Activation of NF-κB by Extracellular Matrix Is Involved in Spreading and Glucose-stimulated Insulin Secretion of Pancreatic Beta Cells*

Laminin-5-rich extracellular matrix derived from 804G cells (804G-ECM) engages β1 integrins to induce spreading, improve glucose-stimulated insulin secretion (GSIS), and increase survival of pancreatic beta cells. The present study examines whether 804G-ECM activates the transcriptional activity of NF-κB and the involvement of NF-κB in those effects of 804G-ECM on pancreatic beta cells. 804G-ECM induces nuclear translocation and the DNA binding activity of the p65 subunit of NF-κB. 804G-ECM-induced nuclear translocation of NF-κB was weak as compared with that induced by interleukin-1β. Transient 804G-ECM-induced DNA binding activity of NF-κB (peak at 2 h) and overexpression of NF-κB target genes IκBα and NF-κB1(p105) (peak at 4 h) were observed. When NF-κB was inhibited by an inhibitor of IκBα phosphorylation (Bay 11-7082) or by a recombinant adenovirus expressing the nonphosphorylatable form of IκBα, 804G-ECM-induced cell spreading and actin cytoskeleton organization were reduced. GSIS from cells on 804G-ECM was inhibited 5-fold, whereas cell survival was not affected. In summary, the results indicate that 804G-ECM induces a transient and moderate NF-κB activity. This study shows for the first time that ECM-induced NF-κB activity is necessary in maintaining GSIS, although it does not affect survival of pancreatic beta cells. The effects of ECM-induced NF-κB activity contrast with the deleterious effects of cytokine-induced NF-κB activity. It is proposed that transient and moderate NF-κB activity is essential for proper function of the pancreatic beta cell.

Engagement of cells by cognate components of the extracellular matrix (ECM) 1 is crucial for various biological processes, including cell adhesion, spreading, proliferation, differentia-tion, migration, apoptosis, and gene induction (1). This engagement involves cell adhesion mediated by integrins, a family of heterodimeric molecules composed of an ␣ and a ␤ subunit, with a long extracellular domain binding to the ECM, and a short cytoplasmic domain associating with the actin cytoskeleton and affiliated proteins (2). ECM is an important component of the pancreatic islet microenvironment. It is a dynamic complex of different molecules that serve as a cellular scaffold regulating both differentiation and survival.
Our group has reported recently (3) that the laminin-5-rich ECM secreted by 804G cells (804G-ECM) induced increased spreading and improved insulin secretion in response to glucose in purified pancreatic beta cells. It was subsequently shown that this ECM has a beneficial effect on cell survival and that it activates intracellular signaling pathways involving the signaling proteins focal adhesion kinase, Akt/PKB, and ERK (4). It has been shown that its effects on pancreatic beta cells (spreading, glucose-stimulated insulin secretion, and survival) are mediated by the engagement of ␤1 integrins to laminin-5 (3,4). 2 Furthermore, it was found that plating pancreatic beta cells on 804G-ECM causes overexpression of the inhibitor of NF-B (IB␣) (4), which is a well established NF-B target gene. This finding led us to hypothesize that 804G-ECM might activate the NF-B signaling pathway, leading to the expression of a subset of genes, which might be involved in the effects of 804G-ECM on the pancreatic beta cell. Engagement of integrins to different ECM molecules (vitronectin, fibronectin, and laminin) has been shown to activate NF-B (5-9) and to promote an NF-B-dependent program of gene expression (10,11). However, this has not yet been reported for the pancreatic beta cell nor indeed for primary, nondividing cells.
NF-B is a transcription factor that plays a pivotal role in many cellular responses to environmental changes, including control of the host immune and inflammatory response, apoptosis, cell cycle progression, cell migration, development, induction of proliferation, and differentiation (12,13). In nonstimulated cells, NF-B remains in the cytoplasm, where it is bound to and inactivated by inhibitory IB subunits (14). Upon stimulation of the cell (with pro-inflammatory cytokines for example), signaling pathways lead to phosphorylation of the IB proteins that are subsequently recognized by the ubiquitin ligase machinery and are targeted for degradation. The freed NF-B heterodimers translocate to the nucleus, where they bind to specific sequences in the promoter or enhancer regions of target genes (reviewed in Ref. 12). Activated NF-B can then be down-regulated by multiple mechanisms, including the well characterized feedback pathway involving newly synthesized IB␣ proteins (12,15,16).
In the pancreatic beta cell, NF-B is considered as an important transcription factor mediating IL-1␤-induced signal transduction and regulating groups of genes contributing to death and dysfunction (17). Indeed, inhibition of NF-B has been shown to protect against the deleterious effects of IL-1␤ in purified mouse islets (18), rat beta cells (19), and human islets (20,21). Given that NF-B has thus been considered previously to be pro-apoptotic in the beta cell, the possibility that activation of NF-B signaling by 804G-ECM may be beneficial for beta cell function and/or survival was quite intriguing. The aims of this study were to investigate whether 804G-ECM activates the NF-B transcriptional activity in primary pancreatic beta cells, to study the kinetics of this activity, and to determine the involvement of this pathway on the effects of 804G-ECM on the pancreatic beta cell. Here we show that 804G-ECM induces transient nuclear translocation of NF-B and its transcriptional activity in pancreatic beta cells, which is followed by overexpression of IB␣ and NF-B mRNAs. Blocking the MAP kinase ERK pathway with PD98059 inhibited the DNA binding activity of NF-B and overexpression of IB␣, whereas the PI3K inhibitor LY294002 did not. We report, for the first time, that long term blockage of NF-B activity disrupts cytoskeleton remodeling, inhibits spreading induced by 804G-ECM, and reduces glucose-stimulated insulin secretion. Furthermore, we show that induction of NF-B activity by 804G-ECM is not involved in its control of pancreatic beta cell survival.

MATERIALS AND METHODS
Reagents and Antibodies-Bay 11-7082 (an inhibitor of IB␣ phosphorylation) was purchased from BioMol Research Laboratories (Hamburg, Germany). PD98059 (a specific MEK1 inhibitor) and LY294002 (a phosphatidylinositol 3-kinase (PI3K) inhibitor) were purchased from Calbiochem, and both were used at final concentration of 50 M as described previously (4). The polyclonal antibodies against the p65 subunit of NF-B (C-20) and against the inhibitors of NF-B (IB␣, C-21; IB␤, C-20) were from Santa Cruz Biotechnology (Santa Cruz, CA). Monoclonal anti-actin was purchased from Chemicon International (Temecula, CA). Anti-mouse horseradish peroxidase and anti-rabbit horseradish peroxidase antibodies were from Amersham Biosciences. Alexa Fluor® 546 phalloidin was purchased from Molecular Probes.
Islet Isolation and Beta Cell Purification-All experiments were performed on primary pancreatic beta cells sorted from adult rat islet cells by autofluorescence-activated flow cytometry. Islets of Langerhans were isolated by collagenase digestion of pancreas from male Wistar rats (weighing 150 -200 g), followed by Ficoll purification using a modification of procedures described previously (22). Cell preparation was performed as described previously (3). Pancreatic beta cells were then separated from non-beta cells by autofluorescence-activated sorting using a fluorescence-activated cell sorter, FACStar-Plus (BD Biosciences), as described previously (22). This purification procedure yields a population consisting of Ͼ95% beta cells (3).
804G-ECM Matrix Preparation-The 804G cells were the kind gift of Desmos (San Diego, CA). They were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 5.6 mM glucose. At confluence, cells were rinsed and maintained for a further 3 days in the same medium in the absence of fetal calf serum. Conditioned medium (referred to hereafter as 804G-ECM) was collected, centrifuged at 120 ϫ g for 10 min to remove any detached cells and debris, filtered through a 0.22-m Millipore filter, and frozen at Ϫ20°C for later use.
Coating of Plastic Dishes with pLL and 804G-ECM-Aliquots (60 l) of pLL (0.1 mg/ml) or of 804G-ECM were layered at the center of 35-mm culture Petri dishes (adherent dishes for mammalian cell culture). Dishes were kept in a damp box at 37°C for 18 -20 h before being rinsed three times with sterile H 2 O and air-dried. Dishes coated with pLL were used as controls.
Cell Culture-Sorted beta cells were washed twice in 10 -15 ml of sterile Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal calf serum, 11.2 mM glucose, 110 g/ml sodium pyruvate and supplemented with 110 units/ml penicillin, 110 g/ml streptomycin, and 50 g/ml gentamycin, followed by centrifugation for 10 min at 130 ϫ g. Aliquots of 3 ϫ 10 5 cells were seeded in nonadherent 100-mm diameter Petri dishes containing 9 ml of medium. Cells were then incubated for 20 h at 37°C to allow full recovery of any cell surface molecules that may have been lost or damaged during islet isolation or cell purification. After recovery, cells were resuspended at a density of 4 ϫ 10 5 cells/ml, and aliquots of 50 l were plated as droplets on plastic dishes coated with pLL or with 804G-ECM.
NF-B Activation-The cellular localization of NF-B in various conditions was analyzed by immunofluorescence. Cells were fixed with 4% paraformaldehyde (20 min, room temperature) and permeabilized with 0.5% Triton X-100 (5 min, room temperature), and immunofluorescence for NF-B was performed. The percentage of cells with dominant nuclear NF-B staining was determined (five different fields were examined for each condition). DNA binding activity of NF-B was quantified using an ELISA-based kit (Trans-AM NF-B p65 from Active Motif, Rixensart, Belgium) using attached oligonucleotides corresponding to an NF-B consensus site and detected by an anti-p65 subunit antibody according to the manufacturer's instructions.
Western Blot Analysis-To analyze IB␣ and IB␤ protein expression, attached cells were washed with ice-cold PBS supplemented with 1 mM sodium vanadate and protease inhibitors and lysed in sample buffer 1ϫ (62 mM Tris-Cl, pH 6.8, 2% SDS, 5% glycerol, 1% 2-mercaptoethanol). Protein concentrations were determined with the Amido Black method (23), and equal amounts of total protein were loaded for SDS-PAGE. All samples, after separation on an SDS-polyacrylamide gel, were electroblotted onto nitrocellulose membranes (Schleicher & Schuell) for immunoblotting with the appropriate antibody. An ECL protein detection kit (Amersham Biosciences) and a Kodak image station were used for visualization of the bands.
Quantitative Real Time PCR-RNA was isolated using RNeasy mini kit (Qiagen). cDNA was synthesized with Superscript II (Invitrogen), using 1 g of total RNA in a 20-l reaction volume. The double-stranded DNA-specific dye SYBR Green I (Eurogentech, Belgium) and fluorescein (Bio-Rad) were incorporated into the PCR buffer (qPCR core kit, Eurogentech) to allow for quantitative detection of the PCR product. The results were analyzed using the iCycler iQ System (Bio-Rad). The housekeeping gene L3 was used as an internal control. The primers used were as follows: NF-B forward 5Ј GTC TAG CAA TCA CGG CTG CA 3Ј, NF-B reverse 5Ј CTC AAG CCA CCA TAC CCC AA 3Ј; IB␣ forward 5Ј TGC TGA GGC ACT TCT GAA AGC 3Ј, IB␣ reverse 5Ј TCC TCG AAA GTC TCG GAG GTC 3Ј.
Inhibition of NF-B Activity-Two approaches were used to inhibit NF-B activity as follows: use of recombinant adenoviruses expressing GFP (control virus) or mutated nondegradable IB␣ (IB␣np), and pharmacological inhibition of IB␣ phosphorylation with Bay 11-7082 (24). The recombinant adenoviruses were a kind gift from Dr. Wrede, University of Regensburg, Germany (19,25). The adenoviruses were titrated using Adeno-X TM Rapid Titer kit (Clontech) according to the manufacturer's instructions. Cells (10 5 cells/ml) were infected in suspension for 2 h at 37°C with 2300 infectious units/ml, washed twice with culture medium, and left in suspension for 24 h before plating them on pLL or on 804G-ECM-coated dishes. To investigate the effect of the inhibition of NF-B activity using Bay 11-7082, cells were pretreated for 1 h with 2.5 and 5 M Bay 11-7082 before plating them on pLL-or on 804G-ECM-coated dishes. Me 2 SO was added to the controls at the same final concentration as used for the inhibitors. Cells were then cultured in the continued presence of the inhibitor. The analyses of the effects of IB␣np-expressing virus and of Bay 11-7082 on NF-B nuclear translocation, cell spreading, cell function (insulin secretion), and cellular death were performed in parallel.
Measurement of Spreading of Cells-After 24 h of culture and treatment of cells as described above, cells were fixed with 4% paraformaldehyde (20 min, room temperature) and stained with Evans Blue. Phasecontrast views of different fields were photographed, and spreading of cells was quantified using ScionImage TM software (Frederick, MD).
F-actin Cytoskeleton Visualization-After 24 h of culture and treatment of cells as described above, cells were fixed with 4% paraformaldehyde (20 min, room temperature) and permeabilized with 0.5% Triton X-100 (4 min at room temperature). After blocking with PBS ϩ 0.5% bovine serum albumin, cells were incubated for 30 min with Alexa Fluor® 546 phalloidin (5 units/ml), subsequently rinsed, and mounted under glass coverslips. The preparations were observed with a confocal microscope.
Analysis of Cell Death by TUNEL-Attached cells were washed with PBS and fixed with 4% paraformaldehyde (20 min, room temperature). After permeabilization with 0.5% Triton X-100 (5 min, room temperature), the TUNEL assay (with fluorescein-dUTP detecting the free 3-OH strand breaks resulting from DNA degradation) was performed with the In Situ Cell Death Detection kit, according to the manufacturer's instructions (Roche Applied Science). The preparations were then rinsed with PBS and incubated (15 min, room temperature) with 1 g/ml Hoechst 33342 to allow detection of nuclei and to facilitate the analysis. The quantification of dead cells was performed by using an Axiocam fluorescence microscope.
Insulin Secretion Assay and Insulin Content-Cells were washed three times with a modified Krebs-Ringer bicarbonate HEPES buffer (KRBH: 125 mM NaCl, 4.74 mM KCl, 1 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 5 mM NaHCO 3 , 25 mM HEPES, pH 7.4, 0.1% bovine serum albumin) supplemented with 2.8 mM glucose and preincubated with this same buffer for 1 h at 37°C. Cells were then incubated for 1 h at 37°C with KRBH containing 2.8 mM glucose (basal secretion), followed by 1 h at 37°C with KRBH containing 16.7 mM glucose (stimulated secretion). The incubation buffer was recovered, and the cells were extracted with acid/ethanol. The buffer was centrifuged to remove any detached cells and debris. Aliquots were stored at Ϫ20°C for subsequent insulin measurement performed by radioimmunoassay using the charcoal separation technique described previously (26). GSIS is expressed as secreted insulin upon glucose challenge (16.7 mM glucose, for 1 h), as a percentage of insulin content. Insulin content is expressed as the quantity of insulin extracted at the end of the stimulation test. The total protein levels after 24 h of culture in all conditions used were quantified using the Amido Black method (23), as described above.
Presentation of Data and Statistical Analysis-Unless stated otherwise, data are presented as mean Ϯ S.E. for "n" independent experiments, and levels of significance for differences between groups were assessed by Student's t test for unpaired groups.

804G-ECM Induces NF-B Nuclear Translocation-This
study was performed with laminin-5-rich extracellular matrix derived from 804G rat bladder carcinoma cells (804G-ECM). 804G-ECM induces attachment and spreading of rat primary beta cells, because of engagement of ␤1 integrins by laminin-5 contained in this ECM (3,4). 804G-ECM is thus used as a validated model to study signaling pathways activated by this engagement.
In the canonical NF-B pathway, one of the first steps involved in NF-B-dependent gene regulation is the phosphorylation and subsequent degradation of the inhibitor of NF-B (IB␣)), leading to the nuclear translocation of the p65 subunit of NF-B (RelA). To investigate a possible involvement of 804G-ECM in NF-B transcriptional activity, the cellular localization of NF-B after short term exposure (1 h) to 804G-ECM compared with control (pLL-coated dishes) was assessed by immunofluorescence for the p65 subunit of NF-B (Fig. 1A). The number of cells with nuclear NF-B localization was quantified (Fig. 1B). 804G-ECM induced a significant increase in the number of cells with nuclear NF-B (19%) compared with control (less than 5%, Fig. 1B). The cytoplasmic staining for NF-B was still apparent in all cells on 804G-ECM, suggesting that only a fraction of the total pool of cytoplasmic NF-B is translocated into the nucleus in response to 804G-ECM (Fig. 1A). By contrast, treatment of cells for 20 min with 2 ng/ml IL-1␤ induced the nuclear translocation of p65 in all treated cells, and this was complete as there was no more cytoplasmic staining for NF-B (not shown).
804G-ECM Induces NF-B Transcriptional Activity-To assess whether DNA binding activity of NF-B is induced by 804G-ECM, the amount of p65 complexes binding to oligonucleotides containing an NF-B consensus binding site was quantified by ELISA. The DNA binding activity of NF-B in cells cultured on 804G-ECM-coated dishes as compared with control was increased after 1 h of culture, and this increase was significant after 2 h of culture (Fig. 2). This activity was decreased after 4 h of culture suggesting that 804G-ECM-induced NF-B DNA binding activity is transient.
It is well established that activated NF-B induces expression of IB␣. Newly synthesized IB␣ binds to NF-B, provid-ing a negative feedback loop for the NF-B signaling pathway (12). Activated NF-B is also able to induce the expression of NF-B1 (p105) in several systems (11,27). To investigate whether 804G-ECM is able to induce NF-B transcriptional activity, IB␣ and NF-B1 mRNA levels were quantified by real time PCR in cells cultured on pLL-or on 804G-ECM-coated dishes. IB␣ mRNA levels were increased after 2 h on 804G-ECM, and both IB␣ and NF-B1 mRNA levels were significantly increased by 804G-ECM after 4 and 8 h of culture to return to control (pLL) levels at 24 h (Fig. 3, A and B). The kinetics of 804G-ECM-induced IB␣ and NF-B1 gene expression thus correlate with those of NF-B DNA binding activity (Fig. 2). In summary, these data indicate that 804G-ECM induces NF-B transcriptional activity transiently, most probably as a consequence of feedback inhibition resulting from increased expression of IB␣.
Regulation of IB␣ and IB␤ Protein Levels by 804G-ECM-To investigate whether 804G-ECM induces IB␣ and/or IB␤ degradation, Western blots were performed with protein extracts from cells cultured for 1 h on pLL or on 804G-ECMcoated dishes. Quantification of the resulting bands normalized to actin showed that there was no significant decrease of IB␣ and IB␤ protein levels on 804G-ECM as compared with pLL after 1 h of culture (Fig. 4).
IB␣ and IB␤ protein levels were also analyzed after 24 h of culture on pLL-or on 804G-ECM-coated dishes. Levels of IB␣ protein were significantly increased on 804G-ECM after 24 h of culture as compared with pLL and as compared with 804G-ECM after 1 h of culture (Fig. 4). By contrast, IB␤ protein levels were not altered with time or by the substrate used. Although NF-B induces expression of IB␣, it does not control expression of IB␤ (28). Therefore, these results suggest that the 804G-ECM-induced IB␣ overexpression occurs specifically through NF-B.
Inhibition of NF-B Activity; Effects on Spreading and Actin Cytoskeleton Remodeling-To investigate the function of the 804G-ECM-induced NF-B activity, Bay 11-7082, an inhibitor of IB␣ phosphorylation and thus of NF-B nuclear translocation (24), was used. Bay 11-7082 induced a dose-dependent decrease of 804G-ECM-induced NF-B nuclear translocation, and a final concentration of 5 M of Bay 11-7082 was necessary to induce a complete and significant decrease compared with control condition (Fig. 5A). Similar results were observed when the effect of this inhibitor on 804G-ECM-induced NF-B DNA binding activity was assessed by ELISA (not shown). Furthermore, Bay 11-7082 inhibited overexpression of IB␣ mRNA induced by 804G-ECM, thus confirming that it inhibits the transcriptional activity of NF-B (Fig. 8B). We have shown previously (3) that 804G-ECM induces spreading of the pancreatic beta cell, and this effect is glucose-dependent. NF-B has been shown to be involved in cell motility through its effect on gene expression (reviewed in Ref. 29), leading us to hypoth-esize that 804G-ECM-induced NF-B activity might be involved in spreading of beta cells by 804G-ECM. Treatment of cells with Bay 11-7082 significantly reduced spreading of cells on 804G-ECM as assessed after 24 h of culture (Fig. 5B).
In order to visualize actin cytoskeleton organization in cells after 24 h of culture, F-actin was labeled with phalloidin. As reported by our group (30), and as shown in Fig. 5C, cells cultured on 804G-ECM organize their actin to form distinct fibers in the cell interior, whereas no clear actin organization could be seen in cells plated on pLL. When treated with Bay 11-7082, the intracellular actin fibers were disrupted in cells cultured on 804G-ECM. Although evaluation of cortical actin organization can be hampered by the shape of a cell, and cells on 804G-ECM in the presence of the inhibitor were less spread, close inspection of the confocal images (Fig. 5C) suggests a better developed and more consequent cortical actin ring in the treated cells.
Recombinant adenovirus expressing nondegradable IB␣ (IB␣np) has been shown to be an effective inhibitor of NF-B nuclear translocation and transcriptional activity (19,20). 804G-ECM-induced nuclear translocation of NF-B was decreased in cells transduced with the recombinant IB␣np virus as compared with cells transduced with the control (GFP) adenovirus (data not shown). The effect of IB␣np on cell spreading was analyzed after 24 h of culture on pLL or on 804G-ECM. Spreading of cells infected with IB␣np was reduced compared with cells infected with the control virus (Fig. 6), confirming that 804G-ECM-induced NF-B activity is involved in the spreading of pancreatic beta cells induced by 804G-ECM.
Inhibition of NF-B Activity; Effects on Insulin Secretion-The importance of cell-matrix interactions for optimal GSIS (insulin secreted at 16.7 mM glucose) by pancreatic beta cells in culture has been demonstrated by us (3) and by other groups (31)(32)(33)(34)(35)(36). However, the mechanism involved has yet to be clarified. It has been suggested that there is a correlation between increased spreading of cells with their ability to respond optimally to glucose (3). As NF-B activity seems to be involved in spreading of cells and organization of the actin cytoskeleton induced by 804G-ECM (Figs. 5 and 6), we hypothesized that it might be involved in the beneficial effects of 804G-ECM on GSIS. As shown in Fig. 7A and in Table I Table I). Furthermore, the fold stimulation of insulin secretion (amount of insulin secreted at 16.7 mM glucose compared with that secreted at 2.8 mM glucose) was significantly decreased in treated cells, both on pLL and on 804G-ECM (Table I). In addition, Bay 11-7082 induced a significant decrease in "absolute" stimulated insulin secretion (i.e. amount of secreted insulin at 16.7 mM glucose, per dish and hour) as well as an increase in total insulin content of cells plated on 804G-ECM, as compared with the control condition (Table I). Cellular protein content was similar in all conditions (Table I). Bay 11-7082 thus inhibits insulin secretion. In summary, these results indicate that 804G-ECM-induced NF-B activity is involved in spreading, actin cytoskeleton organization, and GSIS of pancreatic beta cells.
Inhibition of ECM-induced NF-B Activity Does Not Affect Cell Survival-804G-ECM protects pancreatic beta cells against apoptosis (4), so we hypothesized that NF-B might mediate its pro-survival effect. TUNEL assays were performed on cells cultured for 24 h on pLL or on 804G-ECM-coated dishes, in the absence or in the presence of Bay 11-7082. Treatment of cells with 5 M Bay 11-7082 did not affect cell survival either on pLL or on 804G-ECM (Fig. 7B). Similar results were obtained using adenoviruses expressing nonphosphorylatable IB␣ (data not shown). Therefore, it seems that 804G-ECMinduced activity of NF-B is not involved in the control of survival and/or apoptosis of pancreatic beta cells.

Involvement of the MAP Kinase ERK Pathway in 804G-ECMinduced NF-B Activity-We have shown previously that 804G-ECM induces phosphorylation of the MAP kinases ERK1
and ERK2, as well as of Akt/PKB, and that both pathways seemed to be involved in the anti-apoptotic effect of 804G-ECM (4). The signaling pathways involving ERK and Akt/PKB have both been reported to be able to activate NF-B (5,10,29,37). To get a preliminary insight into how 804G-ECM activates NF-B transcriptional activity, the effects of PD98059 (inhibitor of MEK1 and thereby of the MAP kinase ERK pathway) and of LY294002 (inhibitor of PI3K and thereby of Akt/PKB signaling) on the DNA binding activity of NF-B and on the overexpression of IB␣ induced by 804G-ECM were analyzed. As shown in Fig. 8A, PD98059 inhibited the NF-B DNA binding activity by 40% (p Ͻ 0.0005). We have shown previously that PD98059 inhibits the overexpression of IB␣ induced by 804G-ECM (4). Taken together, these results suggest that the MAP kinase ERK pathway might be involved in the activation of NF-B induced by 804G-ECM. By contrast, the inhibitor LY294002 (50 M) had no effect either on the DNA binding activity of NF-B (Fig. 8A) or on the expression of IB␣ (Fig.  8B), suggesting that the PI3K-Akt/PKB pathway is not involved in the regulation of the transcriptional activity of NF-B by 804G-ECM. DISCUSSION In this report we show that 804G-ECM induces nuclear translocation of NF-B, that it increases binding of NF-B to DNA, and that it induces overexpression of the well established NF-B target genes NF-B1 (p105) and IB␣. Collectively, these data indicate that 804G-ECM induces NF-B transcriptional activity in primary pancreatic beta cells. Others have reported that various ECMs are able to activate NF-B in different cell types; however, most studies have been performed in cell lines (transformed or not), and to our knowledge there has been no such study in primary, nondividing cells.
In the canonical pathway leading to NF-B activation, others have observed a substantial degradation of IB␣ prior to NF-B nuclear localization. In our system, the degradation of IB␣ after 1 h of culture on 804G-ECM was not significant as compared with control (pLL), suggesting that only a small fraction of the total IB␣ pool is degraded. The consequence would be that only a minor fraction of NF-B cytosolic dimers is translocated to the nucleus. The fact that the cytoplasmic staining for NF-B was still apparent in all cells exposed for 1 h to 804G-ECM sustains the latter hypothesis.
After 24 h of culture, IB␣ protein levels were significantly increased on 804G-ECM as compared with pLL, and this increase was sustained for at least 48 h (4). This prolonged overexpression of IB␣ is most interesting, because increased IB␣ expression will both prevent additional NF-B nuclear translocation and remove NF-B from its nuclear binding sites (12) and could thus be responsible for the observed transience of 804G-ECM-induced NF-B activity.
Several different signaling proteins have been reported to mediate signaling from ECM to NF-B, including MAP kinase ERK (37), PI3K (5, 10), and Rho GTPase Rac (5,10,38). We have reported previously (4) that the MAP kinase-ERK and PI3K-Akt/PKB pathways are activated by 804G-ECM, suggesting that these latter pathways might be involved in 804G-ECM-induced NF-B activity. Based on the use of selective inhibitors, we now show that the MAP kinase ERK pathway, but not the PI3K-Akt/PKB pathway, mediates the increased DNA binding activity of NF-B as well as overexpression of IB␣ induced by 804G-ECM.
As reported previously by our group, plating beta cells on 804G-ECM induces their attachment, cytoskeleton remodeling, and cell spreading. Here we report that long term blockage of NF-B activity by pharmacological means (Bay 11-7082) or by adenoviral overexpression of nonphosphorylatable IB␣ impairs 804G-ECM-induced spreading of the cells. This inhibitory effect might depend on expression of genes induced by NF-B. Indeed, it has been reported by other groups that NF-B can mediate overexpression of proteins involved in spreading and/or migration of cells (29).
Furthermore, we have observed that blocking NF-B activity with Bay 11-7082 impairs GSIS, but it does not affect cell survival. While this work was in preparation, Norlin et al. (39) reported the effects of attenuating basal NF-B activity on GSIS and on cell survival, using transgenic mice expressing nonphosphorylatable IB␣. Most interestingly, although their  approach was different from ours, the observed phenotypes because of NF-B inhibition were similar in both models; GSIS was impaired, whereas beta cell survival was not affected.
We have shown previously that the MAP kinase ERK pathway could be involved in the pro-survival effect of the 804G-ECM but that it does not affect spreading of beta cells (4). As discussed above, this pathway seems to be involved in 804G-ECM induction of NF-B activity as well. It is intriguing in this context that inhibition of NF-B did not affect beta cell survival but that it did inhibit spreading of cells. ERK is known to affect the activities of a multitude of signaling pathways and transcription factors (40), regulating migration and survival of cells. Therefore, we propose that ERK effector(s) other than NF-B might mediate its pro-survival effect. Spreading and/or migration of cells is an extremely complex process, involving the interplay of several signaling pathways (41). Further studies are therefore mandatory to better dissect the inter-dependence of the MAP kinase ERK and NF-B pathways as well as other candidate signaling proteins (such as the Rho GTPases) involved in the effects of the 804G-ECM on the pancreatic beta cell.
What is the connection between NF-B and GSIS? Different mechanisms may be involved, and in our opinion, they might well complete each other. One possible connection between NF-B and GSIS is the actin cytoskeleton, and another one is regulation of gene expression. The actin filaments are organized in two ways in secretory cells as follows: a cortical F-actin web (rim or ring) underneath the plasma membrane and actin filament fibers distributed throughout the cytosol (42,43). Many studies, including in pancreatic islets, have reported that the actin web limits the access of the secretory granules to the cell boundary; disruption or remodeling of the web is believed to be a prerequisite for exocytosis (42, 44 -46). On the other hand, the function of actin filaments in the cell interior is less well understood. However, dynamic association of insulin-containing granules with actin cytoskeleton is thought to be involved in insulin exocytosis. As shown in this work, 804G-ECM induces remodeling of the actin cytoskeleton, leading to the appearance of actin filament fibers in the cell interior. Upon treatment with Bay 11-7082, these actin fibers were no more apparent, and this may underlie impaired insulin secretion. The cortical actin web appeared more prominent in the treated cells, and this may further impair insulin secretion. Therefore, we suggest that NF-B might be involved in GSIS through the remodeling of the actin cytoskeleton. Depolarization-induced Ca 2ϩ is a required step for GSIS to occur. Most interestingly, it has been reported that depolarization/Ca 2ϩ influx can activate NF-B transcriptional activity and that the ERK inhibitor PD98059 blocked this activation of NF-B (47). However, the consequences of such depolarization-induced NF-B activity are still unknown (47).
It has been shown that expression of genes implicated in glucose uptake, oxidative metabolism, and Ca 2ϩ -triggered exocytosis is perturbed in transgenic mice expressing nonphosphorylatable IB␣ (39). Furthermore, long term treatment of neuronal cells with tumor necrosis factor-␣ enhances depolarization-induced increases of Ca 2ϩ , and it has been suggested that these effects are induced via altered gene expression mediated by NF-B (48). Therefore, we propose that blockage of 804G-ECM-induced NF-B activity might repress expression of genes necessary for well regulated GSIS.
It is now well acknowledged that NF-B plays a major role in cytokine-induced beta cell dysfunction and apoptosis (17). However, this study shows a completely different aspect of NF-B function in beta cells. A critical question arises: how does a given stimulus lead to specific biological end points through regulation of NF-B activity? The functional consequences of NF-B activity seem to depend on several different factors, such as the kinetics, extent, and context of activation of NF-B. 804G-ECM-induced nuclear translocation of NF-B appeared to be moderate when compared with that occurring after cytokine treatment. Furthermore, we show that 804G-ECM-induced binding of NF-B to DNA declines after 2 h of exposure to 804G-ECM and that 804G-ECM-induced overexpression of IB␣ and NF-B1 (p105) mRNAs begins to decay after 8 h of culture, indicating that 804G-ECM-induced NF-B activity is transient. This contrasts with the sustained activity of NF-B induced by cytokines. 3 These differences in kinetics and extent of NF-B activity may explain the different functional outcomes mediated by 804G-ECM-and cytokine-induced NF-B activation. It has been reported that the first phase of IL-1␤induced NF-B activity leads to the beneficial increase of beta cell defense/repair protein expression. By contrast, the second phase of NF-B activity induced by IL-1␤ is harmful, because it leads to a sustained decrease of specific beta cell proteins like insulin, GLUT-2, and PDX-1 with a concomitant increase of other "aspecific" proteins and inducible nitric-oxide synthase transcription (49). In summary, these data combined with our own data support the concept that transient and/or low NF-B activity is beneficial, whereas sustained and/or strong NF-B activity is deleterious for the pancreatic beta cell.
Recently it has emerged that the transactivation potential of NF-B is modulated by post-translational modifications, including phosphorylation and acetylation of NF-B subunits as well as of histones surrounding NF-B target genes (reviewed in Refs. 50 and 51). Site-specific phosphorylation of the p65 subunit of NF-B can occur by a large variety of kinases in response to different stimuli (51) and may target NF-B to a particular subset of genes (52). We hypothesize that IL-1␤ and 804G-ECM might induce phosphorylation of p65 on distinct sites, leading to different biological responses. Experiments aimed at exploring this hypothesis are underway in our laboratory. The next challenge would be to inhibit selectively the NF-B activity leading to its deleterious effects without interfering with its beneficial outcomes.
The procedure leading to pancreatic beta cell isolation prior to pancreatic islet transplantation involves disruption of cellcell and cell-ECM contacts (33,53). The isolated pancreatic beta cells lose glucose responsiveness and eventually die when maintained in culture for a long period of time. By contrast, when layered on appropriate ECM, pancreatic islet cell function and survival can be maintained (3, 4, 31-33, 35, 36, 54). Here we show that ECM induces transient and moderate activity of the transcription factor NF-B (p65) and that ECMinduced NF-B activity is involved in spreading, cytoskeleton organization, and improved function (GSIS) of primary pancreatic beta cells. We propose that transient and moderate NF-B activity is essential for well regulated glucose-stimulated insulin secretion in the pancreatic beta cell.