: The formation of cell
protrusions. Annu Rev Cell Biol 9:411, 1994
7. Crowley CA, Curnutte JT, Rosin RE, Andre-Schwartz J, Gallin
JI, Klempner M, Snyderman R, Southwick FS, Stossel TP, Babior
BM: An inherited abnormality of neutrophil adhesion: Its genetic
transmission and association with a missing protein. N Engl J Med
291:1163, 1980
8. Arfors K-E, Lundberg C, Lindbom L, Lundberg H,
Beatty PG,
Harlan JM: A monoclonal antibody to the membrane glycoprotein
complex CD18 inhibits polymorphonuclear leukocyte accumulation
and plasma leakage in vivo. Blood 69:338, 1987
9. Mtiller O F Animalcula infusoria fluviatilia et marina, quae
detexit systematice descripset et ad vivum delineari curavit. Copenhagen, Denmark, Molleri, 1786
10. Taylor DL, Condeelis JS: Cytoplasmic structure and contractility in amoeboid cells. Int Rev Cytol 5657, 1979
11. Evans E: New physical concepts for cell amoeboid motion.
Biophys J 64: 1306, 1993
12. Almdal K, Dyre J, Hvidt S, Kramer 0: Towards a phenomenological definition of the term ‘gel.’ Polymer Gels Networks 1 5 ,
1993
13. Katchalsky A: Polyelectrolyte gels. Prog Biophys 4:1, 1954
14. Tanaka T: Gels. Sci Am 244:124, 1981
15. Lewis WH: The role of a superficial plasmagel layer in
changes of form, locomotion, and division of cells in tissue cultures.
Arch Exp Zellforsch 23: 1, 1939
16. Bray D, White JG: Cortical flowin animal cells. Science
239:883, 1988
17. Palek J, Sahr KE: Mutations of the red blood cell membrane
proteins: From clinical evaluation to detection ofthe underlying
genetic defect. Blood 80:308, 1992
18. Bennett V, Gilligan DM: The spectrin-based membrane skeleton and micron-scale organization of the plasma membrane. Annu
Rev Biochem 9:27, 1993
19. Huxley HE: Electron microscope studies on the structure of
natural and synthetic protein filaments from striated muscle. J Mol
Biol 7:281, 1963
20. Adelstein RS: Regulation of contractile proteins by phosphorylation. J Clin Invest 72:1863, 1983
2 1. Hammer J: The structure and function of unconventional myosins: A review. J Muscle Res Cell Motil 15:l. 1994
22. Stossel TP, Hartwig JH: Interactions of actin, myosin and an
actin-binding protein of rabbit alveolar macrophages. Macrophage
myosin Mg++-adenosinetriphosphatase requires a cofactor for activation by actin. J Biol Chem 2505706, 1975
23. Stossel TP, Hartwig JH: Interactions of actin, myosin and an
actin-binding protein of rabbit pulmonary macrophages. 11. Role in
cytoplasmic movement and phagocytosis. J Cell Biol 68:602, 1976
23a. Maestrini E, Matroso C, Mancini M, Rivella S, Roechi M,
Repetto M, Villa A, Frattini A, ZopN M, Vezzoni P, Toniolo D:
Mapping of two genes encoding isoforms of the actin-binding protein
ABP-280, a dystrophin-like protein, to Xq28 and to chromosome 7.
Hum Mol Genet 2:761, 1993
23b. Bany CP, Xie J, L e m o n V, Young AP: Molecular characterization of a multipromoter gene encoding a chicken filamin protein. J Biol Chem 268:25577, 1993
24. Suzuki A, Maeda T, Ito T: Formation of liquid crystalline
phase of actin filament solutions and its dependence onfilament
length as studied by optical birefringence. Biophys J 59:25, 1991
25. Furukawa R, Kundra R, Fechheimer M: Formation of liquid
crystals from actin filaments. Biochemistry 32:12346, 1993
26. Grazi E, Cuneo P, Magri E, Schwienbacher CS, Trombetta
G: Diffusion hindrance and geometry of filament crossings account
for the complex interactions of F-actin with a-actin from chicken
gizzard. Biochemistry 32:8898, 1993
27. Matsudaira P: Modular organization of actin crosslinking proteins. Trends Biochem Sci 16:87, 1991
27a. Goldmann WH, Isenberg G: Analysis of filamin and a-actinin binding to actin by the stopped flow method. FEBS Lett
336:408, 1993
28. Hartwig JH, Stossel TP: The structure of actin-binding protein
molecules in solution and interacting with actin filaments. J Mol
Biol 145563, 1981
29. Gorlin J, Yamin R, Egan S, Stewart M, Stossel TP, Kwiatkowski DJ, Hartwig JH: Human endothelial actin-binding protein
(ABP, nonmuscle filamin): A molecular leaf spring. J Cell Biol
111:1089, 1990
30. Flory PJ: Principles of Polymer Chemistry. Ithaca, NY, Cornell, 1953
31. Ito T, Suzuki A, Stossel TP: Regulation ofwater flow by
actin-binding protein-induced actin gelation. Biophys J 61: 1301,
1992
32. Brotschi EA, Hartwig JH, Stossel TP: The gelation of actin
by actin-binding protein. J Biol Chem 253:8988, 1978
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
BLOOD CELL MOVEMENTS
33. Bennett JP, Zaner KS, Stossel TP: Isolation and some properties of macrophage a-actinin. Evidence that it is not an actin gelling
protein. Biochemistry 23:5081, 1984
34. Hartwig JH, Tyler J, Stossel Tp: Actin-binding protein promotes the bipolar and perpendicular branching of actin filaments. J
Cell Biol 87:841, 1980
35. Niederman R, Amrein P, Hartwig JH: The three dimensional
structure of actin filaments in solution and an actin gel made with
actin-binding protein. J Cell Biol 96:1400, 1983
36. Hartwig JH, Shevlin P The architecture of actin filaments
and the ultrastructural location of actin-binding protein in the periphery of lung macrophages. J Cell Biol 103:1007, 1986
37. Conrad PA, Guiliano K A , Fisher G, Collins K, Matsudaira
PT, Taylor DL: Relative distribution of actin, myosin I and myosin
I1 during the wound healing response of fibroblasts. J Cell Biol
120:1381, 1993
38. Wang Y-L: Exchange of actin subunits at the leading edge
of living fibroblasts: Possible role of treadmilling. J Cell Biol
101:597, 1985
39. Theriot JA, Mitchison TJ: Comparison of actin and cell surface dynamics in motile fibroblasts. J Cell Biol 118:367, 1992
40. Byers HR, Etoh T, Doherty JR. Sober AJ, Mihm MC Jr:
Cell migration and actin organization in cultured human primary,
recurrent cutaneous and metastatic melanoma. Am J Pathol 139:423,
1991
41. Gorlin JB, Henske E, Warren ST, Kunst CB, D'Urso M,
Palmieri G , Hartwig JH, Bruns G , Kwiatkowski DJ: Actin-binding
protein (ABP-280) filamin gene (EN)maps telomeric to the color
vision locus (WCGP) and centromeric to G6PD in Xq28. Genomics
17:496, 1993
42. Cunningham CC, Gorlin JB, Kwiatkowski DJ, Hartwig JH,
Janmey PA, Byers HR, Stossel TP: Actin-binding protein requirement for cortical stability and efficient locomotion. Science 255:325,
1992
43. Howard TH, Oresajo J: A method for quantifying F-actin in
chemotactic peptide activated neutrophils: Study of effect of tBoc
peptide. Cell Motil Cytoskeleton 5:545, 1987
44. Sheterline P, Sparrow JC: Actin, in Sheterline P (ed): Protein
Profile. New York, NY, Academic, 1994, p 1
45. Fechheimer M, Zigmond SH: Focusing on unpolymerized
actin. J Cell Biol 123:1, 1993
46. Stossel TP: From signal to pseudopod. How cells control
cytoplasmic actin assembly. J Biol Chem 264:18261, 1989
47. YinHL, Stossel T P Control of cytoplasmic actin gel-sol
transformation by gelsolin, a calcium-dependent regulatory protein.
Nature 281:583, 1979
48. Kwiatkowski DJ, Stossel TP, Orkin SH, Mole JE, Colten HR,
Yin HL: Plasma and cytoplasmic gelsolins are encoded by a single
gene and contain a duplicated actin-binding domain. Nature 323:455,
1986
49. McLaughlin PJ, Gooch JT, Mannherz H-G, Weeds AG: Structure of gelsolin segment l-actin complex and the mechanism of
filament severing. Nature 364:685, 1993
50. Lamb JA, Allen PG, Tuan B, Janmey PA: Modulation of
gelsolin function. Activation at low pH overrides Ca2+requirement.
J Biol Chem 268:8999, 1993
50a. Pantaloni D, Carlier M-F: How prolifin promotes actin filament assembly in the presence of thymosin p4.Cell 75: 1007, 1993
51. Janmey PA: Phosphoinositides and calcium as regulators of
cellular actin assembly and disassembly. Annu Rev Physiol 56:169,
1994
379
52. Lassing I, Lindberg U: Specificity of the interaction between
phosphatidylinositol 4,5-bisphosphate and the profi1in:actin complex. J Cell Biochem 37:255, 1988
53. Downey GP, Chan CK, Trudel S , Grinstein S : Actin assembly
in electropermeabilized neutrophils: Role of intracellular calcium. J
Cell Biol 110:1975, 1990
54. Maxfield FR: Regulation of leukocyte locomotion byCa*'.
Trends Biol Sci 3:386, 1993
55. Hartwig JH, DeSisto M: The cytoskeleton of the resting human blood platelet: Structure of the membrane skeleton and its attachment to actin filaments. J Cell Biol 112:407, 1991
56. Fox JEB: The platelet cytoskeleton. Thromb Haemost 705384,
1993
57. Hartwig JH: Mechanisms of actin rearrangements mediating
platelet activation. J Cell Biol 118:1421, 1992
58. Luna El, Hitt AL: Cytoskeleton-plasma membrane interactions. Science 258:955, 1992
59. Howard TH, Watts RG: Actin polymerization and leukocyte
function. Curr Opin Hematol 1:61, 1994
59a. Stossel TP: On the crawling of animal cells. Science
260:1086, 1993
60. Cunningham GC, Stossel TP, Kwiatkowski DJ: Enhanced
motility in NM 3T3 fibroblasts that overexpress gelsolin. Science
251:1233, 1991
61. Witke W, Sharpe AH, Hartwig J, Azumi T, Stossel T, Kwiatkowski DJ: Transgenic mice lacking gelsolin. Mol Biol Cell 4:256a,
1993 (suppl)
62. Rodriguez Del Castillo A, Vitale AM, Tchakarov L, Trifar6
J-M: Human platelets contain scinderin, a Cazf-dependent actin filament severing protein. Thromb Haemost 67:248, 1992
63. De La Chapelle A, Kere J, Sack GH Jr, Tolvanen R, Maury
CPJ: Familial amyloidosis, Finnish type: G654 --t A mutation of the
gelsolin gene in Finnish families and an unrelated American family.
Genomics 132398, 1992
64. Lee W M , Galbraith RM: The extracellular actin-scavenger
system and actin toxicity. N Engl J Med 326:1335, 1992
65. Grinnell F, Baxter CR. Zhu M,YinHL: Detection of the
actin scavenger system in burn wound fluid. Wound Rep Reg 1:236,
1993
66. Vasconcellos CA, Lind SE: Coordinated inhibition of actininduced platelet aggregation by plasma gelsolin and vitamin D-binding protein. Blood 82:3648, 1993
67. Lind SE, Smith CJ: Actin is a noncompetitive plasmin inhibitor. J Biol Chem 266:5273, 1991
68. Lind SE, Smith CJ: Actin accelerates plasmin generation by
tissue plasminogen activator. J Biol Chem 266:17673, 1991
69. Hubbard RC, McElvaney NG, Birrer P, Shak S , Robinson
WW, Jolley C, Wu M, Chernick MS, Crystal RG: A preliminary
study of aerosolized recombinant deoxyribonuclease I in the treatment of cystic fibrosis. N Engl J Med 326:812, 1992
70. Lazarides E, Lindberg U: Actin is the naturally occumng
inhibitor of deoxyribonuclease I. Proc Natl Acad Sci USA 71:4742,
1974
7 1. Hitchcock SE, Carlsson L, Lindberg U: Depolymerization of
F-actin by deoxyribonuclease I. Cell 7:531, 1976
72. Vasconcellos CA, Allen PG, Wohl ME, Drazen JM, Janmey
PA, Stossel TP: Reduction in viscosity of cystic fibrosis sputum in
vitro by gelsolin. Science 263:969, 1994
73. Janmey PA, Hvidt S , Peetermans J, Lamb J, Ferry JD, Stossel
TP: Viscoelasticity of F-actin and F-actidgelsolin complexes. Biochemistry 272218, 1988
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
1994 84: 367-379
The E. Donnall Thomas Lecture, 1993. The machinery of blood cell
movements
TP Stossel
Updated information and services can be found at:
http://www.bloodjournal.org/content/84/2/367.citation.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.