α2-Macroglobulin Inhibits the Malignant Properties of Astrocytoma Cells by Impeding β-Catenin Signaling

The role of α2-macroglobulin (A2M) in tumorigenesis is closely related to its receptor, the low-density lipoprotein receptor–related protein 1 (LRP1). The possible mechanisms by which LRP1 regulates tumor cell behavior are related to its function as an endocytic receptor for proteases and protease inhibitors and its ability to regulate expression and catabolism of proteins, cell signaling, and the activity of adjacent membrane receptors (1). Nascent A2M is known to inactivate proteases accompanying tumor cell invasion such as plasmin, urokinase-type plasminogen activator, and metalloproteases (MMP; refs. 2, 3). Binding to proteases or small amines in vitro induces a conformational change in the A2M molecule that triggers its interaction with LRP1 (4) and creates distinct binding sites for transforming growth factor (TGF)-β1 (5) and other growth-regulating ligands (6) that can be rapidly cleared by LRP1 (7–9).

Glial tumors are among the most malignant intrinsic brain tumors, characterized by excessive proliferation, angiogenesis, and infiltration (10). Only few reports illuminate the possible function of LRP1 in these aggressive tumors, and results obtained are rather controversial (11). In addition, the effect of the proteinase inhibitor A2M on glial tumors has thus far not been addressed.

The present study aimed at exploring the impact of A2M* on various malignancy-associated properties of 1321N1 astrocytoma cells. Results imply that A2M* inhibits proliferation, migration, invasion, spheroid formation, and anchorage-independent growth of astrocytoma cells possibly through interfering with the Wnt/β-catenin signaling. The canonical pathway is known to play a pivotal role in cell malignant transformation (12). Wnt ligands form a ternary complex with frizzled receptor (FZD) and with the coreceptors LRP5/6, and this complex ultimately prevents the degradation of β-catenin. The latter accumulates in the nucleus and complexes with the transcription factors T-cell factor/lymphoid enhancer factor (TCF/LEF) to activate the transcription of target genes encoding various events of tumorigenesis.

We show that A2M* at nanomolar concentrations induced the expression of LRP1, FZD, E-cadherin, and N-cadherin and inhibited Wnt ligand expression. This provides evidence for the LRP1-mediated suppressive potency of A2M* and offers a promising approach for tumor treatment.

Anti-human LRP1 monoclonal antibody (mAb; clone II2C7 and II4/8), rabbit polyclonal anti-LRP1 antibody (Ab), native A2M, methylamine-transformed A2M*, and FITC-A2M* were purchased from BioMac; the WST-1 proliferation kit was from Roche; cellulose nitrate membranes were from Whatman Schleicher & Schuell; DMEM and RPMI medium, FCS, and Cy3-labeled goat anti-mouse Ab were from Invitrogen; Growth Factor Reduced Matrigel Matrix and polycarbonate mesh, mouse anti-human E-cadherin, N-cadherin, and β-catenin were from BD Biosciences; hematoxylin was from Merck; trypan blue was from Seromed; rabbit anti-human β-actin Ab was from Abcam; horseradish peroxidase (HRP)–labeled goat anti-mouse Ab and real detection system peroxidase/3,3′-diaminobenzidine rabbit/mouse kit were from Dako; protease inhibitor cocktail was from Sigma-Aldrich; and the chemiluminescence detection kit was from Boehringer. Plasmid pCMV-GLUC was obtained from Prolume Nanolight, Inc.; the TCF Reporter Plasmid kit was from Millipore; TransIT-LT1 was from Mirus Corporation; and the Gaussia luciferase transfection kit and coelenterazine were from PJK. A2M (clone α1-IIE7) and prostate-specific antigen (PSA; clone 1B7) mAbs were used as described previously (13, 14).

Human astrocytoma 3121N1 cells (European Collection of Animal Cell Cultures; 86030102), LNCaP (DSMZ; ACC 256), PC-3 (CRL-1435; American Type Culture Collection), and DU-145 (DSMZ; ACC 261) were used for this study.

Human astrocytoma 1321N1cells were cultured in DMEM, 10% FCS (DMEM-FCS), 100 units of penicillin, and 100 μg streptomycin/mL at 37°C/5% CO₂ in a humidified atmosphere. Cell monolayers were dissociated using 0.04% trypsin/0.03% EDTA in 50 mmol/L PBS (pH 7.4). Cellular protein extracts of tumor cells were prepared using an extraction buffer [25 mmol/L Tris, 2 mmol/L EDTA, 10% glycerol, 1% Triton X-100, 2 mmol/L DTT, and 1 mmol/L phenylmethylsulfonyl fluoride (pH 7.0)] containing 0.3% protease inhibitor cocktail. Protein content was determined (15) and human fibroblast cultures were set up (16) as previously described.

Astrocytoma cell cultures (10⁵/2 mL DMEM-FCS/well) were set up and FITC-labeled A2M* (0.1 μmol/L) was added to cavities containing 2 μmol/L A2M*, 2 μmol/L polyclonal anti-LRP1 Ab, or rabbit preimmune immunoglobulin (2 μmol/L) and incubated for 30 min to estimate binding/uptake of labeled A2M*. Plates were chilled on ice, washed twice (each with 2 mL PBS), and supernatants were discarded. Cells were removed by trypsination and incubated with extraction buffer in the absence of protease inhibitors. Samples were disintegrated by sonication (3 × 15 s) and centrifugation at 13,000 rpm for 15 min at 4°C. Supernatants were analyzed using Perkin-Elmer LS50B.

According to standard protocols, immunochemical staining of astrocytoma 1321N1 cells was done using the primary anti-human Abs directed against E-cadherin, N-cadherin, LRP1, β-catenin, and β-actin. Immunofluorescence detection was accomplished with Cy3-labeled goat anti-mouse Ab. Three-micrometer paraffin-embedded sections of spheroids were stained with mouse anti-LRP1 mAb (clone II2C7) in combination with a secondary biotinylated Ab and HRP-labeled streptavidin. Stained cells were evaluated with a confocal laser scanning microscope (LSM510 META, Carl Zeiss).

Following cell lysis, 20 to 100 μg of cell lysate proteins were loaded to 4% to 20% SDS-PAGE and then blotted to membranes. The primary Abs used were rabbit anti-LRP1 (1 μg/mL) and anti–β-actin Ab (0.5 μg/mL), anti–β-catenin, anti–E-cadherin, and anti–N-cadherin mAb (10 μg/mL each). Detection was accomplished with HRP-labeled goat anti-mouse (1:1,000) or anti-rabbit (1:2,000) Ab. Bands were visualized by chemiluminescence detection.

Astrocytoma cells were cultured as described above, putative receptor ligands were added, and cell proliferation was analyzed using the WST-1 assay (17).

The effect of A2M* on viability of astrocytoma cells was inspected by trypan blue exclusion (18).

Lactate dehydrogenase (LDH) activity was assayed in the supernatant of astrocytoma cells cultured for 24 h (17).

Cell movement was evaluated in Boyden chambers using 24-well culture plates equipped with 8-μm pore polycarbonate mesh; the bottom well was filled with 800 μL DMEM-FCS. For invasion analysis, the membranes of inserts were coated with 25 μL Growth Factor Reduced Matrigel Matrix. Cells (4 × 10⁴ in 100 μL DMEM, 0.5% FCS) were seeded in the upper compartment to adhere overnight. Test substances were added and incubation was continued for 6 h (migration) or 42 h (invasion). Cells were removed and the membranes were fixed with Karnowski solution and stained with hematoxylin.

Six-well plates were coated with 1.5 mL of a base layer of 0.6% basal agar in 2× DMEM-FCS. Cells (8 × 10⁵/well) were added in a mixture of 0.4% agar and 2× DMEM/20% FCS. Test substances were initially mixed into the upper agar layer. Cultures were evolved for 14 d and replenished daily with 100 μL DMEM-FCS. Finally, the plates were stained with 0.5 mL of 0.005% crystal violet and colonies of >20 cells were counted.

Well plates were coated with 0.1% gelatin solution before adding 10⁴ cells/100 μL DMEM-FCS in the absence or presence of A2M*, the receptor-associated protein (RAP), or anti-LRP1 Ab. After certain time lapses, nonadherent cells were washed away and the remaining cells were fixed with 4% paraformaldehyde solution, stained with crystal violet (1%), washed, and destained with acetic acid (33%) followed by reading the absorbance at 590 nm against the blank.

Astrocytoma cells (5 × 10⁶) were transferred into 10 mL DMEM-FCS containing test substances in 50-mL plastic tubes and rolled at 5 rpm at 37°C/5% CO₂ for 3 wk. Spheroid diameters (mean of two orthogonal diameters) were determined and the medium was changed daily.

Fab fragments of rabbit anti-LRP1 Ab (10 mg) were prepared using the Fab Preparation Kit from Pierce.

Isolation of total RNA of astrocytoma cells exposed to serial concentrations of A2M*, digestion of genomic DNA, reverse transcription, and purification of cDNA were done as previously described (19). For real-time PCR, the forward (f) and reverse (r) primer sequences used were as follows: (a) GAPDH (NM_002046.3), f: 5′-AGCCACATCGCTCAGACAC-3′ and r: 5′-GCCCAATACGACCAA ATC C-3′; (b) LRP1 (NM_002332.2), f: 5′-GATGAGACACACGCCAACTG-3′and r: 5′-CGGCACTGGAACTCATCA-3′; (c) LRP5 (NM_002335.2), f: 5′-GAACATCAAGCGAGCCAAG-3′and r: 5′-TGGCTCAGAGAGGTCAAAACA-3′; (d) LRP6 (NM_002336.2), f: 5′-ATCCGAAAGGCACAAGAAGA-3′and r: 5′-GACTCGGAACTGAGCTCACAA-3′; (e) FZD (NM_001463.2), f: 5′-TCATGGGCTATGAAGATGAGG-3′ and r: 5′-TCATATCCCAGCGCTTAACTT T-3′; (f) Wnt1 (NM_005430.2), f: 5′-CGCTGGAACTGTCCCACT-3′ and r: 5′-AACGCCGTTTCTCGACAG-3′; (g) Wnt3A (NM_033131.2), f: 5′-AACTGCACCACCGTCCAC-3′ and r: 5′-CCGACTCCCTGGTAGCTTT-3′; (h) Wnt5A (ENST00000264634.4), f: 5′-TAAGCCCAGGAGTTGCTTTG-3′ and r: 5′-CTGAACAGGGTTATTCATACCTAGC-3′); (i) Wnt10B (U81787.1), f: 5′-GCGAATCCACAACAACAGG-3′ and r: 5′-TCCAGCATGTCTTGAACTGG-3′; (j) β-catenin (X87838.1), f: 5′-GCTTTCAGTTGAGCTGACCA-3′ and r: 5′-AAGTCCAAGATCAGCAGTCTC A-3′; (k) E-cadherin (AB025105.1), f: 5′-CCCGGGACAACGTTTATTAC-3′ and r: 5′-GCTGGCTCAAGTCAAAGTCC-3′; and N-cadherin (M34064.1), f: 5′-GGTGGAGGAGAAGAAGACCAG-3′ and r: 5′-GGCATCAGGCTCCACAGT-3′. Reactions started by an initial activation step for 10 min at 95°C, and each following cycle consisted of a denaturation step for 15 s at 94°C, annealing for 5 s at 60°C, and extension for 15 s at 72°C. GAPDH was chosen as a reference housekeeping gene as it showed amplification efficiency similar to those of other cytokine genes. REST software (20) was used to estimate gene expression relative to controls.

The reporter gene TOP-Gau containing a secreted luciferase of Gaussia princeps, controlled by two sets of three copies of the TCF binding site (wild-type) and a thymidine kinase minimal promoter, was derived by combining the reporter gene TOP-flash (TCF Reporter Plasmid kit) with the luciferase gene. The luciferase coding region was derived from the plasmid pCMV-GLUC. The mutated FOP-Gau reporter gene (with two full copies and one incomplete copy of the TCF binding site followed by three copies in the reverse orientation) was derived from FOP-flash and pCMV-GLUC.

Cells, seeded at 10⁵ per 1 mL medium, were transfected with 1 μg DNA using TransIT-LT1, and then incubated with test substances for 36 h. Gaussia luciferase activity was assessed in 10 μL cell supernatants using 50 μL assay reagent on Orion-II microplate luminometer.

Statistics were done using Wilcoxon's rank-sum test, and PCR results were analyzed using pair-wise fixed reallocation randomization test incorporated in REST software; significance were considered at P < 0.05. Unless indicated otherwise, all data indicate means ± SD of three independent experiments.

LRP1 is apparently highly expressed in 1321N1 astrocytoma cells compared with two prostate cancer cell lines (LNCaP and DU-145) or fibroblasts (Fig. 1A). FITC-labeled A2M* was rapidly endocytosed by tumor cells; its uptake is diminished by adding nonlabeled A2M* or anti-LRP1 Ab (Fig. 1B), highlighting the ability of the polyclonal Ab to block the A2M* binding site of LRP1.

A2M* strongly inhibited proliferation of tumor cells under optimal growth-promoting conditions, whereas native A2M exerted only marginal inhibitory effects (Fig. 2A). To explore an alternative pathway other than neutralizing growth factors (21, 22), we blocked the cellular receptor with a polyclonal Ab that presumably does not mediate clearance of growth factors. As depicted in Fig. 2A, anti-LRP1 Ab exhibited a similar influence on cell proliferation compared with A2M*, a further clue for the LRP1 activation by A2M*. We assume that the polyclonal Ab fraction may contain mAb(s) that specifically reacted with the A2M* binding domain of LRP1 and can mimic the binding of A2M* to the receptor.

Inhibition of cell proliferation by A2M* was not due to cell toxicity as indicated by the unaltered LDH release (Fig. 2B). Moreover, the average cell vitality was 95 ± 3%, 93 ± 3.7%, and 96 ± 8% for control, 0.2 μmol/L A2M*, and 0.2 μmol/L A2M-treated cells, respectively.

The migration rate of 1321N1 cells was inversely correlated with the concentration of A2M*. As expected, cell migration was only slightly changed by native A2M (Fig. 3A). The specificity of A2M*-LRP1 interaction was confirmed by applying mAb (α1; clone α1-IIE7) known to react with the COOH-terminal receptor–binding domain of A2M* (13); its presence abrogated the inhibiting effect of A2M*, probably through blocking the binding of A2M* to LRP1. RAP, known to abolish binding of almost all ligands to LRP1 (23), diminished cell migration, similar to the polyclonal anti-LRP1 Ab. As controls, we used mAbs against α (II2C7) and β (II48) chains of LRP1 as well as an irrelevant Ab against PSA (1B7). All of these mAbs showed insignificant effects on cell migration.

Because A2M* and polyclonal anti-LRP1 Abs inhibited cell proliferation and migration, we hypothesize that inhibition may be due to LRP1-dependent interference with malignant cell adhesion factor(s). Results depicted in Fig. 3B showed that, compared with controls, the adhesion of tumor cells was significantly reduced after 120 minutes in the presence of 0.2 μmol/L A2M*, 1 μmol/L RAP, or 1 μmol/L polyclonal anti-LRP1. No difference between control and ligand-treated cells was observed following 200-min incubation. These data indicate that LRP1 might be involved in cell adhesion, and receptor ligands can diminish cell-matrix interaction.

Tumor cells invade extracellular matrix either through protease-dependent processes or amoeboid movement (24). Therefore, inhibiting extracellular proteases would eventually reduce the ability of cells to invade the matrix network. To prove this, a cell invasion assay was conducted to allow cells to migrate through Matrigel-coated 8-μm pore invasion chambers. In the presence of protease inhibitors, the invasion of tumor cells was reduced by ∼50% relative to controls. Similar results were obtained by A2M*, polyclonal anti-LRP1 Ab, or RAP (Fig. 3C). The agonistic LRP1 mAb (clone II2C7) and the irrelevant anti-PSA mAb (1B7) failed to inhibit cell invasion. Because activated A2M* has lost its antiproteolytic activity, its effect on cell invasion may be possibly brought about by direct activation of LRP1. This was corroborated by applying polyvalent and monovalent anti-LRP1 Ab obtained by papain digestion.

Tumor spheroids provide a suitable three-dimensional in vitro model that may mimic the cellular environment of a poorly vascularized, hypoxic solid tumor (25). We monitored the growth of approximately 50 to 100 spheroids per group to study the size distribution over time. The initial aggregation of cells (over 3 days) showed cell clumps of 40 to 120 μm in diameter (Fig. 4A). The size of the spheroids increased progressively with time but size distribution became more heterogeneous due to shedding of tumor cells from surface and formation of subspheroids (Fig. 4B,, left). Immunostaining of LRP1 (Fig. 4B,, right) indicated its presence at the margin, which may constitute the proliferative zone of a spheroid. Results shown in Fig. 4C revealed that the diameter of the spheroids increased steadily up to day 6, then followed by a plateau phase till day 14. Compared with controls, A2M* (0.2 μmol/L) or anti-LRP1 Ab (1 μmol/L) inhibited the growth of tumor spheroids. A2M* was more potent at the initial phase of spheroid accretion comparable with anti-LRP1 Ab. This indicates the implication of LRP1 in cell-cell adherence and that ligation of the A2M* binding domain of LRP1 modulates this process.

Furthermore, we investigated the effect of A2M* and Abs on anchorage-independent growth, which is a good index of tumorigenicity. Astrocytoma cells were found to form colonies in the soft-agar assay (Fig. 4D,, left) and A2M* (0.2 μmol/L) significantly inhibited colony formation by >50%. Compared with the control, polyclonal anti-LRP1 Ab elicited a similar effect (Fig. 4D , right).

Progression of human tumors is often associated with the induction of the canonical Wnt/β-catenin signaling. Using the Topflash/Fopflash assay, we found that the Wnt/β-catenin signaling is constitutively active in 1321N1 cells discernible at the high activity of the Topflash versus Fopflash signal (Fig. 5A,, left); the latter was used as internal control. The functionality of the Wnt/β-catenin signaling cascade was shown by adding 5 mmol/L LiCl₂ that is known to inhibit glycogen synthase kinase-3β, leading to stabilization of β-catenin, which then accumulates around and translocates into the nucleus. This effect was notably abated in the presence of 0.1 μmol/L A2M*. β-Catenin is relocated to the plasma membrane (Fig. 5A,, middle), implying inhibition of the canonical Wnt/β-catenin pathway by A2M* (Fig. 5A,, right). This was additionally affirmed by evaluating the expression of various genes (Fig. 5B). Surprisingly, low doses of A2M* significantly upregulated FZD, LRP1, E-cadherin, and N-cadherin at the mRNA and protein levels, and slightly upregulated β-catenin levels (Fig. 5C); LRP5 and LRP6 mRNA remained unchanged. The expression of Wnt1, Wnt5A, and Wnt10B mRNA was downregulated (Fig. 5B), whereas Wnt3A mRNA was not detected. As expected, anti-LRP1 Ab exerted comparable effects.

As shown in Fig. 5D, A2M* increased the cellular level of LRP1 and induced the relocation of β-catenin to the plasma membrane. Apparently, in the presence of A2M*, the size of the cytosolic compartments and nuclei diminished in a dose-dependent fashion (middle).

In the present study, we show that A2M* in conjunction with its receptor LRP1 inhibits migration, invasion, anchorage-independent growth, and formation of tumor spheroids of 1321N1 astrocytoma cells. Comparable to the native A2M, which does not bind to the receptor, the impact of A2M* implies that the activated protease inhibitor executes strong tumor-suppressive properties. Furthermore, we provide a novel evidence that A2M* mediates its effects through attenuating the Wnt/β-catenin signaling, which is upregulated in many tumors.

At present, the beneficial impact of A2M* on tumors and other diseases have been mainly discussed in terms of clearing growth factors, especially TGF-β1 that is known to strongly promote the malignancy of gliomas (9, 22, 26). However, our experiments with polyclonal Abs may argue against this assumption because growth factors cannot be cleared by them. This alludes that A2M* and A2M*-mimicking Abs may occupy the A2M* binding site of LRP1, as already observed in other cells (4, 27). In cardiomyocytes, A2M* stimulated hypertonic cell growth by activating the extracellular signal-regulated kinase 1,2 and phosphoinositide 3-kinase/Akt pathway, a cellular effect that could be also detected by anti-LRP1 Ab. These findings corroborate our results, indicating that anti-LRP1 Abs mimic A2M* activities.

It remains obscure how A2M* and A2M*-mimicking Abs can stimulate cell proliferation in one cell type and inhibit proliferation and migration in others. Basically, one has to investigate whether these effects can be evoked by A2M*-associated growth factors or cytokines. This issue was addressed by Mantuano and colleagues (28), who convincingly reformulated that the receptor recognition sites of A2M alone can stimulate the Akt pathway. This accords with recent studies indicating that the endocytic and signaling events of LRP1 are functionally interconnected due to the affinity of certain motifs within the cytoplasmic receptor domain (e.g., NPXY and FXNPXY) to signal transduction adaptor protein (29).

A2M* was also reported to ligate signaling membrane receptors other than LRP1, such as the glucose-regulated protein 78 (Grp78). Targeting this receptor induced both cell proliferation and creation of an antiapoptotic milieu (30, 31). The different outcome with respect to cell proliferation and the fact that anti-LRP1 Abs mimic the action of A2M* may not approximate the conclusion that A2M* triggers Grp78 in astrocytoma cells.

A further pathway through which A2M is implicated is its capacity to inhibit and prime the clearance of proteases. Both A2M and LRP1 may affect the proteolytic potential at plasma membrane and in the pericellular space by manipulating major elements of cell migration and invasion, e.g., MMP, plasmin, and plasminogen activators; all are direct ligands of LRP1 and can be endocytosed following complex formation with protease inhibitors. For instance, MMP-2 forms complexes with A2M (32) and conjugates to thrombospondin-2 as convincingly shown by Fears and colleagues (33). We used the activated A2M*, which has a locked conformation that is unable to bind and clear proteases. This excludes the antiproteolytic activity of A2M* as a possible mechanism.

The effect of LRP1 on tumor growth and invasiveness is well acknowledged (34); however, its precise role remains controversial. Inhibiting LRP1 expression and function is commonly reported to increase cell migration and invasion (11, 35–37). LRP1-deficient fibroblasts showed an increased cell migration (38), and restoration of LRP1 in LRP1-null cells markedly reduced cell growth, migration, and anchorage-independent colony formation and reduced tumor development in nude mice (39). Silencing of LRP1 by siRNA decreased cell migration (40) and prevented invasion of thyroid carcinoma cell lines despite increased pericellular proteolytic activity (34). The observed controversial results in response of cells to manipulations of LRP1 expression may depend on the specificity of cell environment, including interaction of LRP1 with other receptors or intracellular scaffold proteins. It becomes obvious that cellular effects induced by ligand-mediated inhibition of LRP1 are different from those created by silencing LRP1. It seems feasible that a genetically induced inactivation or deletion of LRP1 may disrupt its ability to control cell motility through interaction with other cell signaling receptors or membrane proteins (37). Therefore, it was our approach not to manipulate the phenotype of the tumor cells, e.g., by antisense or siRNA strategies, but rather to use specific natural agonists so as not to interfere with the potential role of LRP1 in regulating other possible functions.

Recently, the involvement of LRP1 in cell adhesion of human leukocyte as well as in cell-matrix interaction has been reported (41, 42). Hereby, we corroborate these reports by demonstrating the significant role of LRP1 and its ligand A2M* in tumor spheroid formation, cell adhesion, and cell-cell interaction.

In the present study, we offer a novel mechanism on how A2M*/LRP1 may regulate tumor cell growth, namely, by interfering with the Wnt/β-catenin signaling. Our findings coincide with results obtained by Guo and colleagues (43), who showed that blocking the Wnt/β-catenin signaling by indomethacin treatment caused β-catenin to relocate gradually from the cytoplasm and the nucleus to the plasma membrane. On the plasma membrane, β-catenin is known to form a complex with cadherins, maintaining adhesive junction and cell integrity (44). We propose a similar behavior in astrocytoma cells following exposition to A2M* as revealed by immunochemical studies. Therefore, the observed reduction of migration, invasion, and spheroid formation can be contributed to the induction of cellular cadherins, which act as tumor suppressors (44, 45). Moreover, consistent with our results is the upregulation of FZD, which is found to decrease invasiveness and proliferation of thyroid carcinoma cells (46). In our opinion, the key event in A2M*-mediated attenuation of the Wnt/β-catenin signaling is the overexpression of LRP1. Zilberberg and colleagues (47) convincingly showed that FZD interacts with LRP1. Upregulation of the latter sequesters FZD, disrupts the FZD/LRP5/LRP6 complex, and ultimately represses the canonical Wnt pathway. Because MMPs and various oncogenes are downstream targets of Wnt/β-catenin signaling, the low expression level of LRP1, which was related to an aggressive and invasive phenotype of tumor cells, may therefore enable the canonical pathway (35, 48). Taken together, A2M* seems to exert its beneficial effect through the following: (a) upregulating LRP1 and FZD, (b) inhibiting expression of Wnt ligands and hence the autocrine activation of the Wnt/β-catenin signaling, and (c) the consequent relocation of β-catenin to the plasma membrane, which, together with upregulated cadherins, ameliorates cell integrity and thereby reduces cell proliferation, migration, and invasion potency. Therefore, it seems conceivable that modulation of Wnt/β-catenin pathway by A2M* may depend on the relative expression level of members of the LRP family, FZD, and secretory proteins, including Wnt ligands. How A2M* mediates this potential remains to be elucidated (Fig. 6).

Overall, the potential of A2M* to inhibit the malignancy-associated properties of human astrocytoma cells may warrant evaluating A2M* and anti-LRP1 for targeting Wnt ligands in gliobastoma and probably other tumors (49). In the light of these findings, we assume that the decline in the A2M blood level found at higher age may therefore facilitate tumor development (50).

No potential conflicts of interest were disclosed.

We thank Dr. Mathias Platzer for the invaluable help, Angelika Schaefer and Elke Usbeck for their technical support, Dr. Dudley Strickland for his gift of the RAP-expressing plasmid, Dr. Anselm Berthold (Berthold Detection Systems, Pforzheim, Germany) for the kind disposal of an Orion II luminometer, Dr. Bruce J. Bryan (Prolume, Nanoligh, Inc., Pinetop, AZ) for the kind gift of Gaussia vector, and Dr. Sonya Faber for proofreading the text.

Grant Support: Wilhelm Sander-Stiftung grant 2006.053.2.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.