Differential patterns of NOTCH1-4 receptor expression are markers of glioma cell differentiation

Abstract

Background Notch signaling is deregulated in human gliomas and may play a role in their malignancy. However, the role of each Notch receptor in glioma cell differentiation and progression is not clear. We examined the expression pattern of Notch receptors and compared it with differentiation markers in glioma cell lines, primary human cultures, and biopsies of different grades. Furthermore, the effects of a γ-secretase inhibitor (GSI) on cell survival were assessed. Methods Notch receptors and markers of cellular differentiation were analyzed by reverse transcriptase PCR, Western blotting, immunohistochemistry, and immunocytochemistry. GSI sensitivity was assessed in both cell lines and primary cultures grown as monolayers or tumorspheres, by MTT assay. Results In cell lines, Notch1 and Notch2/4 levels paralleled those of glial fibrillary acidic protein (GFAP) and vimentin, respectively. In human gliomas and primary cultures, Notch1 was moderate/strong in low-grade tumors but weak in glioblastoma multiforme (GBM). Conversely, Notch4 increased from astrocytoma grade II to GBM. Primary GBM cultures grown in serum (monolayer) showed moderate/high levels of CD133, nestin, vimentin, and Notch4 and very low levels of GFAP and Notch1, which were reduced in tumorspheres. This effect was drastic for Notch4. GSI reduced cell survival with stronger effect in serum, whilst human primary cultures showed different sensitivity. Conclusion Data from cell lines and human gliomas suggest a correlation between expression of Notch receptors and cell differentiation. Namely, Notch1 and Notch4 are markers of differentiated and less differentiated glioma cells, respectively. We propose Notch receptors as markers of glioma grading and possible prognostic factors.

Keywords: Notch receptors; cell differentiation; gene expression; glioma cell cultures; gliomas; intermediate filaments.

Fig. 1.

Vimentin and GFAP are differentially expressed in rat primary astroglial cell culture (As) and C6, 9L, U87, and U373 glioma cell lines as assessed by RT-PCR (A) and western blot analysis (B). Differential Notch mRNA expression levels were detected by RT-PCR (C) in As and C6, 9L, U87, and U373 glioma cell lines. GAPDH was analyzed as a positive control of mRNA expression. Notch4 mRNA expression levels were analyzed by using 2 diverse primer pairs as indicated in Table 2. One primer pair was specific for the rat and the second one for human Notch4 (see Table 2). (D) Representative western blot showing Notch1, -2, -3, and -4 expression levels in As and C6, 9L, U87, and U373 glioma cell lines. One hundred micrograms of proteins were loaded in each lane. Experiments were repeated 3 times with similar results.

Fig. 2.

Representative western blot showing Notch1, Notch4, and Hes1 expression levels in homogenates of human glioma biopsies (GII = low-grade glioma; GBM1, GBM2, GBM3 = high-grade gliomas). U87 and C6 glioma cell lines were included as control for the expression levels of Notch1 and Notch4. Fifty micrograms of proteins were loaded in each lane.

Fig. 3.

Expression of Notch1,-2, -3, -4 immunoreactivity (IR) in human gliomas. Notch1 IR panels: A, E, F, K, L, Q, and R; Notch2 IR panels: B, G, M, S and T; Notch3 IR panels: C, H, N, U; Notch4 IR panels: D, I, J, O, P, V, Z. Panels: A, B, C, D normal white matter (WM), showing no detectable glial Notch1,-2, -3, -4 labeling. Panels E, F: representative photomicrographs of Notch1 IR in astrocytoma II (Astr. II), with strong nuclear and cytoplasmic Notch1 staining. Panels K, L: representative photomicrographs of Notch1 IR in anaplastic astrocytoma (Astr. III). Panels Q, R: representative photomicrographs of Notch1 IR in GBM, displaying variable IR and cytoplasmic staining. Panel G: representative photomicrographs of Astr. II, showing weak Notch2 staining. Panel M: Notch2 IR in Astr. III with cytoplasmic labeling. Panels S, T: representative photomicrographs of Notch2 IR in GBM, displaying variable IR and cytoplasmic staining. Panel H: representative photomicrographs of Astr. II, showing weak Notch3 staining. Panel N: Notch3 IR in Astr. III. Panel U: representative photomicrographs of GBM showing weak cytoplasmic staining. Panels I, J: representative photomicrographs of Notch4 IR in Astr. II, showing strong nuclear staining. Panels O, P: representative photomicrographs of Notch4 IR in Astr. III. Panels V, Z: representative photomicrographs of Notch4 IR in GBM displaying variable IR and both nuclear and cytoplasmic staining. Scale bar in A: 100 µm; B, C, D, and G: 100 µm; E, F, H–Z: 50 µm. Panels: A¹, B¹, C¹, and D¹ scatter plots showing the distribution of Notch1–4 IR scores (total score; see for details Method's section) in control WM and in astrocytic tumors: Astr. II, Astr. III, and GBM.

Fig. 4.

(A) Scheme showing the experimental procedure used to obtain primary cultures from GBM**. I Cells were seeded and grown in the presence of 10% FCS; II cells were seeded and grown in 10% FCS and then shifted to NSF medium when 80%–90% confluent; III cells were seeded directly and grown in a serum-free medium with growth factors (NSF). ** GBM cultures were prepared from FL75 listed in Table 3. Images are representative of primary cultures in the conditions already indicated. (B) Western blot analysis showing CD133, nestin, Notch-1 and Notch-4, vimentin, and GFAP expression levels in primary human glioma cell cultures under the different growth conditions listed (AI, AII, and AIII). Last lane is loaded with total proteins obtained from the tissue homogenate (part of the biopsy from where primary cultures were obtained). Actin and GAPDH expression levels are shown as controls of equal amount of total proteins loaded per lane. (C) Confocal images of immunocytochemistry experiments carried out in primary GBM cultures under different growth conditions by using the same antibodies as in B (AI seeded and grown in FCS and AIII seeded and grown in NSF).

Fig. 5.

Representative western blots showing Notch1, Notch4, and Hes1 expression levels in human primary glioma cell cultures (GII = low-grade glioma cultures; GBM1, GBM2, GBM3 = high-grade/glioblastoma cultures. Cultures were obtained from subjects indicated with * in Table 3). U87 and C6 glioma cell lines were included as a reference for the expression levels of Notch1 and Notch4. 1= total proteins obtained from human primary glioma culture seeded and grown in FCS; 2 = total proteins obtained from human primary glioma culture grown in FCS and shifted to NSF. Actin and GAPDH expression levels are shown as controls of equal amount of total proteins loaded per lane. Twenty-five micrograms of total proteins from glioma cell cultures and 50 µg of total proteins from U87 and C6 were loaded.

Fig. 6.

Analysis of 24 h GSI treatment in glioma cell lines. (A) Western blot analysis showing Notch1 and Notch4 expression levels in C6 and U87 grown in the presence of 10% FCS or in NSF. (B) Western blot analysis showing basal levels of Notch1 expression and reduction of NICD after 24 h treatment with 10 μM GSI of C6 and U87 grown in 10% FCS. Actin expression levels are shown as control of equal amount of total protein loaded per lane (100 μg). (C) Histogram showing caspase-3 activity in U87 and 9L glioma cell lines treated with GSI alone or in the presence of Z-VAD-FMK, a pan-caspase inhibitor. Values are the mean of 4 different samples, and error bars represent standard deviation. (D) Dose-response curves obtained after 24 h treatment with 1, 5, 10, 25, and 50 μM GSI of glioma cell lines grown in FCS or in NSF. Two-way ANOVA revealed a statistically significant difference between FCS and NSF growth conditions in all glioma cell lines (*P < .05, ** P < .001 different between FCS and NSF growth condition by pairwise multiple comparison procedure within each GSI concentration, Holm-Sidak method).

Fig. 7.

Dose-response curves obtained after 24 h treatment with 1, 5, 10, 25, and 50 μM GSI of primary human glioma cell cultures grown in FCS or grown in FCS and shifted to NSF. GII = low-grade glioma cultures; GBM groups = high-grade/glioblastoma cultures. In A low-grade gliomas GII the graph is the result of 3 diverse GII cultures; in B group 1 of high-grade/glioblastomas the graph is the result of 3 diverse GBM cultures. FCS cultures are more sensitive to GSI than NSF cultures. In C group 2 of high-grade/glioblastomas the graph is the result of 2 diverse GBM cultures. FCS and NSF cultures are equally sensitive to GSI. In D group 3 of high-grade/glioblastomas the graph is the result of 2 diverse GBM cultures. FCS cultures are less sensitive to GSI than NSF cultures. All experiments have been repeated 3 times. Two-way ANOVA revealed a statistically significant interaction between FCS and NSF growth conditions and GSI doses in group 3 (*P < .05, **P < .001 different between FCS and NSF by pairwise multiple comparison procedure within each GSI concentration, Holm-Sidak method).