Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Mar 4.
Published in final edited form as: Oncogene. 2008 May 12;27(39):5182–5194. doi: 10.1038/onc.2008.157

The progenitor cell marker NG2/MPG promotes chemoresistance by activation of integrin-dependent PI3K/Akt signalling

M Chekenya 1, C Krakstad 2,*, A Svendsen 1,*, IA Netland 1, V Staalesen 3, BB Tysnes 1, F Selheim 2, J Wang 1, PØ Sakariassen 1, T Sandal 2, PE Lønning 4, T Flatmark 5, PØ Enger 1,6, R Bjerkvig 1,7, M Sioud 8,†, WB Stallcup 9,†
PMCID: PMC2832310  NIHMSID: NIHMS106129  PMID: 18469852

Abstract

Chemoresistance represents a major problem in the treatment of many malignancies. Overcoming this obstacle will require improved understanding of the mechanisms responsible for the phenomenon. The progenitor cell marker NG2/MPG is aberrantly expressed by various tumors, but its role in cell death signalling and its potential as a therapeutic target is largely unexplored. We have assessed cytotoxic drug-induced cell death in glioblastoma (GBM) spheroids from fifteen patients, as well as in five cancer cell lines that differ with respect to NG2/MPG expression. The tumors were treated with doxorubicin, etoposide, carboplatin, temodal, cisplatin and TNFα. High NG2/MPG expression correlated with multi-drug resistance mediated by increased activation of α3β1 integrin/PI3K signalling and their downstream targets, promoting cell survival. NG2/MPG knockdown with shRNAs incorporated into lentiviral vectors attenuated β1 integrin signalling revealing potent anti-tumor effects and further sensitized neoplastic cells to cytotoxic treatment in vitro and in vivo. Thus, as a novel regulator of the anti-apoptotic response, NG2/MPG may represent an effective therapeutic target in several cancer subtypes.

Keywords: Apoptosis, Chemoresistance, Integrin, NG2/MPG

INTRODUCTION

Malignant brain tumors belong to the subgroup of cancers with the poorest prognosis, and the introduction of new chemotherapy regimens has only marginally improved survival. Attempts to improve survival must include strategies that identify and target molecules that confer chemoresistance.

NG2 is a transmembrane chondroitin sulfate proteoglycan that is expressed by progenitor cells in several types of tissues (Stallcup, 2002), including oligodendrocyte progenitors in the CNS (Belachew et al, 2003; Nishiyama et al., 1996b). NG2 is the rat homologue of the human melanoma proteoglycan (MPG), also known as the high molecular weight melanoma associated antigen (Campoli et al., 2004; Pluschke et al., 1996). NG2/MPG is over-expressed by several tumor types that fail to respond to conventional chemotherapy, including glioblastomas, most melanomas, and some leukemias (Behm et al., 1996; Chekenya et al., 2002a; Li et al., 2003; Mauvieux et al., 1999; Schrappe et al., 1991; Shoshan et al., 1999; Smith et al., 1996).

NG2/MPG potentiates cell motility (Burg et al., 1997; Eisenmann et al., 1999; Fang et al., 1999; Makagiansar et al., 2004; Makagiansar et al., 2007; Stallcup & Dahlin-Huppe, 2001) and modulates responses to growth factors (Goretzki et al., 1999; Grako et al., 1999; Grako & Stallcup, 1995; Nishiyama et al., 1996a), processes that are critical for the proliferation and migration of both immature progenitor and tumor cells. NG2/MPG expression increases melanoma growth and metastasis (Burg et al., 1998), and enhances glioma growth and angiogenesis (Chekenya et al., 2002b). The expression of NG2/MPG in childhood acute myeloid leukemic (AML) blasts has also been shown to correlate with poor clinical outcome (Hilden et al., 1997; Smith et al, 1996). The role of NG2/MPG in regulating cell death signalling is largely unexplored. We show here that by activating integrins and downstream PI3K/Akt signalling, NG2/MPG promotes tumor resistance to cytotoxic agents.

RESULTS

NG2/MPG expressing glioma cells are resistant to drug induced apoptosis

Since TNFα is a potent and well-characterized inducer of cell death, we compared its ability to induce apoptosis in NG2/MPG-negative U251 glioma cells (U251-Wt) and U251 cells transfected with NG2/MPG (U251-NG2/MPG). Morphological analyses revealed more frequent nuclear condensation and DNA fragmentation in the U251-Wt cells, indicating that they were significantly more sensitive to TNFα than the U251-NG2/MPG cells (Fig. 1A and B, p< 0.024). Since TNFα also stimulates protein synthesis-dependent cell survival (Beg & Baltimore, 1996) we investigated whether protein synthesis inhibition with cycloheximide (CHX) increased sensitivity to TNFα. TNFα increased apoptosis in both U251-Wt and U251-NG2/MPG cells (Fig 1A and B), however, the extent of cell death in CHX-treated U251-NG2/MPG cells remained less than in U251-Wt cells. While these data support the existence of NG2/MPG-mediated survival signals that are protein synthesis independent, they do not rule out a component that is protein synthesis dependent. The requirement for caspases in TNFα induced apoptosis was confirmed by the dramatic cell death inhibition for both U251-NG2/MPG and U251-Wt (Fig 1A and B) cells in the presence of zVAD. We also tested the effects of other drugs known to induce cell death (Fig. 1C). U251-NG2/MPG cells were less sensitive than U251-Wt cells to Etoposide and Vincristine. In the latter case the difference was small but still statistically significant. Our assessment of apoptosis was confirmed by flow cytometric analyses of DNA content during cell cycle progression (Fig ID). TNFα induced apoptosis in U251-Wt cells resulted in a large increase in the sub-G1 population, whereas this population was markedly smaller in the U251-NG2/MPG cells (Fig 1D).

Figure 1. NG2/MPG expressing tumor cells are more resistant than NG2/MPG negative cells to TNFα and chemotherapy induced apoptosis.

Figure 1

Open in a new tab

(A) Hoechst stained nuclei from U251-Wt (top panels) and 251-NG2/MPG (bottom panels) cells exposed to 50ng/ml TNFα with or without 1µg/ml CHX for 3 or 6 hrs, magnification ×200; scale bar= 60µm.

(B) % apoptosis induced by TNFα with or without CHX at 3 hrs in U251-NG2/MPG cells or the U251-Wt cells, assayed by nuclear fragmentation and DNA condensation as shown in A. The data represent the mean ± S.E.M. from three independent experiments.

(C) % cell death in U251-NG2/MPG and U251-Wt glioma cells treated with 0.5 and 10 µM etoposide, or vincristine for 72 hrs and assayed by MTT. Data represent the mean ± S.E.M of three independent experiments.

(D) Effect of TNFα. (50ng/ml).. treatment for 6hr on apotosis of U251-Wt and U251-NG2/MPG cells, as determined by flow cytometric DNA analysis after staining with propidium iodide. In addition to illustrating the cell cycle distribution from G1 to G2-M, the DNA profiles quantify the appearance of apoptotic cells having less than G1 DNA content (Sub-Gl, blue). The histograms are representative of 3 independent experiments.

To assess the functional significance of NG2/MPG, we used rat sequence specific siRNA2 to reduce the proteoglycan’s expression in the U251-NG2 cells that overexpressed the rat orthologue, (Fig 2A and B), and human specific siRNA3 in U87 cells that endogenously expressed the human homologue, (Fig 2A and C). These siRNAs reduced NG2/MPG protein levels in a dose dependent manner and at low concentrations (10–25nM) (Fig 2D and F). The effect of siRNA treatment on NG2 mRNA levels was confirmed by attenuation of transcripts by approximately 75% in siRNA2-treated U251-NG2 cells (Fig 2E). In contrast, a mutant siRNA2 with 4 consecutive mismatched nucleotides did not perturb NG2 mRNA levels, demonstrating that the NG2 siRNA knockdown was both sequence and target specific.

Figure 2. siRNA-mediated suppression of NG2/MPG gene expression.

Figure 2

Open in a new tab

(A) Analysis of NG2/MPG expression in U251-Wt, U251-NG2, and U87MG cells by flow cytometry and immunoblotting with anti-NG2 antibodies. The arrowhead indicates the 300 kDa NG2/MPG core protein.

(B) NG2/MPG expression in U251-NG2 cells (green), (panel (i)), propidium iodide nuclear counterstain, magnification ×630, scale bar=10µm. NG2/MPG expression (panel (ii), red) and knockdown with siRNA2 in U251-NG2 cells (panel (iii), red), Dapi nuclear counterstain (panels (iv) and (v), magnification ×200, scale bar=60µm.

(C) NG2/MPG expression (green) in U87MG cells, (panel (i)), Dapi nuclear counterstain, magnification ×630, scale bar=10µm. NG2/MPG knockdown (panel (ii), red) with siRNA3 in U87MG cells (panel (iii), red), Dapi nuclear counterstain (panels (iv) and (v), magnification ×200, scale bar=60µm.

(D) Dose dependent knockdown of NG2/MPG expression in U251-NG2/MPG cells by siRNA2 after 48 hrs. Protein expression was determined by immunoblot analysis. (*) proteolytic products. The membranes were stripped and re-probed with β-actin antibodies

(E) Northern blotting of NG2 mRNA transcripts after 48 hr treatment with siRNA2 or mutant siRNA.

(F) Dose dependent knockdown of NG2/MPG expression in U87 cells by siRNA3 after 48 hrs. Protein expression was determined by immunoblot analysis.

Next, we investigated the effect of NG2/MPG knock down on TNFα induced apoptosis. As shown in Fig 3A, the sensitivity of U251-NG2/MPG cells to TNFα treatment was significantly increased when NG2/MPG expression was reduced by treatment with siRNA2 (ANOVA F11.65, df= 2, p=0.0015). In contrast, U251-NG2/MPG cells transfected with control siRNA2, maintained their resistance to apoptosis. Intriguingly, siRNA knockdown sensitized U251-NG2/MPG cells to apoptosis beyond the levels seen in U251-Wt cells, indicating that the proteoglyan positive cells may be highly dependent on NG2/MPG mediated survival signals. NG2/MPG-dependent apoptosis resistance was also seen with U87 and A172 cells that endogenously express the proteoglycan (Fig 3B). These cells are highly resistant to TNFα-induced apoptosis, as are U87 and A172 cells transfected with control shRNA In contrast, their sensitivity to TNFα was significantly increased by transfection with the shRNA3 species that effectively knocks down human NG2/MPG expression (Fig 3B, immunoblots). U87wt cells and cells transfected with control shRNA were also resistant to clinically relevant chemotherapy drugs, but became sensitised upon shRNA mediated NG2/MPG knock down (Fig 3C). Flow cytometric analyses of DNA content during cell cycle progression also reflected the Etoposide-induced increase in apoptosis in U87 cells treated with NG2/MPG shRNAs, as visualised by the increased sub-G1 fraction in the DNA histograms. In addition, Etoposide induced greater G2/M phase arrest in the NG2/MPG shRNA knock down cells (Fig 3E).

Figure 3. SiRNA mediated suppression of NG2/MPG expression abrogates resistance to TNFα.

Figure 3

Open in a new tab

(A) %. apoptosis in U251-Wt cells or U251-NG2/MPG cells transfected with 25nM siRNA or control siRNA for 48 hrs. Cells were treated with or without (+/−) TNFα after 20 min pre-incubation with wortmannin (100ng/ml), and assayed by nuclear fragmentation and DNA condensation The data represent the mean ± S.E.M. from three independent experiments.

(B) %. apoptosis in U87MG and A172 cells transfected with 25nM siRNA or control siRNA for 48 hrs. Cells were either untreated or treated with 50ng/ml of TNFα for 6 hrs, and assayed by nuclear fragmentation and DNA condensation. The data represent the mean ± S.E.M. from three independent experiments.

(C) %. Cell death in U87MG cells transfected with control siRNA or NG2/MPG siRNA and treated with 50µM of temodal, carboplatin and etoposide for 72 hrs. Cells were assayed by MTT. The data represent the mean ± S.E.M. from three independent experiments.

(D) % Apoptosis in A375 melanoma cells transfected with NG2/MPG shRNA or control shRNA. Cells were treated with 50ng/ml of TNFα for 6 hrs or with Doxorubicin (3µM), and Cisplatin (5µM) for 72 hrs assayed by nuclear fragmentation and DNA condensation. The data represent the mean ± S.E.M. from three independent experiments.

(E) Effect of Etoposide (0.5 µM) on cell cycle distribution of control U87wt and U87 cells transduced with control shRNA or NG2/MPG shRNA as determined by flow cytometric analysis conducted after staining with propidium iodide. The profiles indicate cell cycle distribution from G1 to G2-M, as well as the appearance of apoptotic cells in the sub-G1 fraction, (blue). Histograms are representative of two independent experiments.

Similar results were also obtained with A375 melanoma cells that endogenously express human NG2/MPG (Fig 3D, and immunoblot). Both Wt and control shRNA-transfected A375 cells were resistant to TNFα Doxorubicin, and Cisplatin, drugs commonly used for treatment of melanomas (Fig 3D). However, shRNA 3-transfected cells were significantly sensitized to these cytostatic agents (ANOVA F12.01, df= 8, p<0.0001). These findings establish NG2/MPG-mediated chemoresistance as a general phenomenon associated with several tumor types.

The NG2/MPG proteoglycan interacts with the α3β1 integrin to mediate apoptosis resistance

Integrin signalling plays a key role in regulating cell death and survival (Damiano et al., 1999; Downward, 2004; Frisch & Ruoslahti, 1997; Khwaja et al, 1997; Kumar, 1998; Wewer et al., 1997). Since NG2/MPG has been shown to interact with β1 integrins (Burg et al., 1998; Eisenmann et al., 1999; Fukushi et al., 2004; Yang et al., 2004), we investigated whether β1 integrin signalling could be involved in NG2/MPG-mediated resistance to cell death. U251-Wt and U251-NG2/MPG cells were immunostained with the HUTS-21 mAb that recognizes an activation-dependent epitope in the human β1 integrin subunit (Lenter et al, 1993; Luque et al, 1996). While activated β1 integrin was barely detectable on U251-Wt cells (Fig 4A, top left), strong labelling was observed in the U251-NG2/MPG cells (Fig 4A top right and 4B top left). Both cell lines exhibited comparable levels of total β1 integrin (Fig 4A and 4B, bottom panels), indicating the involvement of NG2/MPG in β1 activation rather than β expression. Equivalent results were also obtained for A375 melanoma cells (data not shown). To provide evidence for a functional role of β1 integrin signalling in NG2/MPG-mediated apoptosis resistance, we tested the effect of two β1 integrin function blocking antibodies on TNFα-induced apoptosis in U251-NG2/MPG cells. Both antibodies partially abolished the apoptosis resistance in U251-NG2/MPG cells, p = 0.029, Fig 4C, right-hand panel. The Ha2/5 antibody had a small, non-significant effect in U251-Wt cells, perhaps due to blocking low levels of NG2/MPG-independent integrin β1 activation. Next we asked whether activation of β1 integrin via an NG2/MPG-independent mechanism could rescue U251 Wt cells from TNFα-induced apoptosis. Treating U251-Wt cells with the β1 activating antibody TS2/16 (Arroyo et al., 1992; Hemler et al., 1984), significantly protected the U251-Wt cells from apoptosis (p<0.001, Fig 4C, left panel). Similarly, while U87wt cells and U87 cells transfected with control shRNA expressed equal levels of activated β1 integrin subunit, NG2/MPG knock down with NG2 shRNA cells attenuated levels of activated β1 (Fig 4D), further demonstrating the connection between NG2/MPG and integrin activation. Similar results were obtained with A375 cells (supplemenraty Fig 1).

Figure 4. Apoptosis resistance in NG2/MPG cells is mediated by β1 integrin signaling.

Figure 4

Open in a new tab

(A) Activated β1 integrin (HUTS-21, green, top panels) or total β1 integrin TS2/16, green, bottom panels) in U251-Wt (left) and U251-NG2/MPG (right), Magnification ×400, scale bar 60µm.

(B) Activated β1 (HUTS-21, green, top left); total β1 (TS2/16, green, bottom left) subunits in U251-Wt and U251-NG2/MPG cells determined by flow cytometry. % values reflect the total fluorescence intensity relative to an arbitrary window setting. Controls profiles were obtained by using secondary antibody alone.

(C) Left panel: % Apoptosis in U251-Wt cells pre-incubated with β1 activating antibody TS2/16 (10µg/ml) for 30 min or with CD29 blocking antibody Ha2/5 for 2 hrs prior to 6 hr treatment with 50ng/ml TNFα. Right panel: % apoptosis in U251-NG2/MPG cells pre-incubated with β1 function blocking antibodies (CD29 Ha2/5 or AIIB2) for 2 hrs prior to treatment with 50ng/ml TNFα Apoptosis was assayed by nuclear fragmentation and DNA condensation. Each bar represents the mean ± S.E.M. of three independent experiments.

(D) Activated β1 integrin levels in U87Wt, U87 ctrl siRNA and U87 NG2/MPG siRNA cells determined by HUTS-21 flow cytometry. % values reflect the total fluorescence intensity relative to an arbitrary window setting. Controls profiles were obtained by using secondary antibody alone.

(E) α3 integrin subunits in U251-Wt and U251-NG2/MPG cells determined by flow cytometry. % values reflect the total fluorescence intensity relative to an arbitrary window setting. Controls profiles were obtained by using secondary antibody alone.

(F) Detergent extracts of U251-NG2/MPG cells or (G) U251Wt cells or (H) U87 cells immunoprecipitated with two different anti-NG2/MPG antibodies (ectodomain and D2) and blotted with anti-NG2/MPG antibody (top row of bands) and anti-α3 antibody (bottom row of bands). Crude extracts (far left) and α3 immunoprecipitates (far right) were loaded as positive controls. Control rabbit and mouse immunoglobulin did not immunoprecipitate either protein. Arrowheads denote positions of 200 and 116 kDa molecular weight markers, respectively.

α3 integrin subunit levels were also comparable in U251-Wt and U251-NG2/MPG (Fig 4E), while α2, α4, and α6 subunits were barely detectable (data not shown). Thus, the α3β1 heterodimer is the predominant integrin in the U251 cells, but only displays the activated conformation in NG2/MPG expressing cells. In order to provide evidence for a physical interaction between NG2/MPG and α3β1, we performed co-immunoprecipitation studies. The α3 subunit was detected in NG2/MPG immunoprecipitates prepared from both U251-NG2/MPG (Fig 4F) and U87 cells (Fig 4H) with two independent NG2/MPG antibodies but not with control immunoglobulins, demonstrating the specificity of the assays. The same two NG2/MPG antibodies did not immunoprecipitate α3 from the U251-Wt cells (Fig 4G), demonstrating the requirement for NG2/MPG in the co-immunoprecipitation. We were unable to demonstrate co-immunoprecipitation of NG2/MPG using antibodies against α3 integrin. Isotype specific negative controls and a monoclonal α3 integrin blocking antibody had no effect on NG2/MPG-mediated apoptosis resistance (data not shown). Taken together, these findings indicate the existence of an interaction between NG2/MPG and α3β1 integrin that promotes anti-apoptotic signalling via activated β1 integrin.

NG2/MPG-α3β1-mediated resistance to TNFα-induced apoptosis requires PI3K/Akt signalling

Since PI3K is a well-known promoter of cell survival acting downstream of β1 integrin, we studied the effect of the irreversible PI3K inhibitor wortmannin on TNFα-induced apoptosis in the U251-NG2/MPG and U251-Wt glioma cells (Fig 3A). Wortmannin dramatically reduced the cytoprotective effect of NG2/MPG in U251-NG2/MPG cells treated with control siRNA as well as untreated cells (ANOVA F 45.27, df=1, p<0.002). In contrast, the effect was much smaller in U251-Wt or U251-NG2/MPG siRNA knockdown cells already exhibiting high levels of cell death. These results indicate that NG2/MPG expression protects cells from TNFα-induced cell death via a mechanism involving the PI3K/Akt survival pathway.

To generate support for this hypothesis, we compared levels of the PI3K product, Ptdlns (3,4,5) P3 in U251-Wt and U251-NG2/MPG cells. Measurement of 32P -radiolabelled Ptdlns (3,4,5) P3 under basal conditions revealed increased synthesis of this phospholipid in the apoptosis resistant U251-NG2/MPG cells, (t 4.262, df= 4; p =0.01), (Fig 5A), but TNFα increased PIP3 levels in both cell types (t 1.024, df= 2; p =0.413), (Fig 5A). Quantitative assessment of Akt (S473) phosphorylation revealed a 2-fold greater baseline phosphorylation in U251-NG2/MPG cells compared to U251-Wt cells (supplementary Fig 2) adding further support for activation of PI3K/Akt as downstream targets of NG2/MPG dependent α3β1 integrin signaling. These data suggest that the NG2/MPG-positive cells may be primed for drug resistance.

Figure 5. NG2/MPG mediates resistance to TNFα via PI3K/Akt signalling in both glioma and melanoma cells.

Figure 5

Open in a new tab

(A) Synthesis of 32-P radiolabelled PtdIns (3,4,5) P3 in U251-Wt and U251-NG2/MPG cells treated with 50ng/ml TNFα for 15 min. Bars represent mean + S.E.M from three independent experiments.

(B) Immunoblot analysis of Akt and phospho- S473 Akt in U87 wt cells or in cells transfected with siRNA and mutant-siRNA. Cells were untreated or treated with 50ng/ml TNFα (+/−).

(C) Immunoblot detection of Akt and phospho- S473 Akt proteins in untreated A172 cells and A172 cells transfected with control siRNA or NG2/MPG siRNA. β-actin provides an equal loading control. Densitometric quantification of the total Akt (open bars) and P-Akt bands (solid bars).

(D) Immunoblot analysis of Akt and phospho- S473 Akt in A375 wt cells and A375 cells transfected with control siRNA or NG2/MPG siRNA. Cells were treated with 50ng/ml TNFα (+/−) for 15 min Densitometry was used to quantify the decrease in Akt phosphorylation following NG2/MPG knockdown.

Similar to U251-NG2/MPG cells, constitutively high level of S473 phospho-Akt were also demonstrated in U87 (Fig 5B), A172 (Fig 5C) and A375 cells (Fig 5D) that endogenously express NG2/MPG. Transfection of these cells with NG2/MPG siRNA3, but not control siRNA, attenuated levels of S473 phospho-Akt (Fig 5B–D), substantiating the link between NG2/MPG expression and increased PI3K/Akt signalling.

NG2/MPG promotes resistance to TNFα and enhances tumor growth in vivo

To determine whether NG2/MPG-induced chemoresistance affects tumor growth, U87 cells were infected with lentiviruses expressing GFP and either control shRNAs (U87LVcontrol shRNA), or NG2/MPG shRNAs (U87LVNG2/MPG shRNA). 96 hours after a single exposure to the lentivirus-encoding NG2/MPG shRNAs, 65% of the cells expressed GFP compared to uninfected cells (Fig 6A and B). Western blot analyses confirmed that while both the U87 parental and U87LVcontrol shRNA cells contained high levels of NG2/MPG protein, successful down-regulation of the NG2/MPG protein was achieved in the U87LVNG2/MPG shRNA cells (Fig 6C). The knockdown of NG2/MPG was associated with significant tumor growth inhibition in vivo (ANOVA, F20.92 df=3; p = 0.0004, Fig 6D and E).

Figure 6. NG2/MPG promotes glioblastoma growth in vivo and confers resistance to TNFα treatment.

Figure 6

Open in a new tab

(A) Expression of the lentiviral shRNA indicated by eGFP fluorescence. Untreated U87 glioblastoma cells (left panel,) or transfected with shRNA (right panel), magnification 400×, scale bar 80µm.

(B) Flow cytometric analysis reveals eGFP expression in 65% of U87 cells following a single shRNA lentivirus infection (after 96 hrs).

(C) Top: Western blot of NG2/MPG in U87, U87 LVcontrolshRNA and U87LVNG2/MPG cells. β-actin provides a loading control. Bottom: Quantification of signal intensity in a representative blot.

(D) U87 tumor sizes in the 4 different mouse groups at the end of the experiment.

(E) Growth curves showing the effect of NG2/MPG knockdown, and the effect of TNFα treatment on U87LVcontrol shRNA and U87LVNG2/MPG shRNA tumors. Arrow indicates end of TNFα treatment at day 7.

(F) Western blotting of tumor lysates from U87 LVcontrol shRNA and U87LVNG2/MPG shRNA (treated or untreated with TNFα). Upper panels: NG2/MPG levels in the 4 sets of tumors. Arrowhead denotes NG2/MPG core protein. * indicates proteolytic fragments. Middle panels: Immunoblot analysis of total Akt and phosphorylated Akt in lysates from the same 4 sets of tumors treated with TNFα or vehicle in vivo. Bottom panels: β-actin equal loading controls.

To determine whether knocking down NG2/MPG affects tumor sensitivity to TNFα in vivo, the U87LVcontrol shRNA and U87LVNG2/MPG shRNA tumor bearing mice were treated with TNFα. This treatment had little effect on the growth of U87LVcontrol shRNA tumors (Fig 6D and E). However, the growth of U87LVNG2/MPG shRNA tumors was further retarded, indicating that they had been sensitized to TNFα treatment. Immunoblot analyses of tumor lysates revealed that while U87LVcontrol shRNA tumors had detectable levels of NG2/MPG (Fig 6F, right panels), successful knockdown of NG2/MPG mRNA had been achieved in vivo as indicated by the diminished levels of NG2/MPG protein in the U87MGLV NG2/MPG shRNA tumors (Fig 6F, left panels). Moreover, while levels of β-actin and total Akt remained largely unchanged across the groups (Fig 6F), levels of phosphorylated Akt were attenuated in the U87LV NG2/MPG shRNA tumors (Fig 6F). Using TUNEL staining, we were able to document increased tumor cell death after NG2/MPG knockdown, which was further potentiated by TNFα treatment, as indicated by TUNEL positive cells (supplementary Fig 2).

NG2/MPG expression is associated with chemoresistance in human GBM

To validate the physiological relevance of our observations in human tissues, GBM biopsy spheroids derived from tumors with varying NG2/MPG levels (Fig 7A and Table I) were examined for chemosensitivity to doxorubicin, Etoposide and Carboplatin (Fig 7B). Spheroids from glioblastomas with high NG2/MPG expression, as determined by both immunohistochemistry and qPCR, were resistant to Doxorubicin, Etoposide and Carboplatin (Fig 7B). In contrast, the GBM samples with low NG2/MPG exhibited greater sensitivity to these agents. These findings strongly link NG2/MPG expression to chemoresistance in glioma samples.

Figure 7. Chemosensitivity of human GBMs expressing different levels of NG2/MPG.

Figure 7

Open in a new tab

(A) Anti-NG2/MPG IHC staining of human GBM tumors. High expressors are shown in the right-hand panels and low expressors are shown on the left. Magnification 400×, scale bar 100µm.

(B) Chemosensitivity of GBM biopsy spheroids with high and low NG2/MPG expression. Cells were assayed by the MTT method after 96 hr treatment with Doxorubicin, or 120 hr treatment with Etoposide and Carboplatin. For the histograms, data were pooled from the high and low expressors shown in part A. Data represent the mean ± S.E.M. Results for individual tumors are shown in Table I.

Table I. NG2/MPG expression correlates with chemoresistance in GBM biopsies.

Tumor numbers correspond to those presented in Fig 7. NG2/MPG levels were determined by real time- qPCR, except in cases marked N.D.(not determined) in which NG2/MPG levels were assessed semi-quantitatively by IHC. High NG2/MPG expression was defined as being greater than that seen in normal brain (i.e greater than 1.0 fold change). Low expressors (top of the list) are separated from high expressors (bottom of the list) by a space. % Viable cells was determined via MTS assays, as absorbance of non-treated cells - absorbance of treated cells)/absorbance of non-treated cells. Values greater than 100% therefore represent samples, which had increased metabolic activity in the presence of the drug. These results were pooled to yield the histograms in Fig 7B.

Tumor number Fold change NG2 mRNA, relative to normal brain % viable cells after 96h Doxorubicin % viable cells after 120h Etoposide % viable cells after 120h Carboplatin
NG2 negative
1 −3.5 56.3
2 −8 56.8
3 −1.1 87.5 47.3
4 −1.1 61.2
5 *N.D 35.4
6 *N.D 83
13 −1.1 71.40 64.20
NG2 positive
7 *1N.D 156
8 1.3 137.4
9 2 159.1
10 4.6 127.4
11 4.3 177 116
12 9.2 89.2 111.3
14 26.6 157.7 163.7
15 43.2 123.1 115.1
Open in a new tab
*

negative on IHC; N.D = not determined by qPCR

*1

positive on IHC

∫

The tumors are arranged in the Table in ascending order of NG2/MPG expression.

Discussion

The present study demonstrates a novel role for NG2/MPG in mediating protection from apoptosis induced by TNFα and by other cytotoxic drugs with different modes of action. This effect is seen in a panel of cancer cell lines, as well as in biopsy material from brain tumor patients. Furthermore, knockdown of NG2/MPG sensitized malignant cells to chemotherapy and suppressed the growth rates of gliomas in vivo.

We provide evidence via co-immunoprecipitation and immunoblotting studies that the mechanism responsible for these effects involves complexing of NG2/MPG with α3β1 integrin on the cell surface, with subsequent activation of the PI3K/Akt signalling pathway. While levels of total β1 and α3 integrin subunits were unchanged in both U251-Wt and U251-NG2/MPG cells, increased levels of activated β1 integrin were observed in U251-NG2/MPG compared to U251-Wt cells. Furthermore, inhibition of integrin-mediated signalling using β1 function blocking antibodies attenuated apoptosis resistance in the NG2/MPG expressing cells, identifying α3β1 integrin as an important mediator of NG2/MPG -dependent survival signalling in cancer cells. Our results indicate that NG2/MPG-mediated α3β1 activation results in survival signals transduced via stimulation of PI3K/Akt signalling, a pathway widely recognized as a mediator of cell survival signalling (Cantrell, 2001; Vivanco & Sawyers, 2002). It is likely that this same PI3K/Akt signalling cascade mediates the integrin-dependent increase in cell motility that occurs in response to NG2/MPG expression (Fukushi et al, 2004; Makagiansar et al., 2007). Integrin-dependent effects of NG2 on apoptosis may vary according to the cell type and the nature of the environmental stimuli received by the cell. For example, under inflammatory conditions, NG2 can promote anoikis in fibroblasts by opposing fibronectin-stimulated α4β1 integrin signaling via a PKCα-dependent mechanism (Joo et al, 2008).

Our initial experiments utilized U251 cells overexpressing NG2/MPG in order to assess the proteoglycan´s effect on chemosensitivity. From a therapeutic point of view, it was equally informative to study the effects of downregulating NG2/MPG in tumor cell lines that constitutively express the proteoglycan. Together, these studies demonstrated that heterologous expression of NG2/MPG confers resistance to apoptosis in cells that are normally NG2/MPG -negative, while downregulation of endogenous NG2/MPG expression leads to increased sensitivity to TNFα-induced apoptosis. In addition, the relevance of these findings was validated in human GBM biopsy spheroids where low NG2/MPG expression correlated with chemosensitivity to doxorubicin, etoposide and carboplatin. In contrast, tumours with high levels of NG2/HMP were more resistant to, and in some cases were even stimulated to higher metabolic activity, by this cytotoxic treatment. This increased metabolic activity may be a result of energy consuming processes such as drug efflux or active DNA repair processes. The chemotherapeutic drugs used in this study serve as proof of principle to demonstrate that NG2/MPG expressing cells can be sensitised to various drugs after knockdown of proteoglycan expression.

We have confirmed that the U251-Wt and U251- NG2/MPG cell lines carry identical alterations in PTEN and p53 (supplementary figure 3 and 4, respectively), while these genes were differentially altered in the endogenously expressing cells (supplementary Table I). This demonstrates that differences in apoptosis resistance between these lines are likely to be due to the effects of NG2/MPG rather than to different genetic backgrounds. The over-expression studies and the siRNA knockdown studies clearly demonstrate the cause-effect relationship between the presence of the proteoglycan and resistance to apoptosis.

The in vivo studies reveal two distinct effects of NG2/MPG on tumor progression. First, the differential effects of TNFα on U87 tumors transfected with control or NG2/MPG shRNAs demonstrate the protective effect of NG2/MPG against TNFα -induced apoptosis. Further comparisons of these tumors revealed higher levels of phospho-Akt in the NG2/MPG -expressing tumors, indicative of increased PI3K/Akt signalling in support of the in vitro data. Second, stable downregulation of NG2/MPG in U87 (mediated by lentivirally delivered shRNAs) led to marked reduction of tumor growth rates even in the absence of TNFα. This effect may be due to NG2/MPG ´s role in functions that are unrelated to apoptosis. For example, NG2/MPG potentiates cell proliferation, possibly as a result of its participation in growth factor signalling (Goretzki et al., 1999; Grako et al, 1999). NG2/MPG also stimulates angiogenesis, by sequestering angiostatin and neutralizing its inhibitory effects on angiogenesis (Chekenya et al., 2002b; Goretzki et al., 2000).

In conclusion, we have applied a variety of experimental conditions to both in vitro and in vivo model systems, all of which establish NG2/MPG as a mediator of multi-drug resistance in the tumors examined. The mechanisms responsible for these effects involve increased survival signals that counteract cell death. Future work will be directed at elucidating in more depth the genetic or epigenetic mechanisms involved in the NG2/MPG induced chemoresistance. This involvement of NG2/MPG in multiple aspects of tumor biology makes the proteoglycan an attractive candidate for future therapies against cancer.

MATERIALS AND METHODS

Cell culture

These studies utilized the human glioblastoma multiforme cell lines U251N (U251-Wt), U87, and A172 and the human A375 melanoma (American Type Culture Collection, Rockville, Maryland; ATCC). U251 cells were transfected with the rat NG2/MPG cDNA (U251-NG2/MPG), as previously described (Chekenya et al., 2002b). Cells were exposed to 50ng/ml TNFα for 15 min with or without 30 min pre-treatment with the PI3K inhibitor Wortmannin (Sigma). The chemotherapy agents Vincristine, Etoposide, Temodal, Doxorubicin, Cisplatin and Carboplatin were also used. In some experiments cells were pre-treated for 20 min with the irreversible caspase inhibitor zVAD-FMK (BIOMOL, Plymouth Meeting, PA). Cells were also pre-incubated with 10µg/ml of TS2/16 β1 activating antibody (ATCC), or CD29 (/Ha2/5) β1 function blocking antibody (BD Pharmingen, San Diego, CA), or the β1 integrin function blocking monoclonal antibody AIIB2 (a gift from M. Bissell) at 37°C for 2hr prior to treatment with TNFα. Prior to analysis of apoptosis, cells were fixed in 2% (v/v) glutaraldehyde containing 1µg/ml Hoechst 33342 (Sigma).

Scoring of apoptosis

Apoptosis was determined by morphological changes in Hoechst stained cell nuclei using light and UV microscopy as previously described (Sandal et al., 2002). Apoptosis (sub G1 population) was also assessed by flow cytometric analysis of DNA content in cells stained with propidium iodide using ModFit LTTM version 3.0 software. For fluorescence, goat anti-mouse quantum dots 565 (Invitrogen, Carlsbad, CA) were used. Negative controls were secondary antibody only.

Sensitivity of human glioblastoma biopsy spheroids to doxorubicin

To assess viability/metabolic activity in NG2/MPG positive (n=8) and negative (n=7) human glioma specimens, spheroids of equal sizes (250µm) were treated with 3 µM Doxorubicin (96hr), or 0.5 and 50 µM Etoposide (120hr) or 0.5 and 10 µM Carboplatin (120hr) depending on the nearest IC50 value for the NG2/MPG negative tumors. The 3-(4-5-dimethylthiazol-2-yl)-5-(3carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt (MTS) assay was used as described by the manufacturer (Promega, Madison, WI). Untreated spheroids were used as controls. The effects of drugs (% of control) were calculated as absorbance of non-treated cells - absorbance of treated cells)/absorbance of non-treated cells (Miura et al, 2006). The ethical board at Haukeland University Hospital, Bergen, Norway, approved the collection of tumor tissue. The patients gave their informed consent to specimen collection.

Real-time qPCR

The snap frozen GBM tissue was crushed in liquid nitrogen, total RNA extracted and cDNA reverse transcribed using iScript™ (Bio-Rad Laboratories, Hercules, CA), with iQ SYBR Green in the Real time qPCR using iCycler™ Thermal Cycler fitted with iCycler™ Optical Module (Bio-Rad Laboratories). Primers directed against 18S RNA were used as an internal control. 3 adult normal brain cDNAs used as reference were either prepared from cerebral cortical tissue after radical resection and verified by a neuropathologist, or were purchased from Ambion (Austin, Tx).

siRNA synthesis

siRNAs were transcribed in vitro using chemically synthesised DNA oligonucleotides and T7 RNA polymerase. The following siRNAs were designed: siRNA2 5’´GUAGAUCAAUACCCUACACUU-3’ (positions 970 to 989) in the rat NG2/MPG cDNA. A non–functional mutant version of siRNA2 with four consecutive nucleotide mismatches was used as a control (5’GUAGAUCAAUUGGGUACACUU-3’). The expression of human NG2/MPG was targeted with siRNA3 (5’GUGGACCAGUACCCUACGG-3’), corresponding to Homo sapiens melanoma-associated chondroitin sulfate proteoglycan (cspg4).

Construction of the shRNA lentiviral vectors and lentiviral infection

Lentiviral particles harboring pRNAt-U6.1 (SD 1257, Genscript, Scotch Plains, NJ), encoding a U6 promoter driving shRNA expression were produced using the Virapower lentiviral packaging mix (Invitrogen, Carlsbad, CA) according to the manufacturers descriptions. Briefly, packaging and transfer vectors were co-transfected into 293FT cells, and viral particles were harvested 48h post-infection. Viral particles were concentrated by ultra-sentrifugation and the infected cells were monitored for GFP expression using flow cytometry. Two weeks post infection the cells were harvested and monitored for GFP fluorescence by flow cytometry.

Western blotting

Proteins from cells or tumor tissue were harvested in ice-cold lysis buffer (10mM K2HPO4, 1mM EDTA (pH 6.8) containing 10mM Chaps, 50 µM NaF, and 0.3µM NaVO3, supplemented with Complete Protease Inhibitor (Roche Molecular Biochemicals) and 0.02U of chondroitinase ABC (Sigma). Rabbit anti-total Akt and anti-phospho Akt (Ser 473) (diluted 1:1000) (Cell Signalling Technology, Beverly, MA), and Rabbit anti-NG2/MPG (553) antibody was used at 1:2000 dilution. For normalization, blots were stripped and re-probed with mouse anti-β-actin (diluted 1:5000, Abeam Limited, Cambridge, UK). Immunoprecipitations were performed as previously described (Fukushi et al., 2004). Immunocyto/histochemistry was performed using standard procedures previously described (Chekenya et al., 2002b).

Extraction and analysis of 3-phosphorylated glycerophosphoinositides

U251-Wt and U251-NG2/MPG glioma and cells treated with 50ng/ml TNFα for 15 min were propagated in the presence of [32P] Pi to produce [32P] labelled phospholipids. Extraction and thin-layer chromatography of polyphosphoinositides were performed as described previously (Tysnes et al., 1985).

Animals

Twenty-four, 5–6 week old, severe combined immunodeficient (NOD-SCID) mice (C.B.-Igh-1b/lcrTac-Prkdc) purchased from Taconic Europe, (Ry, Denmark) received 5×106 U87MG tumor cells stably expressing LVNG2/MPG shRNAs and LVCtr shRNAs lentiviruses into the hind leg. The animals were divided into four groups, U87MGLV-control siRNA (n=6), U87MG LVNG2/MPG siRNA (n=6), U87MG LV control siRNA+ TNFα (n=6), U87MG LVNG2/MPG siRNA + TNFα (n=6). 48 hrs after tumor implantation, some animals were treated with intraperitoneal injections (IP) of 150µg/kg rhTNFα, (Peprotech LTD, UK), once a day for 7 days. The animals were anaesthetised with 1–2% isoflurane in a 70%/30% N2/O2 mix and all procedures were in accordance with protocols approved by The National Animal Research Authority (Oslo, Norway). Tumor growth was determined by measuring two perpendicular diameters. Tumor volumes (V) were calculated using the formula (V)=(a×b×c)/2 that was derived for an ellipsoid (πd3/6)(Tan et al, 2004), where a, b and c are the long axis, short axis and the depth, respectively.

Statistical Analyses

Data were analyzed using the Student’s paired T-test (two-tailed), and One-way (StatViewSE+) or Two-way ANOVA (Graphpad Instat software Inc version 3.05, San Diego, CA).

Acknowledgements

This work was supported by research grants from The Norwegian Cancer Society, The Bergen translational Research Group, Familien Blix Fond, The University of Bergen, and a RO1 CA95287 grant from the National Institutes of Health. The work was also supported by the Sixth EU Framework Programme and Helse-Vest. We wish to thank Narve Brekkå, Tove Johansen, Christine Eriksen, Erna Finsås, and Nina Lied Larsen for their technical assistance. We thank Jesus Planaguma for assistance with his excellent image analysis.

References

  1. Arroyo AG, Sanchez-Mateos P, Campanero MR, Martin-Padura I, Dejana E, Sanchez-Madrid F. Regulation of the VLA integrin-ligand interactions through the beta 1 subunit. J Cell Biol. 1992;117:659–670. doi: 10.1083/jcb.117.3.659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beg AA, Baltimore D. An essential role for NF-kappaB in preventing TNF-alpha-induced cell death. Science. 1996;274:782–784. doi: 10.1126/science.274.5288.782. [DOI] [PubMed] [Google Scholar]
  3. Behm FG, Smith FO, Raimondi SC, Pui CH, Bernstein ID. Human homologue of the rat chondroitin sulfate proteoglycan, NG2, detected by monoclonal antibody 7.1, identifies childhood acute lymphoblastic leukemias with t(4;11)(q21;q23) or t(11;19)(q23;p13) and MLL gene rearrangements. Blood. 1996;87:1134–1139. [PubMed] [Google Scholar]
  4. Belachew S, Chittajallu R, Aguirre AA, Yuan X, Kirby M, Anderson S, Gallo V. Postnatal NG2 proteoglycan-expressing progenitor cells are intrinsically multipotent and generate functional neurons. J Cell Biol. 2003;161:169–186. doi: 10.1083/jcb.200210110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Burg MA, Grako KA, Stallcup WB. Expression of the NG2 proteoglycan enhances the growth and metastatic properties of melanoma cells. J Cell Physiol. 1998;177:299–312. doi: 10.1002/(SICI)1097-4652(199811)177:2<299::AID-JCP12>3.0.CO;2-5. [DOI] [PubMed] [Google Scholar]
  6. Burg MA, Nishiyama A, Stallcup WB. A central segment of the NG2 proteoglycan is critical for the ability of glioma cells to bind and migrate toward type VI collagen. Exp Cell Res. 1997;235:254–264. doi: 10.1006/excr.1997.3674. [DOI] [PubMed] [Google Scholar]
  7. Campoli MR, Chang CC, Kageshita T, Wang X, McCarthy JB, Ferrone S. Human high molecular weight-melanoma-associated antigen (HMW-MAA): a melanoma cell surface chondroitin sulfate proteoglycan (MSCP) with biological and clinical significance. Crit Rev Immunol. 2004;24:267–296. doi: 10.1615/critrevimmunol.v24.i4.40. [DOI] [PubMed] [Google Scholar]
  8. Cantrell DA. Phosphoinositide 3-kinase signalling pathways. J Cell Sci. 2001;114:1439–1445. doi: 10.1242/jcs.114.8.1439. [DOI] [PubMed] [Google Scholar]
  9. Chekenya M, Enger PO, Thorsen F, Tysnes BB, Al-Sarraj S, Read TA, Furmanek T, Mahesparan R, Levine JM, Butt AM, Pilkington GJ, Bjerkvig R. The glial precursor proteoglycan, NG2, is expressed on tumour neovasculature by vascular pericytes in human malignant brain tumours. Neuropathol Appl Neurobiol. 2002a;28:367–380. doi: 10.1046/j.1365-2990.2002.00412.x. [DOI] [PubMed] [Google Scholar]
  10. Chekenya M, Hjelstuen M, Enger PO, Thorsen F, Jacob AL, Probst B, Haraldseth O, Pilkington G, Butt A, Levine JM, Bjerkvig R. NG2 proteoglycan promotes angiogenesis-dependent tumor growth in CNS by sequestering angiostatin. Faseb J. 2002b;16:586–588. doi: 10.1096/fj.01-0632fje. [DOI] [PubMed] [Google Scholar]
  11. Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell adhesion mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in human myeloma cell lines. Blood. 1999;93:1658–1667. [PMC free article] [PubMed] [Google Scholar]
  12. Downward J. PI 3-kinase, Akt and cell survival. Semin Cell Dev Biol. 2004;15:177–182. doi: 10.1016/j.semcdb.2004.01.002. [DOI] [PubMed] [Google Scholar]
  13. Eisenmann KM, McCarthy JB, Simpson MA, Keely PJ, Guan JL, Tachibana K, Lim L, Manser E, Furcht LT, Iida J. Melanoma chondroitin sulphate proteoglycan regulates cell spreading through Cdc42, Ack-1 and p130cas. Nat Cell Biol. 1999;1:507–513. doi: 10.1038/70302. [DOI] [PubMed] [Google Scholar]
  14. Fang X, Burg MA, Barritt D, Dahlin-Huppe K, Nishiyama A, Stallcup WB. Cytoskeletal reorganization induced by engagement of the NG2 proteoglycan leads to cell spreading and migration. Mol Biol Cell. 1999;10:3373–3387. doi: 10.1091/mbc.10.10.3373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Frisch SM, Ruoslahti E. Integrins and anoikis. Curr Opin Cell Biol. 1997;9:701–706. doi: 10.1016/s0955-0674(97)80124-x. [DOI] [PubMed] [Google Scholar]
  16. Fukushi J, Makagiansar IT, Stallcup WB. NG2 proteoglycan promotes endothelial cell motility and angiogenesis via engagement of galectin-3 and alpha3beta1 integrin. Mol Biol Cell. 2004;15:3580–3590. doi: 10.1091/mbc.E04-03-0236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Goretzki L, Burg MA, Grako KA, Stallcup WB. High-affinity binding of basic fibroblast growth factor and platelet-derived growth factor-AA to the core protein of the NG2 proteoglycan. J Biol Chem. 1999;274:16831–16837. doi: 10.1074/jbc.274.24.16831. [DOI] [PubMed] [Google Scholar]
  18. Goretzki L, Lombardo CR, Stallcup WB. Binding of the NG2 proteoglycan to kringle domains modulates the functional properties of angiostatin and plasmin(ogen) J Biol Chem. 2000;275:28625–28633. doi: 10.1074/jbc.M002290200. [DOI] [PubMed] [Google Scholar]
  19. Grako KA, Ochiya T, Barritt D, Nishiyama A, Stallcup WB. PDGF (alpha)-receptor is unresponsive to PDGF-AA in aortic smooth muscle cells from the NG2 knockout mouse. J Cell Sci. 1999;111(Pt 6):905–915. doi: 10.1242/jcs.112.6.905. [DOI] [PubMed] [Google Scholar]
  20. Grako KA, Stallcup WB. Participation of the NG2 proteoglycan in rat aortic smooth muscle cell responses to platelet-derived growth factor. Exp Cell Res. 1995;221:231–240. doi: 10.1006/excr.1995.1371. [DOI] [PubMed] [Google Scholar]
  21. Hemler ME, Sanchez-Madrid F, Flotte TJ, Krensky AM, Burakoff SJ, Bhan AK, Springer TA, Strominger JL. Glycoproteins of 210,000 and 130,000 m.w. on activated T cells: cell distribution and antigenic relation to components on resting cells and T cell lines. J Immunol. 1984;132:3011–3018. [PubMed] [Google Scholar]
  22. Hilden JM, Smith FO, Frestedt JL, McGlennen R, Howells WB, Sorensen PH, Arthur DC, Woods WG, Buckley J, Bernstein ID, Kersey JH. MLL gene rearrangement, cytogenetic 11q23 abnormalities, and expression of the NG2 molecule in infant acute myeloid leukemia. Blood. 1997;89:3801–3805. [PubMed] [Google Scholar]
  23. Joo NE, Watanabe T, Chen C, Chekenya M, Stallcup WB, Kapila YL. NG2, a novel proapoptotic receptor, opposes integrin alpha4 to mediate anoikis through PKCalpha-dependent suppression of FAK phosphorylation. Cell Death Differ. 2008 Feb 22; doi: 10.1038/cdd.2008.22. 2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kumar CC. Signaling by integrin receptors. Oncogene. 1998;17:1365–1373. doi: 10.1038/sj.onc.1202172. [DOI] [PubMed] [Google Scholar]
  25. Lenter M, Uhlig H, Hamann A, Jeno P, Imhof B, Vestweber D. A monoclonal antibody against an activation epitope on mouse integrin chain beta 1 blocks adhesion of lymphocytes to the endothelial integrin alpha 6 beta 1. Proc Natl Acad Sci USA. 1993;90:9051–9055. doi: 10.1073/pnas.90.19.9051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li Y, Madigan MC, Lai K, Conway RM, Billson FA, Crouch R, Allen BJ. Human uveal melanoma expresses NG2 immunoreactivity. Br J Ophthalmol. 2003;87:629–632. doi: 10.1136/bjo.87.5.629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Luque A, Gomez M, Puzon W, Takada Y, Sanchez-Madrid F, Cabanas C. Activated conformations of very late activation integrins detected by a group of antibodies (HUTS) specific for a novel regulatory region (355–425) of the common beta 1 chain. J Biol Chem. 1996;271:11067–11075. doi: 10.1074/jbc.271.19.11067. [DOI] [PubMed] [Google Scholar]
  28. Makagiansar IT, Williams S, Dahlin-Huppe K, Fukushi JI, Mustelin T, Stallcup WB. Phosphorylation of NG2 proteoglycan by PKC-alpha regulates polarized membrane distribution and cell motility. J Biol Chem. 2004;279:55262–55270. doi: 10.1074/jbc.M411045200. [DOI] [PubMed] [Google Scholar]
  29. Makagiansar IT, Williams S, Mustelin T, Stallcup WB. Differential phosphorylation of NG2 proteoglycan by ERK and PKCalpha helps balance cell proliferation and migration. J Cell Biol. 2007;178:155–165. doi: 10.1083/jcb.200612084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mauvieux L, Delabesse E, Bourquelot P, Radford-Weiss I, Bennaceur A, Flandrin G, Valensi F, MacIntyre EA. NG2 expression in MLL rearranged acute myeloid leukaemia is restricted to monoblastic cases. Br J Haematol. 1999;107:674–676. doi: 10.1046/j.1365-2141.1999.01730.x. [DOI] [PubMed] [Google Scholar]
  31. Miura M, Miura Y, Padilla-Nash HM, Molinolo AA, Fu B, Patel V, Seo BM, Sonoyama W, Zheng JJ, Baker CC, Chen W, Ried T, Shi S. Accumulated chromosomal instability in murine bone marrow mesenchymal stem cells leads to malignant transformation. Stem Cells. 2006;24:1095–1103. doi: 10.1634/stemcells.2005-0403. [DOI] [PubMed] [Google Scholar]
  32. Nishiyama A, Lin XH, Giese N, Heldin CH, Stallcup WB. Co-localization of NG2 proteoglycan and PDGF alpha-receptor on O2A progenitor cells in the developing rat brain. J Neurosci Res. 1996a;43:299–314. doi: 10.1002/(SICI)1097-4547(19960201)43:3<299::AID-JNR5>3.0.CO;2-E. [DOI] [PubMed] [Google Scholar]
  33. Nishiyama A, Lin XH, Giese N, Heldin CH, Stallcup WB. Interaction between NG2 proteoglycan and PDGF alpha-receptor on O2A progenitor cells is required for optimal response to PDGF. J Neurosci Res. 1996b;43:315–330. doi: 10.1002/(SICI)1097-4547(19960201)43:3<315::AID-JNR6>3.0.CO;2-M. [DOI] [PubMed] [Google Scholar]
  34. Pluschke G, Vanek M, Evans A, Dittmar T, Schmid P, Itin P, Filardo EJ, Reisfeld RA. Molecular cloning of a human melanoma-associated chondroitin sulfate proteoglycan. Proc Natl Acad Sci U S A. 1996;93:9710–9715. doi: 10.1073/pnas.93.18.9710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sandal T, Stapnes C, Kleivdal H, Hedin L, Doskeland SO. A novel, extraneuronal role for cyclin-dependent protein kinase 5 (CDK5): modulation of cAMP-induced apoptosis in rat leukemia cells. J Biol Chem. 2002;277:20783–20793. doi: 10.1074/jbc.M112248200. [DOI] [PubMed] [Google Scholar]
  36. Schrappe M, Klier FG, Spiro RC, Waltz TA, Reisfeld RA, Gladson CL. Correlation of chondroitin sulfate proteoglycan expression on proliferating brain capillary endothelial cells with the malignant phenotype of astroglial cells. Cancer Res. 1991;51:4986–4993. [PubMed] [Google Scholar]
  37. Shoshan Y, Nishiyama A, Chang A, Mork S, Barnett GH, Cowell JK, Trapp BD, Staugaitis SM. Expression of oligodendrocyte progenitor cell antigens by gliomas: implications for the histogenesis of brain tumors. Proc Natl Acad Sci U S A. 1999;96:10361–10366. doi: 10.1073/pnas.96.18.10361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Smith FO, Rauch C, Williams DE, March CJ, Arthur D, Hilden J, Lampkin BC, Buckley JD, Buckley CV, Woods WG, Dinndorf PA, Sorensen P, Kersey J, Hammond D, Bernstein ID. The human homologue of rat NG2, a chondroitin sulfate proteoglycan, is not expressed on the cell surface of normal hematopoietic cells but is expressed by acute myeloid leukemia blasts from poor-prognosis patients with abnormalities of chromosome band 11q23. Blood. 1996;87:1123–1133. [PubMed] [Google Scholar]
  39. Stallcup WB. The NG2 proteoglycan: past insights and future prospects. J Neurocytol. 2002;31:423–435. doi: 10.1023/a:1025731428581. [DOI] [PubMed] [Google Scholar]
  40. Stallcup WB, Dahlin-Huppe K. Chondroitin sulfate and cytoplasmic domain-dependent membrane targeting of the NG2 proteoglycan promotes retraction fiber formation and cell polarization. J Cell Sci. 2001;114:2315–2325. doi: 10.1242/jcs.114.12.2315. [DOI] [PubMed] [Google Scholar]
  41. Tan C, Cruet-Hennequart S, Troussard A, Fazli L, Costello P, Sutton K, Wheeler J, Gleave M, Sanghera J, Dedhar S. Regulation of tumor angiogenesis by integrin-linked kinase (ILK) Cancer Cell. 2004;5:79–90. doi: 10.1016/s1535-6108(03)00281-2. [DOI] [PubMed] [Google Scholar]
  42. Tysnes OB, Aarbakke GM, Verhoeven AJ, Holmsen H. Thin-layer chromatography of polyphosphoinositides from platelet extracts: interference by an unknown phosphohpid. Thromb Res. 1985;40:329–338. doi: 10.1016/0049-3848(85)90268-3. [DOI] [PubMed] [Google Scholar]
  43. Vivanco I, Sawyers CL. The phosphatidylinositol 3-Kinase AKT pathway in human cancer. Nat Rev Cancer. 2002;2:489–501. doi: 10.1038/nrc839. [DOI] [PubMed] [Google Scholar]
  44. Wewer UM, Shaw LM, Albrechtsen R, Mercurio AM. The integrin alpha 6 beta 1 promotes the survival of metastatic human breast carcinoma cells in mice. Am J Pathol. 1997;151:1191–1198. [PMC free article] [PubMed] [Google Scholar]
  45. Yang J, Price MA, Neudauer CL, Wilson C, Ferrone S, Xia H, Iida J, Simpson MA, McCarthy JB. Melanoma chondroitin sulfate proteoglycan enhances FAK and ERK activation by distinct mechanisms. J Cell Biol. 2004;165:881–891. doi: 10.1083/jcb.200403174. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES