Potentiation of temozolomide activity against glioblastoma cells by aromatase inhibitor letrozole

Abstract

Purpose

The DNA alkylating agent temozolomide (TMZ), is the first-line therapeutic for the treatment of glioblastoma (GBM). However, its use is confounded by the occurrence of drug resistance and debilitating adverse effects. Previously, we observed that letrozole (LTZ), an aromatase inhibitor, has potent activity against GBM in pre-clinical models. Here, we evaluated the effect of LTZ on TMZ activity against patient-derived GBM cells.

Methods

Employing patient-derived G76 (TMZ-sensitive), BT142 (TMZ-intermediately sensitive) and G43 and G75 (TMZ-resistant) GBM lines we assessed the influence of LTZ and TMZ on cell viability and neurosphere growth. Combination Index (CI) analysis was performed to gain quantitative insights of this interaction. We then assessed DNA damaging effects by conducting flow-cytometric analysis of ˠH2A.X formation and induction of apoptotic signaling pathways (caspase3/7 activity). The effects of adding estradiol on LTZ-induced cytotoxicity and DNA damage were also evaluated.

Results

Co-treatment with LTZ at a non-cytotoxic concentration (40 nM) reduced TMZ IC₅₀ by 8, 37, 240 and 640 folds in G76, BT-142, G43 and G75 cells, respectively. The interaction was deemed to be synergistic based on CI analysis. LTZ co-treatment also significantly increased DNA damaging effects of TMZ. Addition of estradiol abrogated these LTZ effects.

Conclusions

LTZ increases DNA damage and synergistically enhances TMZ activity in TMZ sensitive and TMZ-resistant GBM lines. These effects are abrogated by the addition of exogenous estradiol underscoring that the observed effects of LTZ may be mediated by estrogen deprivation. Our study provides a strong rationale for investigating the clinical potential of combining LTZ and TMZ for GBM therapy.

Introduction

Glioblastoma (GBM) is a highly aggressive and lethal cancer of the central nervous system (CNS) [1]. The current standard therapy for GBM patients comprises surgical resection followed by concomitant fractionated radiotherapy and chemotherapy with the DNA methylating agent temozolomide (TMZ), 7 days a week for 6 weeks at the dose of 75 mg/m². This is followed by six cycles of adjuvant TMZ administration at the dose of 150–200 mg/m², with each 28-day cycle consisting of 5 days of TMZ administration [2]. Unfortunately, this multimodal approach is not curative, and only extends the median survival from 4 months to approximately 15 months [3]. Even initially TMZ-sensitive tumors acquire drug resistance and recur in almost all patients. TMZ is an alkylating agent prodrug, delivering a methyl group to the N⁷ or O⁶ position of guanine residues of DNA with the primary lesion formed being O⁶-methylguanine (O⁶-MeG). The methylation leads to DNA strand breaks and triggers apoptotic tumor cell death [4]. The mechanisms of resistance to TMZ are complex and may result from selection of pre-existing resistant clones or genetic and epigenetic alterations of neoplastic cells. One of the primary mechanisms of resistance entails the overexpression of the repair protein O⁶-methlyguanine-DNA-methyltransferase (MGMT), which protects the cellular genome by removing the O⁶-alkylguanine DNA adduct [5]. Hypermethylation of the MGMT promoter results in decreased expression of the MGMT protein and is associated with increased sensitivity to TMZ, whereas a reduction in MGMT promoter methylation leads to the development of resistance to DNA alkylating agents and tumor recurrence. In the majority of cases recurrent GBMs are resistant to radiation and chemotherapy at the currently approved standard dosage [6]. In addition to this clinical limitation, TMZ use in GBM patients is associated with serious adverse effects. These include commonly observed gastro-intestinal (nausea, vomiting and diarrhea), CNS (convulsions and headaches) and hematologic (neutropenia and thrombocytopenia) toxicities [3]. Thus, there is a critical unmet need to devise strategies to enhance TMZ activity against GBM cells and overcome the development of resistance to improve outcomes and quality of life for patients and their families.

Considerable efforts have been expended on the discovery and development of chemosensitizers that may serve as clinically useful enhancers of TMZ activity by modulating oncogenic signaling pathways implicated in the development of TMZ resistance in GBM cells [7, 8]. Examples of such compounds include inhibitors of Wnt signaling such as salinomycin, celecoxib and Wnt-C59, JAK/STAT inhibitors such as SH-4–54, a salicylic acid-based STAT3 inhibitor and, HDAC inhibitors such as RGFP109 [9]. Several therapeutic agents such as thioridazine, an agonist of AMPK; mometinib, an ATP-competitive inhibitor of JAK-1 and JAK-2, and chloroquine and its derivatives such as hydroxychloroquine, which work to increase lysosomal pH and inhibit fusion of autophagosomes and lysosomes, have been shown to enhance TMZ cytotoxicity through induction of autophagic cell death [10–13]. There are several other targeted agents that have been advanced to the clinical trials stage to investigate their role as TMZ sensitizers. These include PARP inhibitors olaparib and veliparib; RTK inhibitors, lorlatinib and osimeritinib; p38-MAPK inhibitor, ralimetinib and other miscellaneous drugs such as anti-angiogenesis inhibitor, thaliodomide, and the pro-apoptotic agent, isotretinoin [14–21]. In general, these approaches have had limited clinical success. For instance, the use of hydroxychloroquine was found to be ineffective in phase I/II trials at the highest dose of 600 mg/day when used in conjunction with radiation and TMZ [4]. In addition to the risk of serious adverse side effects, some of these compounds such as olaparib, veliparib lorlatinib, and osimertinib are substrates for multidrug-resistance mediated drug efflux transporters such as P-gp and BCRP which may limit their CNS penetration [14–21].

In this regard, recent discoveries from our laboratory provide a strong rationale for combining TMZ with the aromatase inhibitor, letrozole (LTZ). Global gene expression analysis and immunohistochemical analysis show that the expression of CYP19A1, and its encoded protein aromatase, is significantly higher in GBM relative to normal brain tissue and low-grade gliomas [22–24]. Furthermore, in pre-clinical studies we observed that LTZ is strikingly effective in inhibiting GBM cell growth in vitro and in vivo, and exhibits facile transport across the BBB, accumulating in the tumor tissue at concentrations exceeding cytotoxic levels observed in vitro [25, 26]. In this study, we examined if LTZ has the potential to enhance TMZ activity against a diverse array of patient-derived GBM cell lines. In addition to performing quantitative analysis of the combined cytotoxic effects, we have also attempted to gain insights into the mechanistic bases for the observed interactions.

Materials and methods

Materials

Patient-derived GBM cells, G43, G75, and G76 were kindly provided by Dr. Jann Sarkaria, (Mayo Clinic Brain Tumor Patient-Derived Xenograft National Resource) [27]. BT-142 gfp luc cells were obtained from Cincinnati Children’s Hospital and Medical Center. Neural stem cell culture medium, Gibco^™ StemPro^™ NSC SFM (Catalog # A1050901), penicillin (50 U/mL), and streptomycin (50 mg/mL) and letrozole (99% pure), CyQuantTM LDH assay kit were purchased from Fisher Scientific. Temozolomide was purchased from Toronto Research Chemicals Inc. Enzyme linked immunoassay kit for estradiol was purchased from Fisher Scientific (Invitrogen Human Estradiol ELISA Kit REF #: KAQ0621). Caspase-Glo 3/7^® Assay kit was purchased from Promega Inc. Rabbit monoclonal anti-gamma H2A.X (phospho S139) antibody [9F3] (ab26350) was obtained from Abcam Inc.

Methods

Cell culture and maintenance

Patient-derived GBM cells were grown and maintained in neural stem cell growth medium (StemPro^™ NSC SFM; Gibco^™ Cat# A1050901) supplemented with NSC growth supplement and hEFG and hFGF and 10% penicillin & streptomycin. Serum-free medium enables spheroid formation in the patient-derived cells and therefore, use of any serum containing material was avoided. Cells were cultured in T-25 flasks till 70–80% confluence was reached, at which point the spheroids formed were disrupted into single cells using trituration with a P200 pipette. The confluent single cell suspensions were used for cell counting using a hemocytometer and trypan blue exclusion and seeded into appropriate plates at appropriate densities for further assays.

Aromatase expression analysis

GBM cell pellets were collected following centrifugation of the cell suspension and lysed with RIPA buffer (1 M Tis pH 7.4, 5 M NaCl, 1% NP-40, 50 mM NaF, 0.5 M EDTA, 0.1% SDS, 0.5% Na-deoxycholate) and protease inhibitor cocktail (1:100). The lysates were centrifuged at 14,000 rpm at 4 °C for 10 min and the supernatant was collected. The supernatant was analyzed for protein concentration using the Pierce BCA protein assay kit. Protein lysates were mixed with 1X Lamelli buffer. The samples were heated at 95 °C for 10 min.

Western blot analysis was performed for aromatase expression using 10% pre-cast gel. Equal quantity (30 μg) was loaded, and the gel was allowed to run for 45 min at 110 V. Protein transfer was then performed using TransBlot Turbo^™ PVDF membrane transfer packs, at 25 V for 10 min. Protein transfer on the membrane was assessed using Ponceau S staining. The membrane was blocked at room temperature using 5% non-fat dry milk in 1 X Tris-buffered saline tween (TBST) for 1 h. The membrane was incubated at 4 °C for 12 h with rabbit anti-CYP19A1 antibody (Abcam: ab18995) (1:1000) and rabbit anti-human vinculin antibody (Abcam: ab155120) (1:2000). Membrane was washed with 1X TBST 3-times followed by mouse anti-rabbit HRP linked antibody (Abcam: ab) for 1 h at room temperature. This was followed by the washing step with 1X TBST. The membrane was then exposed to the ECL buffer, and the protein expression bands were captured using an X-ray film [23]. The western blot image for aromatase expression analysis is shown in Supplementary Fig. 1.

Cell viability assay

We assessed the cytotoxic effects of TMZ, LTZ and various combinations of the two agents using the CyQuantTM Lactate Dehydrogenase (LDH) assay (ThermoFisher Scientific). This assay measures the release of LDH, a cytosolic enzyme, into the extracellular medium resulting from damage to the plasma membrane. Cells (1.4 X 10⁴) were plated in 96 well plates in the above-mentioned serum-free medium. Twenty-four hours after seeding cells were treated with TMZ (0–100 μM), LTZ (0–2 μM), or TMZ (0–10 μM) + LTZ (40 nM). Dimethyl sulfoxide (0.1%) was used as a vehicle control. The effect of adding estradiol (250 pg/ml) on cell viability was also determined [28].

Neurosphere growth inhibition

Since G43 and BT-142 cells form neurospheres in serum free culture medium, we also determined the effects of LTZ and TMZ on the growth of the neurosphere. Cells (1 × 1 0⁴) were plated and 24 h later treated with LTZ (0–2 μM), TMZ (0–100 μM) or LTZ (40 nM) + TMZ (0–10 μM) for 72 h cells and DMSO (0.1%) serving as vehicle control. Neurosphere dimensions were measured using Leica Stereomicroscope in bright field setting. The effect of addition of estradiol (250 pg/ml) on the neurosphere growth was also determined [28].

Combination index analysis

To evaluate the nature of the interaction between LTZ and TMZ we employed Combination Index (CI) analyses, first suggested by Chou and Talalay (1984) [29]. For each cell-line, 1.4 × 1 0⁴ cells were seeded in each well of a 96-well plate and allowed to grow for 24 h. Cells were treated with LTZ, TMZ and LTZ + TMZ (0–10 μM) at a fixed LTZ:TMZ ratio across the concentration range. The cytotoxicity of the combination was determined using the LDH assay and the fraction affected (% cell kill relative to control) at each combination was plotted using the Compusyn^® software. For CI analysis of the neurosphere growth inhibition by LTZ and TMZ, BT-142-gfp-luc and G43 cells 1.4 × 1 0⁴ cells were plated per well of a 96-well plate and allowed to grow for 24 h. Cells were treated with LTZ, TMZ and LTZ + TMZ (0–10 μM) at a fixed LTZ:TMZ ratio across the concentration range. Microscopic images were captured 72 h after the treatment and the fraction affected (% decrease in neurosphere size relative to control) at each concentration of the combination was plotted using Compusyn^® Software. The CI values were determined using the following equation: $\text{CI}=(\text{C})1/(\text{C}x)1+(\text{C})2/(\text{C}x)2$, where (Cx)1, (Cx)2 = the concentration of the tested substance 1 and the tested substance 2 used in the single treatment that was required to decrease the cell number by x% and (C)1, (C)2 = the concentration of the tested substance 1 in combination with the concentration of the tested substance 2 that together decreased the cell number by x%. The CI Value quantitatively defines synergism (CI < 1), additive effect (CI = 1) and antagonism (CI > 1). These experiments were repeated in the presence of estradiol to assess the extent to which estradiol impacts the cytotoxic effects of the combination.

Assessment of gamma-H2A.X

Induction of ˠH2A.X in all four GBM lines (1 × 1 0⁶) treated with LTZ, TMZ and LTZ + TMZ (0–10 μM) in the presence and the absence of estradiol (250 pg/ml) was assessed using flow cytometric analysis as described earlier [30]. Based on initial assessment of the time course of ˠH2A.X induction, a 5-h treatment was used to capture maximal effects (the time-course of ˠH2A.X induction is shown in supplementary Fig. 2). Briefly, cell pellets were collected by centrifugation of the cells at 2000 × g for 5 min at room temperature. Cells were then washed with ice-cold PBS, followed by addition of 1:1 mixture of methanol and NP40 (300 μL) for fixation and permeabilization of the cells. Following centrifugation at 4 °C at 290 × g, primary unconjugated antibody for gamma-H2A.X (100 μL of 1:250 dilution) was added. The cells were then incubated for 20 min and washed with 600 μL of 1:1 methanol and NP40 mixture. Secondary PerCP conjugated antibody was added to the cells at a dilution of 1:1000 and incubated at room temperature in the dark for 20 min. Cells were washed again with 600 μL of 1:1 methanol and NP40 mixture and re-suspended in 1X PBS and subjected to flow cytometric analysis at an excitation wavelength of (488 nm) employing Beckman CytoFlex flow cytometer. Gating was done using Human IgG-isotype control as outlined in the manufacturer’s protocol. The scatter plots were analyzed using CytExpert version 2.3, to assess the percentage of cells positive for gamma-H2A.X versus the isotype control.

Apoptosis induction assay using Caspase-Glo^® 3/7 system

The induction of apoptosis by LTZ, TMZ and LTZ + TMZ combination as evident by caspase 3/7 activation was assessed in all four GBM lines, G76, BT142, G43 and G75. Cells were seeded at a density of 1.5 × 10⁴ cells per well in a translucent walled, clear bottom, 96-well plate. Promega Caspase-Glo^® 3/7 Assay system was employed to assess the Caspase 3/7 levels in cells treated with LTZ, TMZ, LTZ (40 nM) + TMZ (0–5 μM) and vehicle (0.1% DMSO) [31]. Cells were treated for 24 h after plating. Caspase-Glo^® 3/7 3D substrate and Caspase-Glo^® 3/7 3D buffer was mixed to form a 10 mL reaction mixture and 100 μL of the mixture was added to each well. The reaction was allowed to proceed at room temperature in the dark for 30 min as described in the Promega Caspase-Glo^® 3/7 3D-Assay technical manual. The luminescence readings were carried out 24 h after the treatment using Synergy^™ HTX Multi-Mode Microplate Reader. The experiment was repeated in presence of estradiol to assess the effect of estradiol on caspase induction in in four GBM lines used in this study. The data were normalized to control using Promega CellTiter- Glo^® assay.

Statistical analysis

The results were expressed as mean values ± standard deviation (S.D). Statistical comparisons between the treatments were analyzed by repeated measure (RM) one-way analysis of variance (ANOVA) and multi-parametric statistical testing was performed to assess statistical significance within the treatment groups using multi-variate correlation matrix using GraphPad^® Prism 8.0.2. A significance level of P < 0.05 was used for all tests.

Results

Potentiation of cytotoxicity and neurosphere growth inhibition of TMZ by LTZ

We first examined the cytotoxic effects of TMZ and LTZ on various patient-derived GBM lines grown in serum-free medium. In initial experiments we assessed the cytotoxic effects of each compound individually. We then repeated the experiments with cells treated with TMZ and LTZ and TMZ combination with LTZ at a relatively non-cytotoxic concentration of 40 nM. Simultaneously, viability of cells treated with TMZ and LTZ (40 nM) in the presence of exogenously added E2 (250 pg/ml) was also determined. Cells were treated for 72 h and the IC₅₀ values for LTZ and TMZ cytotoxicity were determined. The data for percentage of cell survival as function of drug concentrations (0 to 100 μM) are shown in Fig. 1. Based on this initial cell viability assessment, G76 cells are TMZ-sensitive (IC₅₀ = 1.31 μM) (Fig. 1A), BT-142 cells have intermediate TMZ sensitivity (IC₅₀ = 15.33 μM) (Fig. 1B), whereas G43 and G75 cells with TMZ IC₅₀ values of 165.43 and 106.73, respectively are clearly TMZ-resistant (Figs. 1C, D). All four lines were, however, fairly sensitive to LTZ with I C₅₀ values ranging from 0.08 to 0.87 μM. To evaluate the effects of the combination, we then assessed the cytotoxicity of TMZ in the presence of 40 nM LTZ as at this concentration LTZ was non-cytotoxic. As shown in Fig. 1, LTZ significantly enhanced TMZ cytotoxicity in all four GBM lines. The TMZ IC₅₀ values in the presence of LTZ (40 nM) ranged from 0.16 to 0.69 μM (Table 1).

Figure 1:

Effect of drug treatment on viability of patient-derived TMZ-sensitive and TMZ-resistant cells determined using WST-1 LDH assays. Cells (1.4 × 103 /well) were treated with the indicated drug concentrations for 72 h (N = 3 per treatment group) with or without the addition of estradiol (250 pg/mL). A G76, B BT-142, C G43 and D G75 cells. (data: mean ± SD). Multi-parametric correlation matrix analysis and repeated measures one-way ANOVA were used to analyze statistical differences across the treatment groups. p value: * = < 0.05; ** < 0.01; *** < 0.001; n.s. > 0.05

Table 1.

IC50 and C.I. values for LTZ, TMZ and LTZ (40 nM) + TMZ with and without exogenous E2 against Patient-Derived GBM lines and EC50 values for caspase induction

Patient-Derived GBM Line	MGMT methylation	IC50				Predicted EC50 for caspase induction				Combination Index (for IC50)
		LTZ (μM)	TMZ (μM)			LTZ (μM)	TMZ (μM)			Without exogenous E2	With exogenous E2
			TMZ alone	LTZ (40 nM) without exogenous E2	LTZ (40 nM) with exogenous E2		TMZ alone	LTZ (40 nM) without exogenous E2	LTZ (40 nM) with exogenous E2
G76	M	0.078	1.31	0.16*	2.7	1.8	6.4	0.6*	356.2	0.62*	2.08
G75	U	0.448	106.73	0.167***	91.24	5.1	93.8	0.7**	445.8	0.09***	204.5
G43	U	0.870	165.43	0.690***	109.7	8.2	135.9	0.83**	512.4	0.05***	126.8
BT-142	U	0.670	15.33	0.413*	14.84	3.9	35.2	0.65**	402.9	0.09*	23.1

U= Unmethylated, M = Methylated

RM One-way ANOVA was performed to assess statistical differences across the treatment groups using GraphPad Prism 8.0.2.

^*= p < 0.05;

^***= p < 0.0001

The potentiation of TMZ cytotoxicity by LTZ was also evident in the neurosphere growth assay. TMZ-sensitive BT142 and TMZ-resistant G43 drug-treated for 72 h, the neurospheres were visualized using a microscope. Figure 2A, B provide representative pictures of neurospheres treated with vehicle control, LTZ (40 nM), TMZ (1 μM) and TMZ (1 μM) + LTZ (40 nM). In BT-142 cells, the median neurosphere diameter in wells treated with the solvent vehicle, LTZ (40 nM), TMZ (1 μM) or TMZ (1 μM) + LTZ (40 nM) were, 1.3 ± 0.09, 1.1 ± 0.077, 0.9 ± 0.023 and 0.23 ± 0.05 mm, respectively. Likewise, in G43 cells the neurospheres diameters were, 1.9 ± 0.11, 1.85 ± 0.13, 1.6 ± 0.085 and 0.05 ± 0.01, respectively.

Figure 2.

Microscopic analysis of the cytotoxic effects of LTZ and TMZ on neurosphere growth of BT-142 and G43 cells. Cells (1.4 × 10³/well) were treated with solvent control (0.1% DMSO) and the indicated concentration of LTZ and TMZ (N = 3 per treatment group) with or without the addition of exogenous estradiol (250 pg/mL) for 72 h. Representative microscopic images are shown for BT-142-gfp-luc cells (A) and G43 cells (B) for cells treated with vehicle, LTZ (40 nM), TMZ (1 μM), TMZ (1 μM) + LTZ (40 nM) and TMZ (1 μM) + LTZ (40 nM) with E2. The relative diameters were determined and plotted as shown above for BT-142-gfp-luc (C) and G43 (D) cells (data: mean ± SD). Scale displayed on each image = 0.2 mm

As evident, while LTZ at 40 nM minimally impacted the neurosphere growth, when combined with TMZ, it markedly enhanced the neurosphere growth inhibition by TMZ in both cell lines. The effect of varying TMZ and LTZ concentration and TMZ in the presence of LTZ (40 nM) on the reduction of neurosphere size is shown in Fig. 2C, D. Also, as shown in this Fig. 2, the addition of exogenous E2 (250 pg/ml) counteracted the potentiating effects of LTZ in BT-142 and G43 cells.

We then performed combination index analysis of both cell viability and neurosphere growth inhibition data to determine whether the observed potentiation of TMZ activity was additive or synergistic. Using Compusyn^® software we plotted the combination index (CI) plot (Fig. 3). The CI analysis for cell viability assessed in all four lines is shown in Fig. 3A and for the neurosphere growth inhibition using BT-142 and G43 cells is shown in Fig. 3B. The CI values at the IC₅₀ are included in Table 1. In each case the CI values were consistently below the line of additivity (CI = 1) indicative of strongly synergistic interaction between LTZ and TMZ. The CI values at the I C₅₀ ranged from 0.05 and 0.09 for TMZ-resistant G43 and G75 cells which is consistent with the striking reduction in TMZ IC₅₀ values underscoring that the extent of synergy was considerably greater for the TMZ-resistant G43 and G75 lines.

Figure 3.

Combination index plots to assess the nature of interaction of TMZ and LTZ combination and the effect of estradiol addition (250 pg/mL) to the cell culture medium on the interaction. A CI plots for cell viability: G76, G75, BT-142-gfp-luc, G43 (± E2) B CI plots for neurosphere growth inhibition assays in BT-142-gfp-luc, G43 (± E2) based on neurosphere size measurement

Assessment of DNA damage and apoptosis by TMZ and LTZ

Since the mechanism of cell death by DNA alkylating agents such as TMZ entails increased DNA damage leading to double strand DNA breaks and subsequent tumor cell apoptosis, we investigated the DNA damaging and apoptotic effects of these compounds in all four GBM lines. The phosphorylation of histone H2AX following double strand DNA breaks to form γ-H2AX was assessed as a marker for DNA damage. The detection and visualization of γ-H2AX was carried out using flow cytometry. An initial assessment of the time course revealed that maximal γ-H2AX levels were detected approximately 5 h following drug treatment (data included in Supplementary Fig. 2). As such, γ-H2AX formation was measured following a 5-h treatment with TMZ and LTZ combinations. The effect of LTZ (40 nM) and the addition of estradiol to the cell culture medium was also investigated. As shown in Fig. 4, TMZ at concentrations up to 0.1 μM had noticeable induction of γ-H2AX only in TMZ-sensitive G76 cells (Panel A), but not in BT-142 (Panel B), G43 (Panel C) and G75 (Panel D). On the other hand, LTZ alone (500 nM and 1 μM) considerably increased γ-H2AX levels in all four lines(data not shown). Most importantly, LTZ (40 nM) when combined with TMZ (0.1 μM) had a dramatic impact on the number of cells that stained positive for γ-H2AX levels. The effects of LTZ-mediated induction of γ-H2AX and were abrogated in the presence of estradiol.

Figure 4.

Flow cytometry dot plots exhibiting induction of γH2A.X in G76 (A), BT-142 (B), G43 (C) and G75 (D) cells treated with LTZ, TMZ and LTZ + TMZ (± E2) for 5 h. The EC50 for DNA damage for each treatment were predicted using Phoenix^® WinNonLin v8.3 using a standard Emax model and are described in Table 2 (Data; mean ± SD)

Next, we investigated caspase-3/7 activity as a read out for cells undergoing apoptosis following treatment with TMZ and/or LTZ. Caspase 3 and Caspase 7 are cysteine-aspartic acid proteases which can directly execute apoptosis following sequential activation of caspase-8 or caspase-9. As such, induction of caspase-3/7 is an accepted approach to measure apoptosis. We employed a Caspase-Glo^® 3/7 luminescence assay to determine induction of apoptosis following 24 h drug exposure in all four GBM lines used here, data for all GBM lines were normalized using Promega CellTiter-Glo^® assay. As shown in Fig. 5, LTZ induced apoptosis in a concentration-dependent manner. An Emax model was employed to determine the EC₅₀ values, which ranged from 1.8 ± 0.31 and 8.2 ± 2.18 μM for cells treated with LTZ alone. More importantly, LTZ (40 nM) caused a striking increase in TMZ-induced apoptosis in all four lines resulting in marked reduction in the EC₅₀ of TMZ for caspase-3/caspase-7 induction. These observed effects of LTZ were diminished in the presence of exogenously added E2. The EC₅₀ values in the presence and the absence of E2 are also included in the Table 2.

Figure 5.

Induction of caspase 3/7 activity as a marker for apoptosis. G76 (A), BT-142 (B), G43 (C) and G75 (D). Cells were treated with LTZ, TMZ, LTZ (40 nM) + TMZ and LTZ (40 nM) + TMZ + E2 (250 pg/mL) for 24 h (N = 3 per treatment group) and caspase activity was determined using the Caspase-Glo 3/7 assay kit (normalized with Promega CellTiter-Glo^® Luminescent Cell Viability assay). The EC50 for caspase induction for each treatment were predicted using Phoenix^® WinNonLin v8.3 using a standard Emax model and are described in Table 2

Table 2.

IC50 and C.I. values for LTZ, TMZ and LTZ (40 nM) + TMZ with and without exogenous E2 for inhibition of neurosphere growth in G43 and BT-142-gfp-luc cells

Patient-Derived GBM Line	IC50				Combination Index
	LTZ (μM)	TMZ (μM)			Without exogenous E2	With exogenous E2
		TMZ alone	LTZ (40 nM) without exogenous E2	LTZ (40 nM) with exogenous E2
G43	0.750	148.91	0.5104***	160.11	0.003***	214.14
BT-142	0.614	33.37	0.412*	30.92	0.08*	60.11

RM One-way ANOVA was performed to assess statistical differences across the treatment groups using GraphPad Prism 8.0.2.

^*= p < 0.05;

^***= p < 0.0001

Discussion

In this study we tested the hypothesis that LTZ enhances the cytotoxic effects of TMZ, a DNA alkylating agent that is the first line treatment for GBM. LTZ at nanomolar concentrations markedly potentiated TMZ activity in patient-derived cell lines with varying sensitivity to TMZ. Whereas the IC₅₀ values in the G76 (TMZ-sensitive) and BT142 (intermediate TMZ sensitivity) were 1.76 μM and 15.33 μM, respectively, in the TMZ-resistant G75 and G43 the I C₅₀ values were 106.73 μM and 165.43 μM, respectively. This magnitude of loss of TMZ sensitivity is well documented and underscores the clinical failure of TMZ in the recurrent setting [1, 2, 4]. At doses of 100–200 mg/m² the peak plasma levels of TMZ in humans are in the range of 5–11 μg/ml (25–50 μM) [29]. Studies conducted using intracerebral microdialysis in GBM patients to assess the peritumoral TMZ levels suggest that the partitioning of the drug from plasma to the brain extracellular fluid (ECF) is only 13–17% resulting in concentrations ranging from approximately 3–8 μM [32]. Clearly, with the development of resistance, these tumoral concentrations of TMZ are sub-therapeutic.

Another reason for the development of resistance to TMZ is related to the presence of glioma stem cells (GSCs) within GBM. The GSC subpopulations are important for the maintenance of tumor heterogeneity, as well as, for the tumor initiation, maintenance, and invasion because of their capacity for self-renewal and differentiation [33]. GSCs drive resistance to pharmacology, radiation, and surgery, and are thus a key therapeutic target. Therefore, we employed the neurosphere growth inhibition assay since the growth of the patient-derived cells in neural stem cell growth medium facilitates formation of floating spheres comprising GSCs. Microscopic examination of BT-142 and G43 cells treated with LTZ and TMZ combination clearly demonstrate that LTZ potentiated the effects of TMZ against cells required to initiate or sustain neurosphere growth. However, a limitation of our approach is that the neurosphere comprises a heterogenous mass and since we did not use a standardized method such as the limited dilution assay, we were not able to confirm the inactivation or removal of clonogenic GSCs in drug-treated neurospheres [34].

As mentioned earlier, most investigational compounds and therapeutic agents investigated as sensitizers of TMZ activity have had limited clinical success primarily due to significant risks of adverse events. In that regard, we believe that LTZ use has significant advantages. Firstly, LTZ is an FDA approved third generation non-steroidal aromatase inhibitor with a sustained record of safe use. During its development in the phase I clinical trials the compound was found to be generally well tolerated in both and women and men at multiple doses of 10 mg per day [35, 36]. LTZ is primarily metabolized by CYP2A6, a relatively minor liver enzyme in humans, which is consistent with lack of any significant drug-drug interactions due enzyme inhibition or enzyme induction or other processes altering the pharmacokinetics of other co-administered drugs [37, 38]. Our previous studies have comprehensively quantitated the plasma to brain ECF partitioning of LTZ in normal and tumor-bearing rats. These studies show that the brain partitioning, determined comparing the ratio of plasma peak concentrations or AUC of the LTZ in the brain ECF vs unbound plasma concentration is in the range of 0.8–1.2 indicating facile transport of the drug to the brain. The brain tumoral levels were about 1.5-fold higher compared to the tumor-free hemisphere in the brain [26, 39, 40]. Thus, with a record of safe use, desirable pharmacokinetic attributes including excellent BBB permeability, LTZ is a clinically viable combination agent with TMZ.

Furthermore, directly relevant to the pharmacological basis of combining TMZ with LTZ, in a previous study we investigated the potential for drug-drug interactions between these two agents in Sprague Dawley rats. In addition to assessing plasma pharmacokinetics, employing intracerebral microdialysis we also measured the brain ECF concentrations of both agents. Co-administration of these agents did not modulate the plasma or brain ECF peak levels, AUC and elimination half-life of either drug, suggesting a lack of drug-drug interactions [39]. This is consistent with the fact that these compounds do not share clearance pathways. TMZ is an imidazotetrazine which is completely bioavailable after per oral administration. TMZ undergoes non-enzymatic spontaneous hydrolysis to 5-(3-methyltriazen-1-yl) imidazole-4-carboxamide (MTIC). A short-lived intermediate with an elimination half-life of 2 min, MTIC rapidly breaks down to form the reactive methyldiazonium ion that alkylates DNA. MTIC produced in plasma is not able to cross BBB, due to its extremely short half-life and as such it is formed locally in the brain [41]. Given its short half-life, the pharmacokinetics of MTIC in the brain is governed by that of TMZ [41]. Thus, determining TMZ pharmacokinetics is reflective of MTIC formation and clearance. Overall, it appears that the combination of these two agents is unlikely to have any pharmacokinetics based drug-drug interactions [39, 40].

In this study we have also explored the mechanistic basis for LTZ mediated potentiation of TMZ cytotoxicity. Based on the understanding of the function of aromatase as an estrogen synthase in breast cancer tissue, multiple lines of evidence suggest that LTZ treatment results in inhibition of estrogen synthesis which then triggers a cascade of downstream effects leading to apoptotic cell death. The apoptotic signaling pathways induced involve down-regulation of Bcl2, up-regulation of Bax, and the activation of caspase-9, caspase-6, and caspase-7 [42]. As indicated earlier, TMZ mediates its activity by methylating DNA, which mostly occurs at the N⁷ or O⁶ position of guanine residues. The methylation leads to DNA strand breaks and triggers apoptotic tumor cell death [4]. Thus, to gleam the mechanism of perceived synergistic potentiation to TMZ activity by LTZ, we determined the extent of DNA double strand break by assessing the formation of phosphorylated histone product ˠ-H2Ax and induction of caspase3/7 as a marker of apoptosis. We measured the ˠ-H2Ax levels 5 h after drug treatment and as anticipated did not observed any measurable production of ˠ-H2Ax in TMZ-resistant G43 and G75 lines. LTZ, on the other hand caused DNA damage in all four cell lines used. More importantly, LTZ at 40 nM concentrations strikingly enhanced DNA damage by TMZ at concentrations as low as 0.1 μM. For instance, we noted striking differences in caspase 3/7 induction by TMZ in G75 (EC₅₀ = 93.8 μM) and G76 (EC₅₀ = 6.4 μM) cells, which reduced to 5.1 and 1.8 μM, respectively, in the presence of LTZ (40 nM). There was good concordance of the observed I C₅₀ in cell viability assays and the projected E C₅₀ values for caspase 3/7 induction of TMZ in the presence and the absence of LTZ. Thus, taken together, it appears that the potentiation of TMZ activity by LTZ is mediated by increased DNA damage resulting in apoptotic cell death in TMZ-sensitive and TMZ-resistant lines.

Furthermore, the observed effect of LTZ appears to result, at least partially, from its anti-estrogenic activity. In a previous study we noted that the I C₅₀ of LTZ cytotoxicity against several malignant brain tumor cell lines correlated well with the IC₅₀ for the inhibition of in situ estradiol production and the addition of exogenous estradiol averted the cytotoxic effects of LTZ [25]. In this study also we observed that the addition of estradiol to the cell culture media abrogated the LTZ effects in cell viability and neurosphere growth inhibition assays. Similarly, the presence of estradiol prevented LTZ-mediated DNA damage and induction of apoptosis and negated the synergistic potentiation of TMZ activity. In the presence of exogenous estradiol, the CI analysis plots for the cell viability were above the line of additivity suggesting that that estradiol not only canceled LTZ effects but may have protected cells from TMZ-mediated damage.

Consistent with beneficial effects of interfering with estrogen action, compounds known to have anti-estrogen activity such as tamoxifen have also been shown to inhibit glioma growth, and high dose tamoxifen has been investigated in clinical trials in combination with TMZ [43]. However, a fully coherent understanding of estrogen in brain tumorigenesis is still to be achieved. The demographic/epidemiological data are conflicting and are suggestive of sexual dimorphism in GBM. The incidence of GBM is higher in males and male patients have a worse prognosis than females [44]. Furthermore, the incidence rate of gliomagenesis in pre-menopausal women is also significantly lower suggesting that peripheral ovarian-derived estrogen may have a protective role [44]. Sareddy et.al (2012) have shown that ERβ, which is considered a tumor suppressor, is expressed in GBM cells and that ERβ agonist, LY500307, may be a novel agent to treat GBM [45]. Importantly, LTZ therapy would be expected to affect both ERα and ERβ pathways; the sum of the negative or positive impact on GBM growth would, therefore, be theoretically dependent on the relative expression of these two ER receptors.

In addition to the confounding tissue-specific expression and function of estrogen and androgen receptors, several other mechanistic bases for the observed gender dependence of GBM incidence and prognosis have been postulated by recent studies. For instance, Kfoury and colleagues (2018) observed that male astrocytes have a lower malignant transformation threshold compared to females [46]. Recent work by Bayik and colleagues (2020) suggested that there are sex-dependent differences in the immune response in GBM due to differences in myeloid-derived suppressor cell (MDSC) subsets and localization [47]. In mouse models of GBM, they observed that monocytic MDSCs (mMDSCs) were enriched in the male tumors, whereas granulocytic MDSCs (gMDSCs) were elevated in the circulating blood in females. Depletion of gMDSCs extended survival only in female mice. They further demonstrated in preclinical models that mMDSCs could be targeted with antiproliferative agents in males, whereas gMDSC function could be inhibited by IL1β blockade in females. Further mechanistic studies on LTZ effects in glioma cells and the significance of baseline and post LTZ treatment aromatase expression in the glioma setting is currently being investigated. Additionally, studies in patient derived xenograft animal models are warranted to discern the LTZ doses needed for synergistic potentiation of TMZ activity in drug-sensitive and drug-resistant cells.

In summary, this study documents a robust potentiation of TMZ cytotoxic effects in patient-derived TMZ-sensitive and TMZ-resistant cells by LTZ. Using the quantitative approach of Combination Index analysis this interaction was deemed to be highly synergistic. LTZ at low and non-cytotoxic concentrations caused substantial reduction in the IC₅₀ values of TMZ in drug-resistant cells, which suggests that LTZ co-treatment results in restoring TMZ sensitivity. While further mechanistic and confirmatory in vivo studies are warranted, our study strongly suggests that combining LTZ will lead to robust increase in anti-tumor efficacy of TMZ. Thus, combining the two agents may facilitate a reduction in the therapeutic doses of TMZ and thereby minimize the debilitating adverse effects of this compound. Given that LTZ is an FDA approved compound with a relatively benign record of safe use, its repurposing as combination partner for TMZ may be achieved expeditiously to address the urgent need for effective therapeutic treatment of GBM.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.