Reelin is a serine protease of the extracellular matrix.

Reelin is an extracellular matrix protein that plays a pivotal role in development of the central nervous system. Reelin is also expressed in the adult brain, notably in the cerebral cortex, where it might play a role in synaptic plasticity. The mechanism of action of reelin at the molecular level has been the subject of several hypotheses. Here we show that reelin is a serine protease and that proteolytic activity is relevant to its function, since (i) Reelin expression in HEK 293T cells impairs their ability to adhere to fibronectin-coated surfaces, and adhesion to fibronectin is restored by micromolar concentrations of diisopropyl phosphorofluoridate, a serine hydrolase inhibitor; (ii) purified Reelin binds FP-Peg-biotin, a trap probe which irreversibly binds to serine residues located in active catalytic sites of serine hydrolases; (iii) purified Reelin rapidly degrades fibronectin and laminin, while collagen IV is degraded at a much slower rate; fibronectin degradation is inhibited by inhibitors of serine proteases, and by monoclonal antibody CR-50, an antibody known to block the function of Reelin both in vitro and in vivo. The proteolytic activity of Reelin on adhesion molecules of the extracellular matrix and/or receptors on neurons may explain how Reelin regulates neuronal migration and synaptic plasticity.

Reelin is an extracellular matrix protein that plays a pivotal role in development of the central nervous system. Reelin is also expressed in the adult brain, notably in the cerebral cortex, where it might play a role in synaptic plasticity. The mechanism of action of reelin at the molecular level has been the subject of several hypotheses. Here we show that reelin is a serine protease and that proteolytic activity is relevant to its function, since (i) Reelin expression in HEK 293T cells impairs their ability to adhere to fibronectin-coated surfaces, and adhesion to fibronectin is restored by micromolar concentrations of diisopropyl phosphorofluoridate, a serine hydrolase inhibitor; (ii) purified Reelin binds FP-Peg-biotin, a trap probe which irreversibly binds to serine residues located in active catalytic sites of serine hydrolases; (iii) purified Reelin rapidly degrades fibronectin and laminin, while collagen IV is degraded at a much slower rate; fibronectin degradation is inhibited by inhibitors of serine proteases, and by monoclonal antibody CR-50, an antibody known to block the function of Reelin both in vitro and in vivo. The proteolytic activity of Reelin on adhesion molecules of the extracellular matrix and/or receptors on neurons may explain how Reelin regulates neuronal migration and synaptic plasticity.
Reelin (1,2) is an extracellular matrix protein that plays a pivotal role in neuronal migration during development of laminar structures of the mammalian brain including the cerebral cortex, hippocampus, cerebellum, and several brainstem nuclei, as shown by spontaneous Reelin null mutations (i.e. the reeler mouse) (3,4). In the developing cerebral cortex, Reelin is secreted by Cajal-Retzius cells, located in the marginal zone. Reelin must be secreted into the extracellular matrix to exert its biological effect (5).
In the reeler mouse, migrating neurons fail to pass through earlier-generated neurons, possibly because they are unable to penetrate the subplate, or because they maintain extensive contacts with the radial glial fibers (6). Several hypotheses have been suggested regarding the function of Reelin: (i) Reelin may act as an attractant molecule for migrating neurons; (ii) it may act as a repulsive molecule; or (iii) Reelin may interrupt the association between migrating neurons and radial glia (7,8), thus allowing migrating neurons to switch from a "gliophilic" to a "neurophilic" state (9). Furthermore, Reelin has been recently shown to be expressed in several adult neuronal cells, including glutamatergic cerebellar granule neurons and specific GABAergic interneurons of the cerebral cortex and hippocampus (10), and in the adult mammalian blood, liver, pituitary pars intermedia, and adrenal chromaffin cells (11,12). The cellular function of Reelin in the adult organism is unknown. Evidence is accumulating for involvement of Reelin in human diseases such as autosomal recessive lissencephaly (13), schizophrenia (14), and autistic disorder (15).
The mouse Reelin sequence (1) encompasses 3461 amino acids and possesses a signal peptide followed by a domain with 28% sequence identity with F-spondin (as assessed by ⌿-Blast software), a protein secreted by floor plate cells and promoting cell adhesion and neurite growth (16). This region is followed by a unique region with no sequence homology, and then by eight internal repeats of 350 -390 amino acids, each repeat containing two related subdomains flanking a cystein-rich sequence similar to the epidermal growth factor-like motif. The carboxyl terminus region contains many positively charged amino acids required for secretion (5). Human Reelin (2) is 94.8% identical to the mouse protein at the amino acid level, indicating strong functional conservation. Recent findings suggest that the Reelin signal transduction involves binding to the very low density lipoprotein receptor and to apoE receptor 2 followed by intracellular activation of the adapter protein disabled-1 (17)(18)(19). Other possible Reelin signal transduction pathways may involve interaction with the ␣ 3 ␤ 1 integrin receptor (8,9) and with cadherin-related neuronal receptors (20).
We have analyzed the primary amino acid sequence of human Reelin, and found several hints that Reelin might be a serine protease, since: (i) Reelin contains the sequence GKSDG (amino acids 1280 -1284 of human Reelin) (2), corresponding to the serine hydrolase consensus sequence GXSXG; this sequence is 100% conserved among mouse, chicken, and human Reelin; (ii) Reelin shows significant structural similarities with serine hydrolases, such as the extracellular serine protease precursor (EC 3.4.21) of Serratia marcescens, and the probable ubiquitin carboxyl-terminal hydrolases FAM and FAF-Y (EC 3.1.2.15); (iii) Reelin contains eight epidermal growth factorlike repeats; epidermal growth factor-like repeats are observed in serine proteases, for example, coagulation factors VII, IX, and X, and protein C, Z (21), calcium-dependent serine protein-ase (22) which degrades extracellular matrix proteins, and complement C1s and C1r components (23); (iv) several serine hydrolases, such as lipoprotein lipase and the urokinase-type plasminogen activator, bind very low density lipoprotein receptor and apoE receptor 2 (24,25). In this study we present converging evidence that purified Reelin acts as a serine protease, and that this enzymatic activity may be relevant for its cellular function.

EXPERIMENTAL PROCEDURES
Chemicals-Cell culture media, antibiotics, and media supplements were purchased from Invitrogen (Gaithersburg, MD). All other chemicals were from Sigma, unless otherwise specified.
Amino acid Sequence Analysis of Reelin-All sequence analyses are based on human Reelin (accession number: NP005036) (2). Homologies were evaluated by using ⌿-Blast 2 sequences software (www.ncbi.nih. gov/blast). Also, PropSearch software (EMBL, Heidelberg) was used to find homologies with other serine hydrolases. PattinProt software (PBIL, NPSA, Lyon) was used to identify putative consensus sequences for serine proteases.
Cell Culture, Transfection, and Expression of Recombinant Reelin-Human embryonic kidney (HEK) 293T cells (ATCC, Q401) were grown in Dulbecco's modified Eagle's medium supplemented with 10% heatinactivated fetal bovine serum, penicillin/streptomycin (100 international units/ml and 100 g/ml, respectively), and 10 mM L-glutamine. All cells were cultured at 37°C, 5% CO 2 and 97% relative humidity. Passaging was routinely performed with trypsin-EDTA. Cells were stably transfected with 2 g of plasmid DNA pCrl (5), which contains the entire mouse Reelin open reading frame (accession number NP_035391), or with 2 g of pCDNA3 empty vector in 10 l of Lipo-fectAMINE in 4 ml of Opti-MEM serum-free medium, according to the manufacturer's instructions. Cells were transfected in 60-mm plates at a density of 5 ϫ 10 5 /dish and transferred to selection medium (0.6 mg/ml G418) 48 h after transfection. For Reelin purification, a stable cell line (CER) was generated by stable transfection of 293-EBNA cells (Invitrogen) with pCER followed by selection in medium containing 0.25 mg/ml G418 and 0.4 mg/ml hygromycin B. The pCER episomal plasmid contains a reelin insert identical to that of pCrl, cloned into the pCEP4 vector (Invitrogen).
Cultures were harvested for RNA preparation (26). Total RNA (3 g) was incubated in reaction buffer containing 5 M random hexamer (Amersham Bioscience, Inc.), Moloney murine leukemia virus reverse transcriptase, reaction buffer, and RNase inhibitor, according to the manufacturer's specifications (Invitrogen). Reverse transcription was performed for 1 h at 37°C, and stopped by incubating the samples for 10 min at 95°C. Expression of Reelin was checked by PCR amplification with specific primers: forward 5Ј-GGAAAGTCAGATGGAGAC-3Ј, reverse 5Ј-CATCTAAGCCAAACG-3Ј corresponding to nucleotides 4123-4498 of mouse Reelin mRNA (Ref. (1), accession number U24703). PCR amplification was carried out in a total volume of 50 l, with 10 l of reverse transcription reaction and 1.25 units of Taq polymerase (M-Medical Genenco-Life Science). Amplification was performed for 35 cycles (1 min denaturing at 94°C, 1 min annealing at 58°C, and 1 min extension at 72°C) in a thermal cycler (Oracle BioSystems, Delphi 1500). The 376-bp amplified product was analyzed by 2% agarose gel electrophoresis and visualized under UV illumination after staining with ethidium bromide. Reelin secretion into supernatants was analyzed by plating 1 ϫ 10 6 cells onto 90-mm dishes. After 24 h conditioned media were removed, cells were washed twice with phosphate-buffered saline (PBS), and serum-free Dulbecco's modified Eagle's medium was added. Forty-eight hours later, the supernatants were collected, cleared by a brief centrifugation (10 min, 1000 rpm) at 4°C, concentrated by Microcon YM-100 (Millipore Corp. Bedford, MA) and stored at Ϫ80°C for analysis by SDS-PAGE and immunoblotting. Reelin secretion into the cell culture medium was assessed by Western blots using mAb 142, an antibody that recognizes the NH 2 -terminal sequence of Reelin (27).
Cell Adhesion Assay-Cell adhesion assay was performed according to published protocols (28). Ninety-six-well plates were coated overnight with fibronectin (2.5 g/ml) in carbonate/bicarbonate buffer, pH 9.7. Cells were harvested, washed three times with serum-free Dulbecco's modified Eagle's medium, and then resuspended in sterile attachment solution (calcium-and magnesium-free Hanks' balanced salt solution, 20 mM HEPES, 1 mg/ml heat-inactivated bovine serum albumin, 1 mM CaCl 2 , 1 mM MgCl 2 . Mock-and pCrl-transfected 293T cells (1 ϫ 10 4 in 200 l/well) were allowed to attach for 2 h at 37°C in a humidified 5% CO 2 incubator. Unattached cells were removed with Hanks' balanced salt solution. Attached cells were fixed with 4% formaldehyde in PBS, pH 7.4, and counted manually using an inverted microscope by two different observers. The counting area was defined by a grid (12 mm 2 area) placed under the wells. Attached cells were defined as cells that had spread and grown at least 1 process. All data are expressed as mean Ϯ S.E. Differences between groups were tested by one-way ANOVA followed by the LSD post-hoc test, using the SPSS statistics package (SPSS, Chicago, IL, version 9.0).
Labeling of Reelin with FP-Peg-biotin-Fluorophosphate biotin (FP-Peg-biotin) (a gift of Dr. Benjamin Cravatt) (30), stored as a 100 mM stock solution in Me 2 SO at Ϫ20°C, was added directly to protein samples to a final concentration of 2-4 M. The reaction mixture was incubated at room temperature for 30 min, and stopped by adding an equal volume of 2 ϫ reducing sample buffer. As a control for the specificity of FP-Peg-biotin labeling, replica samples were incubated for 1 h with 11.4 M diisopropylphosphofluoridate (DFP), 1 a potent and specific serine-hydrolase inhibitor, before incubation with FP-Peg-biotin. Samples were separated by SDS-PAGE and transferred by electroblotting to nitrocellulose membranes; the membranes were blocked in TBST with 3% bovine serum albumin for 1 h at 25°C or overnight at 4°C, and then incubated for 15 min with an avidin-horseradish peroxidase conjugate (Pierce) diluted 1:300 in blocking solution. The labeled bands were revealed by chemiluminescence (see above).
Purification of Reelin-For gel filtration chromatography purification, 50 l of supernatant from Reelin secreting CER cells were concentrated first by osmotic dyalisis with AQUACIDE I (Calbiochem, La Jolla, CA). A 5-ml concentrated sample was loaded on a FPLC system Ä KTAprime and passed through a HiLoad Superdex 200 26/60 column (Amersham Bioscience, Inc., Uppsala, Sweden). The run was performed at a constant flow rate of 3 ml/min and maximal pressure limit of 0.5 Pa. The eluate was collected in 1.5-ml fractions. After dot immunoblot screening and Western blot, positive fractions for Reelin were pooled and concentrated again as described above.
For purification by SDS-PAGE, Reelin-containing supernatant from transfected 293T cells was concentrated as above and separated on a 5% gel. The gel was run for 3 h 30 min at 125 V to achieve a good separation in the Ͼ250 kDa range. Thereafter a thin vertical slice of gel was cut and stained with silver to reveal the 400-kDa Reelin band. A 5-mm wide horizontal slice was cut from the remaining gel, using the stained slice as a reference, and 400-kDa Reelin was electroeluted at 60 V into 25 mM Tris, 250 mM glycine, 0,1% SDS buffer. Electroelution was carried out overnight at 4°C. Purified Reelin was, then, transferred to PBS, pH 7.2, by overnight dialysis in a Slide-A-Lyzer cassette (Pierce). Finally the protein was concentrated. The final protein concentration was estimated to be 0.2 g/ml (Bradford Reagent, Sigma). Purified Reelin was re-electrophoresed on a SDS gel to check the purity of the protein.
Degradation of Extracellular Matrix Proteins-Reelin aliquots (10 ng) were incubated with 1 g of fibronectin from human plasma (Sigma), or with laminin or collagen type IV from basement membrane of Engelbreth-Holm-Swarm mouse sarcoma (Sigma) for 0, 10, 30, or 120 min at 37°C, in PBS, pH 7.9. The reaction was stopped by adding sample buffer and heating the samples at 100°C for 2 min. Samples were separated in a 8% SDS gel. After electrophoresis, the gel was fixed, and silver-stained.
Reelin Inhibits Cell Attachment in Vitro-To obtain recombinant Reelin protein, HEK 293T cells were transfected with pCrl plasmid, and Reelin mRNA expression was assessed by RT-PCR, using primers complementary to exon 27 sequences. RT-PCR of pCrl-transfected cells revealed the expected 376-bp band (Fig. 2A, lane 3), while the band was absent in mocktransfected cells (lane 2). Reelin secretion into the supernatant was confirmed by Western blotting. The supernatant of pCrltransfected cells showed a major Reelin band at approximately 400 kDa, and minor bands at 350 and 140 kDa (Fig. 2B, lane 2). The supernatant of mock-transfected cells did not show any stained bands (lane 1).
HEK 293T cells endogenously express ␣ 5 ␤ 1 integrin which specifically mediates adhesion to fibronectin (28). Titration experiments showed dose-dependent adhesion to fibronectin, with maximal adhesion occurring at 2.5-5 g/ml fibronectin (data not shown). Reelin-transfected HEK 293T showed significantly less adhesion as compared with mock-transfected HEK 293T cells. After a 2-h incubation, 44.7 Ϯ 6.3 (mean Ϯ S.E.) pCrl-transfected cells were attached to the substrate, as compared with 96.3 Ϯ 10.8 mock-transfected cells (Fig. 3A, p Ͻ 0.001). Furthermore, cell morphology was markedly different: Reelin-secreting cells appeared unable to spread on fibronectin and their processes were diminished in number and length, as compared with mock-transfected cells (Fig. 3B). To assess the biological significance of the enzymatic activity of Reelin, the effect of DFP, a potent and specific inhibitor of serine hydrolases, on cell adhesion was studied. Micromolar concentrations of DFP partially restored adhesion of Reelin-expressing cells on fibronectin, without affecting mock-transfected cells (Fig. 3, A  and B). The effect of DFP was dose-dependent, starting at concentrations Ն0.54 M; maximal increase in adhesion of Reelin-expressing cells was seen at 5.4 M DFP, while 5.4 mM was equally toxic for pCrl-and mock-transfected cells, inhibiting adhesion of Ͼ99% of the cells (Fig. 3C).
Reelin Binds a Serine Hydrolase Probe-FP-Peg-biotin is described to behave as a specific and irreversible probe for serine hydrolases, showing properties similar to those of common FP inhibitors, such as DFP (30). To explore Reelin labeling with FP-Peg-biotin, aliquots of transfected 293T cell culture supernatants incubated with 5 M FP-Peg-biotin were separated on standard SDS-PAGE gels, blotted, and probed with avidin peroxidase; replica samples were stained with the monoclonal antibody 142. The supernatant of Reelin-expressing cells showed distinct Reelin bands at approximately 400, 300, and 140 kDa (Fig. 4A, lane 2). The 400-and 300-kDa band showed faint labeling with FP-Peg-biotin, while the 140-kDa band showed strong labeling with FP-Peg-biotin (Fig. 4B, lane 1; arrows indicate corresponding bands in the two blots). Labeling of these three bands was inhibited by DFP (Fig. 4B, lane 2). The supernatant of mock-transfected cells showed a completely different labeling pattern with FP-Peg-biotin: the most evidently labeled band was a 150-kDa band, while no labeled bands were visible at 400 and 140 kDa (Fig. 4B, lane 3). Given the complex pattern of labeling with FP-Peg-biotin in supernatants, we decided to perform FP-Peg-biotin labeling on partially purified Reelin. The supernatant of the stable cell line CER, expressing high levels of Reelin, was concentrated and partially purified by gel filtration chromatography. The Reelin-positive eluate from the Superdex 200 gel filtration column was concentrated and then incubated in the absence or presence of the serine protease inhibitor phenylmethylsulfonyl fluoride (PMSF). The samples were separated by SDS-PAGE on a 4 -12% gradient gel, blotted, and incubated with the monoclonal antibody E4 to reveal Reelin (Fig. 5A, lanes 1 and 2). The blot was then stripped and incubated with streptavidin to reveal the binding to FP-Peg-biotin (Fig. 5A, lanes 3 and 4). As with the crude supernatant (Figs. 2 and 4), the immunoblot showed two major Reelin-positive bands, one higher than the 250-kDa marker, corresponding to the 400-and 300-kDa isoforms, and a smaller band at about 140 kDa (Fig. 5A, lanes 1 and 2). FP-Peg-biotin binds to both major Reelin bands in the absence, but not in the presence of PMSF (Fig. 5A, lanes 3 and 4). The smaller band appeared to bind FP-Peg-biotin with a higher affinity than the higher molecular weight isoform (Fig. 5A, lane 3).
Reelin Shows Protease Activity on Extracellular Matrix Proteins in Vitro-To investigate the catalytic activity of Reelin, we first further purified the high molecular weight isoforms by SDS-PAGE and electroelution. As shown in Fig. 5B, separation of the concentrated cell culture supernatant using a 5% SDS gel resulted in a good separation of the Reelin isoforms. Silver nitrate staining indicated that only the supernatant of Reelin expressing cells contains a band at approximately 400 kDa corresponding to the Reelin isoforms of 400 (and 300) kDa (Fig.  5B, lanes 2 and 4). Therefore, the high molecular weight Reelin band was electroeluted from the gel to achieve a high degree of purification. The electroeluted sample was reanalyzed by SDS-PAGE and Western blotting (Fig. 5B, lane 5). However, we found that the purified high molecular weight Reelin protein quickly disappeared and smaller bands appeared around 180 and 140 kDa, probably corresponding to self-degradation products. The major proteolytic product of 140 kDa that we observed in this study may correspond to the 180-kDa degradation product that has been described by other investigators (8,35).
To test for proteolytic activity on extracellular matrix proteins, purified Reelin was incubated with pure fibronectin, laminin, or collagen IV, and breakdown products were analyzed by SDS-PAGE and silver staining of gels. Fibronectin and laminin breakdown fragments were seen already after 10 min incubation (Fig. 6, A and D), while collagen IV was degraded at a much slower rate (Fig. 6E). Fibronectin degradation was blocked by inhibitors of serine proteases (DFP, PMSF, and aprotinin), but not by inhibitors of other families of proteases (Fig. 6B). Fibronectin degradation was also partially inhibited by monoclonal antibody CR-50 (Fig. 6C), an antibody directed against the NH 2 -terminal portion of Reelin that has been demonstrated to inhibit Reelin function both in vitro and in vivo (see Ref. 31 and references quoted therein). Inhibition of fi-bronectin degradation was seen at a CR-50 concentration of 9.8 g/ml, the highest concentration tested. Interestingly, this concentration is comparable with the concentrations that have been reported in the literature to inhibit Reelin function Ref. 31). DISCUSSION In this paper we present converging biochemical and cellular evidence that Reelin is a serine protease of the extracellular matrix, and that its enzymatic activity is important for the modulation of cell adhesion. The fact that CR-50, a monoclonal antibody known to inhibit Reelin function both in vitro and in vivo, blocks the proteolytic activity of Reelin further supports the hypothesis that proteolytic activity is of fundamental importance for the function of Reelin. These findings appear interesting in view of the fact that serine proteases, such as tissue plasminogen activator (tPA), are already known to be important modulators of cell migration and axon growth (32).
HEK 293T cells adhere quickly to fibronectin, due to endogenous expression of ␣ 5 ␤ 1 integrin (28), which is a selective fibronectin receptor (33). In this paper we show that expression of Reelin leads to a marked decrease of adhesion of HEK 293T cells to fibronectin. These data, together with the demonstration that purified Reelin degrades fibronectin in vitro, are consistent with the hypothesis that Reelin, secreted by HEK 293T cells, inhibits cell adhesion by degrading the fibronectin substrate. Alternatively, Reelin might activate other targets, for example, cell membrane receptors or other proteases, which are in turn directly responsible for cell detachment. A third hypothesis is that Reelin induces the expression of another serine protease, which is in turn responsible for cell detachment. This hypothesis appears remote, since we have demonstrated that Reelin binds FP-Peg-biotin, and degrades fibronectin in vitro.
We found that, after purification, Reelin appears to undergo rapid self-degradation. Our data suggest that the major 140-kDa fragment is enzymatically active, since its binding to FP-Peg-biotin is even stronger than that of full-length Reelin. Interestingly, we also observed strong labeling with FP-Pegbiotin of the smaller fragment after immunoaffinity purification of Reelin from mouse brain. 2 These data support the idea that the proteolytic processing of Reelin is functionally important, and that full activity of Reelin might require degradation of the 400-kDa full-length precursor to generate smaller, more active isoforms.
Reelin appears to behave as a specific serine protease, as collagen IV is degraded at a much slower rate than fibronectin or laminin. However, this hypothesis needs further confirmation, using model peptide substrates.
Reelin has been suggested to allow migrating neurons to grow past previously migrated cells and to promote detachment of neurons from radial glial fibers (6 -8). The ␣ 5 integrin subunit, expressed in HEK 293T cells, shows high homology with the ␣ 3 subunit, which is expressed on migrating neurons, and appears to be involved in the inhibitory effect of Reelin on neuronal migration along radial glial processes. Reelin has been demonstrated to bind to ␣ 3 ␤ 1 integrin (8). In situ hybridization experiments and double immunolabeling with antibodies against fibronectin and antibodies against radial glia demonstrate transient fibronectin expression on radial glia processes during early stages of cortical development, until completion of corticogenesis (34). On the basis of the available evidence, we propose that ␣ 3 ␤ 1 integrin might immobilize extracellular Reelin on the surface of migrating neurons and thus focus its proteolytic activity on fibronectin expressed on radial glial cells. Alternatively, binding to ␣ 3 ␤ 1 integrin might enhance the activity of Reelin by protecting it from degradation. In this respect it is interesting to notice that elevated levels of cleaved Reelin have been detected in the absence of ␣ 3 ␤ 1 integrin. This finding has been interpreted as evidence that ␣ 3 ␤ 1 integrin inhibits degradation of Reelin by modulating the activity of a zinc-dependent metalloproteinase (8,35). Our data are consistent with the simpler hypothesis that the appearance of Reelin fragments is the result of a self-degradation activity, and that ␣ 3 ␤ 1 integrin might protect Reelin from self-degradation.
Our data are consistent with a role for integrins in Reelinmediated cell adhesion. However, the action of Reelin on neuronal migration may be mediated by other pathways, such as the very low density lipoprotein receptor/apoE receptor 2/disabled-1 pathway (17)(18)(19). It is conceivable, for example, that Reelin may cleave these lipoprotein receptors and thereby activate the disabled-1 signaling pathway. Furthermore, in our cell culture model the action of Reelin is incompletely blocked by DFP, indicating that Reelin might also act via pathways that are independent of serine protease activity. Finally, it should be noted that ␣ 5 ␤ 1 , the integrin expressed by HEK 293T cells, is a highly selective receptor for fibronectin (33). Thus, HEK 293T cells might be much more sensitive to integrinfibronectin interactions than cortical neurons in the developing brain, where other cell adhesion complexes might be predominant.
The discovery of Reelin's function as a serine protease of the extracellular matrix is also intriguing when considering the expression and distribution of Reelin and adhesion molecules of the extracellular matrix in structures other than the developing cerebral cortex. Reelin is co-expressed with integrins in dendritic spines of GABAergic neurons in the adult cerebral cortex, and of glutamatergic neurons in the cerebellum and olfactory bulb (10,11,36). Dendritic spine density is decreased in fronto-parietal cortex and CA1 pyramidal neurons of heterozygous reeler mice (37), and heterozygous reeler mice show abnormalities in complex behavior, like neophobia and increased anxiety (38), a finding that led investigators to propose these mice as an animal model of schizophrenia. We suggest that the synaptic role of Reelin as a serine protease might consist in a rapid and local modulation of adhesive forces between pre-and post-synaptic elements, thus modulating the efficiency of synaptic transmission at the local level. In fact, other serine proteases, like tPA and plasmin, have already been suspected to be involved in synaptic plasticity. tPA contributes to the late phase of long-term potentiation in hippocampal slices and stimulates synapse formation in hippocampal cell culture (39). Plasmin cleaves laminin and appears to regulate long-term potentiation (40). In conclusion, our findings may help to better understand the roles of Reelin both in physiology and in disease, by bringing this protein into the complex and exciting scenario of protease-regulated signaling networks.