CD151, CD81 and (Alpha)3(Beta)1 Integrin Function in Neurite Outgrowth

Home , CD151, CD81, CD82 (gene), Integrin alpha 3

Journal of Cell Science 113, 1871-1882 (2000) 1871 Printed in Great Britain © The Company of Biologists Limited 2000 JCS1409

Transmembrane-4-superfamily proteins CD151 and CD81 associate with α3β1 integrin, and selectively contribute to α3β1-dependent neurite outgrowth

Christopher S. Stipp and Martin E. Hemler* Department of Cancer Immunology and AIDS, Dana-Farber Cancer Institute and Department of Pathology, Harvard Medical School, Boston, MA 02115, USA *Author for correspondence (e-mail: [email protected])

Accepted 22 March; published on WWW 10 May 2000

SUMMARY

Proteins in the transmembrane-4-superfamily (TM4SF) number, length, and rate of extension were all affected by form many different complexes with proteins in the anti-TM4SF antibodies. In summary: (1) these substrate- integrin family, but the functional utility of these complexes dependent inhibition results strongly suggest that CD151 has not yet been demonstrated. Here we show that TM4SF and CD81 associations with α3β1 are functionally relevant, proteins CD151, CD81, and CD63 co-distribute with α3β1 (2) TM4SF proteins CD151 and CD81 make a strong integrin on neurites and growth cones of human NT2N positive contribution toward neurite number, length, and cells. Also, stable CD151-α3β1 and CD81-α3β1 complexes rate of outgrowth, and (3) NT2N cells, a well-established were recovered in NT2N detergent lysates. Total NT2N model of immature central nervous system neurons, can be neurite outgrowth on laminin-5 (a ligand for α3β1 integrin) a powerful system for studies of integrin function in neurite was strongly inhibited by anti-CD151 and -CD81 outgrowth and growth cone motility. antibodies either together (~85% inhibition) or alone (~45% inhibition). Notably, these antibodies had no inhibitory effect on NT2N neurites formed on laminin-1 or Key words: Integrin, Transmembrane-4-superfamily, Neuron, ﬁbronectin, when α3β1 integrin was not engaged. Neurite Neurite, Laminin-5

INTRODUCTION example, α3β1-CD151 association is selectively maintained in 1% Triton X-100; additional interactions of α3β1 and α6β1 TM4SF proteins (also called tetraspanins) are present at with TM4SF proteins are selectively seen in 1% Brij-96 reasonably high levels on nearly every mammalian cell detergent; and many other integrin-TM4SF interactions are and tissue type. There are at least 21 core family members seen only in less stringent detergent conditions (e.g. 1% (Hemler et al., 1996; Maecker et al., 1997; Todd et al., 1998; CHAPS). Wright and Tomlinson, 1994), each containing 4 presumed The role of integrins during cell adhesion and motility is well transmembrane domains, short N- and C-terminal cytoplasmic established (Huttenlocher et al., 1995; Sheetz et al., 1998). domains (containing 5-14 amino acids), a small intracellular TM4SF proteins likewise regulate cell motility, although they loop (typically 4 aa), and two extracellular loops (20-27 aa; 75- generally do not participate in cell adhesion (Hemler, 1998; 130 aa). Maecker et al., 1997). Thus it seems likely that integrin- Like TM4SF proteins, proteins in the integrin family of cell TM4SF protein complexes might play a role in cell motility. adhesion receptors are widespread on many cell types (Hynes Indeed, sometimes in the same experiment, antibodies to both and Lander, 1992). Notably, the α2β1, α3β1, α4β1, α4β7, integrins and TM4SF proteins (CD81, CD151, CD9, CD63) α5β1, α6β1, α6β4, αIIbβ3, and αLβ2 integrins have all been inhibited cell motility (Sincock et al., 1999; Yanez-Mo et al., suggested to associate with various TM4SF proteins including 1998; Yauch et al., 1998). However, in those studies the issue CD9, CD63, CD81, CD82, and/or CD151 (Berditchevski et al., of integrin substrate specificity was not addressed, and it was 1995; Fitter et al., 1999; Hadjiargyrou et al., 1996; Hasegawa not clear the extent to which the integrin-TM4SF proteins et al., 1998; Hemler, 1998; Hemler et al., 1996; Jones et al., complexes needed to be actively engaged on integrin ligand. In 1996; Nakamura et al., 1995; Rubinstein et al., 1994; Schmidt fact, another study has specifically demonstrated that α6Aβ1 et al., 1996; Sincock et al., 1999; Skubitz et al., 1996; Yanez- integrin can induce CD81-dependent motility without Mo et al., 1998; Yauch et al., 1998). However, detection of obviously engaging the integrin substrate (Domanico et al., integrin-TM4SF protein interactions varies considerably 1997). depending on detergent conditions (Berditchevski et al., 1996; To address whether specific α3β1-TM4SF complexes might Serru et al., 1999; Sincock et al., 1999; Yauch et al., 1998). For be functionally relevant, we have chosen to study NT2N cells. 1872 C. S. Stipp and M. E. Hemler

NT2N cells are neuron-like cells derived from the human NT2 integrin cytoplasmic domain (Takada and Hemler, 1989); an embryonic carcinoma cell line (Pleasure et al., 1992). These antiserum raised against intact, denatured β4 integrin (unpublished), cells express many integrins and TM4SF proteins for which and horseradish peroxidase (HRP)-conjugated goat-anti-mouse and antibodies are available, and thus had a strong potential to goat anti-rabbit antibodies (Sigma). provide insight into the roles of TM4SF proteins in neurite Cell culture outgrowth and growth cone motility. In addition, NT2N α β NT2 precursor cells (Stratagene) were maintained in OPTI-MEM neurites grew on many different substrates, including the 3 1 (Gibco-BRL) with 10% fetal bovine serum, penicillin-streptomycin, ligand laminin-5, thus allowing an evaluation of the substrate- and 2 mM glutamine. NT2N neuron-like cells were obtained as dependent inhibitory effects of anti-TM4SF antibodies. described by Pleasure et al. (1992). Briefly, NT2 precursor cells were The roles of TM4SF proteins themselves, or of integrin- treated with 10 µm retinoic acid for 4-5 weeks then split 1:6 into TM4SF complexes during neurite outgrowth have not been medium with mitotic inhibitors and cultured an additional 10-14 days. studied, except for a few reports involving the CD9 protein. In NT2N cells, growing as clusters over a monolayer of non-neuronal one case, CD9 associated with integrin α6β1 under conditions cells, were isolated by rinsing with phosphate-buffered saline (PBS) of relatively low detergent stringency, and anti-CD9 antibody and treating for 2-3 minutes with trypsin/EDTA diluted 30-fold in stimulated neurite outgrowth of small cerebellar neurons on PBS. Purified neurons were maintained on plates coated with Matrigel (Becton Dickinson) diluted 1:50. laminin-1, but not poly-L-lysine (Schmidt et al., 1996). α β Since the exact number of clustered NT2N cells in a culture is However, the functional role of 6 1 was not elucidated. In difficult to determine, NT2N cell purity was estimated as follows: another study, immobilized anti-CD9 antibody supported morphologically non-neuronal cells in 20 random fields were counted, neurite outgrowth from sympathetic neurons, which was partly and the amount of detergent extractable protein contributed by these inhibited by anti-α3β1 mAb (Banerjee et al., 1997). However, cells was calculated using a value of protein/cell obtained from lysates because anti-α3 mAb also inhibited neurite outgrowth on of known numbers of NT2 precursor cells. The protein derived from immobilized anti-Thy1 mAb and on collagen (Banerjee et al., non-neuronal cells was then expressed as a fraction of the total protein 1997), the specific functional involvement of α3β1-CD9 obtained from the NT2N cell lysates, and found to be less than 1% of complexes was not clear. Thus far, antibodies to TM4SF the total. For neurite outgrowth and staining experiments, NT2N cells were proteins had not been shown to inhibit neurite outgrowth. rinsed with PBS and treated with trypsin/EDTA diluted 10-fold in Here we show for the first time strong inhibition of neurite PBS. Cells were dislodged, and trypsin treatment was stopped by outgrowth by antibodies to TM4SF proteins (in this case, adding DME with 5 mg/ml crystalline bovine serum albumin (BSA) CD151 and CD81). Furthermore, we show inhibition by anti- (ICN) and 0.2 mg/ml soybean trypsin inhibitor (Sigma). Cells were TM4SF antibodies that depends on functional engagement of collected by centrifugation, resuspended in a serum-free medium TM4SF-associated α3β1 integrin with its laminin-5 ligand. consisting of DMEM with B27 additives (Brewer et al., 1993), (Gibco Such integrin-substrate dependence had not been previously BRL) and 1 mM glutamine, and plated to substrate-coated coverslips. demonstrated in studies of integrin-TM4SF complexes on Preparation of coverslips; video-microscopy neurons or other cell types. Together these results strongly Acid-washed glass coverslips were coated with 20 µg/ml mouse suggest that integrin-TM4SF complexes can act as a functional laminin-1 (Gibco BRL), diluted in PBS, with 2 µg/ml rat laminin-5 unit, dependent on integrin engagement with ligand, and of (a gift of Desmos Inc.) diluted in PBS with 0.005% Tween-20, or with critical importance during neurite outgrowth. During the 20 µg/ml fibronectin (Becton-Dickinson). Substrates were coated course of these studies we also show that NT2N cell either for 1 hour at 37°C or overnight at 4°C. Coverslips were rinsed differentiation is accompanied by pronounced changes in 3 times in PBS and blocked with 10 mg/ml heat-inactivated BSA for integrin and TM4SF protein expression, and that NT2/NT2N 1 hour at 37°C. Coverslips were rinsed three times and used for cells are an excellent system for studying human neurite experiments described below. differentiation and outgrowth. Video images were acquired using a Zeiss Axiovert 135 microscope equipped with (1) a VS25 electronic shutter driven by a Uniblitz D122 controller (Vincent Associates) and (2) a black and white video camera (TM-7AS, PULNiX America Inc.) connected via a focusing MATERIALS AND METHODS monitor (PVM-137, Sony Corp.) to a Power Macintosh 6500 computer containing a VG-5 frame grabber (Scion Corp.). Image Antibodies acquisition was performed with the program Scion Image 1.60 (Scion Anti-integrin mAbs used were anti-α1, TS2/7 (Hemler et al., 1984), Corp.). For time-lapse experiments, NT2N cells were plated in serum- and FB12 (Chemicon); anti-α2, A2-IIE10 (Bergelson et al., 1994); free medium on coverslips spanning a 12 mm hole drilled in the anti-α3, A3-IIF5 and A3-X8 (Weitzman et al., 1993); anti-α4, B5G10 bottom of a 60 mm Petri dish. A custom-built stage incubator provided (Hemler et al., 1987) and A4-PUJ1 (Pujades et al., 1996); anti-α5, a humidified 10% CO2, 37°C atmosphere. A macro written for Scion A5-PUJ2 (Pujades et al., 1996); anti-α6, A6-ELE (Lee et al., 1995), Image drove shutter control and image acquisition (at 20-30 and GoH3 (Sonnenberg et al., 1988); and anti-β1, TS2/16 (Hemler et frames/hour). al., 1984). Anti-TM4SF antibodies used were anti-CD63, 6H1 (Berditchevski et al., 1995); anti-CD81, M38 (Fukudome et al., 1992); Flow cytometry and immunofluorescence staining and anti-CD151, 5C11 (Yauch et al., 1998), and 11B1.G4 (Sincock et NT2 cells were stained with negative control mAb and specific mAbs al., 1997). Other mAbs used were anti-NCAM, NCAM-OB11 as previously described (Berditchevski et al., 1995), and then analyzed (Sigma); the P3 negative control antibody (Lemke et al., 1978), and using a FACScan flow cytometer (Becton-Dickinson). For 187.1, a rat anti-mouse κ chain antibody (Yelton et al., 1981). immunofluorescence staining, cells were plated in serum-free medium Polyclonal rabbit antisera utilized were 1407 and 0530, raised against on laminin-5-coated coverslips. Cultures were fixed with 4% the cytoplasmic domains of α6A and α6B integrin subunits (the gift paraformaldehyde, 4% sucrose, 1 mM MgCl2 in PBS, rinsed, and of Dr V. Quaranta); an antiserum against the integrin α3A cytoplasmic blocked with 10% goat serum in PBS for 1 hour at room temperature. domain (DiPersio et al., 1995); an antiserum raised against the α2 NT2N cells were stained with 5 µg/ml primary antibodies for 1 hour, CD151, CD81 and α3β1 integrin function in neurite outgrowth 1873 rinsed 3 times, and incubated 1 hour with a CY-3-conjugated goat anti-mouse secondary antibody (Zymed), diluted 1:50. Coverslips were mounted in Antifade reagent (Molecular Probes) and examined on a Zeiss Axiophot microscope. Images were recorded with TMAX 3200 or Kodak Royal Gold ASA 1000 film, or with a cooled CCD camera (DC330E, DAGE-MTI). Photographs were digitized using a ScanJet 4C digital scanner (Hewlet Packard). Color images were converted to black and white using the program Adobe Photoshop 4.0.

Immunoprecipitation and immunoblotting NT2 cells and NT2N cells were biotinylated, or metabolically labeled overnight with [35S]methionine as previously described (Yauch et al., 1998). Cells were lysed for 1 hour in 50 mM Tris-HCl, pH 8.0, 5 mM MgCl2, 150 mM NaCl, with detergent, 1 mM phenylmethylsulfonyl ﬂuoride, 20 µg/ml aprotinin, and 10 µg/ml leupeptin. The detergents used were 1% Brij 96 (Fluka), or 1% Triton X-100 (Sigma), alone or in combination with 0.2% sodium dodecyl sulfate (Gibco BRL). Insoluble material was removed by centrifugation, and protein concentrations were determined by amido black dye binding (Schaffner and Weissman, 1973). Lysates were immunoprecipitated as previously described (Yauch et al., 1998) with anti-integrin and anti-TM4SF protein mAbs followed by Protein A-Sepharose pre- bound with the 187.1 rat anti-mouse mAb or by Protein G-Sepharose. Metabolically labeled proteins were detected with a phosphor storage screen (Molecular Dynamics), and quantiﬁed using Image Quant software. For immunoblotting, lysates from unlabeled cells were immunoprecipitated with Sepharose-coupled TS2/16 anti-β1 integrin or 5C11 anti-CD151, resolved on 11% acrylamide SDS-PAGE gels, and transferred to nitrocellulose. Blots were probed with diluted polyclonal rabbit anti-integrin or with the 11B1.G4 anti-CD151 antibody (diluted 1:500). Blots were developed with HRP-conjugated goat-anti-rabbit or goat-anti-mouse antisera (diluted 1:30,000) followed by chemiluminescence.

Neurite outgrowth assays NT2N cells were harvested as described above and resuspended in serum-free medium. Then 0.4 ml aliquots (with antibodies) were plated to substrate-coated, 12 mm diameter coverslips in the wells of a 24 well plate. Two to three coverslips were plated per condition, and Fig. 1. NT2N cells were plated in serum-free medium on a laminin- ~1×104 cells and clusters of cells were plated per well. In most 5-coated coverslip and individual growth cones were observed by experiments, cultures were fixed (20% formalin, 4% sucrose, 1 mM time-lapse video-microscopy. Migration rates ranged from 25 to 60 µm/hour. Note the large, well-spread lamellipodia and highly motile MgCl2 in PBS) after 4-8 hours of outgrowth at 37°C, 6% CO2. Coverslips were rinsed with PBS and mounted on glass slides. Images filopodia. were acquired by video-microscopy. At least 3 fields were analyzed per coverslip, including both edges and center fields. Neurite lengths were measured with Scion Image 1.60 (Scion Corp.), after calibration identification of serum-free conditions that support neurite of pixels/µm using hemocytometer squares. To measure total outgrowth. Since NT2N cells resemble immature central outgrowth, the total length of all the neuritic material in each field was nervous system (CNS) neurons (Pleasure et al., 1992), we divided by the total perimeter of all the clusters of neurons in the field. tested serum-free conditions that support primary, immature CNS neurons. As shown by time-lapse microscopy in Fig. 1, NT2N cells extended neurites, and possessed large, highly RESULTS motile growth cones in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with B27 additives. One culture of NT2N cells for study of neurite outgrowth and NT2N cells was maintained in this medium for 4 weeks with growth cone motility no loss of viability (not shown). Thus, the minimal media The differentiation of NT2 precursor cells into neuronal requirements of NT2N cells resemble those of embryonic NT2N cells has been utilized extensively in studies of hippocampal neurons (Brewer et al., 1993). neurotransmitter receptor physiology, gene regulation, the processing and secretion of the amyloid beta precursor protein, Dramatic changes in integrin expression accompany and in transplantation experiments (Abraham et al., 1991; NT2 cell differentiation Bani-Yaghoub et al., 1997; Borlongan et al., 1998; Chyung To enable studies of integrin function in NT2N cells, we et al., 1997; Kleppner et al., 1995; Neelands et al., 1998; defined the integrin expression profiles of NT2 precursor cells Skovronsky et al., 1998). However, few NT2N studies have and neuronal NT2N cells by cell surface labeling and specifically focused on regulation of neurite outgrowth and immunoprecipitation (Fig. 2A). Levels of α1β1 and α2β1 growth cone motility. A prerequisite for such studies is the integrins were comparable in both cell types (lanes 1-4), while 1874 C. S. Stipp and M. E. Hemler

Fig. 2. Integrin expression is dynamically regulated during NT2 cell differentiation. (A) Integrins were immunoprecipitated from NT2 precursor cells (P lanes) and NT2N cells (N lanes) that were cell surface labeled with biotin and lysed in 1% Triton X-100, 0.2% SDS. 100 µg total protein were loaded in each lane. Antibodies were anti-α1, TS2/7 (1 and 2); anti-α2, A2-IIE10 (3 and 4); anti-α3, A3-IIF5 (5 and 6); anti- α4, A4-PUJ1 (7 and 8); anti-α5, A5-PUJ2 (9 and 10); anti-α6, GoH3 (11 and 12); and anti-β1, TS2/16 (13 and 14). Left arrows indicate integrin α subunits, and the right arrow indicates the integrin β1 subunit. (B) Biotin-labeled lysates from NT2 precursors (lane 1) and NT2N cells (lane 2) were immunoprecipitated with an anti-β4 integrin polyclonal antiserum. The upper and lower bands in lane 2 represent intact β4 and α6 subunits, respectively. the α4 subunit was absent from both neuronal cells and neurites and growth cones were completely unstained precursors (lanes 7 and 8). The α3 integrin subunit was the (Fig. 3H). only subunit more highly expressed on neuronal cells than on Although CD9 and CD82 were both abundant on NT2 the precursors (lanes 5 and 6). Integrin α5, abundant on NT2 precursors as determined by flow cytometry, they were cells, was virtually absent from NT2N neuronal cells (lanes 9 minimally expressed by the NT2N cells (not shown). In and 10). Integrin α6 was also highly abundant on precursor contrast, CD81 was absent on the precursors (Table 1) but was cells and downregulated on NT2N cells (lanes 11 and 12). clearly present on the neuronal cell surface (Fig. 3B). Thus The larger protein appearing only in the NT2N α6 TM4SF protein expression profiles, like those of the integrins, immunoprecipitates (Fig. 2A, lane 12) was identified as the shift dramatically during the differentiation of NT2 precursors integrin β4 subunit (Fig. 2B). Thus the α6 subunit appears to into NT2N cells. switch from α6β1 to α6β4 heterodimers during differentiation. A similar overall pattern of integrin expression was observed Stable association of CD151, CD81 and α3 integrin in two experiments, performed months apart. Cell surface in NT2N cells labeling and flow cytometry results for NT2 cells were in Next we tested whether α3 integrin, CD151, and CD81 form general agreement: α2, α5, and α6 were the most highly expressed subunits, while α4 was absent (Table 1). NT2N Table 1. Integrin and TM4SF protein expression by NT2 cultures were not examined by flow cytometry because the cells as determined by flow cytometry majority of cells were heterogeneously clustered and difficult to dissociate without compromising cellular integrity. Mean fluorescence Antibody Specificity intensity* Transmembrane-4-superfamily (TM4SF) proteins and Integrin alpha subunits integrins on neurites and growth cones of NT2N TS2/7 Anti-α1 integrin 6.7 α cells A2-IIE10 Anti- 2 integrin 16 A3-X8 Anti-α3 integrin 8.0 TM4SF proteins and integrins α2β1 and α3β1 were localized B5G10 Anti-α4 integrin 0.0 by immunocytochemistry on non-permeabilized NT2N cells A5-PUJ2 Anti-α5 integrin 17 (Fig. 3). The TM4SF proteins CD63, CD81, and CD151, as A6-ELE Anti-α6 integrin 26 well as the integrins, were distributed in a punctate pattern over TM4SF proteins the entire neuronal cell surface, including the neurite shaft and C9BB CD9 16 HD28 CD37 0.0 the growth cone (Fig. 3A-E). In contrast, NCAM staining was HD77 CD53 0.8 smoother and very bright on the cell bodies and neurites of 6H1 CD63 52 NT2N cells, but negative or nearly negative on most growth M38 CD81 1.3 cones (Fig. 3F). Within individual growth cones, TM4SF M104 CD82 43 proteins were often found at the leading edge, especially at the 5C11 CD151 20 tips of filopodia, as shown for CD151 in Fig. 3G. The negative *Background fluorescence of ~3 fluorescence intensity units, determined control P3 monoclonal antibody produced low levels of with the P3 negative control antibody, was subtracted from the values background staining on clusters of neuronal cell bodies, but reported here. CD151, CD81 and α3β1 integrin function in neurite outgrowth 1875 stable complexes in NT2N cells as seen in other cells stringency detergent conditions yielded CD81 (lane 2) and (Berditchevski et al., 1996; Yauch et al., 1998). α3β1 integrin (lane 3) in the absence of associated proteins. Immunoprecipitation of CD151 from metabolically labeled CD151, which is only weakly labeled by biotin (Berditchevski Triton X-100 lysates yielded a protein (Fig. 4A, lane 2) et al., 1995), was not observed in the α3β1 immunoprecipitate precisely comigrating with α3 integrin (lane 3, closed arrow). in this experiment. Besides the α3β1 integrin, an additional The anti-CD151 antibody co-precipitated 60% of the α3 protein of ~75 kDa was also co-immunoprecipitated with CD81 obtained using an anti-α3 antibody (average of two separate (Fig. 4, lane 1), but that protein has not yet been identified. experiments). In the converse experiment, it was difficult to Together the experiments in Fig. 4 indicate that α3β1 can form detect CD151 (open arrow) in the α3 immunoprecipitate, likely stable and specific complexes with CD151 (under very stringent due to slow metabolic turnover of the fraction of CD151 associated with α3 integrin (R. Yauch, C. Stipp and M. E. Hemler, unpublished observations). To confirm that α3 integrin was indeed co-precipitating with CD151, CD151 immunoprecipitates from NT2N cells were analyzed by immunoblotting. As shown in Fig. 4B, immunoprecipitation of comparable amounts of CD151 (lanes 3, 4, lower panels) yielded co-immunoprecipitation of α3 (lane 4), but not α2 (lane 3). A control experiment showed that comparable levels of α2 and α3 subunit could be detected upon immunoblotting of a β1 integrin immunoprecipitate (lanes 1 and 2). As a further test of α3β1- CD151 interaction specificity, we analyzed Triton X-100 lysates of NT2 precursor cells. Immunoblotting of a β1 integrin immunoprecipitate showed that α2 (Fig. 4C, lane 1) and α6B (lane 4) integrin subunits were both more abundant than the α3 subunit (lane 2). Nonetheless, only the α3 subunit was detected in the CD151 immunoprecipitate (Fig. 4C, lanes 5-8). In previous experiments, CD81- α3β1 and CD81-α6β1 complexes were stable in 1% Brij 96 lysates, whereas associations with other integrins such as α2β1, α5β1 Fig. 3. TM4SF proteins and integrins on growth and α6β4 were not detected cones of NT2N cells. NT2N cells in serum-free (Berditchevski et al., 1996). Thus, medium on laminin-5-coated coverslips were to test for CD81-α3β1 complexes fixed and stained without permeabilization. in NT2N cells, they were cell- Primary antibodies were anti-CD151, 5C11 (A surface biotinylated, lysed in and G); anti-CD81, M38 (B); anti-CD63, 6H1 α α 1% Brij 96, and CD81 was (C); anti- 3 integrin, A3-IIF5 (D); anti- 2 immunoprecipitated. As seen (Fig. integrin, A2-IIE10 (E); anti-NCAM, NCAM- OB11 (F); and a negative control antibody, P3 4D, lane 1), biotinylated CD81 (H). Integrins and TM4SF proteins were found was obtained (open arrow) along over the entire neuronal cell surface, including with a co-immunoprecipitated the growth cone. In panel G, CD151 staining protein that comigrated with (G1) and phase contrast (G2) are overlaid to α3β1 integrin (closed arrow). show how individual puncta were often seen to localize to the tips of filopodia (G3). Arrows Immunoprecipitation under high highlight the tips of filopodia that were CD151-positive. 1876 C. S. Stipp and M. E. Hemler conditions) and with CD81 (under moderately stringent statistically significant for the number of coverslips analyzed conditions), whereas associations with other integrins (e.g. (Fig. 7A). Thus, although outgrowth on laminin-1 was α2β1) were not seen. completely β1-integrin-dependent (Fig. 5F) the alpha integrin TM4SF proteins specifically regulate α3 integrin-dependent neurite outgrowth To explore the functional relevance of the CD151- α3 integrin association, we tested anti-α3 and anti-CD151 function-blocking antibody effects on neurite outgrowth on two different isoforms of laminin, laminin-1 and laminin-5. Both isoforms are expressed transiently in the developing central nervous system (Galliano et al., 1995; Hunter et al., 1992), and both have been described as ligands for α3β1 integrin (Carter et al., 1991; Elices et al., 1991; Smith et al., 1996; Tomaselli et al., 1993), although laminin-5 is a much stronger ligand (Eble et al., 1998). Function blocking anti-α3 antibody, A3-IIF5 (Weitzman et al., 1993) was a potent inhibitor of NT2N neurite outgrowth on laminin-5 (Fig. 5A,C). Quantitation revealed a ~90% decrease in total neurite outgrowth (Fig. 7A). Inhibition by A3-IIF5 was nearly as complete as the inhibition caused by mAb 13, an anti-β1 integrin antibody (Fig. 5E). Thus most of the outgrowth on laminin- 5 is mediated by α3β1 integrin. In sharp contrast, A3-IIF5 had no effect on outgrowth on laminin-1 (Figs 5B,D, 7A), although mAb 13 completely abolished outgrowth (Fig. 5F). Thus NT2N cells utilize a β1 integrin other than α3β1 as a major receptor for laminin-1. To test for CD151 function, we used 5C11, an anti-CD151 antibody that inhibits neutrophil migration (Yauch et al., 1998). mAb 5C11 also inhibited outgrowth on laminin-5, resulting in an obvious decrease in neurite number as well as a more modest decrease in neurite length (Fig. 6A,C). Quantitation of several coverslips revealed ~50% Fig. 4. TM4SF-α3β1 complexes in NT2N cells and NT2 precursor cells. reduction in total neurite outgrowth (Fig. 7A). In (A) Metabolically labeled NT2N cells were lysed in 1% Triton X-100 and contrast, mAb 5C11 failed to inhibit outgrowth on immunoprecipitated with the P3 negative control antibody (lane 1); anti-CD151, laminin-1 (Figs 6B,D, 7A). In addition, neither anti- 5C11 (lane 2); or anti-α3 integrin, A3-IIF5 (lane 3). The closed arrow indicates the α3 nor anti-CD151 antibodies had any inhibitory α3 integrin subunit; the open arrow indicates the CD151 protein band. The absence effects on fibronectin (Fig. 7A). of a β1 integrin band in lanes 2 or 3 is the result of the slow turnover of β1 integrin (B) Unlabeled NT2N cells were extracted with 1% Triton X-100 and An anti-CD81 antibody, M38, also inhibited β NT2N neurite outgrowth specifically on laminin-5 immunoprecipitated with Sepharose-coupled antibodies to 1 integrin (TS2/16; lanes 1 and 2), or CD151 (5C11; lanes 3 and 4). Immunoprecipitates were to an extent similar to 5C11 (Fig. 7A). Used in separated by SDS-PAGE and blotted with polyclonal antisera to α2 integrin (lanes combination, 5C11 and M38 inhibited total 1 and 3) or α3 integrin (lanes 2 and 4). Lower panels in lanes 3 and 4 show the outgrowth on laminin-5 by nearly 85% (Figs 6A,E, CD151 protein band from each immunoprecipitate, detected with the 11B1.G4 7A), but outgrowth on laminin-1 was unaffected anti-CD151 antibody. (C) Unlabeled NT2 precursor cells were extracted and (Figs 6B,F, 7A). Thus, for the substrates tested, immunoprecipitated as in (B). Lanes 1-4 show β1 integrin immunoprecipitates and inhibition of NT2N neurite outgrowth by anti- lanes 5-8 show CD151 immunoprecipitates. Immune complexes were blotted with TM4SF antibodies coincides with inhibition by an polyclonal antisera to α2 integrin (lanes 1 and 5); α3 integrin (lanes 2 and 6); α6A anti-α3 integrin antibody. In control experiments, integrin (lanes 3 and 7); or α6B integrin (lanes 4 and 8). Lower panels in lanes 5-8 antibodies recognizing either the CD63 TM4SF show the CD151 bands from each immunoprecipitate, detected as in (B). protein, or α6 integrin (mAb GoH3) did not inhibit (D) NT2N cells were surface-labeled with biotin and lysed and immunoprecipitated with the anti-CD81 antibody, M38 or with the A3IIF5 anti-α3 mAb. outgrowth on laminin-5 (Fig. 7A), despite abundant α β Immunoprecipitates were separated on an 11% SDS-PAGE gel and blotted with expression of CD63 and 6 4 integrin on NT2N HRP-conjugated streptavidin. Lane 1: CD81 immunoprecipitation in 1% Brij 96; cells (Fig. 2). The GoH3 mAb and an anti-α1 Lanes 2 and 3: CD81 and α3 integrin immunoprecipitations, respectively, in 1% integrin antibody may have had a modest inhibitory TX-100, 0.2% SDS. Closed arrow indicates α3β1 integrin bands; open arrow effect on laminin-1, although the difference was not indicates CD81 band. CD151, CD81 and α3β1 integrin function in neurite outgrowth 1877 subunit or subunits required for outgrowth on laminin-1 were not outgrowth. Here, antibodies to TM4SF proteins CD81 and identified in this study. We were unable to test the role of TM4SF CD151, either alone (~45%) or together (~85% inhibition), proteins during α2β1-mediated outgrowth since NT2N cells did strongly inhibited NT2N cell neurite outgrowth. In the only not extend neurites when plated on collagen, the major ligand previous studies of TM4SF proteins and neurite outgrowth, mAb for α2β1 integrin (data not shown). to CD9 promoted rather than inhibited neurite formation (Banerjee et al., 1997; Schmidt et al., 1996). We did not study A role for CD151 and CD81 in growth cone motility CD9 here because it was downregulated on NT2N cells. Anti- on laminin-5 CD81 and anti-CD151 mAb’s inhibited not only neurite The inhibition experiments above do not distinguish between differentiation and initiation, but also elongation, thus suggesting neurite initiation and elongation. Indeed, the decreased number that multiple inhibitory mechanisms could be involved. With of neurites upon treatment with anti-TM4SF antibodies (Fig. 6E), respect to differentiation, TM4SF antibodies elsewhere raised the possibility that these antibodies primarily inhibit perturbed C2C12 cell differentiation into myotubes (Tachibana neurite initiation. To test for effects on growth cone motility, and Hemler, 1999), and drosophila motoneurons lacking the NT2N cells were allowed to attach and extend neurites on TM4SF protein, ‘latebloomer,’, were deficient in differentiation laminin-5 for 2 hours in the presence or absence of anti-CD151 at the level of synapse formation (Kopczynski et al., 1996). With (5C11) and anti-CD81 (M38) (10 µg/ml each). The migration regard to elongation/motility, TM4SF proteins influence the rates of individual growth cones were then measured by time- motility of several cell types (Anton et al., 1995; Ikeyama et al., lapse video-microscopy. As previously observed, cultures treated 1993; Lagaudriere-Gesbert et al., 1997; Miyake et al., 1991; with 5C11 and M38 generated fewer growth cones (not shown). Radford et al., 1997; Yanez-Mo et al., 1998; Yauch et al., 1998). The growth cones that did form migrated significantly slower: an Consistent with a role in motility, TM4SF were located at the average of 35.5±2.5 µm/hour (n=49) in the absence of antibodies tips of NT2N filopodia in this study, and in cell lamellipodia and was reduced to 17.8±2.3 µm/hour (n=38; Fig. 7B). These results filopodia elsewhere (Berditchevski et al., 1997; Hemler et al., indicate that 5C11 and M38 inhibit not only neurite initiation but 1996; Indig et al., 1997; Maecker et al., 1997). also growth cone motility on laminin-5. Consistent with inhibition of elongation, anti-CD151 and anti-CD81 antibodies reduced mean neurite lengths by ~20% (P<0.001; n=10 coverslips per condition) when used individually and by ~33% (P<0.001; n=7 coverslips) when used in combination. The anti-α3 antibody reduced mean length by ~60% (P<0.001; data not shown). Thus, anti-TM4SF inhibition of total neurite outgrowth results from a combination of diminished initiation and elongation. In several additional experiments, we found no evidence for mAb 5C11 having an effect on (1) the amount of α3 integrin on cell surface-labeled NT2N cells or on HT1080 cells as seen by flow cytometry; (2) the detergent extractability of α3 integrin on HT1080 cells (Triton X-100 or Brij 99); (3) the profile of tyrosine- phosphorylated proteins in NT2N cells; (4) the pattern of phosphorylated proteins that co-precipitate with α3β1 integrin from 32 P04-labeled NT2N cells; (5) the amount of PI-4 kinase activity that co-precipitates with α3β1 integrin in NT2N cells or HT1080 cells; or (6) the fractionation of α3β1 in sucrose density gradients prepared from lysates of HT1080 cells or NT2N cells (data not shown).

DISCUSSION α β The role of TM4SF proteins in Fig. 5. Laminin-5 supports 3 1-dependent NT2N neurite outgrowth. NT2N cells were plated in serum-free medium on laminin-5 coated coverslips (A,C,E), or on laminin-1 neurite outgrowth coated coverslips (B,D,F). Cells were plated in the absence of antibody (A and B); in the Antibodies to TM4SF proteins were not presence of 10 µg/ml anti-α3 integrin, A3-IIF5 (C and D); or in the presence of 10 µg/ml previously shown to inhibit neurite anti-β1 integrin, mAb 13 (E and F). 1878 C. S. Stipp and M. E. Hemler

Inhibitory effects of anti-CD151 and -CD81 antibodies on complexes have been observed, but so far only α3β1-CD151 neurites may not involve inhibition of integrin adhesion and/or complexes are maintained under highly stringent detergent ligand binding functions. First, our anti-CD151 and CD81 conditions (Yauch et al., 1998). Also, the α3β1 integrin is one antibodies typically fail to inhibit cell adhesion (Mannion et of few integrins that associates with TM4SF proteins (such as al., 1996; Yauch et al., 1998; and unpublished results). Second, CD81) in relatively stringent 1% Brij 96 conditions. Under the region of α3 integrin that interacts with CD151 is quite similar conditions, TM4SF proteins did not associate with distinct from the ligand binding site (Yauch et al., 2000). α5β1, α2β1, α4β1, αV integrins or β2 integrins (Berditchevski Finally, when neurite outgrowth was partially inhibited by a et al., 1996; Hemler, 1998; Mannion et al., 1996; Nakamura et saturating level of anti-CD151 mAb, an anti-α3 function al., 1995). Our current results now suggest that the most robust blocking antibody caused further inhibition, indicating that the integrin-TM4SF associations (such as α3β1-CD151, α3β1- two antibodies are likely acting by different mechanisms (data CD81) may have the most obvious functional relevance. In this not shown). regard, antibodies to CD81 and CD151 had no effect on neurite outgrowth on laminin-1 or fibronectin, which might depend on CD81-α3β1 and CD151-α3β1 complexes as α6β4, α1β1, and other undetermined β1 integrins. functional units Do α3β1-TM4SF complexes function during neurite TM4SF-α3β1 complexes on NT2N cells resemble those seen outgrowth in vivo? The CD151 protein is present in on other cell types (Berditchevski et al., 1996; Sincock et al., immortalized neuronal precursors (Fitter et al., 1998) and 1999; Yanez-Mo et al., 1998; Yauch et al., 1998). However our CD151 mRNA appears in the developing brain (S. Fitter, S. results now show that anti-TM4SF mAb functional effects Kumar, and L. K. Ashman, personal communication). Also, strongly depend on engagement of TM4SF-associated integrin CD81 is highly expressed by glial cells in the developing rat with ligand. Anti-CD151 and anti-CD81 inhibition was not brain (Sullivan and Geisert, 1998), but its role in developing seen unless neurite outgrowth was almost completely α3β1 neurons has not yet been analyzed. The α3 integrin subunit is dependent, as it occurred on laminin-5 substrate. prominent in the developing Xenopus forebrain (Whittaker and Previous studies hinted at functions for integrin-TM4SF DeSimone, 1993); and supports motility of immature cortical complexes, without addressing the issue of substrate specificity. For example, an anti-α3 mAb and several anti-TM4SF antibodies inhibited chemotactic neutrophil migration on fibronectin (Yauch et al., 1998), while antibodies to β1 and α3 integrins and to TM4SF members (CD81 and CD151) caused reduced endothelial cell migration (Sincock et al., 1999; Yanez- Mo et al., 1998). Elsewhere, neurite outgrowth promoted by an immobilized anti-CD9 antibody was inhibited by an α3 function blocking antibody (Banerjee et al., 1997). However, anti-α3 antibody also inhibited outgrowth on collagen and on immobilized anti-Thy 1 mAb (Banerjee et al., 1997), thus clouding the issue of substrate specificity. How do anti-TM4SF antibodies inhibit in an integrin substrate-dependent manner? Our TM4SF antibodies themselves do not disrupt integrin-TM4SF interactions since we routinely used these antibodies in co- precipitation experiments. Also, our anti- TM4SF antibodies failed to alter α3β1 complexes with respect to expression levels, detergent extractability, protein tyrosine phosphorylation, association of 32 PO4-labeled proteins, association of PI-4 kinase activity, or distribution in sucrose density gradients. We suspect that anti- TM4SF antibodies may perturb critical signaling functions localized to the site of Fig. 6. CD151 and CD81 participate in NT2N neurite outgrowth specifically on laminin-5. integrin engagement with ligand. Such a NT2N cells were plated in serum-free medium on laminin-5-coated coverslips (A,C,E) or localized effect would perhaps not be laminin-1-coated coverslips (B,D,F), in the absence of antibody (A and B), in the presence detectable in whole cell lysates. of 20 µg/ml anti-CD151 antibody, 5C11 (C and D), or in the presence of both 5C11 and the Many different integrin-TM4SF M38 anti-CD81 antibody, 20 µg/ml each (E and F). CD151, CD81 and α3β1 integrin function in neurite outgrowth 1879

Fig. 7. Effects of anti-integrin and anti- A laminin-5 laminin-1 fibronectin TM4SF antibodies on NT2N neurite 140 outgrowth and growth cone motility. (A) Total neurite outgrowth (see Materials 120 and Methods) in the presence of 100 antibodies to the indicated integrins and TM4SF proteins was measured in several 80 independent experiments. Values are 60 ****** expressed as a percentage of the total 40 outgrowth on control coverslips with no antibodies added. Outgrowth was total outgrowth 20 ***

(% positive control) *** quantiﬁed on laminin-5 (black bars), 0 α α CD81 (8)CD151α α α2 α α no Ab (23)CD151 CD81(6) (6)CD151 + CD81CD63 (5) (3) no Ab CD151 3 (5) 6 (3) no Ab (2) CD151 (2) laminin-1 (horizontal striped bars) and 3 (3) 6 (3) (3) 1 (3) 3 (2)

fibronectin (hatched bars). In parentheses (27) (5) + CD81 (4) are the number of coverslips quantified for each condition. Error bars show the standard error of the mean. Outgrowth was for 4-8 hours, except for fibronectin, which was for 18 hours. For laminin-1 and laminin-5, experimental samples were B compared to the control sample by 200 200 no antibody anti-CD151 + anti-CD81 ANOVA. Samples significantly different 35.5 ± 2.5 µm/hour 17.8 ± 2.3 µm/ hour from controls are indicated by asterisks: ***P<0.001 (Bonferroni t-test). 150 150 (B) NT2N cells were plated in serum-free medium on laminin-5 in the presence or absence of anti-CD151 and anti-CD81 antibodies, 10 µg/ml each. After 2 hours, 100 100 cultures were transferred to a microscope stage incubator and the migration of several individual growth cones was micrometers monitored for 2 hours at 20 to 30 50 50 frames/hour. The migration of individual growth cones versus time is plotted. Negative slopes correspond to periods of 0 0 neurite retraction. Results of 3 separate experiments were pooled. In all 3 0 20 40 60 80 100 120 0 20 40 60 80 100 120 experiments, the average migration rate of minutes minutes antibody-treated growth cones was significantly lower than that of untreated control growth cones. The mean migration rate of 35.5±2.5 µm/hour in the absence of antibodies (n=49) was reduced to 17.8±2.3 µm/hour (n=38) in the presence of 5C11 and M38. Antibodies used in (A) and (B) were A3-IIF5 (anti-α3 integrin); 5C11 (anti-CD151); M38 (anti-CD81); GoH3 (anti-α6 integrin); 6H1 (anti-CD63); A2-IIE10 (anti-α2 integrin); and FB12 (anti-α1 integrin). neurons (Anton et al., 1999), and embryonic retinal neurons morphology and migration rate, NT2N growth cones resemble (Ivins et al., 1998). Brain neuronal α3 integrin is likely growth cones of chick dorsal root ganglion (DRG) cells (see, accompanied by CD151, since we have not yet observed α3 for example, Letourneau, 1992). The large size of NT2N integrin to be present on any tissue or cell line in the absence growth cones facilitates immunolocalization (c.f. Fig. 3), as of CD151 (Yauch et al., 1998, and unpublished results). well as visualization of lamellipodia and filopodia during time- lapse motility studies. Utility of NT2N cells As NT2 precursors differentiated to NT2N cells, the α5 and Human NT2N cells express markers characteristic of immature α6 integrin subunits were dramatically downregulated, while central nervous system neurons (Pleasure et al., 1992), develop the α3 subunit was markedly upregulated. This is fully a polarized morphology with immunologically distinct axons consistent with an important role for α3β1 integrin in neuronal and dendrites (Kleppner et al., 1995; Pleasure et al., 1992), and differentiation (Anton et al., 1999; Ivins et al., 1998; Whittaker they integrate and acquire characteristics of fully mature and DeSimone, 1993). While it is unclear whether NT2N cells neurons, when grafted into the central nervous system correspond more closely to any particular class of CNS neuron (Kleppner et al., 1995; Trojanowski et al., 1997). Here we in vivo, each of the other integrin α subunits expressed by found that NT2N cells readily extended neurites and possessed NT2N cells (α1, α2, α6) appears in the developing CNS large, well-formed growth cones in serum-free medium. (Bradshaw et al., 1995; Bronner-Fraser et al., 1992; Cann et Outgrowth was rapid, with a substantial number of neurites al., 1996; de Curtis, 1993; Duband et al., 1992). The absence seen by 4-8 hours. Simple media requirements and rapid of the α5 subunit from NT2N is consistent with its absence outgrowth exclude exogenous attachment factors and give cells from neurons of the developing CNS (Muschler and Horwitz, little time to deposit their own substrate. Regarding size, 1991). 1880 C. S. Stipp and M. E. Hemler

Three TM4SF proteins (CD63, CD81, and CD151), were Bergelson, J. M., St, J. N., Kawaguchi, S., Pasqualini, R., Berdichevsky, expressed on the neurites and growth cones of CNS neuron- F., Hemler, M. E. and Finberg, R. W. (1994). The I domain is essential like NT2N cells. Compared to NCAM, the TM4SF proteins for echovirus 1 interaction with VLA-2. Cell Adhes. Commun. 2, 455- 464. had a more punctate staining pattern, and were often localized Borlongan, C. V., Tajima, Y., Trojanowski, J. Q., Lee, V. M. and Sanberg, at the leading edge and at the tips of filopodia. The minimal P. R. (1998). Transplantation of cryopreserved human embryonal staining of NCAM in NT2N growth cones in Fig. 3 is similar carcinoma-derived neurons (NT2N cells) promotes functional recovery in to that described for in vivo-grafted NT2N cells, in which distal ischemic rats. Exp. Neurol. 149, 310-321. portions of axons were also weakly stained (Kleppner et al., Bradshaw, A. D., McNagny, K. M., Gervin, D. B., Cann, G. M., Graf, T. α α and Clegg, D. O. (1995). Integrin alpha 2 beta 1 mediates interactions 1995). The distribution of the 3 and 2 integrin subunits was between developing embryonic retinal cells and collagen. Development 121, similar to that of the TM4SF proteins examined. Thus, routine 3593-3602. immunofluorescence appears incapable of distinguishing Brewer, G. J., Torricelli, J. R., Evege, E. K. and Price, P. J. (1993). between the more robust TM4SF-α3β1 complexes and weaker Optimized survival of hippocampal neurons in B27-supplemented α β Neurobasal, a new serum-free medium combination. J. Neurosci. Res. 35, TM4SF- 2 1 complexes. Overall, staining patterns observed 567-576. for TM4SF proteins and integrins are consistent with a role in Bronner-Fraser, M., Artinger, M., Muschler, J. and Horwitz, A. F. (1992). growth cone motility, and are reminiscent of β1 integrin Developmentally regulated expression of alpha 6 integrin in avian embryos. staining of chick DRG neurons on growth cones, including the Development 115, 197-211. tips of filopodia (Letourneau, 1992). Cann, G. M., Bradshaw, A. D., Gervin, D. B., Hunter, A. W. and Clegg, D. O. (1996). Widespread expression of beta1 integrins in the developing In summary, our results provide three important advances chick retina: evidence for a role in migration of retinal ganglion cells. Dev. beyond previous studies of neurite outgrowth and TM4SF- Biol. 180, 82-96. integrin complexes. First, we have established that antibodies Carter, W. G., Ryan, M. C. and Gahr, P. J. (1991). Epiligrin, a new cell to multiple TM4SF proteins can strongly inhibit neurite adhesion ligand for integrin alpha 3 beta 1 in epithelial basement membranes. Cell 65, 599-610. outgrowth. Second, we find that specific TM4SF-integrin Chyung, A., Greenberg, B. D., Cook, D. G., Doms, R. W. and Lee, V. M. complexes may function in an integrin substrate-dependent (1997). Novel beta-secretase cleavage of beta-amyloid precursor protein in manner. Third, we demonstrate the utility of the NT2/NT2N the endoplasmic reticulum/intermediate compartment of NT2N cells. J. Cell cell system as a convenient and highly reproducible tool, well Biol. 138, 671-680. suited for studies of human neurite outgrowth. de Curtis, I. (1993). The alpha 6 beta 1 integrin is a laminin receptor for developing retinal neurons. Cytotechnology 11, S41-43. α DiPersio, C., Shah, S. and Hynes, R. (1995). alpha3Abeta1 integrin localizes The authors gratefully acknowledge Dr Mike DiPersio for the 3 to focal contacts in response to diverse extracellular matrix proteins. J. Cell integrin cytoplasmic domain antiserum, Dr Vito Quaranta for α6 Sci. 108, 2321-2336. integrin cytoplasmic domain antisera, Dr Leonie Ashman for Domanico, S. Z., Pelletier, A. J., Havran, W. L. and Quaranta, V. (1997). providing the monoclonal antibody 11B1.G4, and for contributing Integrin alpha 6A beta 1 induces CD81-dependent cell motility without unpublished observations on CD151 expression in the developing engaging the extracellular matrix migration substrate. Mol. Biol. Cell 8, nervous system, and Desmos Inc. for providing rat laminin-5. We also 2253-2265. thank Dr Jon Ivins for helpful discussions and Dr Mary Herndon and Duband, J. L., Belkin, A. M., Syfrig, J., Thiery, J. P. and Koteliansky, V. Dr Robert Yauch for a critical reading of the manuscript. This work E. (1992). Expression of alpha 1 integrin, a laminin-collagen receptor, during myogenesis and neurogenesis in the avian embryo. Development 116, was supported by N.I.H. grants GM38903 to M.E.H. and NS10344 to 585-600. C.S.S. Eble, J. A., Wucherpfennig, K. W., Gauthier, L., Dersch, P., Krukonis, E., Isberg, R. R. and Hemler, M. E. (1998). Recombinant soluble human alpha 3 beta 1 integrin: purification, processing, regulation, and specific binding REFERENCES to laminin-5 and invasin in a mutually exclusive manner. Biochemistry 37, 10945-10955. Abraham, I., Sampson, K. E., Powers, E. A., Mayo, J. K., Ruff, V. A. and Elices, M. J., Urry, L. A. and Hemler, M. E. (1991). Receptor functions for Leach, K. L. (1991). Increased PKA and PKC activities accompany the integrin VLA-3: fibronectin, collagen, and laminin binding are neuronal differentiation of NT2/D1 cells. J. Neurosci. Res. 28, 29-39. differentially influenced by Arg-Gly-Asp peptide and by divalent cations. J. Anton, E. S., Hadjiargyrou, M., Patterson, P. H. and Matthew, W. D. Cell Biol. 112, 169-181. (1995). CD9 plays a role in Schwann cell migration in vitro. J. Neurosci. Fitter, S., Seldin, M. F. and Ashman, L. K. (1998). Characterisation of the 15, 584-595. mouse homologue of CD151 (PETA-3/SFA-1); genomic structure, Anton, E. S., Kreidberg, J. A. and Rakic, P. (1999). Distinct functions of chromosomal localisation and identification of 2 novel splice forms. alpha3 and alpha(v) integrin receptors in neuronal migration and laminar Biochim. Biophys. Acta 1398, 75-85. organization of the cerebral cortex. Neuron 22, 277-289. Fitter, S., Sincock, P. M., Jolliffe, C. N. and Ashman, L. K. (1999). Banerjee, S. A., Hadjiargyrou, M. and Patterson, P. H. (1997). An antibody Transmembrane 4 superfamily protein CD151 (PETA-3) associates with to the tetraspan membrane protein CD9 promotes neurite formation in a beta 1 and alpha IIb beta 3 integrins in haemopoietic cell lines and modulates partially alpha3beta1 integrin-dependent manner. J. Neurosci. 17, 2756- cell-cell adhesion. Biochem. J. 338, 61-70. 2765. Fukudome, K., Furuse, M., Imai, T., Nishimura, M., Takagi, S., Hinuma, Bani-Yaghoub, M., Bechberger, J. F. and Naus, C. C. (1997). Reduction of Y. and Yoshie, O. (1992). Identification of membrane antigen C33 connexin43 expression and dye-coupling during neuronal differentiation of recognized by monoclonal antibodies inhibitory to human T-cell leukemia human NTera2/clone D1 cells. J. Neurosci. Res. 49, 19-31. virus type 1 (HTLV-1)- induced syncytium formation: altered Berditchevski, F., Bazzoni, G. and Hemler, M. E. (1995). Specific glycosylation of C33 antigen in HTLV-1-positive T cells. J. Virol. 66, association of CD63 with the VLA-3 and VLA-6 integrins. J. Biol. Chem. 1394-1401. 270, 17784-17790. Galliano, M. F., Aberdam, D., Aguzzi, A., Ortonne, J. P. and Meneguzzi, Berditchevski, F., Zutter, M. M. and Hemler, M. E. (1996). Characterization G. (1995). Cloning and complete primary structure of the mouse laminin of novel complexes on the cell surface between integrins and proteins with alpha 3 chain. Distinct expression pattern of the laminin alpha 3A and alpha 4 transmembrane domains (TM4 proteins). Mol. Biol. Cell 7, 193-207. 3B chain isoforms. J. Biol. Chem. 270, 21820-21826. Berditchevski, F., Tolias, K. F., Wong, K., Carpenter, C. L. and Hemler, Hadjiargyrou, M., Kaprielian, Z., Kato, N. and Patterson, P. H. (1996). M. E. (1997). A novel link between integrins, transmembrane-4 superfamily Association of the tetraspan protein CD9 with integrins on the surface of S- proteins (CD63 and CD81), and phosphatidylinositol 4-kinase. J. Biol. 16 Schwann cells. J. Neurochem. 67, 2505-2513. Chem. 272, 2595-2598. Hasegawa, H., Nomura, T., Kishimoto, K., Yanagisawa, K. and Fujita, S. CD151, CD81 and α3β1 integrin function in neurite outgrowth 1881

(1998). SFA-1/PETA-3 (CD151), a member of the transmembrane 4 heparin-binding EGF-like growth factor (HB-EGF) and diphtheria toxin superfamily, associates preferentially with alpha 5 beta 1 integrin and receptor-associated protein (DRAP27)/CD9 form a complex with integrin regulates adhesion of human T cell leukemia virus type 1-infected T cells alpha 3 beta 1 at cell-cell contact sites. J. Cell Biol. 129, 1691-1705. to fibronectin. J. Immunol. 161, 3087-3095. Neelands, T. R., Greenfield, L. Jr, Zhang, J., Turner, R. S. and Macdonald, Hemler, M. E., Sanchez-Madrid, F., Flotte, T. J., Krensky, A. M., Burakoff, R. L. (1998). GABAA receptor pharmacology and subtype mRNA S. J., Bhan, A. K., Springer, T. A. and Strominger, J. L. (1984). expression in human neuronal NT2-N cells. J. Neurosci. 18, 4993-5007. Glycoproteins of 210, 000 and 130, 000 m. w. on activated T cells: cell Pleasure, S. J., Page, C. and Lee, V. M. (1992). Pure, postmitotic, polarized distribution and antigenic relation to components on resting cells and T cell human neurons derived from NTera 2 cells provide a system for expressing lines. J. Immunol. 132, 3011-3018. exogenous proteins in terminally differentiated neurons. J. Neurosci. 12, Hemler, M. E., Huang, C., Takada, Y., Schwarz, L., Strominger, J. L. and 1802-1815. Clabby, M. L. (1987). Characterization of the cell surface heterodimer Pujades, C., Teixido, J., Bazzoni, G. and Hemler, M. E. (1996). Integrin VLA-4 and related peptides. J. Biol. Chem. 262, 11478-11485. alpha 4 cysteines 278 and 717 modulate VLA-4 ligand binding and also Hemler, M. E., Mannion, B. A. and Berditchevski, F. (1996). Association contribute to alpha 4/180 formation. Biochem. J. 313, 899-908. of TM4SF proteins with integrins: relevance to cancer. Biochim. Biophys. Radford, K. J., Thorne, R. F. and Hersey, P. (1997). Regulation of tumor Acta 1287, 67-71. cell motility and migration by CD63 in a human melanoma cell line. J. Hemler, M. E. (1998). Integrin associated proteins. Curr. Opin. Cell Biol. 10, Immunol. 158, 3353-3358. 578-585. Rubinstein, E., Le Naour, F., Billard, M., Prenant, M. and Boucheix, C. Hunter, D. D., Llinas, R., Ard, M., Merlie, J. P. and Sanes, J. R. (1992). (1994). CD9 antigen is an accessory subunit of the VLA integrin complexes. Expression of s-laminin and laminin in the developing rat central nervous Eur. J. Immunol. 24, 3005-3013. system. J. Comp. Neurol. 323, 238-251. Schaffner, W. and Weissman, C. (1973). A rapid, sensitive, and specific Huttenlocher, A., Sandborg, R. R. and Horwitz, A. F. (1995). Adhesion in method for the determination of protein in dilute solution. Anal. Biochem. cell migration. Curr. Opin. Cell Biol. 7, 697-706. 56, 502-514. Hynes, R. O. and Lander, A. D. (1992). Contact and adhesive specificities Schmidt, C., Kunemund, V., Wintergerst, E. S., Schmitz, B. and in the associations, migrations, and targeting of cells and axons. Cell 68, Schachner, M. (1996). CD9 of mouse brain is implicated in neurite 303-322. outgrowth and cell migration in vitro and is associated with the alpha 6/beta Ikeyama, S., Koyama, M., Yamaoko, M., Sasada, R. and Miyake, M. 1 integrin and the neural adhesion molecule L1. J. Neurosci. Res. 43, 12- (1993). Suppression of cell motility and metastasis by transfection with 31. human motility-related protein (MRP-1/CD9) DNA. J. Exp. Med. 177, Serru, V., Le Naour, F., Billard, M., Azorsa, D. O., Lanza, F., Boucheix, C. 1231-1237. and Rubinstein, E. (1999). Selective tetraspan-integrin complexes Indig, F. E., Diaz-Gonzalez, F. and Ginsberg, M. H. (1997). Analysis of the (CD81/alpha4beta1, CD151/alpha3beta1, CD151/alpha6beta1) under tetraspanin CD9-integrin alphaIIbbeta3 (GPIIb-IIIa) complex in platelet conditions disrupting tetraspan interactions. Biochem. J. 340, 103-111. membranes and transfected cells. Biochem. J. 327, 291-298. Sheetz, M. P., Felsenfeld, D. P. and Galbraith, C. G. (1998). Cell migration: Ivins, J. K., Colognato, H., Kreidberg, J. A., Yurchenco, P. D. and Lander, regulation of force on extracellular-matrix-integrin complexes. Trends Cell A. D. (1998). Neuronal receptors mediating responses to antibodyactivated Biol. 8, 51-54. laminin-1. J. Neurosci. 18, 9703-9715. Sincock, P. M., Mayrhofer, G. and Ashman, L. K. (1997). Localization of Jones, P. H., Bishop, L. A. and Watt, F. M. (1996). Functional significance the transmembrane 4 superfamily (TM4SF) member PETA-3 (CD151) in of CD9 association with beta 1 integrins in human epidermal keratinocytes. normal human tissues: comparison with CD9, CD63, and α5β1 integrin. J. Cell Adhes. Commun. 4, 297-305. Histochem. Cytochem. 45, 515-525. Kleppner, S. R., Robinson, K. A., Trojanowski, J. Q. and Lee, V. M. (1995). Sincock, P. M., Fitter, S., Parton, R. G., Berndt, M. C., Gamble, J. R. and Transplanted human neurons derived from a teratocarcinoma cell line Ashman, L. K. (1999). PETA-3/CD151, a member of the transmembrane (NTera-2) mature, integrate, and survive for over 1 year in the nude mouse 4 superfamily, is localised to the plasma membrane and endocytic system brain. J. Comp. Neurol. 357, 618-632. of endothelial cells, associates with multiple integrins and modulates cell Kopczynski, C. C., Davis, G. W. and Goodman, C. S. (1996). A neural function. J. Cell Sci. 112, 833-844. tetraspanin, encoded by late bloomer, that facilitates synapse formation. Skovronsky, D. M., Doms, R. W. and Lee, V. M. (1998). Detection of a novel Science 271, 1867-1870. intraneuronal pool of insoluble amyloid beta protein that accumulates with Lagaudriere-Gesbert, C., Le Naour, F., Lebel-Binay, S., Billard, M., time in culture. J. Cell Biol. 141, 1031-1039. Lemichez, E., Boquet, P., Boucheix, C., Conjeaud, H. and Rubinstein, E. Skubitz, K. M., Campbell, K. D., Iida, J. and Skubitz, A. P. (1996). CD63 (1997). Functional analysis of four tetraspans, CD9, CD53, CD81, and CD82, associates with tyrosine kinase activity and CD11/CD18, and transmits an suggests a common role in costimulation, cell adhesion, and migration: only activation signal in neutrophils. J. Immunol. 157, 3617-3626. CD9 upregulates HB-EGF activity. Cell. Immunol. 182, 105-112. Smith, B. E., Bradshaw, A. D., Choi, E. S., Rouselle, P., Wayner, E. A. and Lee, R. T., Berditchevski, F., Cheng, G. C. and Hemler, M. E. (1995). Clegg, D. O. (1996). Human SY5Y neuroblastoma cell interactions with Integrin-mediated collagen matrix reorganization by cultured human laminin isoforms: neurite outgrowth on laminin-5 is mediated by integrin vascular smooth muscle cells. Circ. Res. 76, 209-214. alpha 3 beta 1. Cell Adhes. Commun. 3, 451-462. Lemke, H., Hammerling, G. J., Hohmann, C. and Rajewsky, K. (1978). Sonnenberg, A., Modderman, P. W. and Hogervorst, F. (1988). Laminin Hybrid cell lines secreting monoclonal antibody specific for major receptor on platelets is the integrin VLA-6. Nature 336, 487-489. histocompatibility antigens of the mouse. Nature 271, 249-251. Sullivan, C. D. and Geisert, E. Jr (1998). Expression of rat target of the Letourneau, P. C. (1992). Integrins and N-cadherin are adhesive molecules antiproliferative antibody (TAPA) in the developing brain. J. Comp. Neurol. involved in growth cone migration. In The Nerve Growth Cone (ed. P. C. 396, 366-380. Letourneau, S. B. Kater and E. R. Macagno), pp. 181-193. New York: Raven Tachibana, I. and Hemler, M. E. (1999). Role of transmembrane-4 Press, Ltd. superfamily (TM4SF) proteins CD9 and CD81 in muscle cell fusion an Maecker, H. T., Todd, S. C. and Levy, S. (1997). The tetraspanin superfamily: myoblast maintenance. J. Cell Biol. 146, 893-904. molecular facilitators. FASEB J. 11, 428-442. Takada, Y. and Hemler, M. E. (1989). The primary structure of the VLA- Mannion, B. A., Berditchevski, F., Kraeft, S. K., Chen, L. B. and Hemler, 2/collagen receptor alpha 2 subunit (platelet GPIa): homology to other M. E. (1996). Transmembrane-4 superfamily proteins CD81 (TAPA-1), integrins and the presence of a possible collagen-binding domain. J. Cell CD82, CD63, and CD53 specifically associated with integrin alpha 4 beta Biol. 109, 397-407. 1 (CD49d/CD29). J. Immunol. 157, 2039-2047. Todd, S. C., Doctor, V. S. and Levy, S. (1998). Sequences and expression of Miyake, M., Koyama, M., Seno, M. and Ikeyama, S. (1991). Identification six new members of the tetraspanin/TM4SF family. Biochim. Biophys. Acta of the motility-related protein (MRP-1), recognized by monoclonal 1399, 101-104. antibody M31-15, which inhibits cell motility. J. Exp. Med. 174, 1347- Tomaselli, K. J., Doherty, P., Emmett, C. J., Damsky, C. H., Walsh, F. S. 1354. and Reichardt, L. F. (1993). Expression of beta 1 integrins in sensory Muschler, J. L. and Horwitz, A. F. (1991). Down-regulation of the chicken neurons of the dorsal root ganglion and their functions in neurite outgrowth alpha 5 beta 1 integrin fibronectin receptor during development. on two laminin isoforms. J. Neurosci. 13, 4880-4888. Development 113, 327-337. Trojanowski, J. Q., Kleppner, S. R., Hartley, R. S., Miyazono, M., Fraser, Nakamura, K., Iwamoto, R. and Mekada, E. (1995). Membrane-anchored N. W., Kesari, S. and Lee, V. M. (1997). Transfectable and transplantable 1882 C. S. Stipp and M. E. Hemler

postmitotic human neurons: a potential ‘platform’ for gene therapy of molecules CD81/TAPA-1 and CD151/PETA-3 with alpha3 beta1 integrin nervous system diseases. Exp. Neurol. 144, 92-97. localized at endothelial lateral junctions. J. Cell Biol. 141, 791-804. Weitzman, J. B., Pasqualini, R., Takada, Y. and Hemler, M. E. (1993). The Yauch, R. L., Berditchevski, F., Harler, M. B., Reichner, J. and Hemler, function and distinctive regulation of the integrin VLA-3 in cell adhesion, M. E. (1998). Highly stoichiometric, stable, and speciﬁc association of spreading, and homotypic cell aggregation. J. Biol. Chem. 268, 8651-8657. integrin alpha3beta1 with CD151 provides a major link to Whittaker, C. A. and DeSimone, D. W. (1993). Integrin alpha subunit phosphatidylinositol kinase, and may regulate cell migration. Mol. Biol. Cell mRNAs are differentially expressed in early Xenopus embryos. 9, 2751-2765. Development 117, 1239-1249. Yauch, R. L., Kazarov, A., Desai, B., Lee, R. T. and Hemler, M. E. (2000). Wright, M. D. and Tomlinson, M. G. (1994). The ins and outs of the Direct extracellular contact between integrin α3β1 and TM4SF protein transmembrane 4 superfamily. Immunol. Today 15, 588-594. CD151. J. Biol. Chem. 275, 9230-9238. Yanez-Mo, M., Alfranca, A., Cabanas, C., Marazuela, M., Tejedor, R., Yelton, D. E., Desaymard, C. and Scharff, M. D. (1981). Use of monoclonal Ursa, M. A., Ashman, L. K., de Landazuri, M. O. and Sanchez-Madrid, anti-mouse immunoglobulin to detect mouse antibodies. Hybridoma 1, 5- F. (1998). Regulation of endothelial cell motility by complexes of tetraspan 11.