High expression of estrogen receptor alpha and aromatase in glial tumor cells is associated with gender-independent survival benefits in glioblastoma patients

GBM tissue samples were surgically removed, formalin fixed, paraffin embedded and diagnosed as glioblastoma, IDH-wildtype (WHO grade IV), by two independent neuropathologists at our Department of Neuropathology.

Paraffin embedded tumor tissue was cut and placed on microscope slides. After deparaffinising and boiling in 10 mM citrate buffer for 30 min, the endogenous peroxidase was quenched in 1.5% H₂O₂ for 10 min. Avidin/Biotin Blocking Kit (Vector Laboratories Inc.) diluted in blocking buffer (PBS plus 1% BSA, 0.1% Trixon 100, 0.2% Gold Fish Gelatine, 0.02% NA acid, 2.5% normal horse serum) was used for 30 min. After washing in PBS the sections were incubated with an ERα antibody (DCS Innovative Diagnostik-Systeme GmbH, 1:20, clone 6F11) or an aromatase antibody (Sigma-Aldrich Co., 1:150, clone 19A1) diluted in blocking buffer overnight at 4 °C. The slides were washed in PBS and incubated with the secondary antibody for 30 min at room temperature (RT). After washing in PBS and incubation with Vecastan ABC Kit (Vector Laboratories Inc.) for 30 min at RT, ImmPACT DAB (Vector Laboratories Inc.) was used for developing under visual control. Counterstaining was done with Meyer’s Haematoxylin and mounted with Pertex. Placenta and breast tissue were used as positive controls. For negative controls the first antibody was omitted.

The slides were quantitatively assessed (10 fields of vision in 20 fold magnification) and scaled in four staining groups: 1 = 0–10%, 2 = 11–40%, 3 = 41–70%, 4 = 71–100% of immune positive glial tumor cells in the vital tumor spreading regions. The Aromatase immunohistochemistry showed a distinct staining of the cytoplasm of glial tumor cells, which were rated as positive. The used ER antibody stains the alpha subtype. The glial tumor cells showed ERα positive cell nuclei, but also cytoplasmic expression of ERs [15, 22]. For the assessment a tumor cell was rated positive when the cell nuclei alone or combined with the cytoplasm was positively stained.

GBM cell lines LN229 and LN18 were obtained from the American Type Culture Collection and cultivated in 75 cm² flasks in a humidified 5% CO₂/95% air incubator in high glucose DMEM supplemented with 5% fetal calf serum and 1% l-Glutamine at 37 °C.

Cell numbers were investigated using MTT assay. Briefly, cells were seeded in 96 well plates in 100 µl medium. After finishing the experiment 10 µl of the MTT labelling reagent (Sigma-Aldrich Co.) was added and incubated for 4 h in a humidified atmosphere. Then the medium was aspirated and 100 µl of the solubilisation solution DMSO was added and incubated for 30 min shaking at RT. After the incubation period a spectrophotometrically absorbance measurement was done at 595 nm (650 nm reference wavelength). As control a triplet of wells per plate was treated under experimental conditions without any cells.

GBM cell lines LN18 and LN229 were analysed for ERα protein. As positive control Gibco®Human Astrocytes (Life Technologies Corp.) were used, which express ERα under physiological conditions [15]. Cells were lysed in RIPA buffer (10 mM HEPES, 150 mM NaCl, 2 mM EDTA, 1% NP 40, 1% Trixon and 10 µl/ml RIPA PMSF, 10 µl/ml RIPA proteinase-phosphatase-inhibitor), transferred to an Eppendorf tube, sonicated for 30 s, incubated on ice for 15 min and centrifuged at 14,000 U/min for 10 min at 4 °C. The resulting supernate was used for the Bradford calculation and the other part was diluted 5:1 in Lämmli-buffer [10% SDS, 125 mM Tris, 10% Glycerol, 0.2% Bromophenol solution, 250 mM DTT (fresh added)], boiled for 5 min and electrophoresed on a 10% SDS-PAGE gel. The protein transfer to the nitrocellulose membrane was done as a gel sandwich inside a tray in a transfer tank filled with transfer buffer at 4 °C for 1 h with 100 V. The membrane was rinsed in TBS with 0.1% Tween 20 (TBST), blocked in 5% bovine serum albumin in TBST for 1 h at RT and incubated with ERα antibody (6F11, Thermo Fisher Scientific Inc., 1:500) at 4 °C overnight. After several rinses in TBST, the membranes got incubated with HRP-linked horse anti-mouse antibody (1:5000 in 5% BSA diluted in TBST) for 45 min at RT. Incubation with the HRP substrate (Merck KGaA) was performed after washing with TBST. After exposure of a film to the HRP wetted membrane the bands were visualized. Vinculin was used for the loading control.

Cells were seeded into a 96 well plate and treated 24 h later with estradiol (17β-estradiol: E8875, Sigma-Aldrich Co.) in three different dosing regimens: 10 μM, 25 μM and 50 μM diluted as described in the data information. After 48 h of incubation the cell viability was measured with MTT assay. After the estradiol treatment TMZ (Sigma-Aldrich Co.) was added in five different concentrations: 25 µM, 50 µM, 100 µM, 200 µM, 400 µM for 5 days and tested with MTT assay.

Excel 2016 and SPSS statistics 24 were used for all statistical analysis. Survival analysis was plotted on Kaplan–Meier curves according to the REMARK guidelines [23], adjusted to sex and body-mass-index. The Mantel-Cox lag rank test was used to compare survival distributions.

The GBM tissue samples, assessed as described in Materials/Methods and scaled in four groups: 1 = 0–10%, 2 = 11–40%, 3 = 41–70%, 4 = 71–100% of immune positive glial tumor cells. The classification resulted from the fact that many tissue samples showed no or only a small amount of ERα-positive stained tumor cells. Thus a clearer differentiation between less stained tissue samples was possible. For further classifications 0–40% (group 1 and 2) of immune positive stained tumor cells was rated as low expression of the hormones and 41–100% was classified as high expression.

Examples for each group of immunohistochemically stained tissue samples for ERα and aromatase as well as hematoxylin eosin are illustrated in Figs. 1 and 2.

The analysis of ERα staining demonstrated that 22 GBM tissue samples (12 female and 10 male) showed less than 10% immune positive tumor cells. Staining group 2 consisted of 20 tissue samples (8 female, 12 male) with 11–40% immune positive tumor cells. 13 tissue samples (5 female, 8 male) showed 41–70% of ERα immune positive tumor cells. 5 tissue samples demonstrated more than 70% of immunoreactive tumor cells, consisting of 4 female and 1 male patients.

The evaluation of the enzyme aromatase showed 2 tissue samples (2 female) with less than 10% of immune positive tumor cells. 8 GBM tissue samples, consisting of 4 female and 4 male patients, demonstrated 11–40% of immunoreactive cytoplasm. The staining group 3 consisted of 11 GBM tissue samples (8 female, 3 male) and 39 tissue samples showed more than 70% of immoreactive positive tumor cells (15 female, 24 male).

The GBM cohort of the present study consisted of 60 patients (29 female, 31 male), with an age range from 25 to 85 years and an median age of 60 years for female and 58 years for male patients. The survival time of GBM patients after surgery ranged from 22 to 1920 days, with a median survival time for females of 411 days and male patients of 432 days. The patients received cross total resections followed by Stupp protocol consisting of radiation and chemotherapy. The patients had a Karnofsky index greater than 60%. MGMT status is known for 57 of 60 patients. 29 patients have no methylation of the MGMT and 28 patients present a methylation. There is no statistically significant correlation between the expression of ERalpha or Aromatase and the MGMT status.

For survival analysis the staining intentions were judged as low expression with lower 40% of immune positive tumor cells (staining group 1 + 2) and high expression consisting of tissue samples with higher 40% of immune positive tumor cells (staining group 3 + 4). The Kaplan–Meier analysis showed that high ERα expression (> 40% immunoreactive tumor cells) in GBM patients was associated with significantly longer survival (Log rank: p = 0.0111, Fig. 3a), compared to low expression of ERα. Patients in the staining groups 3 + 4 showed a median survival of 492 days, whereas patients rated in staining groups 1 + 2 of ERα lived 371 days on median. GBM patients in high ERα staining groups lived 121 days on median longer than patients rated in staining group 1 + 2. The same effect was seen in GBM patients with high aromatase expression (> 40% immunoreactive tumor cells, Log rank: p = 0.0104, Fig. 3b), compared to low expression of the enzyme. GBM patients with low expression of aromatase lived 236 days on median, compared to high expression with a median survival time of 444 days, which was a difference of 208 days. No gender difference in the survival of female and male GBM patients appeared (Log rank: p = 0.89, Fig. 3c). The median survival time for female patients was 411 days and for male 432 days. Kaplan–Meier analyses were performed under adjusting to sex and BMI.

Before estradiol treatment a Western blot analysis was done, which confirmed that both cell lines LN18 and LN229 express ERα (Fig. 4).

Calculation of the half maximal inhibitory concentration (IC₅₀) of estrogen and TMZ was performed after therapy with estradiol alone and after the combined estradiol and TMZ treatment.

High concentrations of estradiol resulted in significantly lower tumor cell viability, compared to control (Fig. 5a). The IC₅₀ for the cell line LN229 was 43.11 µM 17β-estradiol (188.01 µM for solvent control) and for LN18: 20.96 µM 17β-estradiol (solvent control: − 191.14 µM). After 17β-estradiol and chemotherapy significantly lower cell viability, i.e. stronger sensitivity against TMZ, was seen for the estradiol concentrations 25 µM and 50 µM for both cell lines (Figs. 5b and 5c).