|Home | About | Journals | Submit | Contact Us | Français|
Chemokines participate in cellular processes associated with tumor proliferation, migration, and angiogenesis. We previously demonstrated that stromal cell–derived factor 1 (SDF1) exerts a mitogenic activity in glioblastomas through the activation of its receptor CXCR4. Here we studied the expression of this chemokine in human meningiomas and its possible role in cell proliferation. Reverse transcriptase–PCR analysis for CXCR4 and SDF1 was performed on 55 human meningiomas (47 WHO grade I, 5 WHO II, and 3 WHO III). Immunolabeling for CXCR4 and SDF1 was performed on paraffin-embedded sections of these tumors. [3H]Thymidine uptake and Western blot analyses were performed on primary meningeal cell cultures of tumors to evaluate the proliferative activity of human SDF1α (hSDF1α) in vitro and the involvement of extracellular signal–regulated kinase 1/2 (ERK1/2) activation in this process. CXCR4 mRNA was expressed by 78% of the tumor specimens and SDF1 mRNA by 53%. CXCR4 and SDF1 were often detected in the same tumor tissues and colocalized with epithelial membrane antigen immunostaining. In 9 of 12 primary cultures from meningiomas, hSDF1α induced significant cell proliferation that was strongly reduced by the mitogen-activated protein kinase kinase inhibitor PD98059, involving ERK1/2 activation in the proliferative signal of hSDF1α. In fact, CXCR4 stimulation led to ERK1/2 phosphorylation/activation. In addition, the hSDF1α-induced cell proliferation was significantly correlated with the MIB1 staining index in the corresponding surgical specimen. In conclusion, we found that human meningiomas express CXCR4 and SDF1 and that hSDF1α induces proliferation in primary meningioma cell cultures through the activation of ERK1/2.
Meningiomas constitute about 20% of all primary intracranial tumors in adults and display a wide variety of histopathological subtypes. However, the real incidence of this neoplasm could be higher, because some benign meningiomas do not produce symptoms (Lusis and Gutmann, 2004). Meningothelial, fibrous, and transitional meningiomas are the most common meningiomas and are classified as grade I using the WHO system; they are slowly growing lesions with a benign clinical course and a good prognosis. Other subtypes, such as atypical (grade II) and anaplastic (grade III) meningiomas, are associated with an aggressive clinical behavior and a high risk of recurrence (Kleihues et al., 2002). At present, the tumor grade together with the study of proliferation indices provides the most useful predictor of recurrence (Nakaguchi et al., 1999; Perry et al., 1998). More than 80% of partially resected tumors relapse within 10 years, but in those completely resected by surgery and the benign meningiomas, the recurrence rate reaches 15%–20%.
Chemokines are the most important and the best-known group of chemotactic proteins. They direct leukocyte trafficking in both immune surveillance and inflammation (Murphy, 2002). Chemokines are also involved in many other physiological and pathological processes, including neoplasia, in which they play an important role through multiple mechanisms. Chemokines affect tumor cell proliferation, regulate the angiogenic/angiostatic processes, control cell migration and metastasis, and regulate the recruitment of the immune cells into the tumor mass (Bajetto et al., 2002).
Stromal cell–derived factor 1 (SDF1),3 recently renamed CXCL12, is a chemokine with a role in neurobiology that was long underestimated, until several studies demonstrated its expression and function in the resident CNS cells. The expression of SDF1 and its receptor CXCR4 has been described at the CNS level in neuronal, astroglial, and microglial cells (Bajetto et al., 1999a, 1999b; Ohtani et al., 1998). SDF1 exerts its chemotactic action to direct organogenesis and tissue structure in the developing brain, as demonstrated by studies of both SDF1−/− and CXCR4−/− knockout mice, in which gene deletion resulted in significant abnormalities in cerebellar development (Ma et al., 1998; Nagasawa et al., 1996; Zou et al., 1998).
SDF1 and CXCR4 are involved in the proliferation of different human tumors, including those developing in the brain (Balkwill, 2004). CXCR4 and SDF1 have been detected in adult glioblastoma multiforme, where their expression was reported to increase with tumor grade (Rempel et al., 2000). In vitro, SDF1 acts as a growth factor for glioblastoma cell lines and normal astrocytes, increasing their proliferation (Bajetto et al., 2001; Barbero et al., 2003). In addition, it has been reported that SDF1 stimulates chemotaxis, survival, and cell proliferation in glioblastoma multiforme and medulloblastoma cell lines and xenografted tumors (Bajetto et al., 2006; Rubin et al., 2003).
In this study, we used reverse transcriptase (RT)–PCR and immunohistochemistry (IHC) to analyze the pattern of expression of CXCR4 and SDF1 in a series of surgical samples derived from meningeal tumors. Moreover, we performed in vitro experiments on 12 primary cultures, derived from the surgical meningioma specimens, to evaluate the proliferative activity of human SDF1α (hSDF1α) and the involvement of extracellular signal–regulated kinase 1/2 (ERK1/2) activation in this process.
Fifty-five surgical tumor samples were collected between 1997 and 2005 from the Section of Neurosurgery, Department of Neuroscience, Ophthalmology and Genetics (University of Genova, Genova, Italy). Tissue samples were either immediately frozen and stored at −80°C for RT-PCR analysis or prepared for IHC. In 12 cases, a portion of the sample was dispersed for cell culture studies. Histology and tumor grading were performed at the Section of Pathology, S. Martino Hospital (Genova), according to the WHO classification (Kleihues et al., 2002), following standard diagnostic criteria.
The expression of specific mRNAs for CXCR4 and SDF1 was evaluated by means of RT-PCR. Total RNA from the human meningioma specimens was isolated using the acidic phenol technique, and RT-PCR was performed as previously reported (Bajetto et al., 1999b). Briefly, 10 μg of total RNA was treated with RNase-free DNase at 37°C for 45 min to remove genomic DNA contamination and then extracted by phenol-chloroform and precipitated with ethanol. The RT reaction was performed using oligo-dT(16) primer and the avian myeloblastosis virus RT (Amersham Biosciences, Buckinghamshire, UK) for 40 min at 42°C. PCR was performed on 10 ng of cDNA as follows (final volume, 50 μl): 5 min of denaturation at 94°C, 40 cycles of 1 min at 94°C, 1 min at 60°C, and 1 min at 72°C, followed by 7 min at 72°C, using 2.5 units/reaction Taq DNA polymerase (Roche, Basel, Switzerland). Amplified DNA fragments were then electrophoresed through agarose gels and visualized by ethidium bromide staining. The sequences of the primers used were as follows: CXCR4: sense 5′708-ggccctcaagaccacagtca-7273′, antisense 5′1059-ttagctggagtgaaaacttgaag-1037 (accession number NM_003467); SDF1α and SDF1β: sense 5′80-atgaacgccaaggtcgtggtc100-3′, antisense 5′346-cttgtttaaagctttctccaggtact-3213′ (accession number NM_000609). We assessed β-actin fragment amplification in all the PCR as a reaction efficiency control; in addition, a negative control (PCR amplification before RT) was used to verify the absence of genomic DNA contamination in the cDNA samples.
Immunohistochemical detection of CXCR4 and SDF1 was performed on 4-μm sections of paraffin-embedded tissue stained with the monoclonal anti-CXCR4 (clone 12G5) (BD PharMingen, San Diego, Calif.) or polyclonal anti-CXCR4 (Affinity BioReagents, Golden, Colo.), and polyclonal anti-SDF1 (Torrey Pines BioLabs, Houston, Tex.) or monoclonal anti-SDF1 (R&D Systems, Minneapolis, Minn.) antibodies. Sections were chosen from among positive tissues previously evaluated by RT-PCR, deparaffinized with xylene, and dehydrated through graded alcohols. Slides were rinsed in sodium citrate buffer (pH 5.4), heated in a microwave oven for 2 min, allowed to cool for 10 min, and then treated with 0.3% Triton-X in phosphate-buffered saline (PBS) for 10 min and rinsed in Tris-buffered saline (TBS) (0.05 M Tris-HCl and 0.15 M NaCl). Sections were saturated with 10% normal goat serum in TBS for 20 min and then incubated overnight at 4°C with a 1:100 dilution of anti-CXCR4 antibody or anti-SDF1 antibody in CHEM-MAT solution (Dako, Glostrup, Denmark). After three washes in TBS, the sections were incubated with biotinylated secondary antibodies (antimouse or antirabbit) for 30 min in TBS. Finally, the sections were washed and incubated for 20 min with StreptABComplex/AP (Dako) according to the manufacturer’s instructions, revealed with the 5-bromo-4-chloro-3-indoxyl-phosphate/nitro blue tetrazolium chloride/iodonitrotetrazolium violet substrate system (Dako), and counterstained with hematoxylin before being mounted with Mowiol (Calbiochem, San Diego, Calif.). Negative controls were included in all of the aforementioned IHC analyses by omitting the primary antibodies. Microphotographs were obtained with a Nikon DS digital camera.
Immunohistochemical detection of Ki-67 was performed on 4-μm paraffin-embedded tissue sections stained with undiluted anti-MIB1 mouse monoclonal antibody (Menarini, Milano, Italy). IHC was performed on a Bench Mark XT Automated System (Ventana Medical Systems, Illkirch, France). MIB1 positive–stained cells were evaluated by visual inspection, and an MIB1 staining index (MIB1-SI) was calculated as the percentage of positive cells per total number of cells (at least 20,000) in 100 randomly selected microscopic fields. Hematoxylin was used for the nuclear counterstaining. All of the slides were blindly reviewed and scored independently by two neuropathologists (J.L.R. and A.D.). Nonneoplastic cells were excluded from the counting.
Primary cell cultures were obtained from 12 freshly resected meningioma tissues by mechanical disruption under sterile conditions (Arena et al., 2004). Single-cell suspensions were maintained in minimal essential medium supplemented with 20% fetal calf serum, penicillin-streptomycin, and nonessential amino acids, all purchased from Euroclone (Milano, Italy). Medium was changed every 24 h to remove nonadherent cells. Cells were subcultured at least twice before the experiments were performed. Immunocytochemical characterization of short-term cultures was performed by analyzing the epithelial membrane antigen (EMA) expression. Primary cultures were fixed with cool methanol and immunostained with monoclonal anti-EMA antibody (Menarini, Firenze, Italy) to evaluate the ratio of meningothelial cells to fibroblasts.
[3H]Thymidine incorporation was performed to measure the DNA synthesis of primary culture cells (Arena et al., 2004). Briefly, cells (8 × 105) were plated in quadruplicate in a 24-well plate, serum starved for 24 h, and then treated for 24 h with hSDF1α (Upstate-Biotechnology, Lake Placid, N.Y.). [3H]Methyl-thymidine (1 μCi/ml; Amersham Biosciences) was added to each well during the last 4 h of treatment. Cells were then trypsinized and harvested on glass-fiber filters (Millipore, Bedford, Mass.). [3H]Thymidine incorporation was measured in a scintillation counter after consecutive DNA precipitation with 10% and 5% trichloroacetic acid and 95% ethanol.
Cell cultures were serum starved for 24 h before being treated with hSDF1α for 2 min. Cells were lysed in lysis buffer (1% NP-40, 20 mM Tris-HCl pH 8, 137 mM NaCl, 10% glycerol, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, 1 mM sodium orthovanadate, and 10 mM NaF; Sigma-Aldrich, Milano, Italy) for 10 min at 4°C. Nuclei were removed by centrifugation (5000 rpm at 4°C) for 10 min, and total protein contents were measured using the Brad-ford assay (Bio-Rad, Hercules, Calif.). Proteins (8–10 μg) were resuspended in 2× reducing sample buffer (2% sodium dodecyl sulfate [SDS], 62.5 mM Tris pH 6.8, 0.01% blue bromophenol, and 1.43 mM 2-mercapto-ethanol, 0.1% glycerol), electrophoresed on 10% SDS-polyacrylamide gels, transferred onto polyvinylidene fluoride (PVDF) membrane (Bio-Rad), and probed with anti–phospho-ERK1/2 (New England BioLabs, Beverly, Mass.) and anti–α-tubulin (Sigma-Aldrich). Immuno-complexes were detected by using the ECL kit (Amersham Biosciences). PD98059 compound (used at 10 μM) was purchased from Calbiochem.
Correlations between variables were performed by using the chi-square test and the one-tailed Spearman’s test (SPSS version 9.0; SPSS, Chicago, Ill.). For in vitro studies, the statistical significance was assessed by using Student’s t-test for independent groups. The relationship between the percentages of hSDF1α-induced DNA synthesis and the MIB1-SI scores was assessed by means of linear regression analysis. P 0.05 was accepted as statistically significant.
Clinicopathological variables (age, gender, histology, and tumor grade) were collected for each patient at the time of diagnosis (Table 1). The 55 meningiomas were graded and classed into histological subtypes as follows: 47 WHO grade I (28 meningothelial, 11 fibrous, and 8 transitional), 5 WHO grade II (atypical), and 3 WHO grade III (anaplastic).
All tumors were analyzed for CXCR4 and SDF1 gene expression by RT-PCR. The size of the PCR products corresponded to the predicted length of the synthesized cDNA fragments based on the position of the PCR primers. We assessed β-actin fragment amplification in all the experiments as an efficiency control. Moreover, no PCR products were obtained when RT was omitted (data not shown).
In the 55 meningiomas, CXCR4 mRNA was expressed in 43 specimens (78%), whereas SDF1 mRNA was found in 29 (53%) (Table 2). CXCR4 mRNA was present in 37 (79%) of the 47 WHO grade I tumors, four (80%) of the five grade II tumors, and two (67%) of the three grade III tumors, whereas SDF1 was detectable in 25 (53%) of the 47 grade I tumors, three (60%) of the five grade II tumors, and one (33%) of the three WHO III tumors. CXCR4 and SDF1 were coexpressed in 27 (49%) of the 55 tumors, whereas only 10 (18%) of the 55 expressed neither protein. The chi-square test showed a highly significant association between CXCR4 and SDF1 expression (P = 0.0047). Of the 55 specimens, 16 expressed CXCR4 but were negative for SDF1, whereas only two tumors expressed SDF1 but were negative for the receptor CXCR4.
However, the frequency of tumors positive for CXCR4 and SDF1 was not significantly associated with the tumor grade, when low-grade (WHO I) tumors were compared with high-grade (WHO II/III) meningiomas characterized by increased risk of recurrence and aggressive growth.
Furthermore, we found no significant correlation between the patterns of expression of both CXCR4 and SDF1 and other clinicopathological variables such as gender, age, and histological subtype (data not shown).
To confirm the data obtained in the RT-PCR experiments at the protein level and to determine the cell type that actually expresses CXCR4 and SDF1 in meningioma tumor tissues, we performed IHC analysis on paraffin-embedded sections of 14 specimens of tumor in which the expression of CXCR4 and SDF1 was previously verified by RT-PCR. The sections were stained for CXCR4 and SDF1 by using two different antibodies for each antigen (see Materials and Methods), and similar results were obtained with both antibodies, ensuring the specificity of the staining observed (data not shown). Figure 1 depicts the IHC of three representative meningiomas (two meningothelial and one fibrous) stained with the monoclonal anti-CXCR4 (clone 12G5) and the polyclonal anti-SDF1. CXCR4 and SDF1 immunoreactivity was widely present in the tissues studied. Panels A–F of Fig. 1 show a meningothelial meningioma with tumor infiltration into the dura mater, particularly evident in the hematoxylineosin stain (panel A) and in EMA (panel D), which stains in detail the meningioma tissue but not the leptomeninges. Serial sections of this specimen were analyzed for CXCR4 (panels B and C) and SDF1 (panels E and F) expression, showing that both molecules were specifically expressed in the tumor cells, as well shown in the higher magnifications.
IHC for CXCR4 and SDF1 expression was also shown in another meningothelial tumor (Fig. 1, panels G–I). Similarly in this case, both molecules were present in the sections where the stain appears homogeneous in the tumor cells and was absent in the thin collagenous septae (Fig. 1, panels H and I).
Panels J–L of Fig. 1 show an example of a fibrous meningioma sample in which both CXCR4 and SDF1 were markedly expressed. In particular, in this histological subtype, characterized by spindle-shaped fibroblast-like cells, forming fascicles in a collagen matrix, we observed SDF1 staining (panel L) more homogeneously than did CXCR4 (panel K), which appeared prominent in the cellular whorls.
The expression of Ki-67 protein, which is recognized by the MIB1 antibody, correlates with the proliferative status and meningioma grading and recurrence (Matsuno et al., 1996; Perry, 1998). IHC for MIB1 was performed in all the available samples (40 cases) of the meningioma tissues examined in this study (Table 3). As expected, values for MIB1-SI significantly increased (P = 0.01) with increasing WHO tumor grade: The means were 1.9% (range, 0.1%–7.2%), 4.8% (range, 3.4%–9.8%), and 9.3% (range, 7.8%–10.4%) for WHO grades I, II, and III, respectively.
We therefore correlated the expression of CXCR4 and of SDF1 with the MIB1 staining by using an MIB1-SI of 3% as a cutoff value (Matsuno et al., 1996). The tumors were grouped as 14 cases with MIB1-SI greater than 3% and 26 cases with MIB1-SI less than 3% (Table 3). Analysis of these data, considering either an MIB1-SI cut off of 3% or the individual MIB1-SI (actual counts), found no significant relationship between the MIB1 proliferation marker and the CXCR4/SDF1 expression. However, the statistical analysis found a significant correlation (P = 0.045) between the CXCR4 expression and the MIB1-SI for only the WHO grade I meningiomas, demonstrating that grade I meningiomas lacking CXCR4 had a lower MIB1-SI (Fig. 2). Tumors of WHO grades II and III, on the other hand, showed similar MIB1-SI independently of CXCR4 expression (P = 0.18).
We analyzed the ability of hSDF1α to induce cell proliferation in primary cultures derived from dispersed tumor cells from 12 postsurgical grade I meningioma specimens (seven meningothelial, one fibrous, and four transitional) by means of the [3H]thymidine incorporation assay. In these cultures, immunostaining with EMA, which is used to identify meningeal tumor cells (Arena et al., 2004), demonstrated the presence of neoplastic cells for more than 95% (Fig. 3). Moreover, cytochemical analysis using CXCR4 antibody confirmed the expression of this receptor in the cultured cells at the time of the in vitro experiments: The presence of CXCR4 was confirmed in all the primary cultures as previously demonstrated by RT-PCR on the corresponding surgical specimen (Fig. 3). In these primary cell cultures, anti-CXCR4 mainly immunostained the cell membrane while the EMA positivity appeared diffuse (Fig. 3). To evaluate the effects of SDF1 on meningioma cell proliferation, cultures were serum starved for 48 h before the stimulation with increasing concentrations of hSDF1α for 24 h.
The treatment with hSDF1α significantly increased the [3H]thymidine uptake in 9 (75%) of the 12 meningioma cultures. Figure 4A depicts the evaluation of hSDF1α-induced DNA synthesis in the 12 primary cultures; tumors 1 and 2 reached a maximum increase in DNA synthesis at 25 nM hSDF1α (+323% and +344%, respectively, vs. untreated controls), whereas tumors 3, 4, 5, 7, 8, and 9 showed peak stimulation after exposure to 12.5 nM hSDF1α (+121%, +88%, +109%, +56%, +65%, and +19%, respectively, vs. untreated controls). Tumors 3 and 4 also showed a significant increase in [3H]thymidine uptake at 50 nM hSDF1α stimulation, whereas tumors 1, 2, and 5 showed no further significant increase in DNA synthesis at higher concentrations. In one case (tumor 10), the cell proliferation was significantly stimulated only with 37.5 nM hSDF1α. Among the three cases that did not respond to hSDF1α, tumor 6 showed an increase in DNA synthesis after the chemokine treatment, although the increase was not statistically significant, probably because of the intraexperimental variability (+29%, P = 0.1, and +31%, P = 0.3, at 12.5 and 25 nM hSDF1α, respectively).
The rate of cell proliferation induced by hSDF1α stimulation of these primary tumor meningothelial cells was correlated with the MIB1-SI when available (9 of 12 cases). A significant positive association (P = 0.01) was observed between the two variables: Tumors showing higher percentages of MIB1-SI displayed a greater increase in DNA synthesis in response to 12.5 nM hSDF1α stimulation (Fig. 4B).
ERK1/2 activation is one of the main intracellular pathways involved in cell proliferation induced by hSDF1α treatment in several cell types (Bajetto et al., 2001; Barbero et al., 2003; Porcile et al., 2005). We therefore analyzed ERK1/2 activation after hSDF1α stimulation in four primary meningioma cultures. Cells derived from tumors 1, 4, 5, and 6 were serum starved before being stimulated with 12.5 nM hSDF1α for 2 min, and total cell lysates were analyzed by Western blot for the phosphorylated form of ERK1/2. The same lysates were also probed for the expression of α-tubulin to ensure the equal loading of proteins in the different lanes. We found that hSDF1α induced ERK1/2 phosphorylation in all primary cell cultures tested (Fig. 5A), ranging from 1.7- to 3-fold over basal, as quantified by the densitometric analysis (Fig. 5B).
To verify whether the effects of hSDF1α on proliferation and on ERK1/2 activation were correlated, we analyzed the hSDF1α-induced cell proliferation in the presence of PD98059, which inhibits mitogen-activated protein (MAP) kinase kinase. Primary cultures were treated with PD98059 for 10 min before being treated with hSDF1α in the presence or absence of the inhibitor and then analyzed by [3H]thymidine incorporation. We found that PD98059 pretreatment strongly reduced the uptake of [3H]thymidine induced by hSDF1α in five primary cultures (−72%, −73%, −32%, −73%, and −48%, for tumors 1–5) and also reduced the basal DNA synthesis in two cases (−56% in tumor 4 and −40% in tumor 5) (Fig. 6).
Our findings demonstrated that the ERK1/2-mediated intracellular signaling pathway is involved in the cell proliferation induced by hSDF1α in primary meningioma cultures.
In the present study, we analyzed the expression and functions of CXCR4 and SDF1 in 55 postsurgical specimens from human meningiomas, as possible molecular players in tumor cell proliferation in vitro.
To our knowledge, this is the first report demonstrating the expression of SDF1 and its receptor in meningiomas, although SDF1 was previously shown to be expressed in embryonic and postnatal meninges and to play an important role during cerebellar and hippocampus development (Bagri et al., 2002; Lu et al., 2002; Reiss et al., 2002; Zhu et al., 2002). It has been reported that SDF1 is synthesized and secreted by meningeal cells in the cerebellum and allows the attraction of the stem cells of the external granule layer, which expresses the CXCR4, to the proliferating compartment. The CXCR4 expression is absent in the postmitotic cells that migrate in the inner layer (Reiss et al., 2002; Zhu et al., 2002). The aberrant laminar structure of fetal cerebellum described in SDF1 and CXCR4 knockout mice supports these findings (Ma et al., 1998; Nagasawa et al., 1996; Zou et al., 1998).
Similarly, CXCR4 plays an essential role in the development of hippocampal dentate gyrus in which SDF1, which is highly expressed in the meninges that overlay the hippocampus, has proliferative effects on dentate granule cell precursors and provides a guiding chemoattractant signal for CXCR4-expressing cells that migrate toward the dentate gyrus (Bagri et al., 2002; Lu et al., 2002).
We evaluated CXCR4 and SDF1 expression in meningiomas (using RT-PCR confirmed by IHC) in a series of surgical meningioma specimens. CXCR4 mRNA was identified in 78% of the meningiomas, whereas SDF1 was found in 53%. There was a significant relationship between the CXCR4 and SDF1 mRNA positivities, whereas their expression did not correlate with WHO tumor grade. This lack of correlation may be due in part to the heterogeneous group of samples in the study, which included a large number of WHO grade I tumors (85%) and only few cases of high WHO grade meningiomas, which corresponds to the real incidence of these tumors. Also, MIB1-SI did not correlate with the CXCR4/SDF1 expression when all WHO grades were considered together for the statistical analysis. However, when only the WHO grade I tumors were considered, MIB1-SI was significantly correlated with CXCR4 expression (P = 0.045). Interestingly, this correlation was not found in the grade II and III tumors. These findings support the hypothesis that CXCR4 expression may represent an early factor regulating cell proliferation in benign tumor growth. On this basis, CXCR4 may represent a potential novel target to control the cell proliferation in the early stages of the disease, and it may help to identify a subset of meningiomas with higher proliferative potential. In the high-grade meningiomas, the role of CXCR4 becomes less relevant because cell growth is highly deregulated and many independent factors are involved in this process. To identify which cell types express SDF1 and CXCR4 in meningioma tissues, we performed IHC. The stain appeared to be abundant and homogeneous in all tissues analyzed, either meningothelial or fibrous meningioma, and absent in the collagen matrix. Of interest, the SDF1 localization was often associated with the expression of its receptor and was colocalized with EMA immunostaining. This was strongly evident in the tissues in which meningioma invaded the meninges.
In human glioblastoma multiforme cell lines in which CXCR4 and SDF1 were coexpressed, stimulation with hSDF1α or its overexpression increased cell proliferation, probably through an autocrine/paracrine loop (Barbero et al., 2003). The lack of established meningioma cell lines of human origin prompted us to study the CXCR4 function and its intracellular signaling in primary cultures of dispersed meningioma cells. Importantly, although in many circumstances tumor cells change their pattern of receptor expression when removed from the in vivo setting to in vitro primary cultures, this phenomenon did not occur in meningioma cells for CXCR4 expression, as we demonstrated in our immunocytochemical experiments. We demonstrated that CXCR4 stimulation induces a mitogenic signal also in cultured primary meningeal tumor cells; in fact, hSDF1α induced a statistically significant cell proliferation in 9 (75%) of the 12 cultures obtained from the meningiomas, which in a few cases was extremely pronounced (up to +340%). It is noteworthy that the hSDF1α-induced cell proliferation in the primary meningothelial cells establishes a significant relationship with the MIB1-SI values in the corresponding tumor specimens. Thus, it appears that the expression of CXCR4/SDF1 in the tumors corresponds to different levels of activity, because the proliferative potential of tumor meningothelial cells varied among tumors of the same histological grade. To better correlate these two parameters, a quantitative analysis of the levels of CXCR4/SDF1 mRNAs or proteins will be necessary. However, the significant relationship between the proliferative effects of SDF1 and the tumor MIB1-SI established an important correlation between our in vitro observation and the in vivo behavior of these tumors, thus, on the one hand, validating our in vitro observation and, on the other, strongly supporting the role of CXCR4 activity in meningioma proliferation.
In normal and tumor cells, the transduction of proliferative signals involves the activation of ERK1/2, whose enzymatic activity increases in response to hSDF1α. We showed that hSDF1α/CXCR4 interaction increases ERK1/2 phosphorylation in primary cultures of tumor meningeal cells and that the hSDF1α-induced proliferation is reduced by PD98059 (the inhibitor of MAP kinase kinase), indicating that ERK1/2 is involved in the proliferative signal of hSDF1α. These results agree with the observations that in primary tumor cells the CXCR4 receptor is functional with a variety of signaling pathways activated (Balkwill, 2004).
The expression of CXCR4 and SDF1 in human meningiomas is of particular interest: Its activity in meningioma cells could contribute to the biological features of the neoplasm, such as the ability to survive and to growth autonomously. In particular, because we observed the expression of both SDF1 and its receptor CXCR4 in the same tumor, the occurrence of an autocrine/paracrine loop can be hypothesized, providing a proliferative advantage for the cells sensitive to SDF1 stimulation.
These results strengthen the notion of targeting chemokine networks involved in tumor progression as both a therapeutic approach and a chemopreventive strategy, blocking the development of more aggressive tumors (Balkwill, 2004).
Indeed, it has been proposed that CXCR4 or its ligand could be a potential target for therapeutic intervention. Numerous pharmacological agents exist that can modulate SDF1/CXCR4-induced responses both in vitro and in vivo (Coscia and Biragyn, 2004). Inhibitors of CXCR4 are already available and currently under investigation in preclinical studies of human cancers (Rubin et al., 2003). A better understanding of the role of CXCR4 and SDF1 in brain tumors will enable a greater manipulation of this important biological axis to affect the outcome of this disease. An increased knowledge of the SDF1 interaction with its receptor, and of the different signaling events resulting in SDF1-induced responses, will also enhance future drug design.
In this view, our results provide an interesting basis for further investigations that should be performed to characterize the roles of CXCR4 and SDF1 in the proliferation of meningiomas.
The authors thank Mr. Paolo Pirani for technical support.
1This study was supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC) and the San Paolo Fondazione (Torino, Italy).
3Abbreviations used are as follows: EMA, epithelial membrane antigen; ERK1/2, extracellular signal–regulated kinase 1/2; hSDF1α, human SDF1α; IHC, immunohistochemistry; MAP, mitogen-activated protein; MIB1-SI, MIB1 staining index; PBS, phosphate-buffered saline; RT, reverse transcriptase; SDF1, stromal cell–derived factor 1; SDS, sodium dodecyl sulfate; TBS, Tris-buffered saline.