| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of
*
Biochemistry and
Orthopedic Surgery, Section of Biochemistry and Molecular Biology, Rush University at Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL 60612; and
Immunex Corporation, Seattle, WA 98101
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
converting enzyme (TACE), a sheddase involved in the processing
of several substrates. Conversely, activation of the CD44 protease has
no effect on the release of TNF- from TACE-expressing cells,
although the same metalloprotease inhibitor effectively blocks both
TACE and the CD44 sheddase. Concomitant with anti-CD44 Ab- or
phorbol ester-induced CD44 shedding, dramatic changes are observed in
cell morphology and the structure of the actin cytoskeleton. Disruption
of actin assembly with cytochalasin reduces CD44 shedding, but not the
release of TNF-. Moreover, pharmacological activation of Rho family
GTPases Rac1 and Cdc42, which regulate actin filament assembly into
distinct cytoskeletal structures, has a profound effect on CD44
release. We conclude that the CD44 sheddase and TACE are distinct
enzymes, and that Ab- and phorbol ester-enhanced cleavage of CD44 is
controlled in a cell type-dependent fashion by Rho GTPases through the
cytoskeleton. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Inflammation is generally associated with increased membrane expression of CD44 (10, 13, 14, 15), suggesting a role of this adhesion receptor in the pathology of acute and chronic inflammatory diseases. Previous work in our laboratory has demonstrated that administration of a monoclonal anti-CD44 Ab (mAb IM7) to mice with experimental arthritis abrogated inflammation primarily by inhibiting the extravasation of leukocytes in the synovial joints (11, 12). Anti-CD44 treatment was associated with a dramatic increase in the levels of circulating CD44. Although the therapeutic role of enhanced CD44 release could not be clearly established in these animal models, we found that the anti-CD44 mAb acted similarly on several cell types including leukocytes and fibroblast-like synovial cells (11, 12), which have effector functions in joint destruction. Using three mAbs (IM7, KM201, and IRAWB14) to distinct epitopes of mouse CD44, we noted that the most extensive CD44 shedding was induced by IM7, both in vivo and in vitro (12). In vitro, the CD44-specific Abs were severalfold more effective when they were immobilized on plastic than when present in solution (12).
A diverse group of membrane proteins, including CD44, can undergo
proteolysis to release the extracellular domains from the cell surface
(11, 12, 16, 17, 18, 19). Protein ectodomain shedding from cells
can be commonly induced by treatment with protein kinase C (PKC)
activators such as phorbol esters, and inhibited by treatment with
metalloprotease inhibitors (17, 18, 20), indicating the
direct involvement of metalloproteases in the inducible process of
pericellular cleavage. The first sheddase has been identified as the
TNF- converting enzyme (TACE), which catalyzes the release of
TNF- from its membrane-bound precursor (21, 22). TACE
belongs to the ADAM (a disintegrin and metalloprotease) family of
cell-associated enzymes that share common motifs in their functional
domains, but whose substrate specificities are largely unknown
(23, 24, 25, 26). Intriguingly, TACE has been found to cleave
several substrates besides TNF-, including the p75 TNF- receptor,
TGF-, L-selectin (27), and the Alzheimer amyloid
protein precursor (28).
The initial focus of this study was to compare and analyze constitutive
and CD44-specific mAb-enhanced releases of CD44 from primary
fibroblasts derived from mouse synovium. The availability of cells from
TACE-deficient mice compelled us to determine whether TACE can cleave
CD44. Subsequent experiments, using cells from wild-type and
TACE-deficient mouse embryos, revealed a major difference in the
cellular mechanism(s) by which the proteolytic processing of CD44 and
TNF- is regulated, suggesting a linkage between cell adhesion and
CD44 cleavage.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Rat B cell hybridomas producing mAbs against CD44 (clones IM7
and KM81) (29, 30) were purchased from the American Type
Culture Collection (Manassas, VA). mAb IM7 binds to an epitope outside
of the HA-binding region of both mouse and human CD44 (11, 29, 30, 31). Purification and biotinylation of the mAbs, and
preparation of F(ab')2 and Fab were conducted as
described elsewhere (11, 12, 32, 33). Biotinylated mAb
KM114, recognizing an epitope in the HA-binding domain of mouse CD44
(30), biotinylated normal rat IgG1, purified mAbs against
mouse integrin 5 and 
1 subunits, and a
mouse TNF- ELISA kit were purchased from PharMingen (San Diego, CA).
Normal rat IgG (reagent grade) was obtained from Sigma (St. Louis, MO),
streptavidin-PE from Life Technologies (Grand Island, NY), and
HRP-conjugated streptavidin from Zymed (San Francisco, CA). Ultrapure
HA (Provisc7) was obtained from Alcon Laboratories (Fort Worth, TX),
human plasma fibronectin was a gift from Gene Homandberg (Rush
University, Chicago, IL), and type I collagen, purified from rat tail,
was obtained from Sigma. Recombinant human platelet-derived growth
factor (PDGF) was purchased from R&D Systems (Minneapolis, MN).
1,10-Phenanthroline, ZnCl2, PMA, cytochalasin B,
lysophosphatidic acid (LPA), and bradykinin were obtained from Sigma.
Small molecule protease inhibitor
N-R(2-hydroxaminocarbonyl)methyl-4-methylpentanoyl-tertbutyl-L-alanine
2-aminoethyl amide (Immunex Compound 3; IC3) was synthesized at Immunex
(21). Streptomyces hyaluronate lyase and
trimethylrhodamine isothiocyanate-conjugated phalloidin were purchased
from Sigma.
Cell isolation and culture
Synovial tissues were harvested aseptically from the knee joints of BALB/c mice under a dissecting microscope, and the cells were released enzymatically as described previously (11, 12). Adherent cells were cultured in DMEM in the presence of 10% FBS (HyClone, Logan, UT), and propagated by trypsinization. Primary mouse synovial fibroblast (MSF) monolayer cultures were used for CD44 shedding experiments after the third passage, when all cells showed fibroblast-like morphology. Primary mouse embryonic fibroblasts (MEF) and a ras- and myc-immortalized monocytic cell line derived from Dexter-type culture of mouse bone marrow (DRM) were generated from wild-type and TACE-deficient mice (27). All cells were maintained in DMEM containing 10% FBS, and DRM were cultured in the presence of 20 ng/ml recombinant murine GM-CSF and 10 ng/ml IL-3 (Sigma) (27).
Induction and measurement of CD44 and TNF- shedding
Culture dishes were coated with IM7, normal rat IgG, or HA. F(ab')2 and Fab of IM7 were also used to coat the bottom surface of multiwell plates. All reagents were diluted in sterile coating buffer (15 mM Na2CO3, 35 mM NaHCO3, pH 9.6). Rat IgG and IM7 were diluted to 25 µg/ml (except for the Ab fragment titration experiments), and HA was used at 2 mg/ml. Fibronectin and type I collagen were immobilized at concentrations ranging from 25 to 100 µg/ml coating buffer, and anti-integrin Abs were used at 25–50 µg/ml. After overnight incubation, the coated dishes were washed, and nonspecific binding sites were blocked with 10% FBS in DMEM. After brief treatment with trypsin-EDTA, synovial or embryonic fibroblasts were transferred to the coated dishes. Occasionally, residual cell surface HA was also removed by digestion with 10 U/ml Streptomyces hyaluronidase in DMEM for 30 min. Monocytic cells (DRM), which did not adhere to culture plastic, did not require trypsinization. The cells were seeded onto dishes containing the immobilized reagents and cultured in the presence of 10% FBS in DMEM for up to 48 h. At various time points, the conditioned media were harvested, and the cells in the corresponding wells were lysed with a lysis buffer composed of 50 mM Tris, pH 8.0, and 150 mM NaCl, containing 1% Nonidet P-40, 0.2% deoxycholic acid, 0.1% SDS, and protease inhibitors. The cell lysates were cleared by centrifugation and stored at -20°C. Shedding experiments were also performed using PMA at increasing concentrations up to 1000 ng/ml (18) for 24 h, or 500 ng/ml for 6 h. Cell viability was determined using trypan blue dye exclusion.
Soluble CD44 (sCD44) in conditioned media was quantitated as described
(12) using a murine sCD44 ELISA, which uses a pair of mAbs
(KM81 and IM7) against two distinct epitopes within the ectodomain of
CD44 (30). Soluble TNF- was determined using a
commercial ELISA kit.
Shedding inhibition and activation studies
Fibroblasts or monocytes were seeded onto uncoated dishes or
those containing immobilized rat IgG or IM7 (each coated at 25 µg/ml
concentration) or HA (coated at 2 mg/ml), and allowed to adhere to the
culture plastic surface for 3 h. To inhibit metalloprotease
activity (17, 18, 20, 34), 1,10-phenanthroline (with or
without ZnCl2) or IC3 was added at increasing
concentrations and incubated with the cells for 24 h. In some
experiments, the cells were cultured for 6, 16, or 24 h in the
presence of 75 µM IC3, and solvent (DMSO) was used as a control. To
study the involvement of the actin-based cytoskeleton in the shedding
of CD44 or TNF-, cytochalasin B, which disrupts actin
polymerization, was used (35). Cytochalasin B was added to
the culture media 3 h following cell seeding, and incubated with
the cells for 6 or 16 h. The conditioned media were assayed for
sCD44 and TNF-.
To study the roles of small GTPases RhoA, Rac1, and Cdc42 in CD44
shedding, selective agonists of the GTPases were used. MSF or DRM (wild
type) were seeded onto multiwell plates coated with rat IgG. The
culture medium (containing 5% FBS) was replaced with serum-free DMEM
3 h after plating, and the cells were serum starved for 16 h.
The cells were then incubated for 6 h in the absence or presence
of RhoA activator LPA (35, 36), Rac1 agonist PDGF
(37, 38), or Cdc42 activator bradykinin (37, 38) in serum-free DMEM. MSF and DRM cultured in IM7-coated
plates served as positive controls for CD44 shedding. The conditioned
media were assayed for sCD44 and TNF-.
Diluted samples of conditioned media and cell lysates from duplicate
wells of each treatment were used for the measurement of sCD44 and
TNF- levels. Data were analyzed using the Mann-Whitney U
test and the paired Student t test (34).
Immunoprecipitation of cell-associated CD44, Western blotting, and Northern hybridization
MSF were seeded onto petri dishes coated with rat IgG, IM7, or HA, and cultured for 24 h. The monolayers were lysed and precleared. CD44 was immunoprecipitated by incubation of lysates with IM7 and protein G-Sepharose, respectively, and eluted by boiling in nonreducing electrophoresis sample buffer. Equal volumes of samples were loaded onto 8% polyacrylamide gels, and the proteins were separated and transferred electrophoretically to nitrocellulose membranes (11). The membranes were blotted with biotinylated KM114 and HRP-conjugated streptavidin, and CD44 bands were visualized by ECL (Amersham, Arlington Heights, IL). CD44 was quantitated by densitometry using a PDI scanner and integrated software (Bio-Rad, Hercules, CA).
For mRNA expression studies, synovial fibroblasts were trypsinized, transferred to dishes coated with rat IgG, IM7, or HA, and cultured for 24 h. Total cellular RNA was isolated using the guanidinium thiocyanate-phenol-chloroform extraction method (39), electrophoresed, and transferred to GeneScreen nylon membranes (NEN, Boston, MA).
A 689-bp murine CD44 cDNA probe was generated by RT-PCR from the
CD44-rich T cell hybridoma 5/4E8 (40). Amplification was
conducted using a primer set designed for the standard form of mouse
CD44 cDNA (5'-TTGGGGACTTTGCCTCTTGC-3' from nucleotide 27 of exon 1, and
5'-GTCACAGTGCGGGAACTCC-3' from nucleotide 716 of exon 6) (41, 42). The PCR product was purified, sequenced, and cloned into a
pCR2.1 vector (Invitrogen, San Diego, CA). The purified CD44 cDNA probe
was labeled by the random primed DNA labeling method (Boehringer
Mannheim, Mannheim, Germany), using
[-32P]-dCTP (Amersham). Northern
hybridization was conducted using the QuikHyb hybridization system
(Stratagene, La Jolla, CA). Membranes were stripped and rehybridized
with a 32P-labeled cDNA probe for GAPDH.
Hybridization signals were quantitated using a PhosphorImager (STORM;
Molecular Dynamics, Sunnyvale, CA).
Flow cytometry, fluorescence labeling of F-actin, and microscopy
CD44 expression on the surface of suspended cells was quantitated using biotinylated KM114 mAb followed by PE-conjugated streptavidin, and biotinylated rat IgG1 served as the isotype control. Fibroblasts were detached using a nonenzymatic cell dissociation buffer (Life Technologies) before cell surface labeling. Flow cytometry was performed using a FACScan instrument and CellQuest software (12, 33). For actin labeling, fibroblasts or monocytes were cultured in chamber slides (Nunc, Naperville, IL) in the absence or presence of PMA or immobilized normal rat IgG, IM7, or HA. Additionally, serum-starved monocytes were cultured for 1 h in rat IgG-coated chamber slides in the absence or presence of LPA, PDGF, or bradykinin, as described for the shedding experiments. The cells were fixed, permeabilized, and stained with trimethylrhodamine isothiocyanate-conjugated phalloidin (33), and then viewed using a Nikon Microphot FXA epifluorescence microscope (Nikon, Melville, NY).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Soluble CD44 was detected in increasing amounts in the conditioned
media of MSF during 48 h of culture after trypsinization. Soluble
CD44 levels increased much more rapidly in cell cultures grown on
immobilized anti-CD44 mAb IM7 than in those on immobilized control
rat IgG (Fig. 1
A). However,
there was a constitutive release of CD44 from MSF as indicated by
slowly increasing sCD44 levels in the media of rat IgG-treated cells
(Fig. 1A). Importantly, cells grown on immobilized CD44
ligand, HA, did not exhibit enhanced CD44 release (Fig. 1A).
CD44 protein content in the cell lysates increased rapidly in cultures
grown on rat IgG or HA, whereas it remained low in those exposed to
immobilized IM7 (Fig. 1B). The same results were obtained
when endogenous HA was enzymatically removed from the cells before the
shedding experiment, indicating that pericellular HA had no influence
on CD44 shedding. The amounts of sCD44 released from MSF cultured on
immobilized fibronectin, type I collagen, or immobilized Abs against
5 and 1 integrin
subunits, were identical with the levels of constitutively shed CD44
from the cells in uncoated or rat IgG-coated dishes (results not
shown). This suggested that engagement of major fibroblast integrins,
such as the fibronectin receptor by immobilized ligand or Abs, did not
contribute to CD44 shedding.
|
85% less
cell-associated CD44 in cells grown on IM7-coated surfaces than in
those cultured on rat IgG or HA (Fig. 2A). However, Northern
hybridization, using RNA isolated from identically treated cultures
at 24 h, demonstrated no change in either the levels or the
composition of CD44 mRNA species (Fig. 2B).
|
Previous studies have demonstrated that shedding of various cell
surface proteins can be induced by external stimuli such as
cross-linking (43), or by internal activation of signaling
pathways, especially those mediated through PKC (16, 18, 44). Therefore, it was of interest to determine and compare the
effects of bivalent F(ab')2 and monovalent Fab of
IM7, and those of the PKC activator PMA, on CD44 shedding.
Dose-response studies revealed that the F(ab')2
was nearly as effective as intact IM7, whereas immobilized Fab failed
to increase the concentration of sCD44 above baseline levels (Fig. 3A). Cross-linking of the Fab
with secondary Ab partially restored CD44 shedding (Fig. 3A). Thus, bivalent binding of CD44 by Ab was required for
the accelerated release of CD44 from MSF, but the Fc part of IgG was
not involved in this activity. Surprisingly, MSF produced much lower
amounts of sCD44 upon stimulation with PMA (Fig. 3B) than in
the presence of immobilized IM7 (Fig. 3A).
|
A). The effective doses of
phenanthroline were in the millimolar range, and this compound was
toxic to the cells above 5 mM concentrations. A hydroxamic acid-based
metalloprotease antagonist, IC3, which blocks TACE activity (20, 21), inhibitedIM7-induced CD44 shedding at micromolar
concentrations (Fig. 4B), without an apparent cytotoxic
effect.
|
TACE has been described as a cell-associated metalloprotease that
processes not only TNF- but several other, functionally unrelated,
membrane proteins (27, 28). To determine whether TACE was
involved in constitutive or IM7-induced CD44 release, first, CD44
shedding from wild-type and TACE-deficient (27) MEF was
examined. Despite similar levels of cell surface CD44 expression (data
not shown), TACE-deficient MEF lost slightly less CD44 than the
wild-type fibroblasts on both immobilized rat IgG and IM7 (Fig. 5). However, the differences between
TACE+/+ and TACE-/-
fibroblasts were not significant, and CD44 loss from both cell types
was effectively inhibited by IC3 (Fig. 5). Comparison of CD44 and
TNF- shedding from wild-type and TACE-deficient DRM
(27) provided further evidence that TACE and the CD44
sheddase were distinct enzymes. First, TACE-deficient DRM cells shed
even more CD44 on immobilized IM7 mAb than the wild-type counterparts,
whereas their TNF- release under the same condition was negligible
(Table I). Second, PMA was a potent
inducer of TNF- shedding from wild-type DRM, but these cells
released only slightly higher amounts of the cytokine on immobilized
anti-CD44 Ab than on control rat IgG (Table I). It is important to
note that elevated sCD44 levels in the media of TACE-deficient,
relative to wild-type, monocytes were not due to a higher membrane
expression, because the two cell types demonstrated similar cell
surface CD44 density (data not shown) and comparable levels of
constitutive CD44 release on both immobilized rat IgG and HA (Table I).
In response to PMA stimulation, both TACE+/+ and
TACE-/- DRM shed more CD44 than MSF (see Table I and Fig. 3
B) or wild-type MEF (data not shown).
Furthermore, unlike MSF (see Fig. 3A), DRM released
substantial amounts of CD44 on immobilized IM7 Fab (data not shown).
These results suggested that there were cell type-specific differences
in the regulation of CD44 shedding.
|
|
DRM did not show significant adherence to uncoated or rat
IgG-coated culture plastic (data not shown), but 50% of the cells
attached to immobilized HA (Fig. 6A). Importantly, DRM not only
attached to, but also rapidly spread on, the IM7-coated surface (Fig. 6B). A number of DRM also became adherent to uncoated
surfaces upon PMA treatment (Fig. 6C). This latter finding
was not entirely surprising because PKC activation has been known to
promote integrin-mediated adhesion and spreading in several cell types
(37, 45, 46, 47). We took advantage of the IM7- and
PMA-induced spreading of DRM to investigate a possible relationship
between the assembly of the actin cytoskeleton (determining cell
morphology) and CD44 shedding. Fluorescence staining of F-actin showed
diffuse intracellular distribution within small and round-shaped DRM on
immobilized HA (Fig. 6D). DRM that spread on
immobilized IM7 were enlarged and demonstrated actin in numerous
peripheral microspikes and filopodia (Fig. 6E). Cells
stimulated with PMA were also slightly enlarged, but their fluorescence
patterns suggested actin distribution in broad lamellipodial, rather
than thin filopodial, membrane protrusions (Fig. 6F).
|
). Actin bundles
(stress fibers) showed a predominantly parallel arrangement in cells
exposed to the rat IgG-coated control surface (Fig. 7A), or
formed long intertwining cables in cells attaching to, and
presumably migrating on, immobilized HA (Fig. 7C). In
contrast, the majority of the fibroblasts grown on immobilized
IM7 demonstrated numerous short, radially arranged filaments
at the cell periphery (Fig. 7C), an actin structure
characteristically associated with filopodia, retraction fibers, and
microspikes (35, 37, 48).
|
A) and DRM (Fig. 8B) in a dose-dependent manner. PMA-induced CD44 release
from DRM was also inhibited, albeit to a lesser degree, by cytochalasin
B (Fig. 8C). Importantly, PMA-induced TNF- shedding from
DRM was not affected by treatment with this cytoskeleton-destabilizing
agent (Fig. 8D).
|
A, third bar). However,
activation of Rac1 with PDGF did increase sCD44 levels by >2-fold
(Fig. 9A, fourth bar). In contrast to DRM, MSF responded to
LPA, but not to PDGF, treatment by a slight increase in CD44 shedding
(results not shown). Importantly, CD44 shedding from both DRM (Fig. 9A, fifth bar) and MSF (data not shown) was significantly
accelerated upon incubation with the Cdc42 agonist bradykinin,
suggesting that the Cdc42-dependent pathway was activatable and most
likely contributed to CD44 cleavage in both cell types. DRM, cultured
at low cell density in the absence or presence of Rho family agonists,
exhibited homogenous actin staining in both untreated (Fig. 9B) and LPA-treated (data not shown) wells. The majority of
cells formed membrane ruffles in response to PDGF (Fig. 9C),
or numerous microspikes and filopodia upon bradykinin treatment (Fig. 9D).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
from wild-type DRM is
readily induced by PMA, but not by anti-CD44 Ab treatment. Taken
together, these results indicate that TACE and the CD44 protease are
distinct enzymes. Cross-linking of another adhesion receptor, L-selectin, with bivalent mAb or chemical cross-linkers, has been found to induce the shedding of L-selectin, but not CD44, from leukocytes (43). We observe enhanced CD44 shedding from MSF only when bivalent or cross-linked monovalent IM7 is displayed on the culture plastic surface; however, monovalent mAb is sufficient for the induction of substantial CD44 release from DRM. Monovalent and bivalent mAbs, when immobilized, must both display multiple Ag binding sites on the solid surface. However, CD44 dimerization by bivalent mAb, or a specific distance between Ab paratopes on the solid surface, might be required for the induction of excessive shedding of CD44 in some, but not in other, cell types. We detect only constitutive CD44 release when MSF and DRM are exposed to a surface coated with the multivalent CD44 ligand, HA, suggesting that in addition to repetitive binding sites, other properties of the immobilized ligand are also of great importance in the shedding process. It has been shown that the affinity of CD44-IM7 binding (49) is several orders of magnitude higher than the affinity between CD44 and HA (50). Therefore, one possible explanation for the difference in shedding induction is that cell surface CD44 recognizes immobilized IM7 as a high-affinity, and HA as a low-affinity adhesion ligand (51).
In support of a differential cell response to the CD44 ligand HA and Ab, we find striking differences in cell morphology upon cell binding to immobilized HA and IM7. DRM attach to, but do not spread on, immobilized HA, whereas they display a spread phenotype on the anti-CD44 mAb. MSF do not show an obvious difference in cell morphology on the surface coated with HA and IM7, yet they exhibit profound differences in the organization of filamentous actin. A number of DRM also spread (on an uncoated surface) in response to phorbol ester (PMA) treatment; in the meantime, these cells shed substantial amounts of CD44. As strong cell adhesion, resulting from extended membrane-substratum contact areas, can impede cell movement (51, 52), shedding might be one of the mechanisms used by cells in an attempt to break up or prevent CD44-dependent strong adhesions, especially when a high-affinity Ab is the adhesion ligand. However, if only IM7-bound CD44 molecules were cleaved, sCD44 could not be detected in the media of cells cultured on immobilized IM7. The accumulation of free, immune reactive CD44 in the media of these cells suggests that CD44 proteolysis is not limited to the sites of actual CD44-Ab binding. The generalized nature of the shedding process, involving unengaged substrate molecules, may explain in part why progressive CD44 cleavage occurs following an alternative mechanism of cell stimulation, i.e., with phorbol ester, as seen in monocytes (this study) or in tumor cells (36).
The coincidence of morphologic changes and enhanced CD44 release, both
induced by IM7 or PMA, favors the suggestion that cytoskeletal assembly
plays a role in the regulation of receptor shedding. Indeed,
destabilization of the cortical actin framework with cytochalasin B
results in a dose-dependent inhibition of CD44 release from cells that
are exposed to either IM7 or PMA. In contrast, PMA-induced TNF-
shedding is unaffected by cytochalasin treatment. These results
indicate that the enhancement of CD44, but not TNF-, proteolysis is
linked in some way to the assembly of actin cytoskeleton.
Actin-binding proteins such as ezrin, radixin, and moesin (ERM) play an essential role in actin polymerization in response to activation of Rho family GTPases such as RhoA, Rac1, and Cdc42 (53). The intracellular domain of CD44 has a binding site for ERM proteins (54, 55), suggesting that through a potential cytoplasmic tail-ERM-actin interaction, CD44 can either be controlled by or gain some control over the organization of the cytoskeleton. Okamoto et al. (36) have found that treatment of tumor cells with phorbol ester induces the redistribution of both ERM proteins and CD44 to newly formed actin-rich membrane-ruffling areas (lamellipodia), concomitant with enhanced CD44 cleavage. Expression of active Rac1 construct in the same tumor cells also results in the redistribution of CD44 to membrane ruffles and the enhancement of CD44 proteolysis (36). In this experimental setting, activation of the Rac1 pathway, via either phorbol ester stimulation (37) or Rac1 overexpression, is accompanied by actin reorganization, colocalization of both CD44 and ERM with actin, and CD44 shedding (36). The simultaneous occurrence of these events indicates a causal relationship between the Rac1-regulated assembly of CD44-associated adhesion complexes and CD44 cleavage.
We find that selective pharmacological agonists of Rho family proteins RhoA, Rac1, and Cdc42 are able to accelerate CD44 cleavage. The degree at which each of these Rho family GTPase activators enhances CD44 proteolysis appears to be related to both the regulatory function of the GTPase and the cell-specific pattern of cytoskeletal assembly. For example, RhoA activation with LPA (35, 36) influences CD44 shedding in MSF that display stress fibers, but has essentially no effect on CD44 shedding from DRM that lack the ability to assemble actin into stress fibers. The Rac1 agonist PDGF (37, 38) has a greater effect on DRM, a monocytic cell type that, in the adherent state, exhibits membrane ruffling over the entire cell surface, than on MSF, a fibroblast-type cell that forms lamellipodial extension only at the leading edge of the cell membrane during locomotion (48). Bradykinin, which induces the formation of microspikes/filopodia through Cdc42 activation (37, 38) (also demonstrated here in DRM), has the greatest effect on CD44 cleavage in both DRM and MSF. Accordingly, filopodial protrusions and concomitant enhancement of CD44 release are observed in both cell types during their adherence to immobilized IM7. These findings favor the hypothesis that filopodia and other actin-rich thin membrane protrusions (e.g., microspikes, retraction fibers, and microvilli) are specifically linked to IM7-enhanced CD44 cleavage, thus suggesting a dominant role for Cdc42, among the three Rho family members, in the regulation of Ab-initiated morphological changes and CD44 shedding.
It is an open question as to what drives excessive CD44 cleavage in response to Ab and PMA treatment. As the CD44 gene expression profile remains unchanged upon exposure of cells to immobilized IM7, and CD44 proteolysis is inhibited by specific inhibitors of metalloproteases, it is unlikely that de novo synthesis of sCD44 would account for the rapid accumulation of cleavage products in Ab-treated cultures. The fact that a CD44-specific trigger is required for the putative sheddase to work at a high capacity suggests that in this system CD44 is selectively targeted to become a competent substrate. In the case of PMA stimulation, CD44-unrelated bystander mechanisms (47) may act in concert to attack CD44. For example, PKC-mediated activation of multiple sheddases (18), clustering/capping of cell surface molecules (56), and increased phosphorylation of regulatory proteins involved in the integrin-dependent assembly of the cytoskeleton (46) can ultimately lead to an enhanced cleavage of several membrane-associated proteins (18) including CD44.
It remains to be elucidated how exactly the binding of cell surface CD44 to immobilized Ab leads to Rho family GTPase activation and subsequent actin reorganization in different cell types. Molecular studies, using an experimental approach described here, and focusing on specific interactions among CD44, ERM, and Rho GTPases upon cell adhesion to non-HA CD44 ligand "mimetics" or (as yet unidentified) natural ligands, will uncover common regulatory elements in cytoskeleton assembly and CD44 shedding.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Katalin Mikecz, Section of Biochemistry and Molecular Biology, Department of Orthopedic Surgery, Rush University at Rush-Presbyterian-St. Luke’s Medical Center, 1653 West Congress Parkway, Chicago, IL 60612. E-mail address: kmikecz{at}rush.edu
3 Abbreviations used in this paper: HA, hyaluronan (hyaluronic acid); DRM, ras- and myc-immortalized monocytic cell line derived from Dexter-type culture of mouse bone marrow; ERM, ezrin, radixin, and moesin; IC3, Immunex Compound 3; LPA, lysophosphatidic acid; MEF, mouse embryonic fibroblasts; MSF, mouse synovial fibroblasts; PDGF, platelet-derived growth factor; PKC, protein kinase C; sCD44, soluble CD44; TACE; TNF- converting enzyme.
Received for publication December 11, 2000. Accepted for publication April 23, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
of neutrophil cell surface expression of the sialophorin CD43 and the hyaluronate receptor CD44 through a proteolytic mechanism. Eur. J. Immunol. 21:3045.[Medline]
from cells. Nature 385:729.[Medline]
. Nature 385:733.[Medline]
and notch. Cell 90:589.[Medline]
converting enzyme is involved in regulated -secretase cleavage of the Alzheimer amyloid protein precursor. J. Biol. Chem. 273:27765.
during HeLa cell adhesion to a gelatin substratum. J. Biol. Chem. 271:13008.This article has been cited by other articles:
|
|
 |
|
  N. Borg and M. Holland The effect of glycosaminoglycans on rat gametes in vitro and the associated signal pathway Reproduction, March 1, 2008; 135(3): 311 - 319. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zoller, P. Gupta, R. Marhaba, M. Vitacolonna, and P. Freyschmidt-Paul Anti-CD44-mediated blockade of leukocyte migration in skin-associated immune diseases J. Leukoc. Biol., July 1, 2007; 82(1): 57 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Roda-Navarro and H. T. Reyburn Intercellular protein transfer at the NK cell immune synapse: mechanisms and physiological significance FASEB J, June 1, 2007; 21(8): 1636 - 1646. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Brule, N. Charnaux, A. Sutton, D. Ledoux, T. Chaigneau, L. Saffar, and L. Gattegno The shedding of syndecan-4 and syndecan-1 from HeLa cells and human primary macrophages is accelerated by SDF-1/CXCL12 and mediated by the matrix metalloproteinase-9 Glycobiology, June 1, 2006; 16(6): 488 - 501. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Garton, P. J. Gough, and E. W. Raines Emerging roles for ectodomain shedding in the regulation of inflammatory responses J. Leukoc. Biol., June 1, 2006; 79(6): 1105 - 1116. [Abstract] [Full Text] [PDF]
|
|
