Inhibition of protein glycosylation reverses the MDR phenotype of cancer cell lines

Biomedicine & Pharmacotherapy 74 (2015) 49–56

Available online at

ScienceDirect www.sciencedirect.com

Original Article

Inhibition of protein glycosylation reverses the MDR phenotype of cancer cell lines Karolina Wojtowicz a,*, Radosław Januchowski a, Michał Nowicki a, Maciej Zabel a,b a b

Department of Histology and Embryology, Poznan University of Medical Sciences, 60-781 Poznan, Poland Department of Histology and Embryology, Wroclaw Medical University, 50-368 Wroclaw, Poland

A R T I C L E I N F O

A B S T R A C T

Article history: Received 12 June 2015 Accepted 9 July 2015

Background: Multidrug resistance proteins are one of the most important factors that cause chemotherapy resistance, which in turn reduces therapeutic efficacy and survival for cancer patients. Tunicamycin is one of the most well-known inhibitors of N-glycosylation and is considered a powerful adjunct that can increase the effectiveness of many drugs. Tunicamycin blocks the first step of P-gp (glycoprotein P) and BCRP (breast cancer resistance protein) N-glycosylation, which is a very important modification for the activity and cellular localisation of these proteins. Methods: The effects of tunicamycin on ovarian and colorectal cancer cells were examined in multiple cell lines. The primary ovarian cancer cell line W1 and the established ovarian cancer cell line A2780 were compared against their drug-resistant derivatives W1TR/W1PR (TR: topotecan resistant; PR: paclitaxel resistant) and A2780T1 (topotecan resistant), respectively. We also compared the colorectal cancer cell line LoVo against its doxorubicin-resistant derivative LoVo/Dx. Cell viability was determined by the MTT assay. The glycopeptides were subjected to deglycosylation using the endoglycosidase PNGase F. A2780T1, LoVo/Dx and W1PR cells were treated with the protein degradation inhibitors MG132 and BMA. Protein expression was detected by western blot and immunocytochemistry. Results: In this study, we showed via the MTT assay that tunicamycin significantly decreased the viability of cancer cell lines that were co-treated with a chemotherapeutic drug. Western blot analysis showed that, in LoVo/Dx and W1PR cells, tunicamycin treatment resulted in the expression of a 70 kDa P-gp protein instead of the mature 170 kDa P-gp. Treatment with MG132 or BMA fully suppressed the effect of tunicamycin in the case of W1PR cells only. In tunicamycin-treated W1TR cells, the size of the BCRP protein did not differ from that of its native unglycosylated form. In tunicamycin-treated A2780T1 cells, BCRP expression was completely inhibited, but pre-treatment with MG132 or BMA suppressed the effect of tunicamycin. Immunocytochemistry analysis indicated that tunicamycin only affected the translocation of P-gp but not that of BCRP. After treatment, we observed higher P-gp expression in the cytoplasm than at the cell membrane. Conclusions: Our results indicated that tunicamycin may enhance the effect of chemotherapy by interfering with the localisation and function of transporter proteins that are responsible for multidrug resistance. ß 2015 Elsevier Masson SAS. All rights reserved.

Keywords: Multidrug resistance Tunicamycin Chemotherapy Glycoprotein P BCRP

1. Introduction Abbreviations: ABC, ATP binding cassette superfamily; BCRP/ABCG2, breast cancer resistance protein; BSA, bovine serum albumin; FBS, foetal bovine serum; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; MDR, multidrug resistance; MEME, minimum essential medium eagle; PBS, phosphate buffered saline; P-gp, glycoprotein P; PNGase F, N-glycosidase F; TBS, tris buffered saline; Tun, tunicamycin. * Corresponding author. E-mail addresses: [email protected] (K. Wojtowicz), [email protected] (R. Januchowski), [email protected] (M. Nowicki), [email protected] (M. Zabel). http://dx.doi.org/10.1016/j.biopha.2015.07.001 0753-3322/ß 2015 Elsevier Masson SAS. All rights reserved.

Cancer is one of the major health crises and causes of death in industrialised countries. Only 30% of adult cancer patients can be completely cured [1]. Therefore, we must improve current therapeutic options to reduce cancer mortality rate. At present, the most effective cancer treatment methods are surgery and chemotherapy. Surgery often allows for complete recovery without further complications or recurrences for early stage

50

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

diseases. By contrast, chemotherapy is more effective against latestage, disseminated, or non-operable diseases. Thus, increasing the efficacy of chemotherapy is one of the most sought-after goals in cancer research. A major obstacle to chemotherapy efficacy is the development of multidrug resistance (MDR). MDR is defined as cross-resistance to a broad spectrum of structurally unrelated cytotoxic drugs. The causes of MDR are diverse, but they are mostly connected with the overexpression of drug transporters of the ABC family, which can remove drugs from the cells [2,3]. The most well-studied ABC protein is glycoprotein P (P-gp), which is encoded by the MDR1 (ABCB1) gene. It was the first protein that was identified to be connected with multidrug resistance. Today we know that many more proteins are associated with MDR, one of which being breast cancer resistance protein (BCRP), which is encoded by the ABCG2 gene. The MDR proteins are mainly localised to the cell membrane, but P-gp and BCRP have also been detected at the nuclear envelope [4,5]. Therefore, it is likely that there exist additional mechanisms of resistance specifically associated with the cell nucleus. P-gp comprises two homologous and symmetrical domains, each containing six transmembrane (TM) segments, where the Nterminal is located at the cytoplasmic side of cell membrane [3,4]. P-gp is N-glycosylated at amino acid (aa) positions 91, 94 and 99 in the first extracellular loop [4,6]. The molecular weight of unglycosylated P-gp is 140 kDa based on the aa structure, and glycosylation increases the molecular weight to 170 kDa [7]. BCRP (ABCG2) belongs to the G family of ABC transporters. It is considered a ‘‘half-transporter’’, and it is composed of six transmembrane domains. Homo-dimerisation is necessary to create a functional BCRP efflux pump. BCRP is glycosylated at Asn596 in the third extracellular loop [8–10]. The molecular weight of unglycosylated BCRP is 60 kDa and it is increased to 70 kDa after glycosylation [11]. Glycosylation is one of the major post-translational protein modifications. Sugar residues are covalently attached to BCRP and P-gp via N-type binding (N-linked glycosylation) to asparagine residues. The entire process of sugar synthesis and protein binding occurs at the endoplasmic reticulum (ER). The first stage of Nglycosylation involves the synthesis of a sugar core on ER membrane-anchored dolichol phosphate at the cytoplasmic side of ER. When a premature sugar core is formed, the entire glycan chain is translocated to the luminal side of ER. An activated sugar core is then created and translocated to an asparagine residue on the target protein. The protein with the sugar core attached is translocated to the Golgi apparatus, where terminal glycosylation occurs. The mature protein is then directed to its final destination [6,12]. For many proteins, including BCRP and P-gp [4,13], glycosylation plays a very important role in protein folding and intracellular translocation. The significance of glycosylation can perhaps be reflected from the functions of glycoproteins in a cell [6,14,15]. For proteins such as BCRP and P-gp, one of the questions of interest is whether the accompanying sugar residues have any effect on the cellular localisation and/or activity of these proteins. In addition, among the past studies on the importance of glycosylation for ABC proteins and the potential application of glycosylation inhibitors as chemotherapy adjuncts, most of which focused on P-gp, and tunicamycin (tun) was the most often used glycosylation inhibitor [16,17]. Recently, several studies focused directly on the effect of glycosylation inhibitors on BCRP activity. The BCRP transporter has not been extensively investigated in the past. Therefore, valuable information may be gained from an in-depth look at BCRP activity [8,9]. Furthermore, the specific role of N-glycosylation inhibitors in improving chemotherapy response has not been clearly elucidated. This paper discusses the effect of tun on the cellular localisation of P-gp and BCRP in human ovarian and colorectal cancer cell lines.

We also assessed the effects of glycosylation inhibition on the protein expression of P-gp and BCRP and the sensitivity of the cell lines to cytostatic drugs. 2. Materials and methods 2.1. Reagents and antibodies Doxorubicin, topotecan, paclitaxel, tunicamycin, RIPA Lysis Buffer, BMA, and MG132 were obtained from Sigma (St. Louis, MO). RPMI-1640 and MEME medium, foetal bovine serum (FBS), antibiotic-antimycotic solution, and L-glutamine were also purchased from Sigma (St. Louis, MO). PNGase F was obtained from New England Biolabs (Hitchin, UK). Bradford Dye Reagent was obtained from Bio-Rad Laboratories (Hemel Hempstead, UK). Nitrocellulose membrane was obtained from GE Healthcare (Buckinghamshire, UK). Cell Proliferation Kit I (MTT) and protease inhibitor cocktail were purchased from Roche Diagnostics GmbH (Mannheim, Germany). Rabbit anti-ABCG2 polyclonal Ab (H-70), rabbit anti-GADPH polyclonal Ab (FL-335), goat anti-mouse HRPconjugated Ab, and goat anti-rabbit HRP-conjugated Ab were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse monoclonal anti-P-glycoprotein Ab (C219) was obtained from Alexis Biochemicals (Lo¨rrach, Germany). The MFP488 fluorescent secondary antibodies were obtained from MoBiTec. Mounting medium with DAPI was obtained from Santa Cruz Biotechnology. 2.2. Cell lines and cell culture Two ovarian cancer cell lines were used for this study: (1) W1, which is a primary ovarian cancer cell line, and (2) A2780, which is a commercially available established human ovarian carcinoma cell line. The W1 cell line was established from ovarian cancer tissues that were obtained from an untreated patient in December 2009. From the W1 cell line, we generated the W1TR subline, which is resistant to topotecan (W1 topotecan resistant), and the W1PR subline, which is resistant to paclitaxel (W1 paclitaxel resistant). These cell lines were derived in our laboratory as described previously by Januchowski et al. [18]. From the A2780 cell line (obtained from ATCC, Poland), we generated the topotecan-resistant subline A2780T1. The drug-resistant sublines were generated by exposure of the drug-sensitive cells to incrementally higher concentrations of each drug. The final concentration of top was 24 ng/ml, and that of pac was 1100 ng/ ml. We also used the human colon adenocarcinoma cell line LoVo and its commercially available, doxorubicin-resistant subline LoVo/Dx (obtained from ATCC, Poland), which was cultured in the presence of 200 ng/ml dox to retain its resistant phenotype. The W1TR and A2780TR1 cells were shown to overexpress BCRP, while the W1PR and LoVo/Dx cells were shown to overexpress Pgp [5,18]. The LoVo, LoVo/Dx, A2780, and A2780T1 cell lines were cultured in MEME medium, and the W1, W1PR, and W1TR cell lines were cultured in RPMI-1640 medium. Media were supplemented with 10% FBS, 2 mM L-glutamine and 1% antibioticantimycotic solution. Cells were cultured at 37 8C in a humidified atmosphere of 5% CO2 (v/v). 2.3. MTT assay The MTT survival assay was performed to estimate the extent of cell resistance to chemotherapeutic agents and tunicamycin. Briefly, cells were seeded at 4  103 cells per well into 96-well culture plates and pre-incubated for 48 h. To examine the effect of tun or chemotherapeutic drugs on cell survival, the cells were treated with increasing concentrations of tun or different drugs for

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

72 h. Afterwards, 10 ml of the MTT labelling reagent was added to the medium (the final concentration of MTT was 0.5 mg/ml) for 4 h and 100 ml of the solubilisation solution was then added to each well. After overnight incubation, absorbance was measured in a microplate reader at 570 nm with a reference wavelength of 720 nm, according to the manufacturer’s protocol for the MTT assay. We also examined the effect of tun on cellular resistance to chemotherapeutic drugs. The cells were seeded at 4  103 cells per well into 96-well culture plates. After 48 h, the cells were pretreated with tun for another 48 h, and the medium was then replaced with fresh medium supplemented with a cytostatic drug and tun. Tun concentration was set at a previously established value for each cell line (600 ng/ml for LoVo/Dx, 300 ng/ml for W1PR, 300 ng/ml for W1TR and 400 ng/ml for A2780T1). After 72 h of incubation, cell proliferation was assessed as before. As a control, we used cells that were treated with only the cytostatic drugs for 72 h. There were three experimental replicates for this assay. 2.4. Preparation of cell lysates and glycosidase treatment For immunoblot analysis of protein expression, the cells were seeded into 75 cm2 culture flasks at 1  106 cells per flask and pre-incubated for 48 h. The medium was then replaced by fresh medium only or fresh medium supplemented with tun at different concentrations. After 72 h of incubation, the cells were harvested by trypsinisation and pelleted by centrifugation. The cell pellets were washed once with phosphate buffered saline (PBS). The cell pellets were treated with the RIPA lysis buffer supplemented with the protease inhibitor cocktail. The protein concentration was determined using Bradford Dye Reagent with BSA as the standard. Glycosidase treatment was performed by incubating 35 mg of cell lysate with 5 ml of PNGase F at 37 8C for 10 min. 2.5. Immunoblot analysis Western blot was conducted under reducing conditions. Cell lysate samples (35 mg) were first treated with a denaturing buffer containing 200 mM Tris–HCl (pH 6.8), 5% SDS, 10% glycerol, 0.25% 2-mercaptoethanol, and 0.1% bromophenol blue. The lysates were electrophoretically separated on a 7% polyacrylamide gel according to the Lammeli method and then electroblotted onto a nitrocellulose membrane. The membrane was blocked with 5% (w/ v) dry milk in TBS at RT for 1 h. The membrane was then incubated with the anti-P-gp or anti-BCRP primary antibody (1:500 dilution) at 4 8C overnight. The membrane was also probed with the antiGAPDH antibody (1:1000 dilution) as a loading control. For the secondary antibody, we used horseradish peroxidase (HRP)-linked anti-species antibody at a dilution of 1:2000. HRP-dependent luminescence was developed with Femto Super Signal Reagent and detected using a UVP Imaging System.

51

anti-rabbit antibody. Fluorescent images were collected on an Olympus confocal microscope. The expression of P-gp and BCRP was analysed by pseudo-colour representations of fluorescence intensity. DAPI fluorescence was excited at 365 nm and detected at 420 nm (blue). 2.7. Statistical analysis Statistical analysis was performed using Microsoft Excel software. The statistical significance of differences was determined by Student’s t-test. p-Values of 0.05 or less were considered statistically significant. 3. Results 3.1. Effect of tunicamycin on cell viability To examine the impact of tun on cell viability, the cells were treated with tun at increasing concentrations for 72 h. We observed that tun induced a dose-dependent decrease in cell viability for all drug-resistant cell lines (Fig. 1). For the subsequent experiments, we selected a tun concentration that would only decrease cell viability to 85–95% in each cell line: 600 ng/ml for LoVo/Dx, 300 ng/ml for W1PR, 300 ng/ml for W1TR and 400 ng/ml for A2780T1, because we established that in those concentrations tun is not toxic to the cells. We did not observe any effect on the sensitive cells after tun treatment (data not shown). 3.2. Tunicamycin increases the cytotoxic effect of anticancer drugs In the next set of experiments we tested whether tun and the cytostatic drugs had a synergistic effect on the resistant cell lines. Cells were treated with tun at the indicated concentrations for 48 h. Afterwards, the medium was replaced by fresh medium supplemented with both tun and one of the cytostatic drugs (dox 200 ng/ml, top 24 ng/ml, or pac 1100 ng/ml). The cells were incubated for another 72 h. We also treated cells with the cytostatic drugs alone for 72 h. As a control, we used cells without cytostatic drugs. In all the cell lines that we investigated, we observed a statistically significant, synergistic effect of tun and each cytostatic drug (Fig. 2). Pre-treatment with tun resulted in significantly

2.6. Immunofluorescence analysis Cells were fixed and permeabilised in ice-cold acetone/ methanol (1:1) for 10 min at 20 8C and rinsed with PBS. To detect P-gp, the cells were treated with a mouse monoclonal antibody against P-gp at 1:25 dilution for 1 h at RT. The cells were then rinsed five times with PBS and incubated with an MFP488labelled anti-mouse antibody for 1 h at RT in the dark. To visualise the cell nuclei, the cells were mounted with a DAPI-containing mounting medium. To detect BCRP, cells were treated with a rabbit polyclonal anti-BCRP antibody at 1:100 dilution for 1 h at RT. Afterwards, the cells were incubated with an MFP488-labelled

Fig. 1. Tunicamycin cytotoxicity, MTT cell survival assay. LoVo/Dx, W1PR, W1TR, and A2780T1 cells were seeded at a density of 4000 cells/well in 96-well plates and treated with tun for 72 h. Cell viability was then determined. The experiments were repeated three times, and each tun concentration was tested in triplicate for every experiment. Cell viability for each of the cell lines was expressed as a percentage of untreated control. Data are expressed as mean values  SD from three independent experiments, p < 0.05.

52

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

3.4. Effect of MG132 and bafilomycin A1 (BMA) on the expression of BCRP in A2780TR1 cells and P-gp in LoVo/Dx and W1PR

Fig. 2. Cytotoxicity of tunicamycin and cytostatic drugs, MTT cell survival assay. LoVo/Dx, W1PR, W1TR, and A2780T1 cells were seeded at a density of 4000 cells/ well in 96-well plates. The cells were pre-treated with tun for 48 h and then with a combination of tun and dox (LoVo/Dx)/top (W1TR and A2780T1)/pac (W1PR) for 72 h. The control cells were treated with the cytostatic drugs alone for 72 h. Cell viability was determined in three replicate experiments, and each sample was tested in triplicate in every experiment. Cell viability for each cell line was expressed as a percentage of untreated control. Data are expressed as mean values  SD from three experiments. * means statistically significant, p < 0.05.

reduced cell viability compared with cells that were treated with the drugs alone. The largest decrease in viability was observed in the W1PR cells that were co-treated with pac (1100 ng/ml) and tun (300 ng/ml), where a greater-than-3-fold decrease in cell viability (p < 0.05) was detected relative to the drug-alone control. The viability of the LoVo/Dx cells decreased by approximately 2-fold (p < 0.05) after co-treatment with dox and tun. For the W1TR and A2780T1 cells, viability decreased by 1.5 and 1.3-fold (p < 0.05), respectively, after co-treatment with top and tun. 3.3. Effect of tunicamycin on the expression of BCRP and P-gp We performed western blot analysis to assess whether tun could affect the protein expression of BCRP and P-gp. After treating the cells with tun for 72 h, western blot analysis showed P-gp expression at 70 kDa instead of the characteristic 170 kDa in the LoVo/Dx and W1PR cells (Fig. 3A). Expression of P-gp in the drugsensitive cell lines LoVo and W1 was not observed. The expression of BCRP was not detectable in the A2780TR1 cells after tun treatment. However, a single 60 kDa band was detected by western blot analysis of the tun-treated W1TR cells, which corresponded to the size of unglycosylated BCRP (Fig. 2B). BCRP expression in the drug-sensitive W1 and A2780 cells was not observed. To determine whether the removal of sugar residues would affect the molecular weight of P-gp and BCRP, lysates from LoVo/ Dx, W1PR, W1TR and A2780TR1 cells were incubated with Nglycosidase F (PNGase F) to remove N-linked glycans. After incubation with PNGase F, additional bands at lower molecular weight were detected for both proteins (Fig. 3C). BCRP was detected at both 72 kDa and 60 kDa, P-gp was observed at both 170 kDa and 145 kDa in the LoVo/Dx, while P-gp in the W1PR was a single 145 kDa band. The lower bands were consistent with the sizes of the unglycosylated proteins.

Due to the observation of BCRP degradation in tun-treated A2780TR1 cells, we decided to further investigate the underlying mechanism for this phenomenon. Protein degradation occurs at two major sites, which are the lysosome and the proteasome. To examine whether the lysosomal pathway or the ubiquitinproteasomal pathway was involved in BCRP degradation, we pre-cultured the A2780TR1 cells in the presence or absence of 2 mM MG132 (a proteasomal degradation inhibitor) or 10 nM BMA (a lysosomal degradation inhibitor) for 3 h and subsequently the medium was replaced by fresh medium supplemented with both tun and one of the cytostatic drugs for another 48 h. We verified that the concentrations of the protein degradation inhibitors were nontoxic to A2780TR1 cells. We found that the pre-treatment with MG132 or BMA suppressed the effect of tun and mature BCRP was detected (Fig. 3D). Moreover, BCRP protein expression was significantly higher in the cells that were co-treated with tun and MG132 or BMA than in the cells that were treated with BMA or MG132 or BMA only (Fig. 3D). In similar way we wanted to verify the effect of tun-treated LoVo/Dx and W1PR cells on P-gp expression. To examine whether the tun was involved in P-gp degradation or there was other mechanism, we pre-cultured both cell lines in the presence or absence of 2 mM MG132 or 10 nM BMA for 3 h and then the medium was replaced by fresh medium supplemented with both tun and one of the cytostatic drugs for another 48 h. We showed the 170 kDa P-gp protein in the W1PR cells after pre-treatment with MG132 or BMA (Fig. 3E). In the LoVo/Dx cells, P-gp was observed at several heights, at 170 kDa, 145 kDa and 70 kDa (Fig. 3E). 3.5. Effect of tunicamycin on the cellular localisation of P-gp and BCRP We were interested in whether tun affected the cellular localisation of P-gp in the W1PR and LoVo/Dx cells and the localisation of BCRP in the W1TR and A2780TR1 cells. In the LoVo/ Dx cells, P-gp was mostly detected at the plasma membrane and nuclear envelope without tun treatment. After tun treatment, P-gp was mainly seen within the intracellular compartments (Fig. 4A). Similar observations were made in the W1PR cells, where P-gp was mainly present at the plasma membrane in untreated cells and was translocated to the intracellular compartments after tun treatment (Fig. 4B). In the untreated W1TR cells, BCRP was detected at the plasma membrane, and tun treatment did not affect its localisation (Fig. 4C). The same observation was made in the A2780T1 cells (data not shown).

4. Discussion In eukaryotes, N-linked glycosylation is known to play a very important role in protein stability, localisation and biological activities [19,20]. The goal of our study was to determine if glycosylation affects the expression and localisation of two MDR proteins P-gp and BCRP, and if the drug-resistant phenotype of cancer cells can be reversed by blocking glycosylation. We used an N-glycosylation inhibitor tunicamycin (tun) for this study. Tun is a nucleoside antibiotic isolated from Streptomyces sp. It acts by blocking the enzymatic transfer of 1-phospho-N-acetylglucosamine from UDP-N-acetylglucosamine to dolichol phosphate at the first stage of glycosylation [21–24]. It has been shown that tun can decrease cell viability and cause endoplasmic reticulum stress [25]. In this study, we performed a detailed analysis on the importance of sugar residues for P-gp and BCRP.

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

53

Fig. 3. Effect of tunicamycin, PNGase F and protein degradation inhibitors on the protein expression of P-gp and BCRP in LoVo, LoVo/Dx, W1, W1PR, W1TR, A2780, and A2780T1 cells. Cellular proteins (35 mg) were separated on a 7% gel and transferred to a PVDF membrane, and the membrane was immunoblotted with a primary Ab and then an HRP-conjugated secondary Ab. (A) Tunicamycin (tun) caused degradation of P-gp, as demonstrated by the 70 kDa band instead of the 170 kDa band. (B) BCRP was detected at a lower molecular weight in tun-treated W1TR cells, but it was completely absent in tun-treated A2780T1 cells. (C) PNGase F treatment resulted in additional lower bands on the western blot for BCRP (60 kDa) and for P-gp (145 kDa). (D) Combination treatment with tun and either BMA (inhibitor of lysosomal degradation) or MG132 (inhibitor of proteasomal degradation) led to increased expression of BCRP in A2780T1 cells. (E) Co-treatment with tun and either BMA or MG132 led to expression of mature 170 kDa form of P-gp. GAPDH was used as a loading control for the western blots. The data shown are representative of three independent measurements.

We first showed that tun decreased cell viability in a dosedependent manner. We then selected tun concentrations that would reduce cell viability by no more than 15% for the subsequent experiments. The MTT assay showed that tun increased the efficacy of cytostatic drugs in all the cell lines that were tested. This finding indicated that tun may enhance the effect of chemotherapy by interfering with drug transporters. Our finding also agrees with the data from other reports. Sˇeresˇ et al. showed that 0.01–10 mmol/ L tun induced a dose-dependent decrease in the proliferation of the vincristine-resistant L1210 cells [14]. The anti-proliferative effect of tun on L1210 cells may be associated with a decrease in the

incorporation of sugar moieties into glycoproteins and may be related to the accumulation of one or more nucleotide sugar precursors for asparagine-linked glycoprotein biosynthesis in the endoplasmic reticulum [26]. Hiss et al. also reported that the combination treatment of tun and anticancer drugs enhanced the drug toxicity in multidrug resistant human ovarian cystadenocarcinoma cells [24]. Similarly, Noda et al. showed that tun enhanced cisplatin sensitivity in human head-and-neck carcinoma cells [23]. Kramer et al. showed that tun induced a time-dependent increase in daunorubicin accumulation in human colorectal carcinoma cells, reflecting a significant reduction in P-gp function

54

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

Fig. 4. Immunofluorescent visualisation of P-gp and BCRP expression in LoVo/Dx, W1PR, and W1TR cells in the absence or presence of tunicamycin (tun). (A and B) Pgp was detected using the anti-P-gp antibody and a MFP488-conjugated secondary antibody (green). To visualise the cell nuclei, the cells were mounted with a DAPIcontaining mounting medium (blue) – right column. Tun affected the cellular localisation of P-gp in LoVo/Dx and W1PR cells. P-gp was observed at the plasma membrane (red arrow) and the nucleus (blue arrow) in the untreated cells but mainly in the cytoplasm (white arrow) in tun-treated cells. (C) BCRP was detected using the anti-BCRP antibody and a MFP488-conjugated secondary antibody (green). Tun did not alter BCRP localisation in W1PR (shown here) and A2780T1 cells (data not shown). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

[27]. Madureira et al. also showed that tun induced a significant decrease in colon cancer cell proliferation after 18 and 24 h of incubation [28]. However, Schinkel et al. found that N-glycosylation may not affect P-gp-mediated drug transport. They mutated one, two or all three glycosylation sites on human P-gp and showed that the absence of N-glycosylation did not have an impact on the extent of drug resistance. They further demonstrated that Nglycosylation was not strictly essential for P-gp function but it drastically reduced the efficiency of generating drug-resistant clonal cells [29]. Furthermore, we observed no effect on viability of sensitive cells. Taken together, these data from the literature and our study indicate that tun can sensitise MDR cancer cells to cytostatic drugs. We propose that higher levels of UDP-sugars inside cells may correspond to increased drug sensitivity, because the presence of free sugar molecules imply that transporter proteins such as P-gp and BCRP are unglycosylated thus inactive. Therefore, tun treatment likely results in decreased cell viability by inactivating P-gp and BCRP. We then assessed the stability of P-gp and BCRP by western blot to determine if decreased protein stability was responsible for increased drug sensitivity under tun treatment. Tun caused the degradation of P-gp in the LoVo/Dx and W1PR cells. We observed a 70 kDa band for P-gp, which corresponded to the mass of one half of the unglycosylated protein. To our knowledge, this is the first report on tun-induced degradation of P-gp down to one homologous domain. Kramer’s team has previously shown that a 6 h exposure of human colorectal carcinoma cells to tun was sufficient to completely block the glycosylation of newly synthesised P-gp, resulting in an unglycosylated 140 kDa precursor [27,30]. Our results suggest that tun impairs the formation of the two halves of P-gp. Knowing that P-gp is glycosylated only within a single transmembrane domain at Asn91, Asn94 and Asn99 [4,6], we concluded that tun most likely blocks the synthesis of just this particular half of P-gp. The inability to form both homologous domains may also be due to a dysfunction of the flexible linker that connects the two halves, which can prevent P-gp from functioning properly as an MDR protein. These observations suggest that glycosylation of P-gp is important in establishing a competent MDR phenotype in cells. Another explanation of our results is that immature P-gp is rapidly ubiquitinated and degraded just like any other immature proteins. For example, it has been shown that treatment of HepG2 cells with tun decreased the net production of apolipoproteins by enhancing the co-translational degradation of the proteins. This effect of tun was partially preventable by a proteasome inhibitor [31]. These results, combined with our analysis of P-gp expression, suggest that protein degradation might be initiated when cells are treated with tun. To prove our hypothesis we used two inhibitors of protein degradation: BMA, which inhibits lysosomal degradation, and MG132, which inhibits proteasomal degradation [19]. Co-treatment of W1PR cells with tun and either one of the protein degradation inhibitors resulted in 170 kDa P-gp expression. These data suggested that both the lysosomal and proteasomal degradation pathways were involved in P-gp degradation in W1PR cells. When we made the similar experiment in LoVo/Dx cells, we showed a few P-gp bands, but still 70 kDa band. This may indicate that we used too low concentration of inhibitors or time of pre-incubation with them was too short in this cause. Another explanation is most likely related to our hypothesis that the un-glycosylated P-gp protein exists longer in the cell before it is degraded. When we attempted to remove the N-linked glycans from P-gp with PNGase F in the LoVo/Dx cell line, we obtained both the glycosylated and the unglycosylated form of P-gp, appearing as a double band at 170 kDa and 140 kDa on the western blot. A similar finding was previously made by Ma’s team [32]. They also observed an additional 110 kDa P-gp band on the western blot, but they did

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

not discuss this result. Our data support the hypothesis that tunmediated inhibition of glycosylation results in P-gp degradation, but probably it is not the only one result. We further observed that the molecular weight of BCRP was decreased in tun-treated W1TR cells. Mohrmann et al. have shown a similar result, but they detected two bands at approximately 72 and 62 kDa, where the higher band represented complexglycosylated BCRP and the lower one the core-glycosylated form [8]. Furthermore, Nakagawa’s team has detected additional lowmolecular-weight forms of BCRP in tun-treated cells [9]. On our western blot, we only observed a single low-molecular-weight band that corresponded to unglycosylated BCRP. We hypothesised that this discrepancy between our finding and the others may have been due to the high efficiency of glycosylation inhibition by tun in our experimental design. To test this hypothesis, the protein lysate of W1TR cells was treated with the endoglycosidase PNGase F. Afterwards, western blot analysis of BCRP revealed two bands at 70 kDa and 60 kDa, where the upper band corresponded to the mature form of BCRP and the lower band the unglycosylated form. The lower band is at the same size as the BCRP band that was detected after tun treatment. The lower band was thought to represent the unglycosylated or immature form of BCRP. Unfortunately, Nakagawa’s team showed only one BCRP band at the lower molecular weight but not at the higher molecular weight after digestion with PNGase F [9]. This may indicate that the PNGase F incubation time that we selected for our experiment was too short, despite the recommendations of the manufacturer. Noticeably, we did not detect BCRP expression in tun-treated A2780T1 cells. Nakagawa et al. analysed the effect of tun on Flp-In293 cells expressing BCRP, and they also observed reduced protein expression of BCRP [9]. This finding suggests that N-glycosylation is important for stabilising BCRP and that disruption of this process may enhance ubiquitin-mediated proteasomal degradation of BCRP [9]. Another report has suggested that N-glycosylation at arginine 596 is not essential for the expression of BCRP [11]. However, our result suggested that tun may have interfered with BCRP synthesis in the A2780T1 cells. Therefore, the BCRP protein most likely underwent degradation in the presence of tun. The lack of the sugar residue may be responsible for blocking the dimerisation of BCRP, which is important for the stability of the BCRP protein [20,33,34]. It is known that WT (Wild Type) BCRP is degraded in lysosomes, but that misfolded BCRP undergoes ubiquitin-mediated protein degradation in proteasomes [9,33]. Furthermore, tun is known to indirectly activate the ubiquitin-proteasome proteolytic pathway [31]. Based on this knowledge, we examined which of the two protein degradation pathways was responsible for decreased BCRP protein expression in tun-treated A2780T1 cells. For this experiment we used two inhibitors of protein degradation: BMA, which inhibits lysosomal degradation, and MG132, which inhibits proteasomal degradation [19]. Co-treatment of cells with tun and either one of the protein degradation inhibitors resulted in mature BCRP expression. Similar to the findings from Nakagawa’s team, we showed that pretreatment with MG132 suppressed the inhibitory effect of tun on BCRP protein expression [9]. The same observation was made for BMA. These data suggested that both the lysosomal and proteasomal degradation pathways were involved in BCRP degradation in A2780T1 cells. By contrast, Nakagawa et al.’s data showed that BCRP was mainly degraded through the lysosomal pathway [9]. Glycosylation plays an important role in targeting P-gp to the cell membrane and stabilising the protein against degradation at the cell surface [29]. Our immunofluorescence experiment showed that P-gp was mainly localised to the plasma membrane and the nucleus in LoVo/Dx and W1PR cells. After treatment with tun we observed additional fluorescent signal for P-gp within the

55

intracellular compartments in both LoVo/Dx and W1PR cells. Loo et al. has reported that tun could prevent immature coreglycosylated P-gp from reaching the cell surface [15]. However, Sˇeresˇ et al. showed that tun had no effect on the cellular localisation of P-gp [14]. Another study showed that unglycosylated P-gp was only expressed at a lower level at the cell surface [17]. Kramer et al. suggested that a certain amount of unglycosylated P-gp could still be localised to the cell surface under tun treatment [27]. Interestingly, none of the previous studies have reported the nuclear localisation of P-gp after tun treatment. Our results are mostly consistent with the previous data, showing that unglycosylated P-gp was still present at the cell membrane but at a much less extent. This finding supports the hypothesis that N-glycosylation is important for maintaining proper P-gp localisation. Furthermore, decreased presence of Pgp at the cell membrane also indicates a reduction in P-gpmediated transport. Therefore, the decrease in cell viability after combination treatment with tun and any of the cytostatic drugs may have been due to increased concentration of the drugs inside the cells. Our immunofluorescence data agreed with the MTT assay data. The localisation of BCRP in W1TR and A2780TR1 cells were not greatly affected by tun. In both cell lines the protein was mainly found at the cell membrane and the nuclear envelope. This result suggested that N-linked glycosylation is not essential for BCRP localisation. This result is also consistent with previous reports [8,11]. One study showed that N-glycosylation was not a prerequisite for routing BCRP to the plasma membrane [8]. Diop and Hrycyna reported that N-linked glycosylation an Asn596 was not essential for trafficking BCRP to the plasma membrane [11]. Interestingly, we only detected BCRP expression with immunofluorescence but not with western blot in tun-treated A2780T1 cells. This discrepancy was most likely due to the target epitope of the antibody, which corresponds to amino acids 301– 370 that are mapped to an internal region of the BCRP protein. It is likely that we obtained a false positive signal in the immunofluorescence experiment because the anti-BCRP Ab was able to bind to the exposed internal epitope during tun-induced BCRP degradation in A2780T1 cells. 5. Conclusions Based on our data, we concluded that tun may impair MDR protein functions in tumour cells, resulting in increased cellular sensitivity to cytostatic drugs. Tun caused a decrease in the activity of BCRP and P-gp. The lack of glycan residues caused changes in the intracellular localisation of P-gp. These results suggest that blocking glycosylation may affect the nuclear localisation of Pgp and inhibit any potential mechanisms of drug resistance that are associated with the nucleus. By demonstrating that the transporter function of BCRP and P-gp could be impaired by tun, we propose that tun may be a useful adjunct to existing chemotherapies, thus leading to better quality of life and improved survival for cancer patients. Author contributions KW conceived the study, set up and performed all experiments, performed data analysis and wrote the manuscript. KW, RJ and MN designed and coordinated the study. RJ involved in data analysis and contributed with critical revisions of the manuscript. MN participated in the design of the study and coordination and helped to draft the manuscript. MZ reviewed and edited the manuscript. All authors read and approved the final manuscript.

56

K. Wojtowicz et al. / Biomedicine & Pharmacotherapy 74 (2015) 49–56

Conflict of interest The authors declare that they have no competing interests. Acknowledgements This work was supported by Poznan University of Medical Sciences 502-14-02229373-10079 (KW). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. References [1] P. Vineis, C.P. Wild, Global cancer patterns: causes and prevention, Lancet 383 (9916) (2014) 549–557. [2] J.P. Gillet, M.M. Gottesman, Mechanisms of multidrug resistance in cancer, Methods Mol. Biol. 596 (2010) 47–76. [3] F.J. Sharom, Complex interplay between the P-glycoprotein multidrug efflux pump and the membrane: its role in modulating protein function, Front. Oncol. 4 (2014) 41. [4] A. Molinari, A. Calcabrini, S. Meschini, A. Stringaro, P. Crateri, L. Toccacieli, M. Marra, M. Colone, M. Cianfriglia, G. Arancia, Subcellular detection and localization of the drug transporter P-glycoprotein in cultured tumor cells, Curr. Protein Pept. Sci. 3 (6) (2002) 653–670. [5] W. Szaflarski, P. Sujka-Kordowska, R. Januchowski, K. Wojtowicz, M. Andrzejewska, M. Nowicki, M. Zabel, Nuclear localization of P-glycoprotein is responsible for protection of the nucleus from doxorubicin in the resistant LoVo cell line, Biomed. Pharmacother. 67 (6) (2013) 497–502. [6] A. Varki, R.D. Cummings, J.D. Esko, et al., Essentials of Glycobiology, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2009. [7] L.M. Greenberger, S.S. Williams, E. Georges, V. Ling, S.B. Horwitz, Electrophoretic analysis of P-glycoproteins produced by mouse J774.2 and Chinese hamster ovary multidrug-resistant cells, J. Natl. Cancer Inst. 80 (1988) 506–510. [8] K. Mohrmann, et al., Absence of N-linked glycosylation does not affect plasma membrane localization of breast cancer resistance protein (BCRP/ABCG2), Cancer Chemother. Pharmacol. 56 (2005) 344–350. [9] H. Nakagawa, et al., Disruption of N-linked glycosylation enhances ubiquitinmediates proteasomal degradation of the human ATP-binding cassette transporter ABCG2, FEBS J. 276 (2009) 7237–7252. [10] X. Qian, Y.H. Cheng, D.D. Mruk, C.Y. Cheng, Breast cancer resistance protein (BCRP) and the testis—an unexpected turn of events, Asian J. Androl. 15 (4) (2013) 455–460. [11] N.K. Diop, C.A. Hrycyna, N-linked glycosylation of the human ABC transporter ABCG2 on asparagine 596 is not essential for expression, transport activity, or trafficking to the plasma membrane, Biochemistry 44 (2005) 5420–5429. [12] K. Wojtowicz, W. Szaflarski, R. Januchowski, P. Zawierucha, M. Nowicki, M. Zabel, Inhibitors of N-glycosylation as a potential tool for analysis of the mechanism of action and cellular localisation of glycoprotein P, Acta Biochim. Pol. 59 (4) (2012) 445–450. [13] J. Dennis, M. Granovsky, C. Warren, Protein glycosylation in development and disease, Bioessays 21 (1999) 412–421. [14] M. Sˇeresˇ, D. Cholujova´, T. Bubencˇı´kova, A. Breier, Z. Sulova´, Tunicamycin depresses P-glycoprotein glycosylation without an effect on its membrane localization and drug efflux activity in L1210 cells, Int. J. Mol. Sci. 12 (2011) 7772–7784.

[15] T.W. Loo, D.M. Clarke, The human multidrug resistance P-glycoprotein is inactive when its maturation is inhibited: potential for a role in cancer chemotherapy, FASEB J. 13 (1999) 1724–1732. [16] F.D. Grosso, M.D. Mariano, L. Passoni, R. Luksch, G.P. Tonini, L. Longo, Inhibition of N-linked glycosylation impairs ALK phosphorylation and disrupts pro-survival signaling in neuroblastoma cell lines, BMC Cancer 11 (2011) 525. [17] J.J. Gribar, M. Ramachandra, C.A. Hrycyna, S. Dey, S.V. Ambudkar, Functional characterization of glycosylation-deficient human P-glycoprotein using a vaccinia virus expression system, J. Membr. Biol. 173 (2000) 203–214. [18] R. Januchowski, K. Wojtowicz, P. Sujka-Kordowska, M. Andrzejewska, M. Zabel, MDR gene expression analysis of six drug-resistant ovarian cancer cell lines, Biomed. Res. Int. 2013 (2013) 241763, http://dx.doi.org/10.1155/2013/241763. [19] L. Ellgaard, M. Molinari, A. Heleniu, Setting the standards: quality control in the secretory pathway, Science 286 (1999) 1882–1888. [20] A. Helenius, M. Aebi, Roles of N-linked glycans in the endoplasmic reticulum, Annu. Rev. Biochem 73 (2004) 1019–1049. [21] S. Gerber-Lemaire, L. Juillerat-Jeanneret, Glycosylation pathways as drug targets for cancer: glycosidase inhibitors, Mini Rev. Med. Chem. 6 (2006) 1043–1052. [22] G.S. Jacob, Glycosylation inhibitors in biology and medicine, Curr. Opin. Struct. Biol. 5 (5) (1995) 605–611. [23] I. Noda, S. Fujieda, M. Seki, N. Tanaka, H. Sunaga, T. Ohtsubo, H. Tsuzuki, G.K. Fan, H. Saito, Inhibition of N-linked glycosylation by tunicamycin enhances sensitivity to cisplatin in human head-and neck carcinoma cells, Int. J. Cancer 80 (2) (1999) 279–284. [24] D.C. Hiss, G.A. Gabriels, P.I. Folb, Combination of tunicamycin with anticancer drugs synergistically enhances their toxicity in multidrug-resistant human ovarian cystadenocarcinoma cells, Cancer Cell Int. 7 (2007) 5. [25] L. Song, P. De Sarno, R.S. Jope, Central role of glycogen synthase kinase-3beta in endoplasmic reticulum stress-induced caspase-3 activation, J. Biol. Chem. 277 (47) (2002) 44701–44708. [26] M.J. Morin, C.W. Porter, P. McKernan, R.J. Bernacki, The biochemical and ultrastructural effects of tunicamycin and D-glucosamine in L1210 leukemic cells, J. Cell. Physiol. 114 (1983) 62–172. [27] R. Kramer, T.K. Weber, R. Arceci, N. Ramchurren, M.V. Kastrinakis, G. Steele Jr., I.C. Summerhayes, Inhibition of N-linked glycosylation of P-glycoprotein by tunicamycin results in a reduced multidrug resistance phenotype, Br. J. Cancer 71 (1995) 670–675. [28] J.C. Madureira, et al., Inhibition of N-linked glycosylation by tunicamycin induces E-cadherin-mediated cell-cell adhesion and inhibits cell proliferation in undifferentiated human colon cancer cells, Cancer Chemother. Pharmacol. 68 (2011) 227–238. [29] A.H. Schinkel, S. Kemp, M. Dolle, G. Rudenko, E. Wagenaar, N-Glycosylation and deletion mutants of the human MDR1 P-glycoprotein, J. Biol. Chem. 268 (10) (1992) 7474–7481. [30] R. Kramer, T.K. Weber, B. Morse, R. Arceci, R. Staniunas, Constitutive expression of multidrug resistance in human colorectal tumors and cell lines, Br. J. Cancer 67 (1993) 959–968. [31] W. Liao, L. Chan, Tunicamycin induces ubiquitination and degradation of apolipoprotein B in HepG2 cells, Biochem. J 353 (2001) 493–501. [32] H. Ma, X. Miao, Q. Ma, W. Zheng, H. Zhou, L. Jia, Functional roles of glycogene and N-glycan in multidrug resistance of human breast cancer cells, IUBMB Life 65 (5) (2013) 409–422. [33] K. Wakabayashi, H. Nakagawa, T. Adachi, I. Kii, E. Kobatake, A. Kudo, T. Ishikawa, Identification of cysteine residues critically involved in homodimer formation and protein expression of human ATP-binding cassette transporter ABCG2: a new approach using the flip recombinase system, J. Exp. Ther. Oncol. 5 (2006) 205–222. [34] U. Henriksen, J.U. Fog, T. Litman, U. Gether, Identification of intra- and intermolecular disulfide bridges in the multidrug resistance transporter ABCG2, J. Biol. Chem. 280 (2005) 36926–36934.