Address correspondence and reprint requests to Patricia F. Maness, Department of Biochemistry and Biophysics, 505 Mary Ellen Jones Building, CB#7260, University of North Carolina School of Medicine, Chapel Hill, NC 27599–7260, USA. E-mail: email@example.com
The neural cell adhesion molecule (NCAM) plays a key role in neural development, regeneration and synaptic plasticity. This study describes a novel function of NCAM140 in stimulating integrin-dependent cell migration. Expression of NCAM140 in rat B35 neuroblastoma cells resulted in increased migration toward the extracellular matrix proteins fibronectin, collagen IV, vitronectin, and laminin. NCAM-potentiated cell migration toward fibronectin was dependent on β1 integrins and required extracellular-regulated kinase (ERK)1/2 mitogen-activated protein kinase (MAPK) activity. NCAM140 in B35 neuroblastoma cells was subject to ectodomain cleavage resulting in a 115 kDa soluble fragment released into the media and a 30 kDa cytoplasmic domain fragment remaining in the cell membrane. NCAM140 ectodomain cleavage was stimulated by the tyrosine phosphatase inhibitor pervanadate and inhibited by the broad spectrum metalloprotease inhibitor GM6001, characteristic of a metalloprotease. Moreover, treatment of NCAM140-B35 cells with GM6001 reduced NCAM140-stimulated cell migration toward fibronectin and increased cellular attachment to fibronectin to a small but significant extent. These results suggested that metalloprotease-induced cleavage of NCAM140 from the membrane promotes integrin- and ERK1/2-dependent cell migration to extracellular matrix proteins.
The neural cell adhesion molecule (NCAM) belongs to the immunoglobulin superfamily and is expressed as three major isoforms. Two of them are transmembrane proteins with either a short (NCAM140) or a long (NCAM180) cytoplasmic domain, whereas NCAM120 is a glycosylphosphatidylinositol-anchored protein (for review see Panicker et al. 2003). The extracellular region of each NCAM isoforms consists of five immunoglobulin-like (Ig) domains and two fibronectin type III domains, and mediates homophilic NCAM–NCAM or heterophilic binding to other proteins (Rutishauser 2000). NCAM is expressed in the developing nervous system with a distinct pattern of spatiotemporal distribution that correlates with axon guidance and targeting (for review see Hinsby et al. 2004).
Interactions of the NCAM extracellular domain can result in activation of several signal transduction pathways implicated in NCAM-dependent neurite outgrowth. NCAM140 activates the extracellular-regulated kinase (ERK)/mitogen-activated protein kinase (MAPK) pathway important for neurite outgrowth through activation of focal adhesion kinase (FAK), a key signal transducer of integrin activation (Beggs et al. 1997; Schmid et al. 1999). FAK activation by NCAM is mediated by the Src family kinase p59fyn and culminates in phosphorylation and activation of cAMP-response element binding protein (CREB) (Beggs et al. 1997; Schmid et al. 1999). NCAM140 can also signal via a fibroblast growth factor (FGF) receptor-dependent pathway involving phospholipase C-γ and diacylglycerol lipase (reviewed in Povlsen et al. 2003). The FAK/p59fyn and FGF-receptor pathways can converge at the level of ERK activation (Niethammer et al. 2002). NCAM is also capable of activating phosphatidylinositol 3-kinase/Akt and protein kinase C (Kolkova et al. 2000; Ditlevsen et al. 2003).
Several lines of evidence suggest that NCAM can regulate cell motility but the mechanism remains unclear. Overexpression of polysialylated NCAM (PSA-NCAM) enhances migration of mouse neural precursor cells along the ventral pathway of chick embryos (Franceschini et al. 2004), and removal of PSA from NCAM perturbs the migration of a subset of neurons in chick forebrain (Murakami et al. 2000). PSA-NCAM in the rostral migratory stream is needed for migration of olfactory interneuron precursors from the subventricular zone but it is not certain if this represents a direct effect on neuronal migration or indirect effect of glial tube disruption (Hu and Rutishauser 1996; Chazal et al. 2000). In addition, soluble form of NCAM promotes Schwann cell migration in vitro (Thomaidou et al. 2001), whereas NCAM140 expression reduces migration of BT4Cn glioma cells through heterophilic interactions with a heparin sulfate proteoglycan (Prag et al. 2002). However, overexpression of soluble NCAM plays an adverse role in neuronal development, as demonstrated by abnormal secretion in affected brain regions and cerebrospinal fluid of schizophrenia patients, correlating with the severity of the disease (reviewed in Panicker et al. 2003). Moreover, a transgenic mouse overexpressing the extracellular domain of NCAM with the beginning of terminal neuronal differentiation displays reduced synaptic connectivity and behaviors similar to schizophrenia (Pillai-Nair et al. 2005).
ERK kinase activation is often required for cell motility downstream of cell adhesion molecules including L1 (Mechtersheimer et al. 2001; Thelen et al. 2002; Silletti et al. 2004) and integrins (reviewed in Slack-Davis and Parsons 2004). Because NCAM140 activates a FAK–Fyn–ERK pathway similar to integrin signaling, we investigated the hypothesis that NCAM may regulate migration of cells towards proteins of the extracellular matrix by functional interaction with integrins. Here we report a novel function of NCAM as a stimulator of integrin-dependent neuronal cell migration toward extracellular matrix proteins, and demonstrate that extracellular cleavage of NCAM mediates cell migration potentially by modulating adhesion of cells to extracellular matrix.
Materials and methods
Cell culture conditions and transfection of B35 cells
The B35 neuroblastoma cell line derived from rat central nervous system (Schubert et al. 1974) expresses a variety of neuronal markers, including neurotransmitters, and can be induced to extend long neurites (reviewed in Otey et al. 2003). B35 cells were maintained in Dulbecco's modified Eagle's medium (4500 mg glucose/L) supplemented with 10% fetal calf serum, 100 U/mL penicillin and 100 µg/mL streptomycin plated on 0.01% poly-l-lysine-coated plastic dishes and grown in a humidified chamber containing 5% CO2. For expression of human NCAM140 (NCAM140), the cDNA (kind gift from R. Gerardy-Schahn, Hannover, Germany) was cloned by PCR into the eukaryotic expression plasmid pcDNA3 introducing BamHI restrictions sites framing the coding region of NCAM140 (forward: 5′-CGCGGATCCATGCTGCAAACTAAGGATCTC-3′, reverse: 5′-CGGGATCCCGTCATGCTTTGCTCTCGTTCTC-3′). The PCR was carried out with Pfu polymerase (Stratagene, La Jolla, CA, USA) and the coding region of NCAM140 was verified by direct DNA sequencing (MWG Biotech, Ebersbach, Germany).
B35 cells were grown to 60% confluence and transfected with 10 µg of plasmid DNA and LipofectAMINE PLUS™ reagent (Invitrogen, Karlsruhe, Germany). Cells were selected in Geneticin-supplemented medium (500 µg/mL) added 48 h after transfection. Approximately 37% of the selected clones homogeneously expressed NCAM on the cell surface, as determined by immunofluorescence staining with antibodies against human NCAM (123C3) (provided by R. Michalides, Amsterdam, the Netherlands). Five of the NCAM-positive clones were further analyzed by immunoblotting with these antibodies, and four showed a single band of 140 kDa, indicating that the entire NCAM140 protein was correctly expressed.
Haptotactic cell migration assay
Haptotactic migration assays were performed as described (Thelen et al. 2002; Otey et al. 2003). Briefly, transwell chambers with 8.0-µm diameter pores were coated on the bottom side with extracellular matrix proteins or bovine serum albumin (fatty acid-free, Sigma, St. Louis, MO, USA) at a concentration of 4 µg/mL. The bottom surfaces of the filters were then washed and blocked in 2% bovine serum albumin. Cells were detached using HBSS/Na-EDTA (5 mm) and 30 000 cells plated per chamber. Some cells were preincubated with antibodies (6 µg/100 µL medium) for 20 min at 4°C in serum-free medium.
To score B35 cells, cells from the top or from the bottom side of the filter were removed and the remaining cells were stained for 10–15 min with Gill's formula hematoxylin (Vector Laboratories, Burlingame, CA, USA). Cells from randomly selected fields were counted using a 20 × microscope objective on a Zeiss Axiovert 200 inverted microscope (Carl Zeiss, Inc., Thornwood, NY, USA) and the mean number of cells per field was multiplied by a factor based on the number and size of fields scored and on a filter diameter of 6.5 mm to obtain the total number of cells migrated. Statistical analysis was performed using Student's t-test.
Determination of ectodomain shedding of NCAM140 by immunoblot analysis
Cells were grown on poly-l-lysine-coated 60-mm plastic dishes to 70–80% confluence. For analysis of NCAM cleavage the cells were washed twice with phosphate-buffered saline and serum-free medium was added. When indicated, pervanadate (10 mm stock solution, 200 µm final concentration on cells), GM6001 (10 µm, 0.4% dimethylsulfoxide final concentration, Chemicon, Temecula, CA, USA), or U0126 (10 or 20 µm, 0.1% or 0.2% dimethylsulfoxide final concentration, Promega, Madison, WI, USA) were included in the medium. A stock solution of pervanadate was freshly prepared before use by mixing equimolar amounts of 30% (w/w) H2O2 and sodium orthovanadate (Sigma). If GM6001 or U0126 were added together with pervanadate, they were applied 15 min before the addition of pervanadate. After the indicated time period (2 or 20 h), conditioned media were collected and cells were harvested in ristocetin-induced platelet aglutination lysis buffer (10 mm sodium phosphate buffer pH 7.2, 40 mm NaF, 2 mm EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate) containing 1 × Complete™ protease inhbitor cocktail (Roche Diagnostics, Mannheim Germany). Lysates were clarified by centrifugation and the protein concentration determined by the bicinchoninic acid protein assay (Pierce, Rockford, IL, USA). The collected media were centrifuged for 10 min at 4°C (1000 g) to remove debris, and the supernatants were concentrated on Centricon YM-30 centrifugal filter devices (Millipore, Bedford, MA, USA). Lysates (25 µg) and media (equivalent adjusted volumes based on lysate protein concentration) were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Membranes were blocked for 2 h in blocking solution (5% milk powder in Tris-buffered saline/0.05% Tween-20) prior to incubation for 2 h with OB11 antibodies (Sigma), which recognize the cytoplasmic domain of NCAM, for cell lysates, or H300 antibodies (Santa Cruz Biotechnology, CA, USA), against the extracellular domain of NCAM, for media. Membranes were washed six times with Tris-buffered saline/0.05% Tween-20, and incubated with peroxidase-conjugated secondary antibodies [lysate: goat anti-mouse; media: goat anti-rabbit antibodies (1:10 000), Jackson ImmunoResearch, Westgrove, PA, USA] for 1 h in blocking solution. After six additional washing steps (5 min, Tris-buffered saline/0.05% Tween-20), proteins were detected by enhanced chemiluminescence (PerkinElmer, Inc., Boston, MA, USA).
To measure cellular adhesion, 24-well plates were coated with fibronectin (Invitrogen) overnight at 4°C. The following day the wells were blocked with 2% fatty acid-free bovine serum albumin for at least 30 min and washed with serum-free media. Cells were detached using HBSS/Na-EDTA (5 mm) and 10 000 cells were plated per well in medium with 10% fetal calf serum. The medium was replaced by serum-free medium 3 h after plating, which, in some cases, contained GM6001 (10 µm). After 17 h, the plates were inverted and centrifuged at 1643 g at room temperature for 5 min. The attached cells were fixed with 4%p-formaldehyde for 15 min, washed three times with phosphate-buffered saline, and stained with hematoxylin (Gill's formula) for 10 min. After additional washing with phosphate-buffered saline, cells were counted from randomly selected fields using a 10 × microscope objective on a Zeiss Axiovert 200 inverted microscope (Carl Zeiss, Inc., Thornwood, NY, USA), and the mean number of cells per field was calculated. Statistical analysis was performed using Student's t-test.
NCAM140 potentiates haptotactic cell migration to extracellular matrix proteins through integrins
To determine whether NCAM140 stimulated migration of B35 neuroblastoma cells, haptotactic migration of NCAM140-expressing and non-expressing B35 cells was measured in Transwell assays using modified Boyden chambers in which cells migrated from top to bottom chambers through filters coated on the bottom side with extracellular matrix proteins. Consistent with previous reports (Thelen et al. 2002), non-expressing B35 cells showed greater migration toward all of the extracellular matrix proteins compared to random migration (bovine serum albumin control). NCAM140-expressing B35 cells displayed significantly (p < 0.005) enhanced haptotactic migration toward fibronectin, vitronectin, collagen IV and laminin compared to non-expressing B35 cells (Fig. 1a). Thus, NCAM140 promoted haptotactic cell migration toward a range of extracellular matrix substrates.
To determine whether NCAM140-potentiated migration to extracellular matrix proteins was mediated through an integrin-dependent mechanism, NCAM140-expressing and non-expressing B35 cells were incubated with function-blocking antibodies directed against the extracellular region of β1 integrin and migration toward fibronectin assayed. NCAM140-expressing B35 cells treated with β1 integrin blocking antibodies exhibited significantly reduced migration compared to cells that were untreated or exposed to an equivalent amount of non-immune immunoglobulin (Fig. 1b). However, NCAM-potentiated migration was not completely blocked by the integrin antibodies, suggesting that other integrin subclasses may contribute. These results suggested that NCAM140 in B35 neuroblastoma cells stimulated β1 integrin-dependent migration to fibronectin.
NCAM140 shedding plays a functional role in NCAM140-stimulated haptotactic migration
Ectodomain shedding of adhesion molecules, such as L1, CD44 and SHPS-1, facilitates cell migration (Kajita et al. 2001; Mechtersheimer et al. 2001; Ohnishi et al. 2004). NCAM140 can be cleaved by a metalloprotease in B35 neuroblastoma cells, primary cortical neurons, and transfected l-fibroblasts. The cleavage has been demonstrated to depend on pervanadate and ATP (Hubschmann et al. 2005; Hinkle et al. in preparation). Consistent with our previous results, ectodomain shedding of NCAM140 was strikingly induced in response to pervanadate treatment in B35 neuroblastoma cells to produce a 30 kDa residual cytoplasmic domain fragment that remained associated with the membrane, and a 115 kDa soluble NCAM ectodomain fragment that was released into the culture supernatant (Fig. 2a). To determine whether NCAM140 shedding in B35 cells was metalloprotease-dependent, cultures were pretreated with the broad spectrum metalloprotease inhibitor GM6001 (10 µm). Levels of the 30 kDa cytoplasmic domain fragment and the 115 kDa soluble NCAM ectodomain were greatly reduced in the presence of GM6001 in unstimulated and pervanadate-stimulated cells after 2 h (Fig. 2a). Cultures were also treated with GM6001 for 20 h to resemble the conditions used for the migration assay (below). After 20 h, unstimulated cultures released a similar amount of the 115 kDa NCAM extracellular domain into the media as the pervanadate-induced cultures did in only 2 h, and GM6001 also reduced the levels of cleavage (Fig. 2a).
To determine whether ectodomain shedding affected NCAM-dependent haptotactic migration, cultures were treated with 10 µm GM6001 and migration toward fibronectin was analyzed. As shown in Fig. 2(b), GM6001 significantly reduced migration of NCAM140-expressing B35 cells but did not affect basal migration of non-expressing B35 cells. These results indicated that, like other members of the Ig superfamily, ectodomain shedding of NCAM140 significantly stimulated cellular migration.
Metalloprotease inhibition enhances adhesion of NCAM140-expressing B35 cells
It has been shown that integrin-dependent cell migration is dependent on adhesion to the surrounding extracellular matrix, with an optimal rate of cell migration corresponding to intermediate levels of cell adhesion to extracellular matrix substrate (Palecek et al. 1997). To investigate whether ectodomain shedding of NCAM influenced cellular adhesion to extracellular matrix, NCAM140-expressing B35 cells were treated with 10 µm GM6001 and adhesion to fibronectin was analyzed using a centrifugation assay. As shown in Fig. 3, there was a small but significant increase (p < 0.01) in the number of cells that adhered to fibronectin in GM6001-treated cultures compared to control dimethylsulfoxide-treated cells. These results suggested that inhibition of NCAM shedding resulted in increased cellular adhesion, correlating with decreased migration. Thus ectodomain shedding of NCAM may limit cell adhesivity to fibronectin substrates to promote an optimal rate of haptotactic migration.
Pervanadate-induced ectodomain shedding of NCAM140 has been shown to depend on MEK activity (Hinkle et al. in preparation). Therefore, we tested whether the specific MEK inhibitor U0126 also has an effect on constitutive NCAM140 shedding using migration conditions. Application of 20 µm U0126 for 20 h resulted in a clear reduction of NCAM140 shedding (Fig. 4a), as indicated by the weaker band at 30 kDa compared to dimethylsulfoxide control-treated cells. We could not see this difference in the media supernatant (data not shown), which could be explained by the lower levels of constitutive shedding in these cultures, and thus a smaller difference between untreated and U0126-treated cells compared to the more robust levels of shedding seen in Fig. 2(a) (last two lanes).
Activation of MAPKs can mediate integrin-dependent haptotactic cell migration (Klemke et al. 1997). To determine whether NCAM140-stimulated cell migration was dependent on ERK activation through the dual specificity kinase MEK, NCAM140-expressing B35 cells were treated with the MEK1/2 inhibitor U0126 and migration to fibronectin was assayed. Treatment with U0126 (10–20 µm) progressively reduced NCAM140-stimulated haptotactic cell migration and, to a very small but significant extent, basal NCAM-independent migration to fibronectin (Fig. 4b). The results were consistent with the hypothesis that NCAM140-stimulated cell migration is mediated by ERK1/2 MAPK activity. Since NCAM has been demonstrated to be able to activate ERK1/2 through activation of the FGF receptor after antibody triggering (Niethammer et al. 2002), we also investigated whether inhibition of the FGF receptor by PD 173074, a highly specific inhibitor, influenced NCAM-dependent cell migration. B35 and NCAM140-B35 cells were treated with 50 nm PD 173074, a concentration demonstrated to completely inhibit FGF receptor dependent effects in B35 cells (Diestel et al. 2004). The inhibitor had no effect on B35 or NCAM140-B35 cell migration, respectively (not shown).
In this report, we provide evidence for a novel function of NCAM140 as a stimulator of β1 integrin-dependent cell migration towards extracellular matrix proteins. Our data also demonstrate a role of metalloprotease-dependent, ERK1/2 regulated NCAM140 cleavage in NCAM-stimulated haptotactic cell migration.
Our results suggest a functional interaction between NCAM140 and β1 integrins. Such an interaction may occur at the level of intersecting signaling pathways, since both receptors activate FAK/p59fyn/ERK signaling pathways (Beggs et al. 1997; Schmid et al. 1999; for review see Guo and Giancotti 2004). A limitation of the study was that we were not able to identify the α integrin subunit interacting with β1 integrin for NCAM140-stimulated migration, because rodent-specific function-blocking antibodies against other integrin subunits were not available. Possible β1 integrin receptors for fibronectin expressed in neuronal cell types include α3, α4, α5, αv, and α8. Furthermore, β3 and β5 integrins might account for the small amount of residual migration that was not blocked by β1 integrin antibodies. Other Ig family members, such as L1 (Felsenfeld et al. 1994; Schmidt et al. 1996; Yip et al. 1998; Silletti et al. 2000; Mechtersheimer et al. 2001; Thelen et al. 2002), CHL1 (Buhusi et al. 2003), and NrCAM (Treubert and Brummendorf 1998), have been shown to interact with integrins, and for L1 and CHL1 these interactions have a physiological role in integrin-dependent cell migration (Mechtersheimer et al. 2001; Thelen et al. 2002; Buhusi et al. 2003). Thus, NCAM joins the L1 family as candidate integrin-interacting cell recognition molecules that modulate cell migration in response to extracellular matrix proteins. NCAM140 may not bind directly to β1 integrins to mediate its migration-promoting function. In contrast to L1 and CHL1, NCAM140 does not contain RGD or DGEA integrin binding motifs, which are responsible for cell migration stimulated by these molecules (Thelen et al. 2002; Buhusi et al. 2003). Because NCAM and β1 integrins were not able to co-immunoprecipitate from cell extracts (Beggs et al. 1997), such an interaction may be weak or transient, mediated through an unidentified binding motif or indirectly via an unknown linker protein to facilitate integrin-dependent migration.
Our results also indicate that NCAM140-dependent cell migration is regulated by ectodomain shedding of NCAM140 mediated by a metalloprotease. The ability of the tyrosine phosphatase inhibitor pervandate to further stimulate NCAM140 shedding suggests that the metalloprotease is regulated by a pathway involving protein tyrosine phosphorylation/dephosphorylation. We also show that NCAM-dependent cell migration to fibronectin required an intact MEK–ERK signaling pathway (independently from FGF receptor activity), in accord with the requirement of MEK–ERK in integrin-dependent haptotactic migration stimulated by L1 or CHL1 (Thelen et al. 2002; Buhusi et al. 2003). We have not identified a target for ERK responsible for migration stimulation; however, Klemke has shown that the ERK target myosin light chain kinase (MLCK) is required for integrin-dependent haptotactic migration in other cell systems (Klemke et al. 1997; Cho and Klemke 2000). MAPK activity has also been shown to regulate metalloprotease-dependent ectodomain shedding (Fan and Derynck 1999), and to mediate cell migration by activating different downstream targets such as calpain, FAK, paxillin, and MLCK (for review see Huang et al. 2004). Since NCAM is not polysialylated in B35 cells, our study supports, in addition to the recently demonstrated role of polysialylation of NCAM (Franceschini et al. 2004), a regulatory role of NCAM in cell migration independent from its PSA modification.
NCAM140 shedding could use several mechanisms to promote migration. It has been shown that ectodomain shedding of the related cell adhesion molecule L1 releases the soluble L1 ectodomain, which binds to integrins in trans, thereby stimulating migration (Mechtersheimer et al. 2001). The soluble NCAM140 extracellular domain could similarly interact with integrins to promote migration. Alternatively, removal of the NCAM140 extracellular domain by metalloproteolytic cleavage may reduce integrin-dependent adhesion to levels that promote optimal levels of migration. The latter interpretation is in accord with our results showing that inhibition of NCAM shedding by GM6001 treatment increased B35 cell adhesion to fibronectin. Furthermore, the residual cytoplasmic domain fragment of NCAM140 that remains in the cell membrane after cleavage may amplify the effect of NCAM shedding, as expression of the NCAM140 cytoplasmic domain has been shown to compete with full-length NCAM140 for Fyn/FAK signaling (Buttner et al. 2004). Our results may also explain how NCAM140 expression in BT4Cn rat glioma cells reduces migration and adhesion to fibronectin. In these cells, NCAM140 expression down-regulates metalloproteases (Edvardsen et al. 1993), which would decrease NCAM shedding and reduce cell migration.
This work was supported by NIH grants NS26620 and MH064056 (Silvio Conte Center for Neuroscience of Mental Disorders).