* Division of Medical Microbiology, Department of Health and Environment, Faculty of Health Sciences, Linköping University, Sweden; and
Department of Medicine and Care, Linköping University Hospital, Sweden
Correspondence: Vesa-Matti Loitto, Division of Medical Microbiology, Department of Health and Environment, Faculty of Health Sciences, Linköping University, SE-581 85 Linköping, Sweden. E-mail: veslo{at}ihm.liu.se
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: aquaporins • anti-aquaporin antibodies • microscopy • HgCl2
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The actin-related proteins (Arp) constitute a recently characterized family of proteins. Two family members, Arp2 and Arp3, act as multifunctional organizers of actin filaments in all eukaryotes [3 , 4 ]. The discovery of the Arp2/3 complex and that the Wiskott-Aldrich syndrome proteins (WASP) is involved in regulation of the complex [5 6 7 8 ] may provide a mechanism for the assembly and disassembly of actin filaments in the leading edge of motile cells. This is called the dendritic-nucleation model [9 ], and it involves treadmilling branched actin arrays in regions involved in motility. The model predicts that external stimuli, acting through receptors and multiple signal-transduction pathways, activate WASP family proteins and the Arp2/3 complex [10 ]. The latter creates new, barbed filament ends, which grow rapidly and push the membrane forward [10 ].
In cells, the WASP family proteins have been suggested to target the motility machinery to the proper site [11 ], and it was recently demonstrated that chemoattractant-stimulated neutrophils dynamically redistribute Arp2/3 complexes to the region receiving maximal chemotactic stimulation [12 ]. In the dendritic-nucleation model [9 ], the Arp2/3 complex and actin depolymerizing factor (ADF)/cofilin are supposed to have antagonistic activities to speed up treadmilling [13 ]. Constitutive adenosine 5'-triphosphate (ATP) hydrolysis within actin filaments and dissociation of phosphate trigger severing and depolymerization of older filaments by ADF/cofilins at a rate that is controlled by some of the signals that can also stimulate assembly [10 , 14 ]. Subsequent nucleotide exchange catalyzed by profilin recycles adenosine 5'-diphosphate (ADP)-actin subunits back to the ATP-actin monomer pool. The dendritic-nucleation model suggests that actin polymerization in conjunction with a small number of proteins is sufficient for the generation of protrusion [11 , 15 ]. These molecules seem enough to mediate the rocket-like motility of Listeria monocytogenes in the cytoplasm of infected cells [16 17 18 19 ] and lamellipodial formation.
All living cells must be able to deal with osmotic and hydrostatic pressure changes in the environment [20 ]. However, the molecular mechanisms cells use to transport water and maintain turgor were largely unknown until the discovery of particular membrane proteins, i.e., the aquaporins (AQP). These serve as channels specific for water and other small nonionic molecules [20 ] and permit water to cross the plasma membranes of a wide variety of human tissues and cell types. Nonetheless, rather little is known about water channel physiology in tissues other than red blood cells and kidney [21 ]. The water-selective AQP channels have been suggested to be of fundamental importance as a result of their conservation in bacteria, plants, and mammals [22 ]. They are abundant (about 10 in humans), and the diversity of isoforms indicates specific roles in cells and tissues [22 , 23 ].
Here we have examined the role of water-selective AQP9 channels in neutrophil motility and suggest that opening and closing these provide a novel concept in the generation of cell protrusions and cell motility. By targeting AQP with anti-AQP9 antibodies and with low concentrations of HgCl2 or tetraethyl ammonium (TEA), we blocked chemoattractant-stimulated shape changes and subsequent motility. In immunofluorescence studies, AQP9 channels were preferentially localized to regions involved in motility, such as the leading edge of morphologically polarized cells. To further examine the role of AQP in the regulation of the motility machinery, we made dual labeling of the channels and N-formyl peptide receptors. At the anterior half of the cell toward regions involved in motility, the two structures colocalized. We also provide more direct evidence of water fluxes in adherent neutrophils. By loading cells with high dilution-sensitive, self-quenching concentrations of fluorophores, i.e., 2',7'-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF) and calcein, we imaged water fluxes at the cell edges and newly protruded membrane extensions. Thus, we are able to present a refined model for the formation of membrane protrusions where water fluxes play a pivotal role.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inhibition of water-selective aquaporin channels
To study the effects of various antibodies on neutrophil
morphology and motility, isolated neutrophils were added to
coverslip-bottomed experimental wells containing human 
-globulin
(100 µg/ml; AB Kabi, Stockholm, Sweden), goat normal immunoglobulin G
(IgG; 100 µg/ml; Chemicon International, Temecula, CA), rabbit serum
(10%; Chemicon International), mouse anti-human ß2
integrin antibodies (100 µg/ml; Dako A/S, Copenhagen, Denmark), or
rabbit anti-human AQP9 antibodies (100 µg/ml; Chemicon
International). The cells were allowed to sediment on the microscope
stage for 6 min when 10 nM formyl-methionyl-leucyl-phenylalanine (fMLF;
Sigma Chemical Co., St. Louis, MO) was added to stimulate motility. To
study how the cells behaved in the various solutions, we used a Zeiss
Axiovert 135M (Oberkochen, Germany) microscope with a 63x
oil-immersion Neofluar objective (1.3 numerical aperture).
Background-subtracted images were recorded continuously for 10 min
using a charge-coupled device (CCD) video camera (XC-75CE; Sony Corp.,
Tokyo, Japan), controlled by a real-time image processor (Argus-20;
Hamamatsu Photonics K.K., Hamamatsu City, Japan), and stored on
videocassettes.
Labeling of water-selective aquaporin channels
Isolated neutrophils (105) were allowed to adhere to
glass coverslips for 5 min at 37°C. After adhesion and equilibration,
the cells were exposed to fMLF (10 nM) for 2 min and then fixed for 20
min with 4% paraformaldehyde (PFA; Sigma Chemical Co.) at 37°C,
washed in phosphate-buffered saline (PBS; pH 7.3) with 1% bovine serum
albumin (BSA; Boehringer Mannheim GmbH, Mannheim, Germany), incubated
in blocking buffer, i.e., 10% normal swine serum from Dako A/S in
PBS/BSA (30 min, 37°C), and then stained for 60 min at 37°C with 10
µg/ml rabbit anti-human AQP9 antibodies in blocking buffer. After
thorough washing in PBS/BSA, the cells were incubated with 15 µg/ml
secondary swine anti-rabbit fluorescein isothiocyanate
(FITC)-conjugated antibody (Dako A/S) for 30 min at 37°C in blocking
buffer and finally mounted in ProLong anti-fade (Molecular Probes,
Eugene, OR).
To investigate potential colocalization of water-selective aquaporin channels and N-formyl peptide receptors, dual labeling of the two was performed. First, the fMLF receptors were labeled using the same procedure as described above but with goat serum (Chemicon International), monoclonal mouse anti-human fMLF receptor antibodies (PharMingen, San Diego, CA), and goat anti-mouse Alexa 594 secondary antibodies (Molecular Probes). Second, the AQP9 channels were labeled as described above.
The preparations were studied using the Zeiss Axiovert 135M (Oberkochen) microscope epifluorescence with a 100 W HBO mercury arc lamp, a 95% neutral density filter to reduce photobleaching, and a 100x oil-immersion Neofluar objective (1.3 numerical aperture). Images were captured using a cooled TE4 Astromed 4200 slow-scan CCD camera from LSR (Cambridge, UK).
Primary antibodies used as controls of unspecific binding were monoclonal mouse antibodies to human epithelial membrane antigen (EMA; Dako A/S), and affinity-purified rabbit antioccludin (Zymed, San Francisco, CA). Controls also were treated with blocking buffer instead of primary antibodies to determine the extent of autofluorescence and unspecific labeling.
Determination of cell motility
Chemoattractant stimulated neutrophil motility was measured
using Transwell® filters (Corning Costar Corp., Cambridge, MA) with
3-µm pores. The filters were first put in a flask with PBS (pH 7.3)
and 1% BSA, kept under vacuum for 15 min to evaporate air bubbles from
the pores in the filter, and then used in the transmigration assay. To
determine the effects of the known AQP inhibitors HgCl2
(Merck, Darmstadt, Germany) and tetraethyl ammonium (TEA; Sigma)
on neutrophil motility,
N,N’-((3’,6’-dihydroxy-3-oxospiro(isobenzoFuran-1(3H),9’-(9H)xanthene)-2,7-diyl)bis(methylene))bis(N-carboxymethyl)
(calcein; Molecular Probes Inc.)-loaded (10 µM, 20 min, 37°C)
cells were pre-incubated for 15 min at room temperature with various
concentrations (1-100 µM) of the inhibitors. The cells were then
added to the upper Transwell® compartment. To stimulate migration 10
nM fMLF was added to the lower compartment, except to the controls for
random motility. After 60-180 min (37°C), the upper compartment was
removed and 0.1% Triton X-100 (Sigma Chemical Co.) was added to
release calcein from transmigrated cells. Transmigration studies also
were performed without pretreatment but with Hg2+
containing solutions, both in the wells and the Transwell® inserts.
The calcein concentration in each well was measured with LS-3B
fluorescence spectrometer (Perkin-Elmer, Buckinghamshire, UK). The
calcein fluorescence from cells exposed to fMLF alone was set to 100%.
Ratio imaging of intracellular free Ca2+
([Ca2+]i)
Ratio imaging of the fluorescent calcium indicator
1[2-(5-carboxyoxazol-2-yl)-6-aminobenzoFuran-5-oxyl]-2-(2’-amino-5-methylphenoxy)-ethane-N,N,N,N’,N’-tetraacetic
acid (Fura-2: Molecular Probes, Inc.)-loaded cells (5 µM, 30 min),
was done with a Photon Technology International (Monmouth
Junction, NJ) system and a Zeiss Axiovert 100 M microscope with a 100x
glycerol-immersion Neofluar objective (1.3 numerical aperture). Labeled
cells were allowed to adhere to the coverslip bottomed experimental
well for 5-10 min at 37°C before analysis. After an initial 10 image
pairs (F340/F380), approximately 1 minute, to
obtain a Ca2+-baseline, 1-10 µM HgCl2 was
added to the cells. Images were taken every 6 sec and represent the
mean of eight successive video frames. In all, the cells were
ratio-imaged for 10–15 min. Bright-field images were captured
simultaneously using a Newvicon video camera (Hamamatsu C-2400,
Hamamatsu Photonics K.K.) by passing the transmission light through a
700-nm band-pass filter in front of the halogen lamp to avoid stray
light in the fluorescence channel. The Fura-2 experiments were
performed in CCM and KRG, supplemented with 1 mM ethylene glycol-O,
O'-bis (2-aminoethyl) N,N,N',N'-tetraacetic acid (EGTA;
Fluka Chemie AG, Buchs, Germany).
Measurements of cell size with flow cytometry
Cell size after incubation with HgCl2 was determined
by forward-scattering in a flow cytometer (FACSCalibur, Becton
Dickinson, San Jose, CA). Isolated neutrophils were preincubated with
0–10 µM Hg2+ for 10 min when fMLF (100 nM)
was added, and the cell volume was measured after 0, 0.5, 1, 5, 10, and
15 min.
Labeling and quantification of filamentous actin
Isolated neutrophils were allowed to adhere to
coverslip-bottomed experimental wells for 5 min at 37°C. Then, 10 nM
fMLF was added to induce morphological polarization.
Chemoattractant-stimulated neutrophils were subsequently exposed to
Hg2+ (1–10 µM, 10 min) before fixation with
4% PFA in CCM for 20 min at 37°C. The cells were permeabilized with
0.1% Triton X-100 before labeling with Alexa 594 phallacidin
(Molecular Probes), according to the manufacturer’s protocol. Labeled
specimens were mounted in ProLong anti-fade.
The fluorescently labeled filamentous actin in neutrophils exposed to Hg2+ was studied using the same microscope setup as described for visualization of fluorescently labeled AQP9. Digitized images were exported to Optimas 6.0 (Optimas, Washington, DC), where cell area multiplied with mean fluorescence intensity (MFI) was used as a measure of the total content of filamentous actin.
Dequenching experiments
To directly visualize sites of water influx, cells were loaded
with 2 µM fluorescent pH indicator BCECF (Molecular Probes) for 45
min at 37°C or calcein (10 µM, 60 min) to obtain a high and
dilution-sensitive, self-quenching concentration. The neutrophils were
allowed to adhere to coverlip-bottomed experimental wells and were
imaged using fluorescence microscopy.
For ratio imaging of BCECF, a system based on an Axiovert 35 (Zeiss, Oberkochen) microscope equipped for epifluorescence with a 75 W xenon arc lamp and a 100x glycerol-immersion Neofluar objective (1.3 numerical aperture) was used. During ratio imaging of BCECF [25 , 26 ], the emitted fluorescence, after excitation at 440 nm and 495 nm using a computer-controlled filter wheel with 10 nm band-pass filters, was passed through a 530 nm band-pass filter to remove residual light and nonspecific fluorescence to a micro-channel plate-image intensifier (Hamamatsu Nightviewer C-2100). The resulting image was captured and digitized by a frame grabber (IV AB, Linköping, Sweden) attached to an IBM-compatible personal computer. For ratio imaging, the first image pair was captured within 2 s with subsequent images grabbed every 8 s; each image intensity represents the mean of four successive video frames [26 ]. A 95% neutral density filter was used to reduce photobleaching. Differential interference contrast (DIC) microscopy images and fluorescence images were obtained simultaneously using a 650-nm long-pass filter in front of the halogen lamp to avoid transmission light in the fluorescence channel. A Newvicon video camera (Hamamatsu C-2400) was used to obtain the DIC images. BCECF calibration curves were obtained by incubating cells in 115 mM K+ buffers of known pH in the presence of 10 mM nigericin [25 , 27 ].
The fluorescence of neutrophils loaded for 10 or 45 min with 2 µM BCECF was also measured using a spectrofluorometer (LS-3B, Perkin-Elmer) before and after addition of 0.1% Triton X-100 to 105 cells in 2 ml. The fluorescence was read at 540 nm after excitation at 440 nm to assess only the pH-independent population of BCECF.
Calcein-labeled cells were studied using the microscope setup as described for visualization of fluorescently labeled AQP9.
Loading profile for calcein
A loading profile for calcein against incubation time was
obtained by removing samples from the loading solution at 15-min
intervals. The cells were washed rapidly and added to
coverslip-bottomed experimental wells containing 10 nM fMLF. Directly
following adhesion and spreading, the cells were imaged using the same
microscope setup as described for inhibition of water-selective
aquaporin channels. The rate of calcein loading was assessed from the
MFI, as measured in Optimas 6.0. The experiment was repeated seven
times, imaging five cells from each 15-min sample. The coverslips in
all experiments were precoated with 10% human plasma.
Membrane accumulation
To investigate fluctuations in cell thickness at the cell
periphery, we investigated the extent of membrane ruffling by labeling
neutrophil membranes. The membranes were labeled for 20 min at 37°C
using 5.5 µg/ml
1,1'-dihexadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate
[DiIC16(3); Molecular Probes]. The cells were washed
three times in KRG and kept protected from illumination on melting ice
until used. Samples of 5 x 104 cells were withdrawn
from the cell suspension and transferred to coverslip-bottomed
experimental wells. The neutrophils were allowed to adhere for 5–10
min to the glass surface at room temperature in CCM. Images were
captured directly following adherence and after 15 min using the same
microscope setup as described for visualization of fluorescently
labeled AQP9.
Statistics
Differences were considered to be significant when
P < 0.05, according to a two-tailed, type-3 Student’s
t-test. Error bars represent SE.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-globulin (100 µg/ml), goat IgG (100 µg/ml), or rabbit serum
(10%) exhibited no morphological or migrational impairment upon
addition of chemoattractant. Furthermore, following a 10-min exposure
to AQP9 antibodies, the morphologically nonpolarized cells formed
numerous miniscule blebs along the cell peripheries, indicating reduced
osmotic regulation at the level of the cell cortex (Fig. 1)
. We think
that the extensive bleb formation upon antibody exposure was a result
of blockage of the pores. To investigate whether this morphology could
be induced with antibodies binding to another cell-surface epitope, we
also incubated controls with anti-ß2 integrin antibodies
(100 µg/ml). These cells did not adhere to the protein-coated surface
but responded transiently to addition of chemoattractant by extending
pseudopodia. Following several attempts to polarize toward the source
of chemoattractant, the anti-ß2 antibody-treated cells
returned to the initial, spherical morphology. Bleb formation was
entirely absent in these controls.
 View larger version (91K): [in a new window] |
Figure 1. AQP9 antibodies recognize a functional domain important in the
morphological response to chemotactic stimulus. Preincubation (5 min)
of cells with AQP9 antibodies (100 µg/ml) abolished
chemoattractant-induced (fMLF, 10 nM) spreading and motility. After 10
min, miniscule blebs formed along the cell peripheries, indicating
impaired osmotic regulation, most likely a result of a block of the
water-selective channels. When controls were treated similarly using
antibodies directed against the ß2 part of integrins, the
cells were unable to adhere to the albumin-coated surface but responded
morphologically to subsequent addition of chemoattractant by extending
pseudopodia. Here, the normal response to addition of fMLF is shown
below anti-AQP9 antibody-treated cells.
|
 View larger version (47K): [in a new window] |
Figure 2. The distribution of fluorescently stained aquaporins and
chemoattractant receptors in human neutrophils. AQP9 channels are in an
ideal position to regulate cell motility by extending all the way out
on the extremely thin leading edge and defining the cell periphery
(A–C). The distribution of fluorescently stained aquaporins and
chemoattractant receptors in human neutrophils shows that AQP9 channels
(D) and N-formyl chemoattractant receptors (E) have a
similar distribution when superimposing the images (F). The
colocalization suggests an intricate relationship between
chemoattractant signaling and concomitant morphological responses
regulated by water fluxes through AQP9. Arrows indicate the leading
edge. Images were captured using fluorescence microscopy and a cooled
CCD camera, as described in Materials and Methods. Original bar, 5
µm.
|
Hg2+ and TEA reduced transmigration
To further investigate the role of water-selective channels in
cell motility, we pretreated cells with nontoxic levels (1–25 µM) of
HgCl2 known to block AQP water-channel function(s)
[28
, 29
]. Following a 15-min pretreatment
with Hg2+, the migratory behavior of fMLF (10
nM)-stimulated, calcein-loaded cells over porous filters decreased.
Hg2+ dose-dependently reduced stimulated
neutrophil transmigration through Transwell® inserts (Fig. 3
), which compared with fMLF-stimulated controls not treated with
Hg2+ was 88%, 45%, and 33% for 2.5, 5, and
10 µM Hg2+, respectively. Pretreatment with
Hg2+ concentrations below 2.5 µM had no
measurable effect. Conversely, concentrations of 
10 µM
Hg2+ gradually made the cells leaky to
preloaded fluorescent calcein (Mw<995). Therefore, it is likely that
the effects of pretreatment with 10 µM Hg2+
on transmigration were even larger, because the measured fluorescence
could be a result of leakage rather than transmigrated cells. When
performing the transmigration assay using
Hg2+-containing buffers, the inhibitory effects
were even clearer. In these experiments, 1 µM
Hg2+ already decreased transmigration
considerably. Here, the fluorescence intensity increased above the
100% control when Hg2+ concentrations of >5
µM were used, indicating severe leakage of calcein from the cells.
The Hg2+-induced leakage was confirmed using
ratio imaging of Fura-2-loaded cells.
 View larger version (27K): [in a new window] |
Figure 3. Evidence for the regulatory role of water-selective channels in
chemoattractant-stimulated cell motility. In comparison with cells
exposed only to an fMLF gradient (used as 100% in the experiments,
first column), inhibition of aquaporins with 1–10 µM
Hg2+ or 100 µM TEA reduced the transmigration
of calcein-loaded neutrophils through porous Transwell® filters.
Error bars represent SE.
|
When Hg2+-treated neutrophils were studied using video-enhanced bright-field microscopy, we noticed distinct effects on cell polarization. Although cells rapidly polarized following addition of Hg2+, the cells displayed markedly reduced motility, which is in accordance with observations by Contrino et al. [31 ]. The cells remained motionless, retained the initial polarization, and did not respond to subsequent addition of chemoattractant. Also, the chemotactic response of TEA-treated cells was temporally delayed.
Hg2+ caused leakage
Because prolonged incubation with 10 µM
Hg2+ caused leakage of fluorescent calcein, it
is conceivable that such conditions, in contrast to the blocking
effects of anti-AQP antibodies, forced AQP into an open, less-regulated
configuration. Therefore, we tested whether the membrane also became
less restrictive for fluxes of other small solutes such as
Ca2+. To challenge this idea, we performed
Fura-2 ratio imaging of [Ca2+]i
in neutrophils continuously exposed to Hg2+.
Indeed, adding Hg2+ (10 µM) to adherent
neutrophils caused an initial, small Ca2+
increase, followed by a stable period where
[Ca2+]i was about 100 nM.
However, after 6–7 min, [Ca2+]i
increased, reaching a maximum at about 10 min. At this point, the
Fura-2 fluorescence disappeared below the preset ratio
(F340/F380) threshold, indicating rapid leakage
of the fluorophore (Fig. 4
). In parallel with the addition of Hg2+,
the cells obtained a motility-related shape change, similar to being
exposed to chemoattractants. Then, the cells suddenly and all together
ceased to change morphologically and resided completely
immobile during the remaining experiment (Fig. 4)
. Under the
microscope, the cells remained polarized and visually intact after up
to 30 min incubation with 10 µM Hg2+ at
37°C; there was no signs of lysis nor cell necrosis, but rather, the
cells appeared in a frozen configuration. When cells were exposed to
Hg2+ in a Ca2+-free
environment, i.e., KRG supplemented with 1 mM EGTA, no
Ca2+ transients were observed. However, also
after 10–12 min, the cells in Ca2+-free buffer
displayed leakage of Fura-2. No sustained Ca2+
increases and leakage of Fura-2 were observed in controls not exposed
to Hg2+. Time-lapse video sequences of cellular
behavior under these conditions are available on request.
Hg2+ caused no quenching of Fura-2 fluorescence
even at 100 µM, which has also been confirmed by Marchi et al.
[32
].
 View larger version (59K): [in a new window] |
Figure 4. HgCl2 caused rapid leakage of loaded fluorophores. Exposure
of Fura-2-loaded neutrophils to 10 µM Hg2+
(arrow). After 6–7 min with Hg2+,
[Ca2+]i increased, and at 10 min,
the fluorescence started to disappear out of the cells, indicating
transmembrane flux through AQP (see pseudo-colored ratio images). The
lines indicate mean ratio above a preset threshold in all eight cells
in the field of view. The cells remained morphologically polarized and
visually intact throughout the experiment (see bright-field images),
indicating that the size of the pore was still restricted; i.e.,
excessive and fully unrestricted flux would most likely result in cell
lysis.
|
Hg2+ reduced cell size
To examine the effects of an AQP block on cell volume, we examined
volume changes in nonadherent neutrophils by measuring
forward-scattering in a flow cytometer. Cells were preincubated with
0–10 µM Hg2+ for 10 min, then fMLF (100 nM)
was added, and the effects on cell volume were measured at 0, 0.5, 1,
5, 10, and 15 min after chemoattractant stimulation. Stimulation of
controls induced a gradual volume increase from 100% to 110% 15 min
after activation of fMLF (Fig. 5 A
). Hg2+ treatment reduced the initial cell
size, and pretreated cells were also unable to respond to subsequent
addition of fMLF. After 15 min with 1, 5, and 10 µM
Hg2+, the cell volume, as compared with 0
Hg2+, was 97%, 77%, and 88%, respectively
(Fig. 5A)
. The data represent mean values of three measurements.
 View larger version (19K): [in a new window] |
Figure 5. Hg2+ treatment reduced
chemoattractant-stimulated cell-volume increase (A). Cells were
preincubated with 0 ( ), 1 (•), 5 ( ), and 10 ( ) µM
Hg2+ for 10 min when fMLF (100 nM) was added,
and the cell volume was measured after 0, 0.5, 1, 5, 10, and 15 min by
forward-scattering in a flow cytometer. Perturbation of water-channel
regulation with 2.5–10 µM Hg2+ reduced the
amount of filamentous actin measured with fluorescence microscopy,
suggesting that the inhibition of water-selective channels interferes
with the physiological regulation of actin-filament dynamics following
chemoattractant stimulation (B). Unrestricted influx of extracellular
Ca2+ most likely activates
actin-filament-severing proteins leading to continuous
depolymerization. Hence, the total concentration of intracellular
filamentous actin decreases. Images represent cells with fluorescently
labeled F-actin from controls and samples with
Hg2+-pretreated cells. Error bars represent
SE.
|
Water fluxes visualized by fluorescence dequenching
We further provide direct evidence of changes in cell hydration
derived from dequenching experiments. To visualize water influxes at
regions involved in motility, we loaded neutrophils with high,
self-quenching concentrations of two fluorophores, i.e., BCECF or
calcein. The idea was to directly measure water fluxes at sites of
lamellipodium extension by dilution and thus dequench the fluorophores.
Then these regions, visualized using low light-level video microscopy,
should appear in brighter intensities than perinuclear regions. Indeed,
at sites of membrane extension at the cell edges, there was always
increased fluorescence compared with perinuclear regions (Fig. 6
). In all instances, the very thin, expanding lamellipodium was
strikingly fluorescent, whereas the adjacent cytoplasm appeared dark up
to the slightly autofluorescent nuclear region. In radially spreading
cells, the fluorescence was localized in circumference at the cell
periphery (Fig. 6B)
. Plotting a fluorescence-intensity profile across
the cell displays the uneven distribution of fluorescence (Fig. 6C) . In
bleb-forming cells, fluorescence also increased in the rapidly
protruding extensions (Fig. 6D)
.
 View larger version (30K): [in a new window] |
Figure 6. Visualization of sites of water influx using fluorescence microscopy.
Cells loaded with the fluorescent probes BCECF (A and B) and calcein (C
and D) to obtain a high dilution-sensitive, self-quenching
concentration. Morphologically active regions (triangles) in
chemoattractant-stimulated adherent cells display increased
fluorescence compared with perinuclear regions. In superloaded cells,
bleb formation also correlated with dequenching. The most
straightforward explanation for the increased fluorescence at the cell
periphery is an instantaneous dilution of the self-quenched fluorescent
probe because of water influx in the developing protrusions. Hence,
water influx at the cell edge forces the quenched fluorophore complexes
to split into excitable monomers. An intensity profile through the
calcein-loaded cell shows increased intensities at the cell edges
compared with perinuclear regions (insert in C). Increasing the
intensity threshold shows high intensities only (inserts in C and D).
Original bar, 5 µm.
|
To ascertain whether a self-quenching concentration of fluorophore was obtained intracellularly, we measured the loading profile for calcein. When cells were incubated with 10 µM acetoxymethylester (AM) derivative of calcein, there was a linear increase in fluorescence up to 45 min followed by a decline to a lower stable level (Fig. 7 A ).
 View larger version (25K): [in a new window] |
Figure 7. Evidence for self-quenching of calcein and BCECF. Loading profile for
calcein against incubation time (A). Samples were removed from the
loading solution at 15-min intervals. The fluorescence, measured with
fluorescence microscopy, increases steadily up to 45 min after which it
drops off to a steady state. The fluorescence of neutrophils loaded for
10 or 45 min with 2 µM BCECF also was measured using a
spectrofluorometer before and after addition of 0.1% Triton X-100 to
105 cells in 2 ml (B). With Triton X-100, the total
fluorescence decreased after 10-min loading but increased after 45-min
loading, suggesting a structure-linked latency, and the cytoplasm
fluorescence was partly quenched after the 45-min loading procedure.
Error bars represent SE. *, P < 0.05, as
assessed using a two-tailed, type-3 Student’s t-test.
|
The increased fluorescence at the cell edges of superloaded cells could not be explained by membrane ruffling, for example, as assessed by labeling the membrane with DiIC16(3) and fluorescence microscopy (Fig. 8 ). No major accumulation of plasma membrane could be observed in these cells.
 View larger version (58K): [in a new window] |
Figure 8. The increased fluorescence at the cell edges in superloaded cells was
not a result of accumulation of plasma membrane. When the cell membrane
was labeled with DiIC16(3) and studied using fluorescence
microscopy, no major membrane accumulation could be observed. Original
bar, 5 µm.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The present results provide direct experimental proof for models, suggesting that osmotic activity helps chemotactically active cells to regulate the formation of membrane protrusions [36 37 38 39 40 41 42 ]. In addition, our data suggest that the influx of water involves precisely tuned water-selective channels at motile regions. The mechanism(s) that regulates the opening and closing of AQP9 channels in neutrophil membranes is as yet unknown, but we propose that chemoattractant signaling to the motility machinery and concomitant actin polymerization can mediate stretch-dependent activation; stretch initiates opening of the pores and relief of stress closure. The concept of stretch-activated channels is not a new feature. In fact, osmotic stretch-induced opening of voltage-dependent Ca2+ channels has been observed in myocytes [43 ], and also the movement of fish epithelial keratocytes has been shown to involve stretch-activated Ca2+ channels [44 ]. These observations lend support for stretch-activated regulation of the opening of aquaporin channels as well.
In our experiments, exposure of cells to anti-AQP antibodies effectively blocked all protrusive activity in chemoattractant-stimulated cells. Moreover, after 10 min of incubation with the antibody, these cells formed numerous miniscule blebs around their cell periphery (Fig. 1) . Also the known AQP inhibitors HgCl2 [28 ] and TEA [30 ] diminished chemoattractant-stimulated motility as assessed using contrast-enhanced video microscopy and Transwell® assays (Fig. 3) . Furthermore, Hg2+ treatment of neutrophils gradually made the cells leaky, indicating a less-restricted regulation of and flux through the pores. Rather than blocking the pores as anti-AQP antibodies, we think that Hg2+ forced the water-selective channels into an open conformation, as assessed using ratio imaging of Fura-2-loaded cells (Fig. 4) . A key question when using Hg2+ is, however, whether Hg2+ inhibits motility by blocking the normal regulation of the water pores or through a nondefined toxic effect. Although the loss of osmotic regulation certainly is also a toxic event if not reversed, we did not observe any signs of cell lysis or necrosis when studying Hg2+-treated cells. Even after 30 min in an Hg2+-containing environment, the cells appeared attached and intact, although no movement was observed. Furthermore, cells pretreated with Hg2+ and then assayed in Hg2+-free buffers displayed less inhibitory effects on transmigration over porous filters than when studied in Hg2+-containing buffers. The difference shows that the inhibitory effects of Hg2+ on cell motility were to some extent reversible.
Cells loaded to self-quenching with fluorophore displayed increased fluorescence in emerging membrane blebs compared with perinuclear regions (Fig. 6D) . This is in accordance with findings by Cunningham et al. [45 ] that melanocytes deficient in the 280-kDa actin-binding protein (ABP-280), or filamin-1, have an increased tendency for blebbing [45 ]. They further suggested that blebs occur when the fluid-driven expansion of the cell membrane is rapid enough to outpace the local rate of actin polymerization [46 ]. As a result of increased levels of actin-filament fragments and monomeric actin, the cells were particularly sensitive to osmotic pressure and water influx, resulting in more random lamellipodium activity. Taken together with these previous observations, our observations provide conclusive support for water influx in rapidly forming membrane extensions such as blebs and pseudopodia.
Pollard and collaborators [9 , 10 ] have recently depicted a mechanism, the dendritic-nucleation model, explaining the formation of membrane protrusions based on actin polymerization. In the model, it is hypothesized that actin filaments per se produce force enough to protrude the membrane. However, our observations involve components of the osmotic mechanisms originally proposed by Oster and Perelson [39 ]. We think that localized water influx generates a gap between the filament ends and the membrane, thereby allowing for rapid polymerization of actin. Further, it has been proposed that actin networks can compartmentalize the hydrostatic effect [47 ], thereby directing the evolving force toward the membrane. Thus, our observations complete the model suggested by Pollard and coworkers [9 , 10 ] and also convene the major hypotheses regarding generation of cell protrusions. By showing that cell motility is water-dependent, we propose a mechanism where a regulated water influx allows for gel-to-sol transitions, osmotic forces, and subsequent dendritic nucleation.
In Figure 9 , we present our contribution to the model originally depicted by Mullins et al. [9 ]. When stimulated with chemoattractant (step 1), cells respond rapidly by initiation of several simultaneous, intracellular signaling reactions (step 2). Although Ca2+ has been shown not to be essential for cell motility on weakly adherent surfaces [48 ], we believe it plays an important role in the normal, physiological events of cell motility, e.g., in the activation of protein kinase C (PKC; step a), which uncaps actin filaments by phosphorylation of myristoylated alanine-rich C-kinase substrate (MARCKS). The localized increase in [Ca2+]i can also activate gelsolin and other actin-filament-severing proteins (step b), which generate osmotically active actin-filament fragments and monomeric actin. Chemoattractant signaling also activates phosphatidylinositol-4,5-bisphosphate (PIP2; step c), which uncaps filament barbed ends (step d), inactivates the actin-severing actions by gelsolin (step e), and liberates polymerization-competent ATP-actin monomers from profilin (step f). Chemoattractant stimulation also activates members of the Rho family GTPases (step g), preferentially Cdc42 and Rac2 in neutrophils. Cdc42 activates WASP (step h), which in turn activates the Arp2/3 complex to generate new barbed actin-filament ends. Rac2 will, among its yet rather obscure roles in lamellipodium formation, liberate gelsolin from its capping position (step i), and it has also been shown to inactivate ADF through unknown pathways [49 ]. The concerted activity of various signaling mechanisms enhance actin-filament turnover, and polymerizing filaments will push on the membrane, thereby generating a stretch-activated opening of water-selective channels (step 3). The following increased water influx will cause a gap between the outermost filaments and the membrane and also enhance the local diffusion of polymerization-competent actin monomers and other vital components of polymerization (step 4). Actin-filament polymerization is stopped when filaments become capped again (step 5), which stabilizes the protrusion and relieves the membrane tension leading to closure of the water pores (step 6). Under normal conditions, the process of membrane extension is reversed when Ca;2+ is pumped out again or into stores of organelles, and the levels of monomeric actin decrease because of actin-filament polymerization, e.g., dendritic nucleation in the advancing lamellipodium of motile neutrophils. Constitutive ATP hydrolysis within actin filaments trigger the severing and depolymerization of older filaments by ADF/cofilins (step 7). Profilin then catalyzes nucleotide exchange and recycles ADP-actin monomers back to the ATP-actin pool (step 8). In this model, integrin-mediated adhesion and signaling have been omitted, but it is likely that adhesion-mediated signaling is as important or even more important than chemoattractant signaling during ongoing motility.
 View larger version (37K): [in a new window] |
Figure 9. How aquaporins and water flux contribute to the formation of membrane
protrusions. The aquaporin model presented here is based on the
dendritic-nucleation hypothesis originally suggested by Mullins et al.
[9
]. Here, water influx is a prerequisite in forming a
gap between the filament barbed ends and the membrane and in increasing
the diffusion of actin monomers as well as other vital components of
the motility machinery. The numbered steps are described in the
text.
|
In summary, our results provide definite experimental support for model(s) of cell motility and shape changes, which attribute a pivotal role for water transport across the plasma membrane. This influx contributes to the propulsive force for pseudopodia formation and cell motility. The opening and closing of finely tuned water-selective channels provide various types of cells, such as fibroblasts [50 ], neutrophils [42 ], macrophages [51 ], and keratocytes [2 ], with a tunable mechanism for motility control.
Received June 5, 2001; revised August 6, 2001; accepted August 6, 2001.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. Binet and D. Girard Novel human neutrophil agonistic properties of arsenic trioxide: involvement of p38 mitogen-activated protein kinase and/or c-jun NH2-terminal MAPK but not extracellular signal-regulated kinases-1/2 J. Leukoc. Biol., December 1, 2008; 84(6): 1613 - 1622. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ruiz-Ederra and A. S. Verkman Aquaporin-1 Independent Microvessel Proliferation in a Neonatal Mouse Model of Oxygen-Induced Retinopathy Invest. Ophthalmol. Vis. Sci., October 1, 2007; 48(10): 4802 - 4810. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ehrchen, L. Steinmuller, K. Barczyk, K. Tenbrock, W. Nacken, M. Eisenacher, U. Nordhues, C. Sorg, C. Sunderkotter, and J. Roth Glucocorticoids induce differentiation of a specifically activated, anti-inflammatory subtype of human monocytes Blood, February 1, 2007; 109(3): 1265 - 1274. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Belge and O. Devuyst Aquaporin-1--a water channel on the move Nephrol. Dial. Transplant., August 1, 2006; 21(8): 2069 - 2071. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hara-Chikuma and A.S. Verkman Aquaporin-1 Facilitates Epithelial Cell Migration in Kidney Proximal Tubule J. Am. Soc. Nephrol., January 1, 2006; 17(1): 39 - 45. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X.-M. Chen, S. P. O'Hara, B. Q. Huang, P. L. Splinter, J. B. Nelson, and N. F. LaRusso Localized glucose and water influx facilitates Cryptosporidium parvum cellular invasion by means of modulation of host-cell membrane protrusion PNAS, May 3, 2005; 102(18): 6338 - 6343. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
