Shc3 affects human high-grade astrocytomas survival

Abstract

A selective switch from expression of Shc1 gene to Shc3 occurs with maturation of neuronal precursors into postmitotic neurons. Previous studies showed that in the embryo, Shc1 is maximally expressed in dividing CNS stem cells while it is silenced in mature neurons, where it is replaced by Shc3. Under normal conditions Shc3 is never expressed by glial cells. We now show that in human astrocytomas and glioblastomas, the normal pattern of expression of Shc1/Shc3 is totally subverted, both proteins being present at the same time and in the same cells. Our data indicate that Shc3 is maximally expressed, together with Shc1, in glioblastoma, a highly proliferative tumor with little, if any, indication of neuronal differentiation. In primary cultures of glioblastoma, tumor cells maintain Shc1 expression but downregulate Shc3. Analysis of the phosphorylation status of Shc3 in human glioblastoma tumor samples in vivo indicates that it is tyrosine phosphorylated. Finally, we found that the expression of truncated variants of Shc3 with dominant-negative effects in human high-grade glioma cells that maintain Shc3 expression in vitro leads to a decreased Akt posphorylation and increased apoptosis, thus resulting in impaired survival of the transfected cells. These data suggest that Shc molecules play an important role in glioblastoma cell growth and survival.

Introduction

Combined activation of MAPK and Akt pathways mediated by receptor tyrosine kinases (RTK) plays an important role in the development, growth and survival of malignant astrocytomas. Among the many proteins involved in those signal transduction pathways during normal development of the CNS, the Shc (Src family of adaptor proteins) is exquisitely regulated. Indeed, it has been shown that different genes of this family alternate their expression according to specific spatio-temporal rules and the emergence of cellular specificities. Shc3 gene expression is tightly regulated, starting at the beginning of neurogenesis and being maintained in postmitotic neurons of the brain through the entire life of the animal (Cattaneo and Pelicci, 1998). During brain development, the increase of Shc3 levels is accompanied by a sharp decrease in Shc1 gene expression, the latter shown to be maximal in proliferating neuro-glial precursors (Conti et al., 2001). Glial cells under normal conditions do not show any Shc3 protein, while Shc1 expression is barely detectable in quiescent glial cells (Conti et al., 2001) becoming highly expressed only after glial activation in pathological conditions (Russo et al., 2002). In conditionally immortalized neural precursors, Shc3 expression has been shown to favor persistent activation of MAPK and increased Akt phosphorylation, resulting, respectively, in enhanced differentiation and survival (Conti et al., 2001). Similar data were obtained in neuroblastoma cells (Pelicci et al., 2002) where Shc3 was found to establish stable complexes with the p85 regulatory subunit of PI3K. Data on Shc3 expression in human tumors are scanty and previous studies indicate that Shc3 is expressed in some but not all neuroblastoma derived cell lines (Pelicci et al., 2002).

Results

High-grade astrocytomas express Shc1 and Shc3

In order to investigate whether low- and high-grade astrocytomas express Shc1 and Shc3 proteins, we performed Western blot analysis of multiple samples taken from human adult brain, low-grade astrocytomas, anaplastic astrocytomas and glioblastomas. Shc3 protein isoforms (p52Shc3 and p64Shc3) were present in normal brain, undetectable in low-grade astrocytomas and detectable in the majority of high-grade tumors. In all tumors, p64Shc3 was only detectable when p52Shc3 was also expressed; however, in some samples only p52Shc3 was present (Figure 1). All of the three Shc1 isoforms (p46/52Shc1 and p66Shc1) were present in the tumor samples, albeit at lower level in low-grade astrocytomas, while the same bands were undetectable in non-neoplastic temporal and parietal human cortex (Figure 1). All samples were collected at surgery from an area of the tumor that did not show, at least macroscopically, indications of containing non-neoplastic tissue or clear signs of necrosis. Parallel samples from the same area were submitted to histological analysis to confirm the absence of normal brain tissue and necrosis. Furthermore, in order to exclude the possibility that, due to the large microscopic variations present in the texture of high-grade astrocytomas, the presence of Shc3 immunoreactive bands might be due to contaminating nerve cells engulfed by the growth of the tumor, we performed immunocytochemical analysis using an anti-Shc3 antibody on cryostat sections of selected low- and high-grade tumor samples. Shc3 positive cells were detected in all tumor samples resulted positive to Western blot analysis. Strong Shc3 immunoreactivity was detected in highly displastic multinucleated cells, whose neoplastic nature was obvious (Figure 2). The results of immunocytochemistry clearly showed that Shc3 presence in high-grade astrocytomas was not due to normal neurons entrapped by the invasive nature of the tumor. Furthermore, double staining showed that many cells are immunoreactive for both Shc3 and Shc1 (not shown), as expected considering the elevate number of positive cells in high-grade tumors when a single antibody was used.

Figure 1
figure1

Shc3, Shc1 and β-actin expression in human astrocytomas and nontumoral human brain tissue. (a) Western blot analysis anti-Shc3, Shc1 and β-actin monoclonal antibodies. Shc3 protein is detected both in human brain (C: human brain sample from trauma patient, B: lobectomy peritumoral brain), in anaplastic astrocytoma (1), in glioblastoma (6, 7, 8, 10). In each lane, 50 μg of total proteins were loaded. Shc3 proteins are not detectable under comparable conditions in astrocytomas grade II (11). In all lanes, 50 μg of total proteins were loaded. Both p52Shc3 and p64Shc3 are present in normal brain, but only p52Shc3 is always detectable in anaplastic astrocytomas and glioblastoma. Shc1 protein is detectable in all tumor samples, albeit at lower level in low-grade tumors, but not in peritumoral brain. All protein isoforms of Shc1 are usually detected. (b) Composite diagram showing the intensity of the p52Shc3 band of tumor samples relative to the average intensity of the controls (brain) as obtained by densitometry. The dots refer to 6 glioblastomas (GBM), 4 anaplastic astrocytomas, 4 (AA) astrocytomas grade II samples (LG), and normal astrocyte culture (Astro). Column height: Average intensity of the samples, filled circles: values of each samples (some circles overlaps). RI (relative intensity): ratio between the densitometric value of the band and the average value of control bands. The line in 1 marks the value of control brain, values equal or below 0.01 of controls are indistinguishable from background

Figure 2
figure2

Detection of Shc3 protein in human glioblastoma tumor samples by immunohistochemistry, cryostate section (10 μm thick). The sections were reacted with the anti-Shc3 monoclonal antibody used in Figure 1. The antibody was revealed according to the ABC protocol (Vector lab, CA, USA); the sections were lightly counterstained with hematoxylin. In (a), almost all tumoral cells contain Shc3 immunoreactivity, in (b) only about 30% of the cells are labeled. Scale bars: 10 μm

Primary cell cultures of human glioma maintain Shc1 expression while Shc3 expression is downregulated

We tested human glioblastoma cells freshly dissociated from three surgical specimens and maintained in vitro in medium supplemented with serum or growth factors (bFGF and EGF). Under these conditions, all Shc1 isoforms were detectable and their levels remained unchanged compared to their levels in vivo. On the contrary, under both growth conditions, Shc3 levels were rapidly reduced compared to those exhibited in vivo and becoming undetectable by Western blotting after two culture passages (Figure 3). Established human glioma cell lines were tested for Shc1 and Shc3 expression (Table 1). All of these cell lines showed Shc1 expression, while Shc3 was detectable only in the U87MG cell line. Since Shc3 has been reported to be normally coexpressed with neuronal markers in nerve cells, we studied by Western blotting the presence of neurofilaments and NeuN immunoreactivities, a group of typical neuronal antigens, in U87MG cells that retains Shc3 protein and in the remaining glioma cell lines that lack Shc3 expression (not shown). We did not observe any correlation between Shc3 expression and the presence of neuronal markers, indicating that Shc3 presence in U87MG cells is not sufficient to induce neuronal differentiation.

Figure 3
figure3

Western blot analysis of glioblastoma tissue lysate (0) and lysates of cells dissociated from the same tumor and maintained in vitro in the presence of 10% fetal calf serum for one or two passages. The membrane was probed with antibodies against Shc3 (upper panel) and against β-actin (lower panel). The same amount of total protein was loaded in each lane. Noti the clear reduction in immunoreactivity for Shc3 with increasing passages in vitro

Table 1 The table lists the presence (+) or absence (−) of Shc1 and Shc3 protein isoforms as demonstrated by western blot in 9 human cell lines derived from astrocytic tumors

Shc3 is phosphorylated in high-grade astrocytomas in vivo and in vitro

In order to study the tyrosine phosphorylation status of Shc3 in high-grade astrocytic tumors, we performed immunoprecipitation assays on selected tumor samples (Figure 4a). We found Shc3 to be tyrosine phosphorylated both in vivo and in vitro. This result suggests the involvement/recruitment of p52Shc3 in the mediation of the activity of signal transduction pathways in brain tumor cells. Previous studies demonstrated that both Alk (anaplastic lymphoma kinase) and c-Ret were able to phosphorylate Shc3 in neuroblastoma cell lines (Miyake et al., 2002; Pelicci et al., 2002). We tested by Western blotting the presence of these kinases in those tumors samples expressing both Shc1 and Shc3. In selected high-grade tumor samples that we found positive for Shc3, Alk was detectable but c-Ret was under the detection limits of our technique. Both c-Ret and Alk were detectable in total extracts of U87MG, the only high-grade glioma cell lines tested that maintains strong p52Shc3 expression in vitro. Co-immunoprecipitation assays performed on total extracts of the same cells indicated a direct interaction of Shc3 with activated c-Ret (Figure 4b). No coimmunoprecipitation was detected between Shc3 and activated Alk on the same samples (Figure 4b).

Figure 4
figure4

(a) Total lysates of glioblastoma sample (6) and the human glioblastoma cell line U87MG (U). Upper panel: Western blot analysis of 50 μg of total tissue or cellular lysate probed with a monoclonal antibody against Shc3. In sample U, only the p52Shc3 protein was detected while the less abundant p64Shc3 protein was detected only in 6. Lower panel: total lysates from the same samples were immunoprecipitated with the same antibody against Shc3 and after electrophoresis and blotting probed with an antibody against phospho-tyrosine (pY). p52Shc3 tyrosine phosphorylation was detectable both in vitro and in vivo (6) (U). (b) Total lysate of U87MG cells were immunoprecipitated with Shc3 antibody and after electrophoresis and blotting probed with anti-phospho (Tyr 905) c-Ret (left) and anti-Alk (right). Coimmunoprecipitation is present only with posphoprylated c-Ret protein

Forced p52Shc3 expression has a modest inhibitory effect on growth of those glioblastoma cells that downregulate endogenous Shc3 in vitro

Shc3 expression in normal adult brain is restricted to postmitotic neurons, and our immunohistochemical analysis showed that not all cells in a glioblastoma contain high levels of Shc3. We were interested to evaluate if high level of Shc3 expression in glioblastoma cell lines that in vitro downregulate Shc3 could cause growth arrest or cell death. To test this hypothesis, we transfected the human p52Shc3 cDNA inserted in an EGFP bicistronic vector into Hu197 cell line. Analyses of Shc3-overexpressing clones showed that the transgene expression did not heavily interfere with cell replication. Indeed, the average growth index of Shc3 expressing clones was approximately 0.85 of control clones and of the parental Hu197 cells (Figure 5). Similar results were obtained by using clones of the same cells obtained by infection with a replication defective retroviral construct carrying both p52Shc3 cDNA and EGFP under the control of different promoters (data not shown).

Figure 5
figure5

Growth curves of Hu 197 (Hu 197, densely stippled bars) and subclones of Hu197 derived by stable transfection with retroviral vectors driving either the expression of EGFP alone (Vector, stippled bars) or the expression of p52Shc3 (three clones) and EGFP from different promoters (Shc3a, b, c, darker bars). All data derive from MTT analysis performed 24, 48 and 72 h after plating. Values of absorbance (Abs) were read at 550 nm. The experiments were performed in quadruplicate. Column heights represent average and the error bars refer to s.d. The growth of Shc3 expressing cells was significantly (P<0.05) lower than those of the parental cells (Hu 197) and clones expressing EGFP (vector). Difference in growth of Hu 197 and Vector cells were not significant. Unpaired T-test was employed to estimate significance of the differences between groups. Growth of Shc3 expressing Hu 197 was on average 0.15 times lower than growth of the parental cell line

Expression of dominant-negative variants of Shc3 in U87MG cells induces cell death

In order to investigate the role of p52Shc3 in glioma cells, we expressed separately both the PTB and SH2 domain of p52Shc3 (PTB-Shc3, SH2-Shc3) in U87MG and Hu197 cells. For these experiments we transfected either a plasmid for the coexpression of PTB or the SH2 domain and EGFP from independent promoters or the same domains of p52Shc3 fused with the amino-terminus of ECFP, a cyan variant of EGFP. Expression of the PTB and the SH2 domains of Shc3 separated from the rest of Shc3 protein has been shown to exert dominant-negative effects on Shc3 function (O’Bryan et al., 1998). Clones of Hu197 cells stably expressing the dominant-negative fragments were readily obtained either by multiple FACS enrichments or by antibiotic selection. These clones were similar in their growth potential compared to the wild-type cells, thus indicating the specificity of the PTB dominant-negative effects, since these cells do not express Shc3 at a detectable level. On the contrary, under similar conditions, we were unable to obtain stable clones expressing PTB-Shc3 or SH2-Shc3 dominant-negative molecules in U87MG cells that maintain p52Shc3 expression in vitro. Following transfection with PTB-Shc3 domain, U87MG cells stop dividing and underwent cell death. Similar results were obtained by transfecting U87MG cells with constructs where either the PTB or the SH2 domains were fused to the amino terminal end of ECFP (Figure 6a and b). At 2 days after polyethilenimine mediated transfection of U87MG cells with PTB-ECFP or SH2-ECFP constructs, caspase-3 activation was increased (Figure 6c). This result was further confirmed by propidium iodide staining of the cultures showing that cells expressing the dominant-negative Shc3 constructs fused to ECFP were readily distinguishable by their fluorescence and vital staining with propidium iodide, demonstrating that starting from 2 days after transfection they underwent apoptotic cell death (Figure 6d). Finally, 1 week after transfection, we found more than one order of magnitude less cells expressing PTB-ECFP compared to those expressing ECFP alone (Figure 7). Altogether, these results suggest that abrupt impairment of Shc3 function by expression of a truncated variant of the protein with dominant-negative effects is lethal for those high-grade glioma cells that maintain Shc3 expression in vitro.

Figure 6
figure6

Expression of dominant-negative Shc3 constructs in U87MG cells induces cell death. (a) U87MG cells, which spontaneously express p52Shc3, were induced to express, following PEI based transfection, either ECFP (A24, A96) or a fusion protein between the PTB domain of Shc3 and ECFP (PTB-ECFP) (B24, B96). Each picture represent an electronic overlapping of microphotographs of the same field taken using phase interference optics or fluorescent light in order to show the intrinsic ECFP fluorescence. Microphotographs were taken 24 (A24, B24) and 96 h after transfection (A96, B96). At 24 h after transfection, fluorescent cells (some are indicated by the arrows) were detected both in cells transfected with ECFP (A24) and in cells transfected with the Shc3 dominant-negative construct PTB-ECFP (B24); on the contrary, at 96 h viable transfected cells (again some of these cells are indicated by arrows) were more numerous in cells transfected with ECFP (A96). (b) Western blot analysis showing caspase-3 activation 24 h after transfection with the SH2 domain of Shc3 (U87MG SH2), or a cyan variant of the green fluorescent protein (U87MG ECFP); U87MG is the parental cell line. Despite only 20% of the cells were transfected, caspase-3 activation is higher in SH2–Shc3 cells. Activation of caspase-3 is also present but at lower levels, in the absence of the dominant-negative and it is due to spontaneous apoptosis that is always detected during U87MG growth in vitro. (c) At 48 h after transfection (048) with PTB-ECFP cells were vitally stained with propidium iodide a dye, which under these conditions, stains only the nuclei of dying cells. The arrows and the arrowheads point to cells expressing PTB-ECFP containing apoptotic nuclei. PC: phase contrast view, PTB-ECFP: same field shown under conditions specific for demonstrating ECFP fluorescence, PI same field shown under conditions that demonstrate propidium iodide fluorescence. Scale bars: 10 μm

Figure 7
figure7

Expression of PTBShc3 domain in U87MG cells induces cell death. We counted the number of fluorescent and not fluorescent U87MG cells 24 and 96 h after PEI based transfection with pECFP-N1 (ECFP) or the same plasmid after cloning the PTB-Shc3 domain in frame with the ECFP ORF in order to obtain expression of the fluorescent fusion protein PTB-ECFP. Column heights represent the average ratio between fluorescent and not fluorescent cells, the error bars refer to s.d. The experiments were performed in triplicate. At 96 h after transfection, the number of PTB-ECFP expressing cells was significantly (P<0.05) lower than the number of cells expressing ECFP. Unpaired T-test was employed to estimate significance of the differences between groups of data

Shc3 dominant-negative constructs induced apoptosis in U87MG cells is preceded by a decrease in Akt phosphorylation

Shc3 proteins can form stable complexes with the p85 regulatory subunit of PI3K and increase Akt phosphorylation, resulting in enhanced survival of immature neurons and neuroblastoma cells (Conti et al., 2001; Pelicci et al., 2002). We compared Akt phosphorylation on serine 473 of stable clones of Hu197 after infection with a retroviral vector that induced p52Shc3 expression to other stable clones of Hu197 cells infected with the empty vector that induces only EGFP expression. While total Akt was similar, Akt phosphorylation on serine 473 increased in Hu197 clones expressing p52Shc3 compared to Hu197 cells expressing EGFP (Figure 8a). We also measured Akt phosphorylation 24 h after transfection of U87MG cells with the dominant-negative variants of Shc3 (PTB and SH2). In these cells we found a decrease in phosphorylated Akt compared to U87MG cells transfected with control plasmids (Figure 8b). The decrease preceded the onset of apoptosis that peaks 48–72 h after transfection (Figure 6).

Figure 8
figure8

(a) Western blot analysis showing an increase in Akt phosphorylation on serine 473 (Akt-P) in a stable clone of Hu197 cells expressing p52Shc3 after retroviral infection (HU197 Shc3) compared to another clone of Hu197 cells infected with the empty vector (HU197 Mock). Total Akt (Akt) and α-tubulin immunoreactivities remained stable. (b) Western blot analysis showing Akt phosphorylation 24 h after transfection of U87MG cells with the dominant-negative variants of Shc3 (PTB and SH2) or ECFP. Akt phosphorylation on serine 473 decreased in cells transfected with dominant-negative variants of Shc3 compared to U87MG cells transfected with ECFP. The total amount of Akt remained the same

Discussion

Our data provide evidence of altered expression of Shc genes in high-grade astrocytomas. Analysis of surgical specimens derived from patients operated for glioma resection has revealed that Shc1 gene is expressed in all samples while Shc3 is expressed only in 70% of the tumors. Comparing Shc3 content and histological grade of the tumor, we found that the higher levels of both p52Shc3 and p64Shc3 isoforms are present in high-grade astrocytomas (glioblastoma and anaplastic astrocytoma). In normal brain, Shc3 expression is limited to mature neurons and the protein is either not present in glial cells or it is undetectable by Western blotting. Shc1 is expressed ubiquitously in the body (not in the brain) and its expression has been previously reported in human brain tumors (Cattaneo et al., 1998; Hecker et al., 2002), where it participates in Ras activation by signals originating from activated RTK and focal adhesion kinase (Guha et al., 1997; Hecker et al., 2004). However, Shc1 and Shc3 are not coexpressed in normal developing brain and during postnatal life (Conti et al., 2001). On the contrary, anaplastic astrocytomas and glioblastoma in our series inappropriately contain both Shc1 and Shc3 proteins. In normal neuronal cells, Shc3 has been reported to favor neuronal maturation by activation of Erk/MAPK pathway and to positively influence neural cell survival by recruiting the PI3K–Akt–Bad pathway (Conti et al., 2001). Our results show that, in high-grade astrocytomas, the presence of both Shc3 protein isoforms does not increase the level of expression of neuronal markers either in vivo and in vitro. On the other hand, in vitro overexpression of p52Shc3 in cell lines derived from human malignant astrocytomas not spontaneously expressing Shc3, does not confer any distinct growth advantage over the parental cells. On the contrary, the growth index of stable clones of Hu197 human glioblastoma cell line forced to express p52Shc3 was on average 15% lower than the growth index of the parental cell line. This may account for the rapid downregulation of endogenous Shc3 in vitro in primary glioblastoma cells after few passages under normal growth conditions. Interestingly, Shc3 expression is not the only common alteration present in malignant astrocytomas in vivo to be lost after few passages in vitro: epidermal growth factor receptor (EGFR) amplification is also lost by most high-grade glioma cells in vitro while is maintained in xenografts derived by the same tumors (Pandita et al., 2004). Noteworthy, Shc proteins are substrates for the tyrosine kinase activity of EGFR and their interactions with Grb2 is instrumental for activation of the Erk/MAPK pathway (Guha et al., 1997). Ras activation in glioblastoma cells expressing a truncated EGFR is mediated through Shc-Grb2 interaction (Prigent et al., 1996). The stable presence of high levels of p46/p52Shc1 in human glioblastoma cells makes unlikely that Shc3 will be essential to maintain active this pathway. Furthermore, we showed that T98G and U373MG do not contain detectable Shc3, but their growth in vitro has been reported to be impaired by specifically targeting EGFR functionality (Rebbaa et al., 1997; Rubenstein et al., 2001). Our results show that p52Shc3 is phosphorylated in human glioblastoma samples and U87MG cells. Phosphorylation studies performed in neuroblastoma cell lines showed that Shc3 is constitutively phosphorylated also in these tumor cells, where its tyrosine phosphorylation seems to be mediated by Alk (Miyake et al., 2002), the putative receptor tyrosine kinase for pleiotropin/midkine (Stoica et al., 2001). Alk has been shown to be commonly expressed by glioma derived cell lines like U87MG (Dirks et al., 2002) and in vivo in glioblastoma (Powers et al., 2002). We confirmed Alk expression in U87MG cells but we were unable to coimmunoprecipitate active Alk using an anti Shc3 monoclonal antibody. Previous studies indicate that Alk is not only able to phosphorylate Shc3 but also Shc1 (Fujimoto et al., 1996; Trinei et al., 2000; Stoica et al., 2001), and we speculate that coexpression of both genes in glioblastoma cells may generate competition between the two Alk substrates in favor of Shc1. An alternative explanation is that Alk is weak or not active in U87MG cells under our culture conditions. Shc3 also binds to and is phosphorylated by Ret tyrosine kinase, a component, together with the GFR-alpha1 molecule, of the glial cell line-derived neurotrophic factor (GDNF) receptor complex shown to activate multiple signaling pathways (Trupp et al., 1999; Airaksinen and Saarma, 2002). GDNF is expressed in human glioblastomas and it is constitutively secreted by glioblastoma derived cell lines like U87MG (Wiesenhofer et al., 2000; Nishiguchi et al., 2003). Differently to what we found for Alk, we were able to coimmunoprecipitate phosphorylated Ret by using an anti-Shc3 antibody on total protein extract of U87MG. This suggests that GDNF production by U87MG can stimulate in a paracrine fashion Ret and that Shc3 may be involved in the transduction pathway of this signal. During normal development, the increased content of Shc3 isoforms in maturing neurons is important for their survival (Conti et al., 2001). Our experiments where Shc3 activity was perturbed by expression of truncated variants of p52Shc3 with dominant-negative activity (O’Bryan et al., 1998; Conti et al., 2001) indicated that Shc3 is essential for survival of U87MG glioblastoma cell, the only cell line we found to maintain Shc3 expression. In these cells transient expression of the PTB or the SH2 domains of p52Shc3 either alone or fused to ECFP induced apoptosis as shown by caspase-3 activation and vital staining with propidium iodide. Akt phosphorylation under the same experimental conditions decreased. The decrease in Akt phosphorylation preceded the onset of apoptosis of the transfected cells that become apparent approximately 48 h after transfection. In U87MG cells PTEN is mutated (Li et al., 1997) and PTEN phosphoinositide 3-phosphatase activity is null (Su et al., 2003). PTEN inactivation in U87MG cells results in increased Akt phosphorylation compared to clones where normal PTEN activity is restored by transfection of the appropriate constructs (Pore et al., 2003). Interestingly, our data indicate that, in U87MG cells, expression of dominant-negative molecules that impair p52Shc3 activity reduces the level of Akt phosphorylation despite the lack of functional PTEN. Shc3 stimulate PI3K activity both in the presence and in the absence of p52Shc3 phosphorylation (Pelicci et al., 2002). Transfection of U87MG cells with dominant-negative variants of Shc3 may affect akt phosphorylation by reducing phosphatidylinositol production by PI3K to an extent that is not compensated by the lack of phosphoinositide 3-phosphatase activity of PTEN.

Altogether, the present data and previous evidence suggest that inappropriate in vivo expression of Shc3 in high-grade glioma may contribute to the survival of the cancer cells. We were able to obtain, by transfection, clones expressing p52Shc3 of human high-grade glioma cell lines that do not express Shc3. Interestingly, these cells were slightly impaired in cell growth compared to the parental cell line. Since Shc3 is connected both during development and by the present experiments with a slow down of the proliferative activity, we can speculate that in vivo its role may be more important for those cells that have a slower mitotic activity or are temporarily not cycling. These cells are notoriously very resistant to apoptosis and consequently difficult to treat by radio and chemotherapy. The presence of these cells is thought to be responsible for the high incidence rate of recurrences that makes glioblastoma such an ominous neoplasm. Our results may open to the exploitation of Shc3 as a potential target molecule for the future development of the therapy of glioblastoma even in those tumors where PTEN activity is impaired.

Materials and methods

Cell lines and tissue samples

Tumor and nontumoral brain tissue samples were collected at surgery from patient operated in the Section of Neurosurgey (Department of Surgery IRCCS Policlinico S. Matteo, Pavia). All samples derived from astrocytic tumors (WHO classification 2001) of different malignancy and from an area of the tumor that did not show, at least macroscopically, indications of containing non-neoplastic tissue or clear signs of necrosis. Parallel samples from the same area were submitted to histological analysis to confirm the absence of normal brain tissue and prevalence of necrosis.

Plasmids and transfection experiments

The human p52Shc3 cDNA was cloned in the sense and antisense orientation in the pIRES2-EGFP (Clontech, Palo Alto, CA, USA), an IRES containing vector for bicistronic expression of the insert and EGFP (enhanced green fluorescent protein) or into the amphotrophic retroviral vector PINCO (Conti et al., 2001). The PTB and SH2 domains of p52Shc3 were obtained by PCR amplification of p52Shc3 cDNA as described by O’Bryan et al. (1998) and cloned in frame to a cyan variant of the EGFP protein in the BamHI–NcoI site of pECFP-N1 vector (Clontech, Palo Alto, CA, USA), or into the retroviral vector PINCO. Retroviral vectors construction and production was described before (Conti et al., 2001). All plasmids were grown in the appropriate Escherichia coli strains and endotoxin-free DNA was prepared by a solid phase purification system (Qiagen, Italy). All transfection experiments were performed according to a published polyethylenimmine (Sigma-Aldrich) based protocol (Lemkine et al., 1999). For the pIRES2-EGFP bicistronic vectors and retroviral vectors, clones expressing EGFP and p52Shc3 were obtained by repeated cycles of FACS sorting and expansion after cloning by limiting dilutions. After six passages in culture and three cycles of FACS sorting, homogeneous cell populations were obtained and were stably propagated in the absence of any antibiotic selection.

Tissue culture and immunohistochemistry

Tumor specimens were obtained during surgery and immediately immersed in HEPES (4-[2-hydroxyethyl]-1-piperazine-ethanesulfonic acid)-buffered Dulbecco's modified Eagle's medium (DMEM). Specimens were mechanically minced and small fragments (maximum diameter 50 μm) were placed in culture in DMEM supplemented with 10% fetal bovine (FBS) serum. After 1 week, cells outgrowing from the fragments were trypsinized and replated. All high-grade glioma cell lines and the clones obtained were cultured in DMEM supplemented with 10% heat-inactivated FBS (GIBCO/Invitrogen) and antibiotics (penicillin, streptomycin and amphotericin B) at 37°C in a 5% CO2 atmosphere. We performed the determination of the growth index in 24-well plate maintained under optimal growth conditions (see above) and growth was determined by MTT assay 72, 96 and 120 h after plating. MTT assay (Mossmann, 1983) was performed as described (Nikkhah et al., 1992) (MTT was used at a final concentration of 1 mg/ml). Absorbance readings were performed at 550 nm with 630 nm as a reference wavelength. We obtained cryostate sections (10 μm) of tissue samples collected from the same tumors that were used for Western blotting analysis. Section were mounted and incubated overnight with the appropriate dilutions of Shc1 (1 : 250 rabbit polyclonal, BD-Transduction) and/or Shc3 (1 : 500 mouse monoclonal clone 23, BD-Transduction) antibodies, after washing and reacting with the appropriate biotinylated secondary antibodies bounded antibodies were revealed by the avidin–biotin complex method with biotinylated horseradish-peroxidase or alkaline phosphatase as reporters (ABC, Vector Lab). Diaminobenzidine (Sigma-Aldrich) and NBT (Roche) respectively were used as substrates for the final enzymatic reactions. At the end of the procedure, the slides were counterstained with hematoxylin, dehydrated and mounted.

Western blotting and immunoprecipitation

Protein extraction and Western blot analysis were performed as described elsewhere (Conti et al., 2001). Briefly, each frozen tissue specimen was homogenized using a Teflon-glass homogenizer in lysis buffer (10 μl/mg of tissue: 10% glycerol; 50 mM Tris, pH 7.5; 150 mM NaCl; 5 mM ethylenediaminetetraacetic acid; 1 mM ethyleneglycol-bis [3]-aminoethyleter]N,N′-tetraacetic acid; 1% Triton X-100; 1 mM Na3VO4; 1 mM ZnCl2 in the presence of 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, and 5 μg/ml leupeptin at 4°C). The extracts were cleared by centrifugation. Protein content was measured using a protein assay kit (Biorad, Italy). Aliquots of extracts were diluted with sodium dodecylsulfate (SDS) sample buffer and boiled for 5 min. Equal amounts of proteins (50 μg) were loaded for SDS–polyacrylamide gel electrophoresis. After transfer to nitrocellulose, the blots were blocked in 10% non-fat dry milk in Tris-buffered saline and Tween-20 (TBS-T) (20 mM Tris (pH 7.5); 500 mM NaCl; and 0.01% Tween-20) overnight at 4°C. The blots were then incubated for 1 h at room temperature in the appropriate primary antibody. The following primary antibodies were employed at the indicated dilutions: Shc1 (1 : 5000 rabbit polyclonal, BD-Transduction), Shc3 (1 : 5000 mouse monoclonal clone 23, BD-Transduction), NeuN (1 : 1000 mouse monoclonal, Chemicon), Neurofilament – PAN (1 : 4 mouse monoclonal, Zymed), Alk (1 : 1000 rabbit polyclonal, Cell Signalling), Phospho-c-Ret (Tyrosine 905) (1 : 1000 rabbit polyclonal, Cell Signalling), Akt (1 : 1000 mouse monoclonal, Cell Signalling), Phospho-Akt (Serine 473) (1 : 1000 rabbit polyclonal, Cell Signalling), Caspase-3 (1 : 1000 mouse monoclonal, Cell Signalling), Active Caspase-3 (1 : 1000 rabbit polyclonal, Cell Signalling), α-tubulin (1 : 1000 mouse monoclonal, SIGMA) and β-actin (1 : 1000 mouse monoclonal, SIGMA). After washing with TBS-T, membranes were exposed to horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. Immunoreactivities were detected using the enhanced chemiluminescence method according to the manufacturer's instructions. For immunprecipitation assays, tissue and cell homogenates were reacted in the presence of protease inhibitors with a mouse monoclonal antibody recognizing the carboxyl terminal portion of Shc3 (clone 23 BD, Transduction Laboratories). The immune complexes were purified from the other proteins by protein G coated magnetic micro beads (μmacs, Milteny), the precipitated proteins were resolved by SDS–PAGE electrophoresis and subjected to Western blot analyses with the appropriate antibodies. Phospho-tyrosine containing proteins were detected after immunoprecipitation by immunolabelling with a biotynilated monoclonal antibody against phosphotyrosine (Cell Signaling).

References

  1. Airaksinen MS and Saarma M . (2002). Nat. Rev. Neurosci., 3, 383–394.

  2. Cattaneo E, Magrassi L, De-Fraja C, Conti L, Di Gennaro I, Butti G and Govoni S . (1998). Anticancer Res., 18, 2381–2387.

  3. Cattaneo E and Pelicci PG . (1998). Trends Neurosci., 21, 476–481.

  4. Conti L, Sipione S, Magrassi L, Bonfanti L, Rigamonti D, Pettirossi V, Peschanski M, Haddad B, Pelicci P, Milanesi G, Pelicci G and Cattaneo E . (2001). Nat. Neurosci., 4, 579–586.

  5. Dirks WG, Fahnrich S, Lis Y, Becker E, MacLeod RA and Drexler HG . (2002). Int. J. Cancer, 100, 49–56.

  6. Fujimoto J, Shiota M, Iwahara T, Seki N, Satoh H, Mori S and Yamamoto T . (1996). Proc. Natl. Acad. Sci. USA, 93, 4181–4186.

  7. Guha A, Feldkamp MM, Lau N, Boss G and Pawson A . (1997). Oncogene, 15, 2755–2765.

  8. Hecker TP, Ding Q, Rege TA, Hanks SK and Gladson CL . (2004). Oncogene, 23, 3962–3971.

  9. Hecker TP, Grammer JR, Gillespie GY, Stewart J and Gladson CL . (2002). Cancer Res., 62, 2699–2707.

  10. Lemkine GF, Goula D, Becker N, Paleari L, Levi G and Demeneix BA . (1999). J. Drug Target, 7, 305–312.

  11. Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Wigler MH and Parsons R . (1997). Science, 275, 1943–1947.

  12. Miyake I, Hakomori Y, Shinohara A, Gamou T, Saito M, Iwamatsu A and Sakai R . (2002). Oncogene, 21, 5823–5834.

  13. Mossmann T . (1983). J. Immunol. Methods, 65, 55–63.

  14. Nikkhah G, Tonn JC, Hoffmann O, Kraemer HP, Darling JL, Schonmayr R and Schachenmayr W . (1992). J. Neurooncol., 13, 1–11.

  15. Nishiguchi M, Tokugawa K, Yamamoto K, Akama T, Nozawa Y, Chaki S, Ueki T, Kameo K and Okuyama S . (2003). Neurochem. Int., 42, 493–498.

  16. O’Bryan JP, Lambert QT and Der CJ . (1998). J. Biol. Chem., 273, 20431–20437.

  17. Pandita A, Aldape KD, Zadeh G, Guha A and James CD . (2004). Genes Chromosomes Cancer, 39, 29–36.

  18. Pelicci G, Troglio F, Bodini A, Melillo RM, Pettirossi V, Coda L, De Giuseppe A, Santoro M and Pelicci PG . (2002). Mol. Cell. Biol., 22, 7351–7363.

  19. Pore N, Liu S, Haas-Kogan DA, O’Rourke DM and Maity A . (2003). Cancer Res., 63, 236–241.

  20. Powers C, Aigner A, Stoica GE, McDonnell K and Wellstein A . (2002). J. Biol. Chem., 277, 14153–14158.

  21. Prigent SA, Nagane M, Lin H, Huvar I, Boss GR, Feramisco JR, Cavenee WK and Huang HS . (1996). J. Biol. Chem., 271, 25639–25645.

  22. Rebbaa A, Yamamoto H, Saito T, Meuillet E, Kim P, Kersey DS, Bremer EG, Taniguchi N and Moskal JR . (1997). J. Biol. Chem., 272, 9275–9279.

  23. Rubenstein M, Glick R, Lichtor T, Mirochnik Y, Chou P and Guinan P . (2001). Med. Oncol., 18, 121–130.

  24. Russo C, Dolcini V, Salis S, Venezia V, Zambrano N, Russo T and Schettini G . (2002). J. Biol. Chem., 277, 35282–35288.

  25. Stoica GE, Kuo A, Aigner A, Sunitha I, Souttou B, Malerczyk C, Caughey DJ, Wen D, Karavanov A, Riegel AT and Wellstein A . (2001). J. Biol. Chem., 276, 16772–16779.

  26. Su JD, Mayo LD, Donner DB and Durden DL . (2003). Cancer Res., 63, 3585–3592.

  27. Trinei M, Lanfrancone L, Campo E, Pulford K, Mason DY, Pelicci PG and Falini B . (2000). Cancer Res., 60, 793–798.

  28. Trupp M, Scott R, Whittemore SR and Ibanez CF . (1999). J. Biol. Chem., 274, 20885–20894.

  29. Wiesenhofer B, Stockhammer G, Kostron H, Maier H, Hinterhuber H and Humpel C . (2000). Acta Neuropathol. (Berl), 99, 131–137.

Download references

Acknowledgements

We thank Professor G Milanesi for his hospitality and continuous support, Dr M Ferrari (Centro substrati cellulari, Istituto Zooprofilattico Sperimentale della Lombardia e dell’Emilia) for her generous gift of U87MG, U254, U373MG, T98G, CCF-STTG1, SW1088, cell lines and Dr Comicini (Dipartimento di Genetica e Microbiologia Università di Pavia) for his generous gift of the D384 cell line. This work was supported by grants from M.U.R.S.T. (FIRB 2001, Italy; #RBNE01-08) to LM and M.U.R.S.T. (FIRB 2001, Italy; #RBNE01-01) and Associazione Italiana Ricerca sul Cancro (AIRC, Italy) to EC. We thank P delli Santi and I Marini for their excellent technical support.

Author information

Correspondence to Lorenzo Magrassi.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Magrassi, L., Conti, L., Lanterna, A. et al. Shc3 affects human high-grade astrocytomas survival. Oncogene 24, 5198–5206 (2005) doi:10.1038/sj.onc.1208708

Download citation

Keywords

  • Shc
  • brain tumors
  • signal transduction
  • neuro-oncology
  • growth control
  • gliomas

Further reading