Rabeprazole has efficacy per se and reduces resistance to temozolomide in glioma via EMT inhibition

Abstract

Purpose

Epithelial to mesenchymal transition (EMT) is pivotal in embryonic development and wound healing, whereas in cancer it inflicts malignancy and drug resistance. The recognition of an EMT-like process in glioma is relatively new and its clinical and therapeutic significance has, as yet, not been fully elucidated. Here, we aimed to delineate the clinical significance of the EMT-like process in glioma and its therapeutic relevance to rabeprazole.

Methods

We investigated the expression profiles of EMT-associated proteins in primary glioma biopsies through Western blotting and immunohistochemistry, and correlated them with various clinicopathological features and data listed in the cancer genome atlas (TCGA). In addition, the anticancer efficacy of rabeprazole and its therapeutic relevance to EMT along with temozolomide chemo-sensitization were assessed using multiple cell-based assays, Western blotting and confocal imaging. For in vivo assessment, we used a stereotaxic C6-rat glioma model.

Results

Expression analysis of EMT-associated proteins in glioma biopsies, in conjunction with clinicopathological and TCGA dataset analyses, revealed non-canonical expression of E/N-cadherin and upregulation of GFAP, vimentin and β-catenin. The increased expression of EMT-associated proteins may attribute to glioma malignancy and a poor patient prognosis. Subsequent in vitro studies revealed that rabeprazole treatment attenuated glioma cell growth and migration, and induced apoptosis. Rabeprazole suppressed EMT by impeding AKT/GSK3β phosphorylation and/or NF-κB signaling and sensitized temozolomide resistance. Additional in vivo studies showed restricted tumor growth and inhibited expression of EMT-associated proteins after rabeprazole treatment.

Conclusions

Our data revealed (i) a clinical association of the EMT-like process with glioma malignancy and a poor survival and (ii) an anticancer and temozolomide sensitizing effect of rabeprazole by repressing EMT.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-021-00609-w.

Introduction

Glioblastoma multiforme (GBM) is a highly invasive form of glioma that covers 55.4 % of malignant gliomas and has a poor (5.5 %) five-year post diagnosis survival rate [1]. The reason for this low survival rate is its invasive behavior and resistance to currently available drugs, owing to its complex genetic and phenotypic heterogeneity [2–4].

Epithelial to mesenchymal transition (EMT) is a well-studied developmental regulatory process, characterized by loss of an epithelial phenotype and gain of a partial or complete mesenchymal phenotype [5]. The epithelial phenotype is characterized by low or no motility, and the expression of epithelial markers such as E-cadherin, Zo-1 and γ-cadherin, whereas the mesenchymal phenotype is characterized by high motility and the expression of vimentin, fibronectin and N-cadherin [5, 6]. The phenomenon of EMT in glioma has recently been recognized and, since glial cells acquire mesenchymal properties, it is also termed glial to mesenchymal transition (GMT) or EMT-like process [7–9]. EMT may be triggered by several molecular pathways such as the Wnt/β-catenin, phospho-inositide 3 ­kinase (PI3K), NF-κB and phospho-inositide 3 ­kinase-AKT-mTOR (mammalian target of rapamycin) (PI3K-AKT-mTOR) pathways [5, 10–12]. Activation of these signaling pathways, in turn, induces EMT effectors such as Snail, Slug, ZEB1/ZEB2, STAT3, β-catenin, hypoxia-inducible factor 1α (HIF1α) and TWIST [5, 13, 14]. In addition to molecular stimuli, other factors in the tumor microenvironment such as hypoxia and a low extracellular pH (pHe) may induce EMT and NF-κB signaling, thereby facilitating cancer progression, malignancy and drug resistance [4, 5, 15–18].

Proton pumps such as V-ATPase, Na⁺/H⁺ Exchanger and the carbonic anhydrase are upregulated in cancer cells to maintain a homeostatic intracellular pH (pHi) and to escape form intracellular acidity caused by bioenergetics switching. The increased activity of these pH regulators protects cells from intracellular acidity and apoptosis by extruding H⁺ in extracellular spaces [19–22]. Proton pump inhibitors (PPIs) such as esomeprazole, lansoprazole, pantoprazole and rabeprazole are acid activated prodrugs widely used to reduce stomach acidity and gastric ulcer with minimal adverse effects [23, 24]. PPIs are inactive in their naive form, but in an acidic microenvironment they are converted into active sulfenamides and act as active inhibitors [23, 25]. The selective interference of tumor acidity by PPIs induces apoptosis and/or inhibits EMT, suggesting that they may serve as promising anticancer drugs [26–28].

Rabeprazole is a second-generation PPI that effectively inactivates H⁺/K⁺ ATPase, thereby leaving gastrin levels unaffected, making it more advantageous than previous PPIs. It shows minimal or no adverse effects even when used at a high dose of 60 mg/day for a period of two years [23, 29–31]. The observed anti-proliferative efficacy of rabeprazole in gastric cancer cells suggests that it may also have anticancer effects against other malignancies, including glioma [32]. In addition, recent in vitro studies showed EMT inhibition by pantoprazole in gastric cancer cells [28, 33]. Here, we investigated the EMT-like process in human glioma through the analysis of EMT-associated proteins such as E-cadherin, N-cadherin, vimentin and β-catenin in clinical biopsies. The expression patterns of these proteins were correlated with various clinicopathological features and data listed in the cancer genome atlas (TCGA). We also assessed the anti-glioma and anti-EMT-like potential of rabeprazole using various in vitro assays in multiple glioma cell lines under physiological and acidic conditions. In vivo studies using a C6 rat glioma model were employed to corroborate the in vitro results. We also tested the temozolomide (TMZ) sensitizing potential of rabeprazole using a TMZ resistant cell line.

Materials and methods

Patient specimens

Surgically resected glioma (n = 44) and control (n = 7) tissues were collected from the Krishna Institute of Medical Sciences (KIMS), Secunderabad, India, and three normal brain tissues were procured from the National Institute of Mental Health and Neurosciences (NIMHANS), Bengaluru, India. The study group comprised 44 glioma tissues (16 grade II, 12 grade III and 16 grade IV) with different grades classified according to the WHO histopathological grading system [34] and seven nontumor temporal epilepsy and three normal brain tissues that were considered as controls. Written informed consent was obtained from all the study subjects. The Institutional Ethics Committee (IEC) of the University of Hyderabad and the KFRC-Ethics committee (KIMS) approved use of the human tissues for experimental purposes.

Cell culture

The rat glioma cell line C6 and human GBM cell lines U373, U87 and LN18 were procured from the NCCS-Cell repository, Pune, India. The GBM cell line T98 was a kind gift from Dr. Ellora Sen (National Brain Research Center, Gurgaon, India). The cell lines included in this study were below the 50-passage number, authenticated and free from any contamination. All cell lines were maintained in RPMI-1640 medium, supplemented with 10 % fetal bovine serum (FBS), 100 IU/ml penicillin and 100 µg/ml streptomycin at 37^oC in a 5 % CO₂ chamber. All experiments were performed in buffered medium with sodium bicarbonate and unbuffered medium without sodium bicarbonate to demonstrate rabeprazole efficacy at physiological and acidic pH conditions.

Rabeprazole extraction

Rabeprazole was extracted from aciphex tablets, which were pulverized into fine powder followed by dilution with water and extracted into dichloromethane. After three water washings, the combined organic layer was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure to get a beige colored solid fraction which was characterized by Mass, ¹H and ¹³C NMR as rabeprazole.

Reagents and media

RPMI-1640 medium was purchased from (Hi-Media), FBS (10,270), anti-anti (15240-062) and sodium bicarbonate were purchased (25,080) from Gibco, MTT (M2128), and Trypan blue (T-0776) and crystal violet were obtained from Sigma Aldrich. Cell lysis buffer (CST-9803) and all primary antibodies enlisted as β-actin (#4970), apoptosis antibodies (#9950), EMT antibodies (#9782), NF-κB antibodies (#9936), STAT3 (#4904), AKT (#9272), p-AKT (#9271), GSK-3β (#9315), p-GSK-3β (#9336) and secondary antibodies (#7074, #7076) were procured from Cell Signaling Technology (CST).

Cell proliferation/viability assay

Dose-dependent inhibition of cell proliferation by rabeprazole was evaluated using MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assays as described previously [35, 36]. Six-eight thousand cells were seeded in each well of 96-well plates and kept in the CO₂ chamber for overnight growth. Next, the medium was replaced with freshly prepared buffered and unbuffered medium followed by incubation in the CO₂ chamber for 4 h in order to acidify the medium [37]. Subsequently, the cells were treated with rabeprazole and vehicle (0.1 % DMSO) in a dose- and time-dependent manner. The formazan crystals were dissolved in DMSO after which absorbance was measured at 570 nm using a multi-plate reader (Tecan Infinite-200). The half-effective concentration (EC₅₀) was determined using nonlinear fit equation through GraphPad Prism. The cell death percentage was estimated using a trypan blue exclusion assay after rabeprazole (50 µM, 100 µM and 200 µM) and vehicle treatment. The cell death percentage upon 0.4 % trypan blue staining was calculated using the formula: number of dead cells (stained cells)/total number of cells (stained + unstained) x 100.

Scratch wound healing migration assay

C6 cells (1 × 10⁵) were seeded in each well of six-well plates and allowed to grow overnight in a CO₂ chamber. After overnight growth, a vertical and horizontal scratch (‘wound’) was made. Next, the wound areas were washed with phosphate buffer saline (PBS), after which the incubation setup and media replacement protocol followed was similar to that mentioned for the MTT assay in Section 2.5. After incubation, the cells were treated with three different doses (25 µM, 50 µM and 100 µM) of rabeprazole and vehicle (0.1 %) for 24 h. Snapshots were taken from the same position at different time intervals. The wound-healing percentage was calculated by the ratio of the healing area at each timepoint to the wound area at the zero hour timepoint. Wound-healing coverage was determined from two different fields and the results were expressed as mean ± SEM.

Transwell invasion assay

Cell invasion was assessed using gelatin-coated transwell inserts embedded with polycarbonate membranes (8-micron pore size). Upon rabeprazole treatment, the cells were seeded in the transwell inserts (top level) in serum-free media and kept in a CO₂ chamber for 12 h. FBS was used as a chemoattractant (bottom level). After 12 h incubation, the inserts were carefully removed and washed twice with PBS. Traversed cells were fixed in methanol and stained with 1 % crystal violet. To analyze the results, six individual fields of traversed cells of the membrane were captured and counted.

Clonogenicity assay

Clonogenicity assays were performed by seeding 1000 cells in each well of a six-wells plate and kept in a CO₂ chamber overnight. Next, the cells were treated with two doses (50 µM and 100 µM) of either rabeprazole or temozolamide (TMZ), or combination of both along with a vehicle control (0.1 %) for 24 h. After this period, the media including drugs were replaced with fresh media and the cells were allowed to grow for 7 days or until colonies were visible. The colonies were fixed with 4 % paraformaldehyde (PFA) and stained with 1 % crystal violet in 75 % methanol. Cloning efficiencies were determined by counting the total number of colonies/seeded cells x100, as previously described [35].

RNA isolation and RT-PCR

Total RNA from cell lines was isolated by Trizol reagent (Sigma-T9424) as per the manufacturer’s instructions. The RNA was quantified using NanoDrop, after which 1–5 µg was used for cDNA synthesis using a BluePrint 1st Strand cDNA Synthesis Kit. Next, RT-PCR was performed using green Taq polymerase and an Applied Biosystems-PCR System, after which the PCR products were visualized using 0.5-1 % EtBr stained agarose gels. The primer sequences used are listed in supplementary Table-S3.

Western blotting

Tissues and cells were lysed in lysis buffer containing a protease inhibitor cocktail as described earlier [35, 36]. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. After 1 h of blocking, the membranes were probed with primary antibodies at 4 °C overnight, followed by 1 h incubation with a horseradish peroxidase-conjugated secondary antibody at room temperature. Immunoreactivity was visualized using a Bio-Rad Chemidoc imaging system. Wherever needed, the blots were stripped using a previously described Gn-HCl striping solution and re-probed with different antibodies [38]. Densitometric analyses were carried out using Image J software (NIH). The intensities of the individual protein bands were quantified and normalized to β-actin.

Immunofluorescence

For immunofluorescence (IF), 2 × 10⁵ cells were seeded on glass coverslips in a 24-well plate and kept overnight in a CO₂ chamber. Next, the cells were treated with mentioned concentrations of rabeprazole for 24 h, after which the cells were fixed with 4 % paraformaldehyde and permeabilized using cold acetone. Next, a brief wash with PBS was given and the cells were incubated with a specific primary antibody and a FITC-conjugated secondary antibody for one hour at RT. Following these incubations, the cells were washed with PBS and counterstained using Vectashield mounting medium. Images were analyzed using confocal microscopy.

Stereotaxic glioma model

Male Wistar rats, 4–5 months old and a body weight of 180–220 g, were obtained from the National Institute of Nutrition (NIN), Hyderabad, India. Before stereotaxic implantation, all animals were housed for a minimum of 15 days in an animal house facility of the University of Hyderabad. Stereotaxic implantation was performed as reported before [39]. Briefly, C6 glioma cell suspensions (2 × 10⁵ cells in 10 µl) were injected into the striatum region of anesthetized (ketamine 80 mg/kg and xylazine 10 mg/kg body weight) rat brains through a drilled burr hole using a Hamilton syringe, after which the cavity was filled with dental cement. After 6 days of cell implantation, the rats were divided into two groups of six rats each. Group (a) received five doses of vehicle (5 % DMSO) and group (b) received five doses of rabeprazole (40 mg/kg of body weight) through the tail vein at intervals of 48 h. The animal experiments were performed in accordance with the Institutional Animal Ethics Committee guidelines (UH/IAEC/PPB/2017-I/P7) of the University of Hyderabad.

Histological analysis

For evaluation of the antitumor efficacy of rabeprazole, we extracted rat brains after perfusion [40] and fixed them in 4 % paraformaldehyde for at least 48 h. Next, 5–10 µ thick sections of brain tissue were prepared using a cryotome on silane-coated glass slides, followed by hematoxylin and eosin (H&E) staining. For immunohistochemistry (IHC) and IF, dewaxing, rehydration and staining of the tissue sections were carried out according to standard protocols [41]. The quantification of stained/unstained cells was performed by counting cells in at least six random fields of a tissue section. At least three brain sections were separately stained in duplicates for each study.

TCGA data analysis through UALCAN

To support our clinical study, we performed expression analyses of EMT genes using TCGA data through the ULCAN portal (http://ualcan.path.uab.edu/index.html). The UALCAN portal is an online resource used for expression analysis of various genes in relation to their predictive and survival significance [42]. Expression and survival graphs were downloaded and processed from the UALCAN portal.

Statistical analysis

All the data were analyzed using the statistical software packages SigmaPlot 11.0 and GraphPad Prism 5. The results are expressed as mean ± SEM and the p values calculated using One-way ANOVA for multiple groups or Student t-test for two groups. A p value ≤ 0.05 was considered statistically significant, (*) indicates p ≤ 0.05, (**) indicates p ≤ 0.02 and (***) indicates p ≤ 0.001. Survival data were analyzed using Kaplan-Meier survival curves and compared by Log-rank (Mantel-Cox) test. All experiments were performed in triplicate and repeated thrice, unless otherwise indicated.

Results

EMT in glioma is associated with malignancy and a poor prognosis

EMT in glioma was assessed by expression analysis of EMT-associated proteins using Western blotting and IHC. We found an increased expression of vimentin, β-catenin and GFAP and a negative or weak expression of E-cadherin and N-cadherin in GII, GIII and GIV gliomas compared to control tissues, indicating activation of the EMT-like process in glioma (Fig. 1a and S1a, b). The expression patterns of these EMT-associated proteins in GIV/GBM and low-grade glioma were consistent with those obtained from TCGA datasets (Fig. S2 and S3). IHC staining revealed significantly increased cytosolic vimentin (p = 0.001) and nuclear β-catenin (p = 0.01) levels with increasing glioma grades from GII to GIII and GIV compared to control tissues (Fig. 1b-e). Vimentin expression in tumor cells was confirmed by co-staining with Ki-67 (Fig. S4 b and c). Of the total cohort, 86.36 % tissues showed a negative expression of E-cadherin, while 88.36 %, 65.90 and 72.72 % of the tissues showed positive expression of GFAP, vimentin and β-catenin, respectively (Supplementary Table S1). Additional clinicopathological analyses revealed that variations in the expression levels of E-cadherin, GFAP, vimentin and β-catenin were independent of histological grade, age and sex (Table 1). The correlation studies showed a negative association of E-cadherin and positive associations of GFAP and β-catenin with vimentin expression (Fig. S5 a-c and Supplementary Table S2). Log-rank (Mantel-Cox) survival graphs showed a poor survival with negative E-cadherin (p = 0.7343) expression (Fig. 1f) and positive GFAP (p = 0.2831), vimentin (p = 0.0091) and β-catenin (p = 0.0175) expression (Fig. 1g-i). These clinical features were associated with a poor prognosis and a reduced median survival age, concordant with TCGA survival data (Table 1 and Fig. S6), suggesting that EMT targeting may improve GBM patient survival.

Fig. 1.

Increased EMT-associated protein levels in glioma grades and their association with poor survival. a Representative Western blot images depicting a weak/negative expression of N-cadherin and E-cadherin and increased expression of GFAP, vimentin and β-catenin in different grades of glioma and in cell lines compared to control tissues; β-actin was used as an internal control. b and d IHC showing increased expression of vimentin and β-catenin in GII, GIII and GIV gliomas compared to control tissues. c and e IHC quantification showing significant increases in vimentin (p = 0.001) and β-catenin (p = 0.01) expression in GIV gliomas compared to GII gliomas and control tissues. f-i Kaplan-Meier graphs showing EMT protein associations with glioma (astrocytoma) malignancy and poor prognosis. Log-rank (Mantel-Cox) test indicates E-cadherin negativity associated with poor patient survival (p = 0.010) (f). g-i GFAP (p = 0.001), vimentin (p = 0.012) and β-catenin (p = 0.011) expression associated with poor patient survival

Table 1.

The association of EMT proteins with various clinicopathological features of glioma patients

Features	E-cadherin expression profile (n = 44)		Fisher-exact test/Mantel-Cox survival test
	Positive expression	Negative expression
Gender
Male (n = 28)	4	24	p = 1
Female (n = 16)	2	14
Median survival age	39 months	35 months	p = 0.734, HR-1.2
Histopathological grades (WHO-2007, Classification system)
GII/GIII (n = 28)	4	24	p = 1
GIV (n = 16)	2	14
	GFAP expression profile (N = 44)
	Positive expression	Negative expression
Gender
Male (n = 28)	24	4	p = 0.63
Female (n = 16)	15	1
Median survival age	30 months	Not available	p = 0.283, HR-2.09
Histopathological grades (WHO-2007, Classification system)
GII/GIII (n = 28)	25	3	p = 0.1
GIV (n = 16)	14	2
	Vimentin expression profile (N = 44)
	Positive expression	Negative expression
Gender
Male (n = 28)	16	12	p = 0.18
Female (n = 16)	13	3
Median survival age	24 months	39 months	**p = 0.009, HR-4.03
Histopathological grades (WHO-2007, Classification system)
GII/GIII (n = 28)	17	11	p = 0.51
GIV (n = 16)	12	4
	β-catenin expression profile (N = 44)
	Positive expression	Negative expression
Gender
Male (n = 28)	18	10	p = 0.16
Female (n = 16)	14	2
Median survival age	24 months	39 months	*p = 0.017, HR-3.69
Histopathological grades (WHO-2007, Classification system)
GII/GIII (n = 28)	19	9	p = 0.48
GIV (n = 16)	13	3

p ≤ 0.05 is considered as significant; # HR, hazard ratio

Rabeprazole attenuates glioma cell growth and induces glioma cell death in vitro

To study the anti-cancer efficacy of rabeprazole on glioma cells in vitro, we first determined its effect on C6 glioma cells using a MTT-assay. C6 cells were treated with ten different concentrations of rabeprazole and vehicle (0.1 %) for 24 h in buffered and unbuffered medium. The EC₅₀ value observed was 94.53 µM rabeprazole in the buffered medium and 57.4 µM in unbuffered medium (Fig. 2a). Phase contrast microscopy revealed altered C6 cell size, shape and globularity after rabeprazole (100 µM) treatment in both media (Fig. 2b). In addition to C6, we used three GBM cell lines, T98, U87 and U373, to assess cell survivability and cell death upon rabeprazole treatment. Among all four cell lines, C6 exhibited 50 % cell survivability at 94 µM (Fig. S7 a) while T98, U87 and U373 showed 50 % cell survivability at 163.4 µM, 184 µM and 193.6 µM rabeprazole, respectively, in buffered medium upon 24 h treatment (Fig. S7 b-d). In the unbuffered medium, C6, T98, U87 and U373 showed 50 % cell survivability at 57 µM, 69 µM, 142.9 µM and 155 µM rabeprazole, respectively (Fig. S7 a and b). Cell death analysis using C6 and T98 cells revealed 50 % cell death at ~ 100 µM rabeprazole in the buffered medium, while more than 50 % cell death was observed at ~ 50 µM rabeprazole in the unbuffered medium after 24 h treatment (Fig. 2c and d). U373 and U87 exhibited 50 % cell death at 200 µM rabeprazole in the buffered medium while in the unbuffered medium more than 50 % cell death was observed at 100 µM rabeprazole after 24 h treatment (Fig. 2e and f). The EC₅₀ and cell death values were enhanced upon 48 h rabeprazole treatment, and even more so in unbuffered medium, indicating an increased efficacy of rabeprazole in unbuffered medium or acidic environment. Subsequent IF imaging depicted altered cell morphology and nuclear fragmentation upon exposure to 100 µM rabeprazole in both conditions (Fig. 2g). Western blot analysis revealed increased cleaved caspase-3 and cleaved PARP levels upon rabeprazole treatment compared to vehicle control (Fig. 2h), confirming apoptosis induction.

Fig. 2.

Rabeprazole induces apoptotic cell death. a Cell survival graph showing EC₅₀ at 94.53 µM of rabeprazole in buffered medium and at 57.4 µM in unbuffered medium. b Phase contrast image showing altered cell morphology in terms of cell size, shape and granularity upon rabeprazole treatment compared to vehicle control. c-f Rabeprazole treatment induces C6, T98, U373 and U87 cell death in both conditions with enhanced effects in the acidic condition. g IF images stained with GFAP and DAPI depicting altered cell morphology and nuclear fragmentation at 100 µM rabeprazole compared to control. h Western blots showing increased cleaved PARP and cleaved caspase-3 levels upon rabeprazole treatment in both media. β-actin was used as a loading control. At least three independent experiments were performed in triplicates; bars indicate SEM. (*) = p ≤ 0.05, (**) = p ≤ 0.02, (***) = p ≤ 0.001

Rabeprazole reduces cell migration by targeting EMT through AKT/GSK-3β/β-catenin signal inhibition

To determine the effect of rabeprazole on cell migration and invasion, we performed scratch wound healing and transwell assays, respectively. The wound healing assay revealed a significant reduction in wound closure (migration) of glioma cells at 50 µM rabeprazole in buffered medium and 25 µM in unbuffered medium compared to the vehicle control (Fig. 3a and b). The transwell assay showed a marked reduction in cell invasion at 50 µM rabeprazole in both conditions (Fig. 3c and d). These results prompted us to assess the effect of rabeprazole on EMT, since EMT-associated genes and/or proteins are pivotal for cell migration and invasion [5, 6]. We found that RT-PCR products on agarose gel (Fig. S8) and GAPDH normalized densitometry graphs (Fig. 3e) showed decreased expression levels of several EMT biomarkers and its modulators (GFAP, VIM, CTNNB1, AKT, VEGF, STA3, NFKB1 and MAPK1/3). Subsequent Western blot analysis (Fig. 3f, S9) confirmed attenuation of EMT-associated proteins such as N-cadherin, vimentin, β-catenin, STAT3 and snail upon rabeprazole exposure compared to the control. Rabeprazole treatment also inhibited AKT and GSK-3β phosphorylation in a dose-dependent manner in C6 and U87 cells in both conditions, indicating EMT suppression through the AKT/GSK-3β/β-catenin signaling pathway (Fig. 3f). Densitometric analysis of the Western blots revealed significant inhibition of p-GSK-3β, vimentin and β-catenin in C6 and U87 cells in both media at 100 µM and 200 µM rabeprazole (Fig. S9).

Fig. 3.

Rabeprazole reduces cell migration and invasion by targeting EMT through AKT/Gsk-3β/β-catenin inhibition. a Images depicting wound healing taken at two different time points (0 and 24 h) after scratch formation. b Graph showing significant inhibition of wound healing (migration) at a rabeprazole concentration of 50 µM in buffered medium and 25 µM in unbuffered medium. c Representative images of transwell migration assay showing reduced invasion upon rabeprazole treatment. d Graph indicating significant inhibition of cell migration at 50 µM rabeprazole in both conditions. e GAPDH normalized graphs showing signifcant inhibition of EMT biomarkers and associated transcripts at 100 and/or 200 µM rabeprazole in both cell lines compared to vehicle control. f Western blots upon rabeprazole exposure demonstrating significant inhibition of EMT biomarkers (vimentin, N-cadherin, snail) together with pAKT, pGsk-3β and β-catenin in C6 and U87 cells in both conditions. β-actin was used as an internal control. Three independent experiments were performed, bars indicate SEM. (*) = p ≤ 0.05, (**) = p ≤ 0.02, (***) = p ≤ 0.001

Rabeprazole reduces tumor growth and malignancy along with EMT inhibition

To substantiate our in vitro data, we performed an in vivo study using a C6 rat glioma model. Morphological analyses after five doses of rabeprazole (40 mg/kg of body weight) revealed a better physiology, cognition and brain structure compared to the vehicle (5 % DMSO) treated rats (Fig. 4a). Log-rank (Mantel-Cox) survival curves revealed a significant difference (p = 0.037) between vehicle and rabeprazole treated glioma rats, i.e., an increased median survival of 28.5 days for rabeprazole and 15 days for vehicle treated glioma rats (Fig. 4b). In fact, 3 out of 6 rabeprazole infused rats survived long-term (up to 30 days), whereas 3 out of 6 vehicle treated rats died before 15 days and only one rat survived up to 27 days of the defined survival period (30 days) (Fig. 4b). H&E and IHC analyses revealed restricted tumor growth in rabeprazole treated glioma rats compared to the vehicle treated glioma rats (Fig. 4c and d). The IHC images showed a significantly weaker staining of Ki-67 (p = 0.001) in the rabeprazole treated glioma rats compared to the vehicle treated glioma rats (Fig. 4d and f), indicating reduced tumor growth. STAT3, a master regulator of EMT and known to be linked to malignancy [14], also showed a weaker expression in the rabeprazole treated glioma rats compared to the vehicle treated glioma rats (Fig. 4e and g). IHC staining using anti-GFAP, anti-vimentin and anti-β-catenin antibodies showed a weak GFAP, nuclear β-catenin and vimentin expression in the rabeprazole treated glioma rats compared to the vehicle treated glioma rats (Fig. 4h-j), confirming EMT inhibition. Quantification assessment showed significantly lower expression levels of β-catenin (p = 0.027) and vimentin (p = 0.016) in the rabeprazole treated glioma rats compared to the vehicle treated glioma rats (Fig. 4k-m).

Fig. 4.

Rabeprazole restricts tumor growth and augments survival by targeting EMT in a rat glioma model. a Morphological pictures showing reduced tumor growth in rabeprazole treated glioma rats (n = 6) compared to vehicle treated glioma rats (n = 6). b Kaplan-Meier survival curves depicting significant difference (p = 0.037) between rabeprazole and vehicle treated groups with increased median age. c-e H&E and IHC staining showing a decrease in tumor cells in rabeprazole treated glioma rats compared to vehicle treated glioma rats. d-g IHC images and graphical quantifications revealing significant inhibition of Ki-67 (p = 0.001) and STAT3 (p = 0.057) expression upon rabeprazole treatment compared to vehicle control treatment. h-j IHC images stained with anti-GFAP, anti-β-catenin and anti-vimentin antibodies showing weak staining in rabeprazole treated glioma rats compared to vehicle treated glioma rats. k-m IHC quantified graphs revealing significant inhibition of GFAP (p = 0.010), β-catenin (p = 0.027) and vimentin (p = 0.016) expression upon rabeprazole treatment. At least six random fields were selected for the quantification of IHC stained tissue sections

Rabeprazole suppresses the TNF-α-induced EMT-like process by attenuating NF-κB signaling in vitro

To next assess the EMT suppressing potency of rabeprazole and its concomitant signaling, C6 and U87 glioma cells were exposed to TNF-α (20 ng/ml) before rabeprazole treatment to stimulate the EMT-like process. Subsequent Western blot analysis of the TNF-α treated cells revealed increased vimentin and β-catenin expression compared to the control, which was further inhibited by rabeprazole in both conditions (Fig. 5a). Quantitative assessment of the Western blots by densitometry (Fig. 5b) and IF imaging (Fig. 5c and d) underscored a significant inhibition of vimentin expression at 100 µM rabeprazole, even upon EMT induction by TNF-α. Previously, a role of NF-κB signaling in the TNF-α-induced EMT pathway has been reported [43]. We observed NF-κB-P65 activation and IKBα phosphorylation upon TNF-α (20 ng/ml) exposure and their inhibition by rabeprazole in both the cell lines (Fig. 5e). NF-κB-P65 activation was evident by increased p-IKBα and reduced total IKBα expression levels, respectively (Fig. 5e and f). IF images of C6 cells showed attenuated nuclear translocation of NF-κB-P65 upon rabeprazole treatment compared to the control (Fig. 5g). We also affirmed NF-κB gene suppression using NF-κB luciferase reporter assays, showing significant inhibition of NF-κB expression upon rabeprazole exposure in a dose-dependent manner (Fig. 5h).

Fig. 5.

Rabeprazole suppresses TNF-α-induced EMT by targeting NF-κB signaling in vitro. a Western blots of TNF-α (20 ng/ml) treated cells illustrating increased vimentin and β-catenin expression compared to the control, and suppression by rabeprazole in C6 and U87 cells in both conditions. b Densitometry graph showing significant inhibition of vimentin and β-catenin expression at 100 µM rabeprazole in both media. c and d IF images showing significant inhibition of vimentin at 100 µM rabeprazole upon EMT induction. e Western blots showing increased NF-κB-P50, NF-κB-P65 and p-IKBα protein levels upon TNF-α (20 ng/ml) exposure in both cell lines and further repression by rabeprazole. f NF-κB-P65 activation visualized by densitometry graph showing reverse expression of NF-κB-P65 and total IKBα. g IF images displaying attenuated nuclear translocation of NF-κB-P65 upon rabeprazole treatment. h NF-κB luciferase reporter assay showing significant inhibition of NF-κB expression upon rabeprazole exposure in a dose-dependent manner. Bars indicate SEM. (*) = p ≤ 0.05, (**) = p ≤ 0.02, (***) = p ≤ 0.001

TMZ sensitization by rabeprazole of TMZ resistant LN18 cells

Previous studies have reported a chemo-sensitizing ability of PPI [28, 44]. In order to evaluate the chemo-sensitizing efficiency of rabeprazole, we used a TMZ resistant LN18 GBM cell line [45]. To assess the cytotoxic effect of rabeprazole and TMZ on this cell line, we performed MTT and clonogenicity assays. The MTT results revealed a dose-dependent LN18 cell survival upon rabeprazole treatment. The EC₅₀ of rabeprazole was 138.4 µM and the EC₅₀ of TMZ was 240.6 µM (Fig. 6a and b). Combined treatment of rabeprazole and TMZ of LN18 cells revealed a synergistic effect. The combination index (CI), calculated according to the Chou-Talalay method [46], was < 1.0 in all combined treatments (Fig. 6c). A subsequent clonogenicity assay showed 1.2 and 11 % cloning reductions at 50 µM and 100 µM rabeprazole, respectively, compared to the vehicle control (Fig. 6d and e). We also observed 4.5 and 5.8 % cloning reductions at 50 µM and 100 µM TMZ, respectively, compared to the vehicle control. Combined treatment of 50 µM rabeprazole + 100 µM TMZ showed a 17.2 % cloning reduction and of 100 µM rabeprazole + 50 µM TMZ and 100 µM rabeprazole + 100 µM TMZ 24.9 and 27.9 % cloning reductions, respectively (Fig. 6d and e). A significant reduction in cloning ability was observed in the combined treatment, thus confirming TMZ sensitization. Using Western blotting, we also observed a significant inhibition of vimentin and β-catenin expression after treatment with 50 µM rabeprazole + 100 µM TMZ compared to the vehicle control and 50 µM rabeprazole or 100 µM TMZ in C6 and LN18 cells (Fig. 6f-h).

Fig. 6.

TMZ sensitization by rabeprazole in a TMZ resistant cell line (LN18). a and b Cell survival graphs showing EC₅₀ of rabeprazole (138.4 µM) and TMZ (240.6 µM) in LN18 cells. c Combined treatment of rabeprazole and TMZ reveals synergistic effect on LN18 cells, as indicated by CI < 1.0, in all combinations. d and e Clonogenicity images showing dose-dependent reductions after rabeprazole treatment. TMZ treatment at 50 µM and 100 µM show 4.5 and 5.8 % reductions in colony formation, respectively. Combined treatment of rabeprazole and TMZ show simultaneous 15.4 %, 17.2 %, 24.9 and 27.9 % reductions in cloning efficiency. f Representative Western blots and g and h densitometry quantifications showing significant inhibition of vimentin and β-catenin expression at 50 µM rabeprazole + 100 µM TMZ compared to vehicle and 50 µM rabeprazole or 100 µM TMZ in C6 and LN18 cells. Bars indicate SEM. (*) = p ≤ 0.05, (**) = p ≤ 0.02, (***) = p ≤ 0.001

Discussion

Development of drug resistance and tumor recurrence, specifically resistance to TMZ, is a major challenge in GBM treatment. EMT is pivotal for embryonic development and wound healing processes, and its occurrence in cancer is known to aggravate its invasion, migration and drug resistance [6, 47]. The recent recognition of an EMT-like process in glioma has opened new options for glioma research [7, 8].

In the current study, we investigated the clinical importance of the EMT-like process in human glioma by assessing the expression profiles of EMT-associated genes (through TCGA) and proteins using glioma biopsies followed by correlation analyses of altered expression patterns with various clinicopathological parameters. In addition, its therapeutic relevance was assessed through in vitro and in vivo studies. The expression analyses revealed a negative expression of E-cadherin, a differential weak expression of N-cadherin and an increased expression of GFAP, vimentin and β-catenin in GII, GIII and GIV gliomas compared to control tissues. The negative E-cadherin expression and increased GFAP, vimentin and β-catenin expression was associated with glioma malignancy and a dismal survival, indicating that the transition of glioma cells to a mesenchymal phenotype leads to enhanced invasiveness, which is concordant with an earlier report [48]. Interestingly, we observed a positive correlation of GFAP and β-catenin expression with vimentin expression and a negative correlation between N-cadherin and E-cadherin expression with vimentin expression. These observations support EMT as an ‘all or nothing’ event and underscore the independency of the E/N-cadherin switch or noncanonical EMT process in glioma, which is in agreement with previous reports [7, 48, 49].

Multiple reports support the concept that, in addition to molecular aberrations and hypoxia, proton pumps may facilitate epithelial to mesenchymal transition by modulating the pH of the tumor microenvironment [10–12, 20, 22]. In the past, various inhibitors have been used to target EMT, but due to their high toxicity, they could not be used as therapeutic agents [5]. The first observation of anticancer activities of a proton pump inhibitor (PPI) was reported using a gastric cancer model [50]. Since then several groups, including our group, demonstrated anti-neoplastic and chemo-sensitizing abilities of PPIs [26, 27, 33, 37, 51, 52]. Breedveld et al. [53] and Ferrari et al. [44] showed that pantoprazole inhibits BCRP and has a chemo-sensitizing activity, thereby contributing to an improved CNS delivery of imatinib, suggesting that PPIs may increase the efficacy of chemotherapy through a mechanism other than pH modulation. Recently, an anti-EMT efficacy and chemo-sensitizing ability of a PPI was observed in gastric cancer [28, 33], which triggered us to investigate the anticancer and anti-EMT potentials of rabeprazole in glioma. We found that exposure to rabeprazole altered glioma cell morphology and induced glioma cell death. Rabeprazole also induced nuclear fragmentation with concomitantly increased cleaved caspase-3 and cleaved PARP levels, indicating apoptotic cell death in agreements with previous reports [26, 54].

Emerging studies suggest that EMT or mesenchymal to epithelial transition (MET) do not embody transitions of purely mesenchymal or epithelial phenotypes, but rather represent dynamic processes involving several cellular states, termed partial EMT/MET, hybrid intermediate EMT/MET or EMT-like [5, 48, 55]. In the present study, we showed independency of E/N cadherin expression in glioma tissues, unlike classical EMT, in agreement with previous reports [7, 49, 55]. EMT in cancer cells has been reported to enhance cellular plasticity and resistance to apoptosis and drug treatment [56]. Marcucci et al. reported possible strategies for EMT targeting in cancer [5]. Concordantly, we found that rabeprazole administration induced cell death and reduced cell migration together with EMT by inhibiting Akt and Gsk-3β phosphorylation, which in turns suppressed the EMT inducers snail, slug and β-catenin, in line with earlier studies [28, 57]. Our in vivo results additionally revealed restricted tumor growth/malignancy and augmented glioma rat survival via inhibition of Ki-67, GFAP, vimentin and nuclear β-catenin levels.

We also found that rabeprazole treatment suppressed NF-κB signaling and affirmed EMT inhibition stimulated by TNF-α via attenuating NF-κB-P65 expression and nuclear translocation. These findings are concordant with previous reports, showing TNF-α-induced EMT through NF-κB activation [43]. Collectively, these data show anti-EMT efficacy of rabeprazole by targeting Wnt/β-catenin or NF-κB signaling. We also observed efficient inhibition of STAT3, a master regulator of the EMT pathway in vitro and in vivo, thereby identifying rabeprazole as a potent anticancer and anti-EMT drug.

TMZ resistance impedes GBM therapeutic interventions and hampers positive clinical outcomes [4, 45]. Recent data have shown that pre- or combined treatment of PPIs can sensitize chemoresistance [27, 37]. In parallel with the anti-EMT and anticancer efficacy of rabeprazole, we observed TMZ sensitization by rabeprazole in TMZ resistant cells. The synergistic effect of rabeprazole and TMZ was observed using cytotoxicity and clonogenicity assays. Simultaneously, we observed significant inhibition of the EMT-associated proteins vimentin and β-catenin after combined rabeprazole and TMZ treatment.

In conclusion, we observed an increased expression of EMT-associated proteins in different grades of glioma and their association with malignancy and clinical outcome. Additional in vitro and in vivo data showed that rabeprazole treatment restricted glioma growth and improved survival, and that rabeprazole treatment sensitized TMZ resistance through EMT suppression. Together, our data provide proof of concept that rabeprazole may serve as a safe and effective candidate for therapeutic intervention of malignant glioma.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

GBM

Glioblastoma multiforme

EMT

epithelial to mesenchymal transition

PPI

Proton pump inhibitor

NF-κB

Nuclear factor kappa light chain enhancer of activated B cells

AKT

Protein kinase B

GSK3β

Glycogen synthase kinase-3β

TMZ

Temozolomide

Authors contributions

DB and PPB made hypothesis. DB, AM and NY performed all cell-based experiments. DB performed the animal experiments. DB, AM, NY, ChYBVK, MP and PPB analyzed the data. Patient samples from the Department of Neurosurgery, Krishna Institute of Medical Sciences (KIMS) were collected and analyzed by DB, ChYBVK and MP. DB and AM wrote the original draft of the manuscript. DB, AM, NY, ChYBVK, MP and PPB critically reviewed the manuscript and approved its submission.

Funding

The authors acknowledge the financial support of the Department of Science and Technology (DST- India), CSRI (Grant No. SR/CSRI/196/2016), Science & Engineering Research Board (SERB)(Grant No. CRG/2020/005021, and the Department of Biotechnology (DBT-India) (Grant No. BT/PR18168/MED/29/1064/2016. The authors also thank DST- FIST and UGC-SAP of the DoBB. DB thanks the Department of Biotechnology-India for a student fellowship (Award no: DBT/2013/UOH/79). AM acknowledges funding from DST-Women scientist- A (Grant No. SR/WOS-A/CS-31/2019(G)

Data availability 

All data are available in the paper and/or its supplementary files.

Declarations

Conflict of interest

The authors declare no conflict of interest. 

Study approval

The Institutional Ethics Committee (IEC) (reference number UH/IEC/2016/180) and Institutional Animal Ethics Committee (IAEC) (reference number UH/IAEC/PPB/2017-I/P7), University of Hyderabad, Hyderabad (500 046) approved all procedures involving human tissue and animal-related experiments.

Informed consent

Written informed consent was obtained from all participants included in the study.