Vincristine-Induced Overexpression of P-Glycoprotein in L1210 Cells Is Associated with Remodeling of Cell Surface Saccharides Zdenka Sulova´,† Danica Mislovicˇova´,‡ Lenka Gibalova´,† Zuzana Vajcˇnerova´,‡ Eva Pola´kova´,† Branislav Uhrı´k,† Lucia Tylkova´,† Annama´ria Kova´rova´,‡ Ja´n Sedla´k,§ and Albert Breier*,† Institute of Molecular Physiology and Genetics, Centre of Excellence of the Slovak Research and Development Agency, BIOMEMBRANES2008, Slovak Academy of Sciences, Vla´rska 5 83334, Bratislava, Slovakia, Institute of Chemistry, Center for Glycomics, Slovak Academy of Sciences, Du ´ bravska´ cesta 9, 84538, Bratislava, Slovakia, and Cancer Research Institute, Slovak Academy of Sciences, Vla´rska 7, 833 91, Bratislava, Slovakia Received September 4, 2008
Multidrug resistance of murine leukemic cell line L1210/VCR (R), obtained by adaptation of parental L1210 cells (S) on vincristine, is associated with overexpression of P glycoprotein (P-gp, the ATPdependent drug efflux pump). Previously, we found that cytochemical staining of negatively charged cell surface binding sites (probably sialic acid) by ruthenium red (RR) revealed a compact layer of RR bound to the external coat of S cells. This is in contrast to R cells and L1210/VCR cells cultured in the presence of vincristine during the last cultivation prior to the experiment (V cells), where the RR layer was either reduced or absent. In the current paper, we observed differences in the interactions of S, R and V cells with Concanavalin A (ConA) and tomato lectin (lycopersicum esculentum agglutinin, LEA). ConA bound and induced cell damage more effectively in S cells than in R or V cells. Both of these effects could be prevented by methyl-manopyranose, but not by N-acetylglucosamine. In contrast, LEA lectin preferentially bound to R and V cells. While LEA agglutinated cells more effectively than ConA, it did not cause cell damage comparable to ConA. Binding of LEA to the cell surface could be prevented by chitooligosaccharides. Both LEA and ConA failed to identify P-gp in lectin blots. Thus, changes in ConA and LEA interactions are not caused by massive expression of P-gp in the plasma membrane and the consequent exposure of the inner saccharides to the external side of the plasma membrane. Taken together, the above facts suggest that S cells differ from R and V cells in the composition of cell surface glycosides not directly linked to P-gp. Keywords: p-glycoprotein • multidrug resistance • cell surface saccharides • concanavalin A • lycopersicum esculentum agglutinin • L1210 cell line
Introduction Multidrug resistance (MDR) of neoplastic tissues represents a major obstacle in effective chemotherapy of cancer.1 Several mechanisms involving changes in the expression of drug transporting pumps2 and drug metabolizing enzymes3 as well as changes in apoptosis regulatory pathways4 have been described to induce MDR. Overexpression of P-glycoprotein (P-gp) (ABCB2 member of ABC transporter family), an efflux pump of the plasma membrane with broad substrate specificity5,6 responsible for the elimination of lipophilic drugs from the intracellular space of cells, represents the most common cause of MDR.7 Overexpression of P-gp in neoplastic cell lines may be accomplished by their adaptation to selective pressure from cytotoxic agents-Pgp substrates.2,7 This procedure was applied to generate the * To whom correspondence should be addressed. Institute of Molecular Physiology and Genetics; phone, +421 2 54775266; fax, +421 2 54773666; e-mail,
[email protected]. † Institute of Molecular Physiology and Genetics. ‡ Institute of Chemistry. § Cancer Research Institute. 10.1021/pr8007094 CCC: $40.75
2009 American Chemical Society
P-gp-overexpressing L1210/VCR cell line, which was prepared by stepwise adaptation of the mice leukemic cell line L1210 to vincristine.8 The obtained cell line differs from parental L1210 by the presence of a massive amount of P-glycoprotein;8,9 however, expression and activity of glutathione S-transferase or MRP proteins (other members of the ABC transporter family) remain unchanged.10,11 We have described considerably reduced levels of UDP-sugars in L1210/VCR cells compared to L1210 cells, which reflects the lower contents of glycoproteins and polysaccharides.9 Cell surfaces of resistant cells show markedly reduced staining by ruthenium red (a polycationic dye that interacts with negatively charged groups on cell surfaces). Sialic acid is the main constituent of negatively charged sites on the cell surface, and is presumed to be visualized by ruthenium red.12 This indicates the existence of differences between sensitive and resistant cells in the content and composition of cell surface sugars. These differences potentially detectable by several lectins13 may be considered to reflect changes in the sugar signal code.14,15 Vincristine-colchicine-resistant L1210 cells showed less agglutination by concanavalin A (ConA) Journal of Proteome Research 2009, 8, 513–520 513 Published on Web 12/19/2008
research articles
Sulova´ et al.
Figure 1. Detection of P-gp expression in total mRNA fractions isolated from S, R and V cells (A) by RT-PCR and in crude membrane fractions isolated from S, R and V cells by Western blot (B); detection of ConA- and LEA-interacting material in crude membrane fractions isolated from S, R and V cells by lectin blot (C and D). Both GAPDH and β-actin gave very week signals for the crude membrane fraction in the Western blot; therefore, the accuracy of protein loading in each lane was controlled by Commassie blue staining of a separate gel (B, left record). Data represent a characteristic record from five independent measurements that gave similar results. Standards (St) in panel B and C: MBP-β-galactosidase, 175 kDa; MBP-paramyosin, 83 kDa; MBP-CBD, 62 kDa; Aldolase, 47.5 kDa; triosephosphate isomerase 32, 5 kDa (prestained protein marker from BioLabs, New England).
compared to their sensitive counterpart.16 Tomato lectin LEA (lycopersicum esculentum agglutinin) was described to interact with the 180 kDa glycoform of P-glycoprotein in rat brain capillary endothelia and MDR tumor cells.17 The objective of the present study was to examine any differences between the cellular surfaces of L1210 and L1210/ VCR cells that could be identified by ConA and LEA.
Materials and Methods Cells and Cultivation Conditions. Multidrug-resistant P-gpoverexpressing L1210/VCR cells were obtained by stepwise adaptation of parental L1210 cells (S) to vincristine.8 Cells were cultured for 2 days in a humidified atmosphere supplemented with 5% CO2 at 37 °C, in standard RPMI medium with 514
Journal of Proteome Research • Vol. 8, No. 2, 2009
glutamine containing 4% bovine fetal serum and 20 µg/L gentamycin (Gibco). The last cultivation of L1210/VCR cells prior to the experiments was carried out in the absence (R) or presence (V) of vincristine (0.1 µmol/L, Gedeon Richter Co., Hungary). Detection of mRNA Encoding P-gp. Total cellular RNA was extracted from S, R and V cells using the LiCl/urea method.18 RT-PCR was performed with the Gene Amp RNA PCR kit (Perkin-Elmer via Rocher Molecular Systems, Inc., Branchburg, NJ) as described by Stein et al.19 The following PCR primers were used: 5′-CCC ATC ATT GCA ATA GCA GG-3′ and 5′-GTT CAA ACT TCT GCT CCT GA-3′ for mdr1, which yields a 167 bp product20,21 and 5′-ATC ATG TTT GAG ACC TTC AA-3′ and 5′CAT CTC TTG CTC GAA GTC CA-3′ for β-actin, giving a 316
research articles
Remodeling of Cell Surface Saccharides Table 1. ELLBA Experiment: Binding of ConA and LEA to Glycoprotein in Crude Membrane Fractions Isolated from S, R and V Cellsa
S R V
ConA binding (A480/mg)
LEA binding (A480/mg)
significant differences
P-value ConA
P-value LEA
6.14 ( 0.84 3.85 ( 0.25 2.26 ( 0.11
0.51 ( 0.06 2.04 ( 0.31 4.35 ( 0.73
S vs R S vs V R vs V
0.060 0.010 0.004
0.009 0.007 0.045
a Experimental data represent the slopes of linear parts of dependences of absorbance at 480 nm (A480) and the concentration of membrane protein (see Materials and Methods). Data are expressed as means ( SEM from three independent experiments. Statistical significances were determined by SigmaPlot Graphing Software (version 2.01).
bp product. The PCR products were separated on a 1.5% agarose gel (Gibco) and visualized with ethidium bromide. Detection of P-gp by Western Blot. Cells were disrupted and a crude membrane fraction was prepared as described by Hamada and Tsuruo.22 Membrane proteins (12 µg per lane) were separated by SDS-PAGE23 on polyacrylamide gradient gel (8-16%). Proteins were transferred by electroblotting to nitrocellulose (Amersham, U.K.) according to the method of Towbin et al.24 The signal of GAPDH or β-actin under these conditions was very weak,25 and therefore, the accuracy of equivalent protein loading in each lane was shown by Commassie blue staining of a separate gel. P-gp on the membrane was detected using C219 anti-P-gp monoclonal antibody (Calbiochem) and a secondary anti-mouse antibody linked to horseradish peroxidase with the aid of the ECL detection system (Amersham U.K.) using Kodak scanning system CF 440. Lectin Blots. Glycoproteins interacting with ConA and LEA were detected in crude membrane fractions after SDS-PAGE
Table 2. ELLBA Experiment: Inhibition of ConA Binding by MMP and NAcGlc and LEA Binding by Chitooligosaccharides to Glycoproteins in Crude Membrane Fractions Isolated from S, R or V Cellsa inhibition of ConA binding IC50 for MMP (mmol/L)
inhibitiom of LEA binding
IC50 for IC50 for NAacGlc chitooligosacharides (mmol/L) (mg/L)
S 0.0821 ( 0.0063 4.775 ( 1.575 R 0.0590 ( 0.0068 5.986 ( 1.482 Significance P < 0.05 n.s.
0.798 ( 0.254 2.184 ( 0.357 P < 0.05
a Data are expressed as medians of inhibitory concentrations IC50 and represent means ( SEM from three independent experiments. Statistical significances were determined by SigmaPlot Graphing Software (version 2.01).
and electroblotting (processed similarly to the description above) by specific staining with the respective lectin conjugated with biotin and avidin linked to horseradish peroxidase (Sigma) using the ECL detection system (Amersham U.K.) and Kodak scanning system CF 440. Enzyme-Linked Lectin Binding Assay (ELLBA). Crude membrane fractions (0.03-0.6 µg of membrane protein) in 100 µL of 0.05 mol/L Tris-HCl buffer (containing 8 g NaCl/L, pH 7.6, TBS) were applied to F16 MAXISORP Immunomodules (NUNC A/S, Roskilde, Denmark). The plates were incubated at 10 °C for 24 h and washed three times with 100 µL of TBS containing 0.05% Tween. Then, the wells were blocked with 100 µL of 0.5% BSA in TBS buffer for 45 min at room temperature. After blocking, the biotinylated ConA and LEA (100 and 200 ng, respectively) were applied to each well in 100 µL of TBS with 0.1% BSA, and then incubated for an additional 45 min at room
Figure 2. Lectin-fluorescence confocal microscopy of ConA and LEA binding to S, R and V cells. Cells were incubated without FITClabeled lectins in control experiments. Data represent a characteristic record from six independent measurements that gave similar results. Journal of Proteome Research • Vol. 8, No. 2, 2009 515
research articles
Sulova´ et al.
Figure 3. Lectin-gold electron microscopy of ConA (panels A and B) and LEA (panels C and D), labeling of S (panels A and C) and R (panels B and D) cells. Arrows point to gold particles.
temperature. Plates were washed three times with 100 µL of TBS containing 0.05% Tween and then left to incubate with 100 µL of avidin-peroxidase (25 ng/mL) solution in TBS with 0.1% BSA for 45 min. The plates were washed three times with 100 µL of TBS containing 0.05% Tween and 100 µL of freshly prepared solution (in phosphate-citrate buffer, pH 4.6, containing in 1 mL, 0.5 mg of o-phenylenediamine and 1.25 µL of 30% H2O2) was added to each well and incubated for 15 min in the dark. The reaction was stopped by addition of 50 µL of 3.6 mol/L H2SO4 and absorbance at 490 nm (A490) was measured. Data are expressed as slopes of linear part of dependences between A490 and the amount of membrane proteins. ConA ligands (R-methylmannopyranoside, MMP, and Nacetylglucosamine, NAcGlc; Sigma) and LEA ligand (chitooligosaccharides [kindly provided by Eva Machova from Institute of Chemistry, Slovak Academy of Sciences] containing 2-5 units26) were used for specific inhibition of lectins binding to the membrane protein. Data are expressed as the median inhibitory concentration for the respective ligand. Lectin-Fluorescence Microscopy (LFM). Cells after cultivation were washed three times with phosphate buffered saline (PBS), resuspended in RPMI medium without fetal bovine serum (5 × 105 cells/mL) and incubated 30 min with FITClabeled lectins (Sigma) at a concentration of 0.1 mg/L in a humidified atmosphere supplemented with 5% CO2 in air at 37 °C. After incubation, cells were washed three times with PBS, and finally specific labels were evaluated as green fluorescence in a confocal microscope (Leica TCS SP-2 AOBS) using excitation at 488 nm. 516
Journal of Proteome Research • Vol. 8, No. 2, 2009
Lectin-Gold Microscopy (LGM). Cells were processed similarly to the description in the previous section, but were left to interact with biotin-conjugated lectins (0.1 mg/L, Sigma). After this incubation, the cells were washed three times in PBS and additionally incubated with avidin gold (Sigma) for the next 30 min. Following the washing in PBS (3×) the cells were fixed for 60 min with 2% glutaraldehyde in 0.1 mol/L sodiumcacodylate buffer (CB) at pH 7.2. After fixation, the cells were centrifuged at 1000g for 3 min to form a precipitate. The specimen was then washed three times for 7 min with CB and postfixed with 1% osmium tetroxide in CB for 60 min. The samples were rinsed in bidistilled H2O, stained overnight with 2% uranyl acetate, dehydrated in increasing concentrations of ethanol, cleared in propylene oxide and embedded in Durcupan. Ultrathin sections were cut with a microtome Porter-Blum MT2, stained with lead citrate and examined in a JEOL JEM1200 EX electron microscope at 80 kV. Effect of ConA and LEA on Cell Survival. Cells were cultivated as described above (inoculum 5 × 104 cells in 200 µL per well in 96 wells cell cultivation plates) in the absence or presence of ConA (0-25 mg/L) and LEA (0-125 mg/L). After cultivation, cell viability was checked by Thiazolyl Blue Tetrazolium Bromide (MTT).27 In the case of experiments testing the efficiency of lectin saccharide-ligands to protect cells against lectin-induced cell damage, MMP, NAcGlc and chitooligosaccharides were added directly to the cultivation medium 30 min prior to the addition of lectin. Detection of ConA Induction of Apoptosis or Necrosis of Cells. Induction of apoptosis or necrosis was detected by flow cytometry using FITC labeled Annexin V and propidium iodide
research articles
Remodeling of Cell Surface Saccharides
where N represents proportion of viable cells after cultivation in the presence of a lectin at concentration c. IC50 is the concentration of lectin when N ) 50%, n is order exponent. Experimental data were fitted according to this equation using SigmaPlot Graphing Software (version 2.01).
Results and Discussion
Figure 4. Effects of ConA and LEA on viability of S, R and V cells. The number of cells in the absence of lectins was arbitrarily taken as 100%. Data represent mean ( SEM from six independent experiments. Experimental data were fitted according to eq 1 by nonlinear regression that yields the following IC50 and n values: for ConA effect on S cells, 8.73 ( 0.92 mg/L and 1.96 ( 0.17; R cells, 25.29 ( 1.92 and 1.25 ( 0.14; V cells, 16.55 + 1.72 and 1.03 ( 0.09; and for LEA effect on S cells, 126.58 ( 24.29 and 0.98 ( 0.14; R cells, 267.25 ( 72.24 and 0.99 ( 0.12; V cells, 149.19 ( 32.87 and 1.04 ( 0.09. Nonlinear regression were processed using SigmaPlot Graphing Software (version 2.01).
as specific fluorescent stains using the detection kit supplied by Calbiochem. Cells (S, R or V) were preincubated for 60 min with ConA (5 mg/L) in cultivation medium in a humidified atmosphere supplemented with 5% CO2 at 37 °C. After this incubation, cells were washed two times with PBS and resuspended in 500 µL of binding buffer (10 mM HEPES/NaOH, pH 7.5, containing 140 mM NaCl and 2.5 mM CaCl2) at a concentration of approximately 106 cells/mL. The cell suspensions were transferred to plastic test tubes, and 1.25 µL of FITC labeled Annexin V (resulting concentration: 0.5 µg/mL) was added. Mixtures were incubated at room temperature in the dark for 15 min; cells were washed once and finally resuspended in ice-cold PBS. Then, 10 µL of propidium iodide (resulting concentration: 0.6 µg/mL) was added. Fluorescence of the cells was determined using a Coulter Epics Altra flow cytometer. Statistical Analysis and Data Processing. Data were expressed as mean ( SEM. Statistical significance was assessed using an unpaired Student’s t test using SigmaPlot Graphing Software (version 2.01). The concentration dependence of ConA and LEA cytotoxic effects on the cell viability could be described by eq 1, that is, an equation of exponential decay28 N ) 100% × exp[ln(0.5) × (c/IC50c)n]
(1)
Massive expression of P-gp was observed in R and V cells at the mRNA and protein levels; in contrast, an observable P-gp signal was not detected in S cells, and may be considered the main difference between L1210 and L1210/VCR (Figure 1). We have described elsewhere several other changes in the expression and activity of different protein kinases,29-32 nuclear receptors for retinoids,25 calcium homeostasis33 and expression of proteins involved in calcium regulatory pathways34 associated with the P-gp-mediated MDR phenotype of L1210/VCR cells. The finding that R and V cells also differ from S cells in the content of several oligo- and polysaccharides9 is directly related to the subject of the current paper. P-glycoprotein is glycosylated on its first extracellular loop, and this glycosylation causes molecular weight increase from about 140 kDa (deduced from the sequence of 1280 amino acids) to about 170 kDa.35 Thus, massive expression of P-gp in the plasma membrane of L1210/VCR cells could produce exposure of P-gp specific saccharide parts to the extracellular space of the cells. Lectins are carbohydrate-specific natural reagents36,37 capable of detecting the presence of these saccharides. The idea that the presence of huge amounts of P-glycoprotein in the cell membranes could also be detected by lectins was supported using LEA lectin, which specifically detects the 180 kDa isoform of P-gp.17 In contrast to this, neither ConA nor LEA was able to detect changes in P-gp expression between S, R and V cells. This may be deduced from the following facts: the presence of P-gp in crude membrane fractions isolated from R and V cells but not from S cells was clearly documented by anti-P-gp antibody in the region of the molecular weight characteristic for the P-gp molecular forms (140-180 kDa, Figure 1B); if LEA or ConA will interact with P-gp effectively, similar strong differences in staining of protein bands in this region as by c219 antibody should be observable also by lectins. However, these differences were not found (Figure 1C,D). Thus, glycosylation of P-gp in our cells did not produce saccharide regions recognizable by the applied lectins. The lack of LEA ability to directly bind P-gp in our experiments is in contrast to experiments of Fakla et al.17 and may be related to different glycosylation patterns of P-gp specific for different cell lines. There are more than 150 P-gp isoforms, which could present a range of differences within the glycome.35 More intensive staining of glycoproteins in crude membrane fractions isolated from S cells than from R or V cells by ConA is visible in Figure 1. In contrast, labeling of proteins in crude membrane fractions by LEA appears to be more pronounced in R and V cells than in S cells. However, the lectin blot must be considered a semiquantitative method only; therefore, we still need to use an exact quantitative method such as ELLBA to quantify the different contents of ConA and LEA interacting material in crude membrane fractions isolated from S, R and V cells. Data in Table 1 provide clear evidence that ConA is interacting more effectively with glycoproteins in crude membrane fractions isolated from S cells than from R or V cells. This observation supports old data from Csuka and Sugar16 that demonstrated ConA more effectively agglutinated vincristinesensitive than vincristine-resistant L1210 cells. In our experiJournal of Proteome Research • Vol. 8, No. 2, 2009 517
research articles
Sulova´ et al.
Table 3. Effect of MMP on ConA-Induced Damage to S, R or V Cells
a
ConA (10 mg/L)
S R V
K
MMP 5 mmol/L
100.0 ( 1.1 100.0 ( 1.8 100.0 ( 1.3
98.7 ( 2.5 101.0 ( 3.6 100.6 ( 1.9
68.5 ( 3.4b 82.0 ( 6.0b 78.0 ( 4.8b
ConA (25 mg/L)
MMP 5 mmol/L
p
83.6 ( 2.8b 95.0 ( 4.5 88.0 ( 4.3b
0.03 0.14 0.20
21.4 ( 3.5b 61.0 ( 3.8b 54.0 ( 5.5b
MMP 5 mmol/L
p
54.0 ( 3.6b 83.0 ( 5.1b 65.0 ( 4.3b
