IIp45, an insulin-like growth factor binding protein 2 (IGFBP-2) binding protein, antagonizes IGFBP-2 stimulation of glioma cell invasion

Abstract

Our previous studies have shown that insulin-like growth factor binding protein 2 (IGFBP-2) is frequently overexpressed in the highly invasive glioblastoma multiforme (GBM). By using a yeast two-hybrid system, we identified a gene, invasion inhibitory protein 45 (IIp45), whose protein product bound to IGFBP-2 through the thyroglobulin-RGD region of the C terminus of IGFBP-2. The IIp45 gene is located on chromosome 1p36 and has nine exons. The IIp45 protein has three SEG (segment of low compositional complexity) domains and an integrin-binding RGD motif. The IIp45 protein was not expressed in some GBMs. Functional studies showed that IIp45 inhibited GBM cell invasion both in vitro and in xenograft model. Gene expression profiling studies showed that IIp45 consistently inhibited the expression of cell invasion-associated genes, such as the transcriptional NFκB, and its downstream target gene, intercellular adhesion molecule 1. Thus, we report here the isolation and characterization of a gene, IIp45, whose protein product binds to IGFBP-2 and inhibits glioma cell invasion.

Gliomas are the most common type of primary brain tumor. Although the highest grade of glioma, glioblastoma multiforme (GBM), rarely disseminates beyond the CNS, it is extremely invasive within the CNS. The fact that surgery and radiotherapy are generally ineffective in patients with glioblastomas, and that patients can expect a dismally short median survival of only 8–12 months, underscores the need to better understand glioma invasion and to develop therapies that block this process. Several proteins have already been identified that promote the invasion of GBMs. These proteins include matrix metalloproteinase types 2 and 9, which are activated in many GBMs (1, 2), and epidermal growth factor receptor, whose gene is amplified in >50% of GBMs (3).

During the gene expression profiling of human gliomas, we identified an even more frequent (80%) event in GBMs: the overexpression of insulin-like growth factor binding protein 2 (IGFBP-2) (4). IGFBP-2 is also frequently overexpressed in other cancers at advanced stages (5, 6). This finding suggests that IGFBP-2 and its associated signaling pathways play a fundamental role in tumor malignancy and thus warrant close scrutiny.

Under normal physiological conditions, IGFBP-2 is predominantly expressed in fetal tissues (7). In the CNS, IGFBP-2 is highly expressed in fetal astroglial cells, but, after birth, its expression in this population of cells significantly decreases (8). Initially, IGFBP-2 was thought to mainly bind to IGFs (IGF-I and IGF-II) through its IGF-binding motif, and thereby inhibit IGF-mediated mitogenic activities (9). However, recent studies (10–12) have shown that IGFBP-2 also increases the tumorigenic potential and mitogenesis in some cancer cells. Collectively, these studies therefore suggest that IGFBP-2 possesses multifaceted functions. Identification of the binding partners of IGFBP-2 is therefore a crucial step in understanding IGFBP-2's functions and its associated signaling pathways.

In this study, we used a yeast two-hybrid system to identify a binding partner of IGFBP-2, which we named invasion inhibitory protein 45 (IIp45). The IIp45 protein was not expressed in some GBMs and was posttranslationally modified in most GBMs. Overexpression of IIp45 resulted in decreased invasion of GBM cells, whereas IGFBP-2 overexpression increased cell invasion. These results suggest that IIp45 and IGFBP-2 are opposing regulatory partners in cell invasion. Xenograft model studies showed that tumors formed from IIp45-expressing cells were also less invasive compared to the controls. To gain insight into the downstream molecular events resulting from IIp45 overexpression, we first generated a number of stable cell lines that express IIp45 and then compared their gene expression profiles with those of their parental cells. This process revealed a consistent decrease in the expression of cell invasion-related genes including the transcriptional NFκB and its downstream target gene intercellular adhesion molecule 1 (ICAM-1).

Materials and Methods

Yeast Two-Hybrid Screening and Expression Vector Construction. The human IGFBP-2 cDNA (encoding the mature protein consisting of amino acids 1–289) was inserted into the pGBKT7 bait vector containing the GAL4 DNA-binding domain. The plasmid was used to screen a human fetal brain cDNA library in the pACT2 vector (Clontech). Transformants (5 × 10⁶) were screened and evaluated according to the matchmaker user manual provided by the manufacturer.

To create expression vectors for use in the mammalian cell lines, full-length cDNAs of IIp45 and IGFBP-2 were cloned into the pcDNA3.1/hygromycin and pcDNA3.1/geneticin vectors, respectively. The resultant IIp45 and IGFBP-2 expression vectors were then verified by sequencing.

Cell Culture, Transfection, Normal Adult Brain, and Primary Glioma Tissues. HEK293 cells were purchased from the American Type Culture Collection. The LN-229 GBM cell line, which expresses a barely detectable level of IGFBP-2 protein, was obtained from A. Yung (The University of Texas M. D. Anderson Cancer Center). HEK293 and LN-229 cell lines were maintained, respectively, in DMEM and DMEM/F12 medium supplemented with 10% FCS and incubated in a humidified incubator at 37°C with 5% CO₂. Transfection was performed with a GenePorter transfection system (Gene Therapy Systems, San Diego), which has a transfection efficiency of >40%, as determined by the number of transfected GFP-positive cells counted by using the FACScan flow cytometer (Becton Dickinson). For the generation of stable IIp45 cell lines, transfected LN-229 cells were selected in the presence of hygromycin (200 μg/ml) for a month. Six stable lines were used for both the cell invasion assays and the microarray analysis. For the generation of stable EGFP/IIp45 lines, the stable EGFP line of the LN-229 cells was first selected with G418 (500 μg/ml) and then transfected with the pcDNA3.1/IIp45 plasmid and selected with hygromycin.

All normal adult brain and primary glioma tissues were obtained from the Brain Tumor Tissue Bank of the M. D. Anderson Cancer Center with the approval of the institutional review board.

Northern Blot Analysis. A multiple normal human tissue Northern blot containing 1 μg of poly(A) RNA in the each lane (Clontech) was hybridized with a ³²P-labeled full-length cDNA of IIp45. Labeling was carried out with the RediPrime II random prime labeling system (Amersham Pharmacia).

Antibodies and Western Blotting. A rabbit polyclonal antibody raised against an IIp45 epitope (46 NSETPSTPETSSTSL 60) was generated and affinity-purified by Bethyl Laboratories (Montgomery, TX). Polyclonal antibodies against IGFBP-2, ICAM-1, NFκB/p52, and β-actin were purchased from Santa Cruz Biotechnology. Western blotting was performed as described (13).

Fusion Protein and GST Pull-Down Assays. A fusion construct, GST-IIp45, was constructed in the expression vector pGEX-2T (Amersham Pharmacia). The fusion protein was then induced in BL21 (DE3) cells by 0.4 mM isopropyl β-d-thiogalactoside for 3hat30°C and purified by using glutathione-Sepharose 4B beads (Amersham Pharmacia) according to the manufacturer's instructions. Binding assays were carried out as described (13) with some minor modifications. HEK293 cells transfected with IGFBP-2 were lysed with lysis buffer (20 mM Tris·HCl, pH 7.4/150 mM NaCl/1 mM EDTA/0.5% Nonidet P-40/10% glycerol) containing protease inhibitors (Sigma) for 30 min at 4°C. Recombinant GST-IIp45 or GST was then mixed with 150 μl of cell lysates and 150 μl of binding buffer (20 mM Tris·HCl, pH 7.4/50 mM NaCl/2 mM DTT/20 mM MgCl₂/0.04% Nonidet P-40/10% glycerol) and incubated for 2 h at 4°C with rotation. The binding fraction was washed three times with the binding buffer and loaded on to an SDS/10% PAGE gel, and immunoblotting with anti-IGFBP-2 antibody was performed.

Coimmunoprecipitation (Co-IP) Assays. Co-IP was performed as described (14). Briefly, HEK 293 cells cotransfected with IGFBP-2 and IIp45 were lysed in the lysis buffer. Control IgG-precleared supernatants were incubated with anti-IIp45 or anti-IGFBP2 antibodies at 4°C overnight, after which protein G-agarose was added and the mixture was incubated at 4°C for 4 h. Beads were washed three times in cold lysis buffer and precipitates for immunoprecipitation (IP)-Western blot assays or protein extracts for Western analyses were boiled in SDS loading buffer and loaded on to SDS/10% PAGE gels, which was followed by immunoblotting using anti-IGFBP-2 or anti-IIp45 antibody.

Cell Invasion and Migration Assays. Invasion assays were performed by using a BioCoat Matrigel invasion chamber with a pore size of 8 μm (Becton Dickinson Biosciences, Bedford, MA) as described (15). All experiments were conducted in triplicate and repeated at least three times.

Cell migration assays were performed by using transwell polycarbonate chambers (Becton Dickinson Biosciences). The lower surface of the membrane was coated with 10 μg/ml fibronectin for 2 h at 37°C, washed with PBS, and then blocked with 1% BSA for 1 h at 37°C. Similar to the invasion assays described above, 1 × 10 ⁴ cells and 0.5% FBS-DMEM/F12 were used for the migration assays. The chamber was incubated for 2 h at 37°C, and cells were then fixed and counted. Assays were performed in triplicate and repeated three times.

Adhesion Assays. One hundred microliters of 1 × 10 ⁵ cells in serum-free DMEM/F12 was plated onto the fibronectin-coated wells of a 96-well plate (Becton Dickinson Biosciences) and incubated for 2 h at 37°C. After incubation, plates were washed with cold PBS and fixed in 5% formaldehyde for 15 min. Fixed cells were stained with a 0.5% crystal violet solution and then solubilized with 1% SDS. Absorbance at 570 nm was measured with an automated microtiter plate spectrophotometer (Dynatech Laboratories, Chantilly, VA) and the percentage of attached cells was normalized to that of the control.

Mouse Brain Xenografts and Microscopic Analysis of Invasion. EGFP- or EGFP/IIp45-expressing cells (1 × 10⁶) suspended in 5 μl of PBS were injected into the caudate nucleus of 6-week-old nude mice by using a guide screw system (16). Three IIp45-expressing clones were analyzed and three mice were used for each cell line. Mice were killed on day 10 after cell injection, whole brain was removed, and 6-μm sections were cut. Sections were observed directly under fluorescence microscopy and adjacent sections were stained with hematoxylin/eosin.

Microarray Analysis. Total RNA was extracted with TRI Reagent (Molecular Research Center, Cincinnati). Gene expression profiling of the IIp45-stable cell lines and parental LN-229 cells and data analysis were performed as described (17, 18). The pathway array, which includes 1,146 functionally known genes, was performed in duplicate, and generated by the M. D. Anderson Cancer Center Genomics Core Laboratory (www3.mdanderson.org/~genomics), was used in this study.

Results

Identification of the IGFBP-2-Binding Protein, IIp45. To identify IGFBP-2-interacting proteins, we performed a yeast two-hybrid screening assay of a human fetal brain cDNA library by using mature IGFBP-2 as bait. Among ≈5 million yeast transformants, we identified 42 candidate clones after three rounds of screening. After we performed sequence analyses to exclude out-of-frame clones, we identified one that had a long ORF encoding a 291-aa protein. The cDNA insert, which was 1.25-kb long, carried a 3′ untranslated poly(A) tail, but lacked a 5′ start codon. A blast search performed 2 months after our initial cloning found that our clone matched a cDNA clone (GenBank accession no. AK024020) that had been sequenced and deposited by the Japanese Human cDNA Sequence project; however, it lacked a 5′ 120-bp noncoding and a 225-bp coding region. We subsequently recovered the 1.6-kb cDNA of this gene from human fetal brain total RNA by using RT-PCR. We later named the gene IIp45. A search of the genomic database revealed that IIp45 was located on chromosome 1p36 and had nine exons (Fig. 1 A). The predicted molecular mass of the protein product was 45 kDa, which was consistent with the size of the protein detected by using an anti-IIp45 antibody in normal adult brain and gliomas (see below). The complete coding region of the IIp45 cDNA was subsequently cloned into the pACT2 vector in-frame with the GAL4 AD domain, and the interaction of IIp45 with IGFBP-2 was confirmed by yeast two-hybrid assays (data not shown).

Fig. 1.

Characterization of the IIp45 gene. (A) Genomic location and structure of IIp45. IIp45 is located on chromosome 1p36 and is comprised of nine exons. (B) Amino acid sequence and domain analysis of IIp45. Three SEG domains and an RGD motif are indicated. (C) Potential phosphorylation sites of IIp45. (D) IIp45 expression in normal tissues. An mRNA blot with multiple human normal tissues (1 μg of mRNA per lane) was probed with IIp45 cDNA labeled with [α-³²P]dCTP. Tissue sources of the poly(A)⁺ mRNA and size markers are indicated on the top and to the left. A 1.6- and a 2.4-kb transcript were detected in most of the normal tissues analyzed and their ratios were shown.

Protein analysis revealed that IIp45 was a highly hydrophilic protein and had three SEG (segment of low compositional complexity) domains and an RGD motif (Fig. 1B). IIp45 carried 42 Ser, 6 Thr, and 1 Tyr potential phosphorylation sites (Fig. 1C). Northern blotting analysis by using a commercial normal human tissue mRNA blot showed the ubiquitous expression of a 1.6- and a 2.4-kb transcript (Fig. 1D). A gene finder search (http://argon.cshl.org/genefinder) of ≈2 kb upstream of the genomic sequence from the first ATG codon of IIp45 cDNA did not identify other potential coding regions, which supported our assessment that the 1.6-kb transcript contains a full-length coding sequence of IIp45. Future studies may show whether the 2.4-kb transcript is a product from an upstream promoter or represents a homologous family member.

To confirm the interaction of IGFBP-2 and IIp45, we performed GST pull-down assays. Purified recombinant GST-IIp45 (Fig. 2A) was mixed with HEK293 cell extract, and bound proteins were analyzed by immunoblotting by using anti-IGFBP-2 antibody. As shown in Fig. 2B, IGFBP-2 was pulled down by GST-IIp45, but not by GST, alone or in the absence of GST-IIp45.

Fig. 2.

Interaction between IIp45 and IGFBP-2. (A) Purified recombinant protein used in the GST pull-down assay. Approximately 0.1–1.0 μg of GST-IIp45 and GST was separated by SDS/PAGE and stained with Coomassie brilliant blue. Markers represent molecular masses of proteins. (B) Direct interaction between IIp45 and IGFBP-2 shown by GST pull-down assays. The pcDNA-IGFBP-2 vector was transfected into HEK293 cells, and protein extracts were mixed with GST, with GST-IIp45, or neither of them. Eluted fractions were separated by SDS/PAGE and the IGFBP-2 was detected by immunoblot analysis using anti-IGFBP-2 antibody. HEK293 lysate was also analyzed to confirm the IGFBP-2 expression. (C) Validation of the specificity of an anti-IIp45 antibody. Total protein extracts from LN-229 cells transfected with either empty vector or IIp45-expressing vector were blotted with either IIp45 antibody or preimmune serum. (D) Specific interaction of IIp45 and IGFBP-2 detected by co-IP assays. IIp45 and IGFBP-2 were cotransfected into HEK293 cells, and proteins were extracted. IIp45 was immunoprecipitated (IP) with anti-IIp45 antibody. IGFBP-2 was detected with the anti-IGFBP-2 antibody in the precipitated complex on a Western blot. Cells transfected with vector alone were used as a negative control. Total protein extract (Pre-IP) was also analyzed to monitor the IGFBP-2 expression in transfected cells. (E) Same as D except that the IP was carried out by using anti-IGFBP-2 antibody and the Western blotting was performed with anti-IIp45 antibody. (F) Interaction of IIp45 and IGFBP-2 in GBM detected by co-IP assays. The anti-IIp45 antibody or the antibody preabsorbed with IIp45 peptide was used to immunoprecipitate IIp45 and IGFBP-2 was detected in the IIp-45-precipitated complex. (G) Domain requirement for the IIp45 and IGFBP-2 interaction revealed by yeast two-hybrid assays. Bait constructs encoding IGFBP-2, IGFBP-2 mutant (amino acids 1–258), or its Tg-RGD region were cotransformed into yeast cells with different IIp45 truncation constructs, as indicated in the diagram. Interactions were measured by β-galactosidase activity (+, blue staining; –, no staining). Transformants with control bait (pGBKT7 vector) and prey (pACT2) showed no β-galactosidase activity. The shaded area represents the Tg-RGD region, and the bar inside the box shows the RGD motif. ND, not determined.

To show that IIp45 interacts with IGFBP-2 in mammalian cells, we performed co-IP assays by using a rabbit polyclonal antibody raised against an IIp45 peptide. First, we confirmed the specificity of the anti-IIp45 antibody in LN-229 cells that were transfected with the IIp45 expression vector. The anti-IIp45 antibody, but not the preimmune serum, specifically detected the IIp45 protein in the transfected cells (Fig. 2C). The anti-IIp45 antibody was then used to immunoprecipitate IIp45 complex, followed by immunoblotting using anti-IGFBP-2 antibody. IGFBP-2 was detected in the IIp45-immunoprecipitated complexes (Fig. 2D). Reciprocally, when the anti-IGFBP-2 antibody was used for IP followed by Western blotting with the anti-IIp45 antibody, IIp45 was detected in the precipitates (Fig. 2E). A GBM sample endogenously expressing both IIp45 and IGFBP-2 was also examined in co-IP assays. The anti-IGFBP-2 antibody detected IGFBP-2 in the anti-IIp45-immunoprecipitated complex, but not in the complex brought down by IIp45 peptide, preabsorbed IIp45 antibody (Fig. 2F). Taken together, these data demonstrated that IIp45 was a bona fide binding protein for IGFBP-2.

We further examined the interaction between IIp45 and IGFBP-2 by using truncation constructs in the yeast two-hybrid assays (Fig. 2G). This examination showed that the 77-aa C-terminal region of IGFBP-2 was sufficient in producing as robust an interaction with IIp45 as the full-length IGFBP-2, as shown by the intensity of β-galactosidase activity in colorimetric assays. However, the truncated IGFBP-2 missing the 77-aa C-terminal region did not interact with IIp45. The 44-aa middle region encoded by exon 6 of IIp45, on the other hand, was both necessary and sufficient for the interaction with IGFBP-2, whereas neither the N-terminal region encoded by exons 1–5 nor the C-terminal region encoded by exons 7–9 was able to interact with IGFBP-2.

Expression of IIp45 in Normal Adult Brain and Glioma Tissues. We then examined the expression of the IIp45 protein in normal adult brain and glioma tissues. The anti-IIp45 antibody detected a distinct band of ≈45 kDa in normal adult brains and anaplastic gliomas, and in 85% of GBMs (11 of 13) (Fig. 3 A–C). The expression pattern of the IIp45 protein was consistent with that of the IIp45 mRNA detected by RT-PCR (data not shown). Interestingly, we also observed an extra slow motility band in 73% of the GBMs (8 of 11) (Fig. 3C). These data suggest that IIp45 expression is decreased and the IIp45 protein is posttranslationally modified in GBM.

Fig. 3.

Expression of IIp45 protein in normal adult brain and glioma tissues. Total protein extracts from three normal brain tissues (A), six anaplastic glioma tissues (B), and 13 glioblastoma multiforme tissues (C) were blotted and probed with the anti-IIp45 antibody. Antibody against β-actin was used as a control for protein loading.

Inhibition of Glioma Cell Invasion and Antagonism of IGFBP-2 by IIp45. Our results showed that IIp45 interacted with IGFBP-2 through its thyroglobulin (T)g-RGD region, suggesting that IIp45 and IGFBP-2 modulate cell invasion, migration, and/or adhesion, because the Tg domain may mediate cell-surface binding (19), and the RGD motif is capable of binding to integrin and then initiate integrin-mediated adhesive signaling (20). To test this possibility, we transfected IIp45, IGFBP-2, or IIp45/IGFBP-2 into LN-229 cells, and then analyzed the transfected cells in a BioCoat Matrigel invasion chamber. This analysis showed that the invasiveness of the IIp45-expressing cells was decreased by ≈50%, but that of the IGFBP-2-expressing cells was increased by ≈250%, compared with the invasiveness of the control vector-transfected cells (Fig. 4A). Of particular note, the cells cotransfected with IIp45 and IGFBP-2 had attenuated invasion, compared with that of cells expressing IGFBP-2 alone (Fig. 4B). These results suggest that IIp45 and IGFBP-2 have opposing effects on cell invasion and that the invasiveness of cells depend on their relative levels of the two proteins (Fig. 4B). Cell cycle and cell viability analyzed by flow cytometry and soft agar colony formation assays did not reveal marked differences among any of the transfected cells (data not shown). To further confirm that IIp45 inhibits cell invasion, six stable clones of LN-229 cells expressing IIp45 were also examined in cell invasion assays. These clones had 10–45% of the invasion potential of control parental cells (Fig. 4C).

Fig. 4.

Inhibition or stimulation of cell invasion by IIp45 or IGFBP-2. (A) LN-229 cells were transfected with vector alone, with IIp45, or with IGFBP-2. After 48 h, the invasion potential of the transfected cells was analyzed as described in Materials and Methods. Expression of IIp45 or IGFBP-2 in the transfected cells detected by Western blotting is shown in the invasion assay diagram. (B) LN-229 cells were transfected with vector alone, with IGFBP-2, or with IGFBP-2 plus IIp45, and the cell invasion activity was analyzed. The expression of IIp45 and IGFBP-2 in the transfected cells is shown below the bar graph. (C) LN-229 cells transfected with IIp45 were selected. Invasion activity was analyzed in six stable clones that express IIp45. The expression of IIp45 protein in the six clones was shown below the bar graph and quantified with a densitometer, and the quantities normalized to that of β-actin. Two were also used for the cell migration assays (D) and the adhesion assays (E).

To determine how IIp45 inhibits cell invasion, we analyzed the effect of IIp45 expression on cell migration and adhesion. The migration of IIp45 stable lines was reduced by 62–82%, and their adhesive capacity was reduced by 70–81%, compared with those of the control parental cells (Fig. 4 D and E).

Furthermore, we conducted mouse brain xenograft studies by using stable GFP-expressing LN-229 cells to determine whether IIp45 inhibited invasion in vivo. The expression of GFP allowed us to visualize the invasion of individual cells that may not be apparent by hematoxylin/eosin staining. Under fluorescence microscope, we found that the tumors from control parental cells showed locally invasive foci with clear single-cell invasion into the surrounding normal brain parenchyma, whereas the tumors from all three IIp45-expressing clones showed a more defined margin and the cell invasion was markedly inhibited (Fig. 5 A and B).

Fig. 5.

Reduced cell invasion of LN-229 cells into mouse parenchyma by IIp45 in vivo. Tumors from parental and IIp45-expressing cells in mouse brains were examined by hematoxylin/eosin staining (Upper). Infiltration of cells into the surrounding normal brain parenchyma was examined under fluorescence microscope (Lower). Representative invasion areas of the parental (A) and the IIp45-expressing (B) clones were shown.

Down-Regulation of NFκB and its Target ICAM-1 Expression by IIp45. In an attempt to identify the molecular mechanism of IIp45-inhibited cell invasion, we used cDNA microarray gene expression profiling to determine gene expression alterations affected by IIp45. To address the heterogeneity problem, we profiled five stable IIp45-expressing clones and identified genes that showed the same pattern of expression change among the five clones. Of a total of 1,146 known genes on the pathway array, we detected consistent changes in 50 genes. Among them, the adhesion- and motility-associated genes were all down-regulated (Table 1). This analysis included the following genes: ICAM-1, integrin αL/β8 (ITGAL/ITGB8), cell division cycle 42 (CDC42), (RhoGTPase), IL-10, and E-selectin (SELE). Interestingly, NFκB, a transcriptional factor that activates the expression of ICAM-1, IL-10, and SELE (21–25), was also down-regulated in the IIp45-expressing cells, suggesting that the decreased expression of ICAM-1, IL-10, and SELE in the IIp45-expressing clones was due to the down-regulation of NFκB expression. Western blot analysis confirmed the decreased expression of the NFκB and ICAM-1 proteins (Fig. 6).

Table 1. Down-regulation of adhesion- and motility-associated genes in IIp45-expressing clones.

	Accession no.		T value (fold change)
Gene		Description	N1	N2	N3	N4	N5
ICAM-1	NM000201	Intercellular cell adhesion molecule 1	-5.4 (3.2)	-3.4 (2.0)	-7.0 (3.1)	-2.6 (2.0)	-2.3 (2.0)
ITGAL	NM002209	Integrin, αL	-4.0 (3.0)	-1.2 (1.3)	-3.7 (2.3)	-2.0 (2.0)	-2.4 (2.3)
ITGB8	NM002214	Integrin, β8	-3.8 (2.9)	-1.0 (1.0)	-4.3 (2.5)	-1.4 (1.9)	-2.7 (2.4)
CDC42	L10844	Cell division cycle 42 (RhoGTPase)	-3.9 (2.7)	-2.9 (1.8)	-5.7 (2.6)	-2.3 (1.8)	-2.0 (1.9)
IL-10	NM000572	Interleukin 10	-2.1 (1.5)	-2.6 (1.7)	-5.5 (2.4)	-3.3 (2.2)	-2.8 (2.6)
SELE	NM000450	E-selectin/endothelial adhesion molecule 1	-2.8 (5.5)	-1.0 (1.2)	-1.0 (1.5)	-1.0 (1.6)	-0.2 (1.1)
NFκB	X61499	Nuclear factor κB (p52/p100)	-1.7 (1.4)	-1.2 (1.3)	-2.8 (1.6)	-2.7 (1.9)	-2.2 (2.0)

N1, N2, N3, N4, and N5 are five IIp45-expressing clones. -, decrease in gene expression relative to the parental cells from statistical analysis of microarray data. Fold changes are shown in parentheses.

Fig. 6.

Down-regulation of ICAM-1 and NFκB expression in IIp45-expressing cells. Protein extracts from LN-229 parental cells and from the N1 and N4 clones were separated by SDS/PAGE and blotted with anti-ICAM-1 or anti-NFκB. Anti-β-actin was used to monitor protein loading. The intensities of the bands were normalized to that of β-actin. The values for the parental cells were designated as 1. The relative levels for ICAM-1 and NFκB are shown below.

Discussion

In this study, we attempted to identify a potential novel IGFBP-2-associated pathway to understand the function of IGFBP-2 reactivation in GBM. We searched for IGFBP-2-binding partners by using yeast two-hybrid screening and successfully identified a gene that encodes an IGFBP-2-binding protein that we named IIp45. Studies of IIp45's function showed that it inhibited GBM cell invasion. In contrast, IGFBP-2 stimulated GBM cell invasion. Importantly, in GBM cells coexpressing IIp45 and IGFBP-2, the invasion potential was attenuated, compared with that in cells expressing IGFBP-2 alone. Thus, it is possible in GBM that IIp45 and IGFBP-2 coregulate cell invasion through a common invasion signaling pathway.

Expression analyses have shown that IIp45 and IGFBP-2 are differentially expressed under physiological and pathological conditions. Specifically, IIp45 is expressed in adult brains but its expression is decreased in GBM, whereas IGFBP-2 is expressed at low levels in adult brains and reactivated in GBM (4). Considering the inhibitory role of IIp45 and the promoting role of IGFBP-2 in cell invasion, the relative amounts of IGFBP-2 and IIp45 in cells may therefore dictate whether the cells invade or not.

Interestingly, the IIp45 gene is located on chromosome 1p36, a region that is frequently deleted in oligodendrogliomas and sometimes deleted in astrocytomas (26). In this study, we observed decreased expression of IIp45 in 15% of GBMs, which is consistent with the decreased mRNA expression (17%) detected by RT-PCR (data not shown). In the GBMs expressing IIp45, the IIp45 protein is posttranslationally modified, suggesting the existence of another mechanism that regulates the function of IIp45 in GBM.

To elucidate the molecular mechanism of IIp45-inhibited cell invasion, we profiled the gene expression of IIp45-expressing cell lines. The most consistent gene expression change seen in the five stable clones tested was the reduced expression of cell adhesion- and migration-associated genes such as ICAM-1, ITGAL, CDC42, E-selectin, and IL-10, which supported our observation of inhibited adhesion and migration in the IIp45-expressing cells (Fig. 4 D and E). The roles of integrin aL and CDC42 in cell adhesion and motility have been demonstrated (27, 28), and ICAM-1, a ligand, binds to integrin aL and initiates cell adhesion (27). Interestingly, the expression of ICAM-1, E-selectin, and IL-10 is activated by the same transcription factor, NFκB (21–25), whose expression was also reduced in the IIp45-expressing cells, suggesting that the decreased expression of ICAM-1, E-selectin, and IL-10 resulted from the reduced NFκB. Recent studies (21) have found that NFκB regulates genes important for invasion and metastasis, and is activated by RhoGTPases such as CDC42 (29). Thus, our results suggest that the decreased cell adhesion and motility contribute to IIp45's inhibition of cell invasion, at least in part, through targeting the NFκB-activated invasion pathway probably through CDC42 in GBM cells.

In this study, we also showed that IGFBP-2 stimulated the invasion of GBM cells. Because IGFBP-2 carries the Tg domain and the RGD motif, it is possible that IGFBP-2 stimulates cell invasion via the integrin-mediated signaling pathway, and that IIp45 inhibits IGFBP-2-stimulated invasion by interacting with the Tg-RGD region of IGFBP-2. Future studies will further investigate the IGFBP-2-stimulated and IIp45-inhibited invasion signaling pathways.