Flow cytometry analysis of neural differentiation markers expression in human glioblastomas may predict their response to chemotherapy

Abstract

Glioblastoma multiforme (GBM) represents an extremely chemoresistant tumour type. Here, authors analysed the immunophenotype of GBM tumours by flow cytometry and correlated the immunophenotypic characteristics with sensitivity to chemotherapy. The expression of selected neural and non-neural differentiation markers including A2B5, CD34, CD45, CD56, CD117, CD133, EGFR, GFAP, Her-2/neu, LIFR, nestin, NGFR, Pgp and vimentin was analysed by flow cytometry in eleven GBM (WHO gr.IV) patients. The sensitivity of tumour cells to a panel of chemotherapeutic agents was tested by the MTT assay. All tumours were positive for A2B5, CD56, nestin and vimentin. CD133, EGFR, LIFR, NGFR and Pgp were expressed only by minor tumour cell subpopulations. CD34, CD45, CD117, GFAP and Her-2/neu were constantly negative. Direct correlations were found between the immunophenotypic markers and chemosensitivity: A2B5 vs lomustine (r(2) = 0.642, P = 0.033), CD56 vs cisplatin (r(2) = 0.745, P = 0.013), %Pgp(+) vs vincristine (r(2) = 0.846, P = 0.008), and %NGFR(+) vs daunorubicine (r(2) = 0.672, P = 0.047) and topotecan (r(2) = 0.792, P = 0.011). In contrast, inverse correlations were observed between: EGFR vs paclitaxel (r(2) = -0.676, P = 0.046), CD133 vs dacarbazine (r(2) = -0.636, P = 0.048) and LIFR vs daunorubicine (r(2) = -0.878, P = 0.004). Finally, significant associations were also found among sensitivities to different chemotherapeutic agents and among different immunophenotypic markers. In conclusion, histopathologically identical GBM tumours displayed a marked immunophenotypic heterogeneity. The expression of A2B5, CD56, NGFR and Pgp appeared to be associated with chemoresistance whereas CD133, EGFR and LIFR expression was characteristic of chemosensitive tumours. We suggest that flow cytometric imunophenotypic analysis of GBM may predict chemoresponsiveness and help to identify patients who could potentially benefit from chemotherapy.

Fig. 1

Illustrative photomicrographs showing primary culture of glioblastoma cells isolated from a GBM tumour (a) and results of the MTT assay: cells cultured in the absence of chemotherapeutic agent (b) and after 72 h cultivation with cisplatin at 100 μg ml⁻¹ (c), 6,272 μg ml⁻¹ (d) and 0,392 μg ml⁻¹ (e). The amount of MTT reduced to formazan (blue crystals) is proportional to the number of viable cells (original magnification 10×, Leica DMIL inverted microscope)

Fig. 2

a Chemosensitivity of glioblastoma tumours: statistically significant correlations observed between chemosensitivities to BCNU, CCNU, CDDP, DAU and DTIC. b Immunophenotype of glioblastoma tumours: statistically significant correlations observed between the expression of A2B5, CD56, EGFR and Pgp. r², Spearman’s rho correlation coefficient

Fig. 3

Illustrative bivariate dot plots of the multiparametric five-colour immunophenotypic analysis of a U373 MG glioma cell line (positive control) and b glioblastoma tumour cells (case 1). The upper dot plots show Hoechst 33342-stained cells used to establish the baseline autofluorescence levels of tumour cells (unstained negative control). The lower panels show glioblastoma cells stained simultaneously for Hoechst 33342 and vimentin-FITC, GFAP -PE, A2B5-PE-Cy7 and nestin-Alexa Fluor 647; the cells display bimodal vimentin and nestin staining whereas A2B5 staining is uniform

Fig. 4

Illustrative bivariate dot plots of the immunophenotypic analysis of glioblastoma tumour cells. Examples of tumours positive for CD56 (case 5), CD133 (case 3), Pgp (case 5), EGFR (case 4), LIFR (case 4) and NGFR (case 11) are shown (positive cells are visualised as highlighted black dots). However, all tumours were negative for CD45, CD34 (case 3), Her-2/neu (case 1) and CD117 (case 1). CD45 was only expressed on infiltrating leukocytes