Telephone: +39-02-57489830; Fax: +39-02-94375990
Cancer Stem Cells
Article first published online: 9 APR 2012
Copyright © 2012 AlphaMed Press
Volume 30, Issue 5, pages 817–832, May 2012
How to Cite
Ortensi, B., Osti, D., Pellegatta, S., Pisati, F., Brescia, P., Fornasari, L., Levi, D., Gaetani, P., Colombo, P., Ferri, A., Nicolis, S., Finocchiaro, G. and Pelicci, G. (2012), Rai is a New Regulator of Neural Progenitor Migration and Glioblastoma Invasion. STEM CELLS, 30: 817–832. doi: 10.1002/stem.1056
Author contributions: B.O. and D.O.: conception and design, data collection and/or assembly, data analysis and interpretation, manuscript writing; S.P. and G.F.: data analysis and interpretation; F.P. and P.B.: data collection and/or assembly and data analysis and interpretation; L.F.: statistical analysis; D.L. and P.G.: GBM provision; P.C. and A.F.: data collection; S.N.: data interpretation; G.P.: conception and design, data analysis and interpretation, manuscript writing, and final approval of manuscript. B.O. and D.O. contributed equally to this article.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS February 6, 2012.
- Issue published online: 9 APR 2012
- Article first published online: 9 APR 2012
- Accepted manuscript online: 6 FEB 2012 12:19PM EST
- Manuscript Accepted: 20 JAN 2012
- Manuscript Received: 14 OCT 2011
- FIEO fellowship
- Veronesi Foundation
- Regione Lombardia ASTIL
Additional Supporting Information may be found in the online version of this article.
|STEM_1056_sm_SuppFig1.pdf||207K||Figure S1. Immunohistochemistry validation of the anti-Rai mouse monoclonal antibody. Rai expression in WT and Rai−/− adult mouse brains. (A-B) Representative H&E sections of WT and Rai−/−. (C-F) Immunostaining with the mouse monoclonal antibody raised against Rai: strong reactivity is observed in the cortical neurons of WT mice (C, E), while no Rai expression is detected in Rai−/− mice (D, F).|
|STEM_1056_sm_SuppFig2.tif||297K||Figure S2. Cells grown under different experimental conditions express similar levels of Rai. Western blotting: Rai expression in hGBM NS derived from two patients grown under different experimental conditions: standard growing conditions (G) and after EGF and FGF withdrawal (S). Note that when cells are grown in the absence of growth factors the EGF Receptor (EGFR) is not tyrosine-phosphorylated (upper blot). The blot was cut and reprobed with an anti-EGFR antibody (middle bot) and anti-Rai antibody (lower blot). _-Vinculin was used as protein loading control.|
|STEM_1056_sm_SuppFig3.pdf||517K||Figure S3. Rai silencing in GBM neurospheres results in improved mice survival. (A) Western Blotting: Rai protein levels in NT shRNA and Rai shRNA neurospheres derived from several patients. (B) Kaplan-Meyer survival curves: increased mice survival after injection of Rai-silenced cells into the nucleus caudatus of immunodeficient mice; **p = 0.0081 for 105 injected hGBM#7 cells, n = 6 mice for naϊve cells, n = 6 mice for NT shRNA cells, n = 15 mice for Rai shRNA cells; *p = 0.0246 for 105 hGBM#9 cells, n = 3 mice for each condition; p < 0.0001 for 105 hGBM#10 cells, n = 9 mice for each condition; *p = 0.0455 for 105 hGBM#18 cells, n = 3 mice for each condition; *p = 0.0224 for 105 hGBM#20 cells, n = 3 mice for each condition; **p = 0.0027 for 105 hGBM#22 cells, n = 5 mice for each condition. (C) H&E staining of xenograft tumors derived from NT shRNA and Rai shRNA hGBM#7 cells. Homolateral hemispheres, tumor centre and contralateral hemispheres are shown. Note the invasive tumor cells localized away from the main tumor mass, infiltrating the homolateral and the contralateral periphery in control brains, and the smoother borders of Rai-silenced tumors.|
|STEM_1056_sm_SuppFig4.pdf||834K||Figure S4. Rai silenced tumors show reduced infiltrative growth. (A - P) Immunostaining for anti-human Nuclei on xenograft tumors derived from NT shRNA and Rai shRNA hGBM#7 (A - F) and hGBM#10 (G - P) cells. Low and high magnifications are provided. (Q) Maximal tumor areas measured per section (n = 3 mice per group), *p <0.05.|
|STEM_1056_sm_SuppFig5.pdf||318K||Figure S5. Rai silencing does not affect proliferation in vivo. (A) Percentage of Ki67 positive cells in NT and Rai-silenced xenograft tumors (n = 3 sections per mouse). (B) Immunostaining for Ki67 on xenograft tumors derived from NT shRNA and Rai shRNA hGBM#7, hGBM#8, hGBM#18 and hGBM#22 cells.|
|STEM_1056_sm_SuppFig6.pdf||349K||Figure S6. Rai silencing in GBM neurospheres results in reduced GBM invasiveness. (A) Immunostaining for Dcx and GFP on xenograft tumors derived from orthotopic injection of NT shRNA and Rai shRNA hGBM#7, hGBM#18 and hGBM#20 cells. (B) Immunofluorescence staining for Dcx (green) and human mitochondria (red) on tumors derived from orthotopic injection of NT shRNA and Rai shRNA hGBM#10 NS and on tumors derived from orthotopic injection of hGBM#7 NS isolated from serially transplanted xenografts. Nuclei were counterstained with DAPI (blue). Human mitochondria immunoreactivity identifies human tumor cells.|
|STEM_1056_sm_SuppFig7.pdf||637K||Figure S7. Rai is distributed at the periphery of the tumor. Immunohistochemical staining for Rai on xenograft tumors derived from NT shRNA and Rai shRNA hGBM#7, hGBM#9, hGBM#10, hGBM#18, hGBM#20, hGBM#22 cells confirms its expression in control tumors, both in the tumor centre and at the edge (arrowheads), and its absence in silenced tumors, with the exception of rare cells at the infiltrative margin (arrowheads).|
|STEM_1056_sm_SuppFig8.pdf||423K||Figure S8. The reduced GBM invasiveness associated with Rai silencing is a dosagedependent effect. (A) Western Blotting: Rai protein levels in NT shRNA and Rai shRNA neurospheres derived from GBM#11 patient. (B) Kaplan-Meyer survival curve: partial suppression of Rai expression by siRNA does not affect mice survival, compared to control tumors. n = 3 mice for each condition. (C) Tumors grown by orthotopic transplantation of partially Rai-silenced stem/progenitor cells show similar histology and express Rai at similar degree compared to control tumors.|
|STEM_1056_sm_SuppFig9.pdf||582K||Figure S9. Rai silencing in GBM neurospheres improves mice survival in serially transplanted mice. (A) Western blotting: Rai expression levels in hGBM#7 and hGBM#8 neurospheres isolated from primary xenografts. (B - C) Kaplan-Meyer survival curves: reduced Rai expression in hGBM#7 and hGBM#8 neurospheres isolated from primary xenografts increases mice survival after orthotopic injection. (B) **p = 0.0002 for 105 hGBM#7 cells, n = 6 mice for naϊve cells, n = 6 mice for NT shRNA cells n = 12 mice for Rai shRNA cells. (C) p = 0.0224 for 105 hGBM#8 cells, n = 3 mice for each condition. (D) Secondary xenografts show histological features of the correspondent primary xenotransplanted lesions, and express the Rai protein at similar levels as assessed by H&E and Rai staining.|
|STEM_1056_sm_SuppFig10.pdf||342K||Figure S10. Rai is expressed in the subventricular zone and does not affect either proliferation or apoptosis. (A) X-gal staining on coronal section of mouse Rai-LacZ knock-in brain: Rai_-gal is expressed in the SVZ. Higher magnification of the boxed areas shows Rai positivity (arrowheads) at the beginning of the RMS (box 1) and in some periventricular cells (boxes 2 and 3). (B - C) Cell count of Ki67 positive cells on coronal sections of WT and Rai−/− mice show the same amount of steady-state proliferating cells in the SVZ. (B) Representative image of immunofluorescence staining for Ki67 (green); nuclei were counterstained with DAPI (blue). (C) Quantification of the number of Ki67 positive cells in the SVZ of WT and Rai−/− mice; no differences were observed, as assessed by t-test: p = 0.8, n = 6, mean±SD. (D) Immunofluorescence staining for cleaved caspase-3 (red) on SVZ coronal sections of WT, Rai−/− and irradiated WT mice, which were used as positive control, since no positive cells were detected in WT and Rai−/− brains. Nuclei were counterstained with DAPI (blue). Higher magnification of the boxed areas shows cleaved caspase-3 positive cells at the beginning of the RMS in the irradiated control mice only.|
|STEM_1056_sm_SuppFig11.tif||967K||Figure S11. Rai regulates cell migration and infiltration through novel multiple pathways. (A) Western blotting: phosphorylation status of Akt in control and Rai-silenced GBM neurospheres. (B) Western blotting: phosphorylation status of Akt in WT and Rai−/− normal neurospheres. (C) No association was observed between Rai mRNA expression level and pAKT as measured by IHC. The proportion of elevated pAKT expression is lower than expected in the proneural compared to the other subtypes (÷2 =16.4 *p<0.01 DF=4 after Yates' correction). One way ANOVA (F=1.41; p=0.25; DF=2;61). Multiple t-test with Bonferroni's correction: not significant. No association is observed between CGH ratios for the PTEN locus and Rai mRNA expression levels (F=1.39; p=0.29 with GLM-ANOVA), independently of the tumor subtypes (F=0.63; p=0.54 with GLM-ANOVA). No association is observed between EGFR gain or amplification and Rai mRNA expression levels (F=1.78; p=0.14 with GLM-ANOVA), independently of the tumor subtypes (F=0.06; p=0.94 with GLM-ANOVA). (D) Western blotting: phosphorylation status of _-catenin in control and Rai-silenced GBM neurospheres. (E) Western blotting: reactivation of the NF_B pathway in Rai-silenced GBM neurospheres transduced with a lentiviral vector carrying p65 cDNA and then stimulated with TNF-_ for the indicated time points. (F) Western blotting: Rai overexpression induces NF_B p65 subunit accumulation in the nucleus. T98G cells were transduced either with a retroviral vector carrying RAI cDNA or with an empty vector as a negative control, and then stimulated with TNF-_ for the indicated time points. (G) NF_B promoter driven luciferase reporter assay: Rai overexpression in 293T cells induces a significant increase in NF_B dependent gene transcription. Luciferase activity was measured 48h after transfection and normalized to the co-transfected Renilla luciferase construct; mean±SD, **p < 0.01.|
|STEM_1056_sm_SuppFig12.tif||698K||Figure S12. Rai expression is higher in the Proneural subtype GBM. (A – B) Rai expression is higher in proneural subtype tumors compared to mesenchymal and proliferative subtypes. (A) Analysis of Rai expression in 75 GBMs derived from Phillips' work classified as proneural (PN), proliferative (Prolif) and mesenchymal (Mes). (B) Analysis of Rai expression in 200 GBM samples derived from Verhaak's work classified as proneural (PN), classical (CL), neural (NL) and mesenchymal (Mes). Data from samples assayed on three gene expression platforms (Affymetrix HuEx array-Broad202, Affymetrix U133A array-LBL202, and Agilent 244K array-UNC202) were analyzed. Y axes represent average expression values with a 95% confidence limit. The difference between groups was assessed by one-way ANOVA. To account for multiple comparisons, multiple t-test with Bonferroni's correction was applied. (C) Rai expression in gliomas from Phillips' work significantly correlates with Notch pathway activation: Rai expression is higher in those tumors which show a higher activation of Notch evaluated by IHC; proneural tumors are more frequent among high-Notch tumors (÷2=11.1, *p<0.05, DF=4 after Yates' correction). Statistical analysis: one way ANOVA (F=3.38; *p<0.05; DF=2;40); multiple t-test with Bonferroni's correction, Low vs High: Avg diff value = 0.85*; p<0.05, Medium vs High Avg diff value = 1.15**; p<0.01, Low vs Medium Avg diff value = 0.3 n.s.; p=0,87.|
|STEM_1056_sm_SuppTab1.pdf||14K||Supplementary Table 1.|
|STEM_1056_sm_SuppTab2.pdf||13K||Supplementary Table 2.|
|STEM_1056_sm_SuppTab3.pdf||14K||Supplementary Table 3.|
|STEM_1056_sm_SuppTab4.pdf||39K||Supplementary Table 4.|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.