L1 is sequentially processed by two differently activated metalloproteases and presenilin/gamma-secretase and regulates neural cell adhesion, cell migration, and neurite outgrowth

Abstract

The immunoglobulin superfamily recognition molecule L1 plays important functional roles in the developing and adult nervous system. Metalloprotease-mediated cleavage of this adhesion molecule has been shown to stimulate cellular migration and neurite outgrowth. We demonstrate here that L1 cleavage is mediated by two distinct members of the disintegrin and metalloprotease family, ADAM10 and ADAM17. This cleavage is differently regulated and leads to the generation of a membrane bound C-terminal fragment, which is further processed through gamma-secretase activity. Pharmacological approaches with two hydroxamate-based inhibitors with different preferences in blocking ADAM10 and ADAM17, as well as loss of function and gain of function studies in murine embryonic fibroblasts, showed that constitutive shedding of L1 is mediated by ADAM10 while phorbol ester stimulation or cholesterol depletion led to ADAM17-mediated L1 cleavage. In contrast, N-methyl-d-aspartate treatment of primary neurons stimulated ADAM10-mediated L1 shedding. Both proteases were able to affect L1-mediated adhesion and haptotactic migration of neuronal cells. In particular, both proteases were involved in L1-dependent neurite outgrowth of cerebellar neurons. Thus, our data identify ADAM10 and ADAM17 as differentially regulated L1 membrane sheddases, both critically affecting the physiological functions of this adhesion protein.

FIG. 1.

Different effects of ADAM10 and ADAM17 inhibitors on L1 proteolysis. (A) Schematic map of L1, the metalloprotease cleavage site and antibody binding regions. Full-length 220-kDa L1 is cleaved near the transmembrane domain in N-terminal 200-kDa fragments (NTF) and C-terminal 32-kDa fragments (CTF1). (B) Representative immunoblot of supernatants (top) and cell lysates (bottom) of wild-type MEFs with N-terminal L1 antibodies. MEFs were transfected with an L1 expression vector; 48 h afterwards, cells were incubated with the hydroxamate-based ADAM10 inhibitor GI254023X (5 μM), the ADAM17/ADAM10 inhibitor GW280264X (5 μM), the broad-spectrum metalloprotease inhibitor EDTA (5 mM), or DMSO as control. EDTA, as well as the ADAM10 inhibitor, reduced constitutive L1 cleavage. L1/FL, full-length L1; WT, wild-type-MEFs. (C) Densitometric analysis of L1-NTF generation. To normalize variations in transfection efficiency, NTF generation was calculated as a percentage of total L1 (full-length L1 plus NTF). Results are obtained from three independent experiments and expressed as means ± SEM. The release of soluble L1 was significantly decreased when cells were treated with GI254023X or EDTA compared to vehicle-treated cells (_*).

FIG. 2.

Involvement of ADAM10 in constitutive L1 shedding. (A) Constitutive L1 cleavage is reduced in ADAM10^−/− fibroblasts. (Left) A representative immunoblot with total cell extracts from ADAM9^−/−, ADAM10^−/−, ADAM15^−/−, and ADAM17^−/− fibroblasts and wild-type MEFs stained with C-terminal anti-L1 antibodies is shown. Supernatants of these cells were also subjected to Western blot analysis with N-terminal anti-L1 antibodies (bottom). L1/FL, full-length L1. (Right) CTF generation was calculated as a percentage of total L1 (full-length L1 plus CTF1) by densitometric analysis. Results were obtained from five independent experiments and expressed as means ± SEM. L1-CTF1 generation was significantly decreased only in ADAM10^−/− MEFs compared to WT cells (*). (B) Anti-L1 (C-terminal) immunoblot of proteins from two independent ADAM10^−/− and wild-type MEF cell lines. L1/FL, full-length L1; WT: wild-type-MEFs. (C) ELISA of constitutive L1 shedding. L1 was transiently transfected into ADAM10^−/− or wild-type MEFs. Cells were incubated with or without the metalloprotease inhibitor GM6001 (5 μM) for 3 h, and subsequently conditioned media and cell lysates were analyzed for the presence of soluble and cell-associated L1, respectively. To consider possible variations in transfection efficiency, the data for each single transfection were calculated as the percentage of soluble L1 released in the medium in relation to the total amount of soluble and cell-bound L1 determined in the medium and the lysate. Results were obtained from four independent experiments, each performed in triplicate and are expressed as means ± SEM. L1-NTF generation was significantly lower in ADAM10^−/− MEFs, WT cells, and ADAM10^−/− MEFs treated with GM6001 than in untreated WT cells (*). (D) ADAM10-deficient cells were cotransfected with wild-type mouse ADAM10 (A10^−/−retr) and L1 and compared with L1-transfected wild-type and ADAM10-deficient MEFs for ADAM10 and L1 expression. The immunoblot shows one representative of three independent experiments. L1/FL, full-length L1; WT, wild-type-MEFs; p, precursor of ADAM10; m, mature form of ADAM10. (E) Reduced L1 ectodomain shedding in ADAM10-deficient embryos (embryonic day 9) was demonstrated by immunoblot analysis of whole-embryo extracts of ADAM10^−/− and wild-type mice. Tubulin was used as a loading control.

FIG. 3.

L1-CTF1 is further processed through γ-secretase activity. L1 shedding was analyzed with WT MEFs in the absence or presence of the γ-secretase inhibitor L-685,458 (1 μM, overnight) with ADAM10^−/− MEFs and fibroblasts deficient for the γ-secretase complex components PS1/2. The application of the γ-secretase inhibitor increased the amount of CTF1 and diminished CTF2 production in wild-type fibroblasts, visible after overexposure of the immunoblot. CTF2 was not detectable in the PS1/2-deficient cells or in ADAM10-deficient fibroblasts. One representative out of three immunoblots is shown.

FIG. 4.

ADAM17 mediates L1 shedding after PMA stimulation and cholesterol depletion. (A) Effect of different metalloprotease inhibitors on PMA-stimulated L1 shedding. (Left) Wild-type MEFs were stimulated with PMA (200 ng/ml) or vehicle control (DMSO) for 3 h. Subsequently, cell pellets and supernatants were subjected to L1 (N-terminal) Western blot analysis. (Right) NTF generation was calculated as the percentage of total L1 (full-length L1 plus NTF) by densitometric analysis. Results were obtained from three independent experiments and are expressed as means ± SEM. Compared to the vehicle control, PMA treatment significantly increased L1-NTF generation (+). All inhibitors significantly decreased NTF generation in the presence of PMA compared to the only-PMA-treated cells (*). (B) PMA-induced L1 shedding is abolished in ADAM17-deficient cells. (Top) Different ADAM-deficient fibroblasts and ADAM17^−/− cells stably retransfected with full-length ADAM17 (A17^−/−retr) were stimulated with PMA (200 ng/ml) or vehicle control (DMSO) for 3 h. Cell pellets and supernatants were harvested and analyzed by Western blotting using N-terminal anti-L1 antibodies. L1/FL, full-length L1. (Bottom) The constitutive NTF generation (lower left) and the increase in NTF generation after PMA treatment compared to the constitutive conditions (defined as inducible shedding) were also depicted as densitometric analysis (lower right). Results were obtained from five independent experiments and are expressed as means ± SEM. Constitutive NTF generation and inducible NTF generation were significantly reduced in ADAM10^−/− MEFs and ADAM17^−/− cells, respectively (*). (C) Effect of different stimuli on L1 shedding. L1 was transiently transfected into ADAM10^−/−, ADAM17^−/−, or wild-type MEFs. Cells were stimulated with MCD (10 mM) for 60 min, PMA (200 ng/ml) for 3 h, or vehicle control (DMSO) for 3 h; subsequently, conditioned media and cell lysates were analyzed for the presence of soluble and cell-associated L1, respectively, by ELISA. The generation of soluble L1 was calculated as the percentage of total L1 (L1 in cell lysates and supernatants) by densitometric analysis. Results are shown as means ± SEM of triplicate transfections of four independent experiments. Compared to untreated WT cells, constitutive NTF generation was significantly decreased in ADAM10^−/− MEFs, while MCD and PMA stimulation significantly increased NTF production in WT MEFs (*). Compared to the vehicle-treated control, PMA and MCD treatment significantly increased L1-NTF generation in ADAM10^−/− MEFs (*). This was not the case in ADAM17^−/− cells.

FIG. 5.

ADAM10-mediated L1 shedding in neuronal cells. (A) (Top) Human neuroglioma H4 cells were transiently transfected with pSUPER-hADAM10 siRNA, pSUPER-hADAM17 siRNA, or empty vector. Cell pellets were harvested 48 h and 72 h after transfection and analyzed for ADAM10 and ADAM17 expression by immunoblotting, respectively. Membranes were reprobed with anti-L1 antibodies. For densitometric quantification, ADAM10-L1-CTF values obtained from mock-treated cells after 72 h were considered 100% and compared with the values of ADAM10 siRNA- or ADAM17 siRNA-treated cells after 48 h and 72 h. (Bottom) Results from three independent experiments are shown and expressed as means ± SEM. Compared to mock-transfected control cells, both siRNA treatments led to a significant suppression of the respective gene after 48 h and even more after 72 h. In contrast, L1-CTF1 generation was only significantly decreased after ADAM10 siRNA treatment, compared to mock-treated cells (*). (B) Cell adhesion on recombinant human L1/Fc chimera. H4 cells were preincubated with ADAM10 inhibitor GI254023X (5 μM), the mixed ADAM17/ADAM10 inhibitor GW280264X (5 μM), or DMSO control with (right) or without (left) additional PMA (200 ng/ml) treatment. In another assay, cells were transfected with ADAM10 or ADAM10 siRNA and harvested after 48 h or 72 h, respectively (middle). ADAM10 expression of mock-transfected cells (middle, lane 1), ADAM10-transfected cells (lane 2), and siRNA-transfected cells (lane 3) was controlled by immunoblotting. Results are shown as means ± SEM of three independent experiments, all performed in triplicate. In all assays, the adhesion on L1 substrate was significantly higher than on BSA (+). L1-dependent cell adhesion was significantly increased in the presence of inhibitors (left, *), while PMA treatment (right, *) decreased cell adhesion compared with the respective control on L1 substrate. The PMA effect could be significantly abolished after pretreatment with the inhibitors (right, °). While ADAM10 overexpression significantly decreased cell adhesion, ADAM10 siRNA transfection significantly increased adhesion on L1 compared to the control (middle, *).

FIG. 6.

ADAM10 mediates L1 shedding in the mouse brain and affects haptotactic migration of cerebellar neurons. (A) (Top) Crude membrane fractions prepared from adult wild-type mouse brain were incubated at 37°C for different times with recombinant mouse ADAM10 (rADAM10). As a control, aliquots were incubated without rADAM10 for 60 min. Samples were analyzed by immunoblotting with C-terminal L1 antibodies. (Bottom) CTF generation was also calculated by densitometric analysis of three independent experiments and expressed as means ± SEM. Addition of rADAM10 led to a significant time-dependent increase in CTF generation (*). (B) Overexpression of ADAM10 in primary neurons of mouse cerebellum increases L1 shedding. Cells were prepared by mechanical and enzymatic dissociation of mouse cerebellum. Cells were transiently transfected with ADAM10 or empty vector. Subsequently, cells were lysed and analyzed by Western blotting with anti-ADAM10 antibodies. The same blot was reprobed with anti-L1 antibodies (C terminal). L1/FL, full-length L1; WT, wild-type-MEFs; p, precursor of ADAM10; m, mature form of ADAM10. (C) Metalloprotease inhibitors reduce haptotactic migration of cerebellar neurons on L1 substrate. NCAM L1/Fc chimera at a concentration of 2 μg/ml or BSA for control were coated on the back sides of transwell chambers. Cerebellar neurons were seeded into the top chamber in the presence of GI254023X (5 μM), GW280264X (5 μM), or DMSO as a control and allowed to transmigrate for 48 h at 37°C. Cells on top of the filter were removed, and transmigrated cells were stained with crystal violet solution. The dye was eluted from the filter and measured at 595 nm. The amount of dye was proportional to the number of transmigrated cells. Values for the migration of mock-treated cells on L1 substrate were set at 100% and compared to the values of inhibitor treated cells. Results were obtained from three independent experiments, each performed in triplicate, and are expressed as means ± SEM. The haptotactic migration on L1 substrate was significantly higher than on BSA (+). L1-dependent cell migration was significantly decreased in the presence of inhibitors (*). (D) NMDA stimulation induces ADAM10-mediated L1 shedding in primary cerebellar neurons. Cells were incubated in the presence or absence of NMDA (50 μM) and the inhibitor GI254023X (5 μM) or GW280264X (5 μM) for 30 min. Cell pellets were subjected to Western blot analysis with C-terminal L1 antibodies. (E) NMDA-stimulated increase of L1 shedding was associated with an increase of the mature form of ADAM10. (Left) Western blots of NMDA-stimulated primary neurons were analyzed for L1 expression and afterwards reprobed with ADAM17 and ADAM10 antibodies. (Right) Results were obtained from three independent experiments and are expressed as means ± SEM. L1-CTF1 generation was significantly increased after NMDA treatment compared with the vehicle-treated cells (*). The ratio of the mature form of ADAM10/ADAM17 was calculated as the percentage of mature form to total ADAM10/ADAM17 (precursor form plus mature form). NMDA stimulation significantly decreased the mature form of ADAM17 (°), while the mature form of ADAM10 was significantly increased compared to control cells (+).

FIG. 7.

Effect of different metalloprotease inhibitors on neurite outgrowh of cerebellar neurons. (A) Cerebellar microexplant cultures were plated onto glass coverslips coated with PLL or a combination of PLL and L1. After incubation for 24 h in the presence or absence of different metalloprotease inhibitors (5 μM), the explants were fixed and stained. Neurite outgrowth and also movement of small neuronal cell bodies (arrowheads) from the explants grown on L1 substrate were decreased by the inhibitors applied. Representative images of three independent experiments are shown. Bar, 50 μm. (B) The effect of the metalloprotease inhibitors GI254023X and GW280264X on neurite outgrowth from the explants was quantified by measuring the 10 longest neurites of 10 aggregates in three experiments. Results are shown as means ± SEM of these independent experiments. The values obtained from DMSO-treated explants were set at 100% and compared with the values of inhibitor-treated cells. Neurite outgrowth from the explants grown on L1 substrate was significantly decreased after inhibitor treatment compared to vehicle-treated explants (*). (C) Representative immunoblot with cerebellar (CB) explants stained with C-terminal L1 antibody. Cerebellar explants seeded on PLL were incubated for 24 h in the absence or presence of the metalloprotease inhibitor GI254023X (5 μM) or GW280264X (5 μM) or with DMSO as a control. L1/FL, full-length L1; WT, wild-type-MEFs.