Overexpression of glucosylceramide synthase (GCS), a pivotal enzyme in glycolipid biosynthesis, contributes to cancer cell resistance to chemotherapy. We previously showed that transfection of doxorubicin-resistant MCF-7-AdrR cells with GCS antisense restored cell sensitivity to doxorubicin and greatly enhanced sensitivity to vinblastine and paclitaxel. In that study, doxorubicin promoted generation of ceramide in MCF-7-AdrR/GCS antisense cells; the present study implicates factors in addition to ceramide that augment sensitivity to chemotherapy. Although GCS antisense cells showed enhanced ceramide formation compared with MCF-7-AdrR when challenged with paclitaxel, GCS antisense cells also showed a 10-fold increase in levels of intracellular drug (paclitaxel and vinblastine). In addition, transfected cells had dramatically decreased expression (80%) of P-glycoprotein and a 4-fold decrease in the level of cellular gangliosides. Chemical inhibition of GCS produced the same effects as antisense transfection: exposure of MCF-7-AdrR cells to the GCS inhibitor 1-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP, 5.0 μmol/L, 4 days) decreased ganglioside levels, restored sensitivity to vinblastine, enhanced vinblastine uptake 3-fold, and diminished expression of MDR1 by 58%, compared with untreated controls. A similar effect was shown in vinblastin-resistant KB-V0.01 cells; after 7 days with PPMP (10 μmol/L), MDR1 expression fell by 84% and P-glycoprotein protein levels decreased by 50%. MCF-7-AdrR cells treated with small interfering RNAs to specifically block GCS also showed a dramatic decrease in MDR1 expression. This work shows that limiting GCS activity down-regulates the expression of MDR1, a phenomenon that may drive the chemosensitization associated with blocking ceramide metabolism. The data suggest that lipids play a role in the expression of multidrug resistance.
- Glucosylceramide synthase
Development of resistance to chemotherapeutic agents is a major concern in cancer patients. Resistance to chemotherapy is associated with myriad mechanisms that decrease drug cytotoxicity. Two members of the large family of ABC transporters confer multidrug resistance (MDR) in human cancer cells: P-glycoprotein and multidrug resistance protein (MRP). P-glycoprotein, a membrane-resident glycoprotein encoded by the MDR1 gene, decreases the intracellular concentration of anticancer agents by acting as a drug efflux pump ( 1, 2). P-glycoprotein exports many types of chemotherapy drugs, including Vinca alkaloids, anthracyclines, paclitaxel, actinomycin D, and epipodophyllotoxins. Like P-glycoprotein, MRP is a transport protein ( 3); however, the transport of unconjugated chemotherapeutic agents by MRP seems to require glutathione. MDR in tumors can also be caused by overexpression of proteins such as antiapoptotic proteins belonging to the Bcl-2 family ( 4, 5) and by loss of tumor suppressor protein p53 ( 6, 7). Others factors responsible for chemotherapy resistance include reduction of topoisomerase II activity ( 8), modification of glutathione S-transferase activity ( 9), and up-regulation of rafts and caveolae, which are glycosphingolipid-enriched constituents of microdomains ( 10).
Glucosylceramide synthase (GCS) catalyzes the first glycosylation step in the biosynthesis of glycosphingolipids ( 11, 12). This central enzyme of ceramide metabolism has also been implicated in MDR ( 13). Glycosphingolipids, including glucosylceramide, lactosylceramide, and gangliosides, play an essential role in cell development, cell death, tumor progression, and pathogen/host interaction ( 13, 14). In addition, membrane gangliosides can decrease the sensitivity of human melanoma cells to ionizing radiation ( 15). In that study, radioresistant melanoma cells were made radiosensitive by exposure to either fumonisin B1, which blocks ganglioside biosynthesis at the juncture of ceramide synthase, or Vibrio cholerae neuraminidase, which cleaves cell surface gangliosides. Conversely, adding bovine brain GM1 to radiosensitive melanoma cells conferred radioresistance ( 15). Targeting glycolipid metabolism has proven useful in altering chemotherapy responses in numerous human tumor cell lines ( 13, 16– 18).
In previous studies, we increased the level of MDR by transfecting doxorubicin-resistant human breast cancer cells (MCF-7-AdrR) with GCS, and we enhanced cellular sensitivity to anthracyclines, Vinca alkaloids, and taxanes by transfecting MCF-7-AdrR cells with GCS antisense ( 16, 19). Although doxorubicin treatment of GCS antisense transfectants increased intracellular levels of ceramide ( 16), which is a second messenger of apoptosis, the extremely high sensitivity of MCF-7-AdrR/GCS antisense cells to Vinca alkaloids and taxanes suggested the participation of mechanisms other than ceramide signaling in cellular responses. We have observed equivalent intracellular levels of rhodamine-123 in MCF-7-AdrR/GCS antisense cells and in rhodamine-123–exposed MCF-7 parental cells, which indicates that GCS antisense transfection reverts drug retention in MCF-7-AdrR cells on a par with the drug-sensitive phenotype ( 16, 19). Because rhodamine-123 is a substrate for P-glycoprotein, we began to investigate the influence of glycolipid metabolism on function and expression of MDR1 and P-glycoprotein. The present study shows that MDR1 and P-glycoprotein expression can be down-regulated by GCS antisense transfection or chemical inhibition of GCS. We suggest that this avenue contributes to antisense GCS chemosensitization of drug-resistant cells. Overall, our work shows that lipids play a role in multidrug resistance and that targeting glycolipid biosynthesis could be a promising approach for enhancing chemotherapy.
Materials and Methods
Cell cultures. The MCF-7-AdrR human breast adenocarcinoma cell line, which is resistant to doxorubicin ( 20), was kindly provided by Dr. Kenneth Cowan (UNMC Eppley Cancer Center, Omaha, NE) and Dr. Merrill Goldsmith (NIH, Bethesda, MD). MCF-7-AdrR cells were maintained in RPMI 1640 (Invitrogen, Chicago, IL) containing 10% (v/v) fetal bovine serum (FBS; HyClone, Logan, UT), 100 units/mL penicillin, 100 μg/mL streptomycin, and 584 mg/L l-glutamine. Cells were cultured in a humidified, 5% CO2 atmosphere tissue culture incubator and subcultured weekly using trypsin-EDTA (0.05%, 0.53 mmol/L) solution. MCF-7-AdrR cells transfected with GCS antisense (MCF-7-AdrR/GCS antisense) were cultured in RPMI 1640 containing the above components and G418 (400 μg/mL; ref. 16).
The KB-V0.01 multidrug-resistant human epidermoid carcinoma cell line ( 21) was grown in high glucose (4.5 g/L) DMEM with 10% fetal bovine serum and vinblastine (10 ng/mL). Cells were a gift from Dr. Michael Gottesman (National Cancer Institute, Bethesda, MD).
Cytotoxicity assays. Assays were done as described previously ( 13). Briefly, cells were seeded in 96-well plates (5,000 cells per well) in 0.1 mL RPMI 1640 containing 10% FBS and cultured at 37°C for 24 hours before addition of vinblastine sulfate or paclitaxel (Taxol), both of which were obtained from Sigma (St. Louis, MO). Drugs were added in FBS-free medium (0.1 mL), and cells were grown at 37°C for the indicated periods. Drug cytotoxicity was determined by the CellTiter 96 AQueous One Solution cell proliferation assay (Promega, Madison, WI). Absorbance at 490 nm was measured by a fluorescent reader (Microplate FL600, Bio-Tek, Winooski, VT).
Analysis of lipids. Analysis was done as described previously ( 13, 22). Briefly, for ceramide, sphingomyelin, phospholipid, and neutral lipid analysis, cells were seeded 6-well plates (60,000 cells per well) in 10% FBS RPMI 1640. After 24 hours, medium was replaced with 5% FBS medium containing [3H]palmitic acid (1.0 μCi/mL culture medium; DuPont/NEN, Boston, MA) for 24 hours. Total lipids were extracted as described ( 22), and the resulting organic lower phase was withdrawn and evaporated under a stream of nitrogen. Lipids were resuspended in 0.1 mL chloroform/methanol (2:1, v/v), and resolution was by TLC using solvent systems described ( 23). Commercial lipid standards (Avanti Polar Lipids, Alabaster, AL) were cochromatographed. After development, lipids were visualized by iodine vapor staining, and the area of interest was scraped into 0.5 mL water. EcoLume scintillation fluid (4.5 mL; ICN, Costa Mesa, CA) was added, samples were mixed, and radioactivity was quantitated by liquid scintillation spectrometry. For ceramide quantitation, cells were grown in 15-cm dishes, and total lipids were extracted and analyzed by TLC using a solvent system containing chloroform/acetic acid (90:10, v/v). Approximately 880 μg of lipid were loaded per lane. After development, TLC plates were sprayed with conc. sulfuric acid, and heated in an oven (30 minutes, 180°C). The ceramide chars were quantitated by scanning densitometry, using a ceramide standard curve (1.0-6.0 μg). Silica Gel G prescored TLC plates were purchased from Analtech (Newark, DE). Before the above procedure, plates were acid washed by running in methanol/HCl (90:10, v/v), and thoroughly dried before use.
Western blot for P-glycoprotein. Confluent monolayers of MCF-7-AdrR and MCF-7-AdrR/GCS antisense cells were rinsed, harvested in PBS, and lysed in a PBS buffer containing 10% glycerol, 1% Triton X-100, 1.0 mmol/L Na3VO4, 10 mmol/L β-glycerophosphate, 50 mmol/L NaF, 0.1 mmol/L phenylmethylsulfonyl fluoride, 2 μg/mL leupeptin, and 10 μg/mL aprotinin for 30 minutes on ice. The mixture was centrifuged at 11,000 × g for 15 minutes at 4°C. Equal aliquots of protein (25 μg) were resolved using 4% to 20% gradient SDS-PAGE (Invitrogen, Chicago, IL). The transferred nitrocellulose blot was blocked with 5% fat-free milk powder in PBS containing 0.1% Tween 20, at room temperature for 1 hour. The membrane was immunoblotted with 0.7 μg/mL of C219 murine monoclonal antibody against human P-glycoprotein (Calbiochem, Pasadena, CA) in the same blocking solution. Detection was done using enhanced chemoluminescence (Amersham Pharmacia Biotech, Piscataway, NJ).
RNA analysis. Total RNA was isolated using the RNeasy Protect Mini Kit from Qiagen, Inc. (Los Angeles, CA). MDR1 reverse transcription-PCR (RT-PCR) was carried out by a one-step method (SuperScript One-Step RT-PCR with Platinum Taq; Invitrogen, Chicago, IL). Total RNA (25 ng) was added to buffer containing 0.2 mM deoxynucleotide triphosphate, 1.2 mmol/L MgSO4, 1.0 μL SuperScript II RT/Platinum Taq mix (containing reverse transcriptase and platinum Taq DNA polymerase), 0.2 μg of MDR1 upstream primer 5′-CCATCATTGCAATAGCAGG-3′, and 0.2 μg of MDR1 downstream primer 5′-GAGCATACATATGTTCAAACTTC-3′. RT-PCR, in a total volume of 50 μL, was done for 35 cycles in a thermocycler (Eppendorf Scientific, Westbury, NY); each cycle comprised denaturation at 94°C for 30 seconds, annealing at 55°C for 30 seconds, and elongation at 72°C for 1 minute. RT-PCR products were analyzed by 1% agarose gel electrophoresis stained with ethidium bromide. β-Actin primer (Stratagene, Cedar Creek, TX) was used as a housekeeping gene.
Real-time PCR. A real-time quantitative PCR analysis was done using the Rotor-Gene 3000 (Corbett Research, Sydney, Australia). Primers and probe sequences for MDR1 were chosen as follows: MDR1 forward 5′-GGTTTATAGTAGGATTTACACGTGGTTG-3′, MDR1 reverse 5′-AAGATAGTATCTTTGCCCAGACAGC-3′, and MDR1 probe 5′ FAM CTAACCCTTGTGATTTTGGCCATCAGTCC Tamra 3′. Human β-actin was used as endogenous control. Both assays used the SuperScript III Platinium one-step quantitative RT-PCR system (Invitrogen, Chicago, IL).
Glucosylceramide synthase gene silencing by small interfering RNA. The small interfering (siRNA) sequence targeting human GCS was selected using the BLOCK-IT RNAi Designer program; reagents were synthesized by Invitrogen (Carlsbad, CA). The siRNA duplex with the following sense and antisense sequences was used: 5′-CCAGGAUAUGAAGUUGCAA (sense) and 5′-UUGCAACUUCAUAUCCUGG (antisense). Established protocols were followed ( 24, 25). Briefly, siRNA was introduced into cells using LipofectAMINE 2000 in serum-free medium for 4 hours. FBS was added, and after 48 hours, total RNA was extracted and used to examine GCS and MDR1 mRNA levels. LipofectAMINE 2000 alone and expression of β-actin were used as controls.
Purification and analysis of gangliosides. Cells harvested in PBS were homogenized in 6 mL chloroform/methanol (1:1, v/v); the mixture remained overnight at room temperature. After centrifugation, the supernatant was dried and the lipid residue was taken up in chloroform/methanol (1:1) and centrifuged to remove all solid particles. Addition of PBS in a volume ratio of 1:1:0.7 (chloroform/methanol/PBS) separated the organic phase from the ganglioside-containing aqueous phase, as previously described ( 26). After thorough vortex mixing, the tube was centrifuged, and the upper phase containing gangliosides was withdrawn. Partitioning was repeated twice, each time by adding methanol/PBS (1:0.7, v/v) to the lower phase, followed by centrifugation. The upper phases were pooled, and gangliosides were recovered from the aqueous solution by column chromatography on C18-bonded silica gel. Ganglioside profiles were determined by high-performance TLC on Silica Gel 60 plates (Merck, Darmstadt, Germany) developed in chloroform/methanol/0.2% aqueous calcium chloride (55:45:10, v/v/v) and sprayed with resorcinol-HCl reagent. Ganglioside sialic acid content was determined by the periodate-resorcinol method ( 27).
[3H]Paclitaxel and [3H]vinblastine uptake studies. Cells were seeded into 12-well plates at 100,000 cells per well in 1.0 mL of complete medium. After 24 hours at 37°C, the medium was removed; cells were rinsed with PBS and incubated for 10 to 90 minutes with 0.5 mL of 5% FBS RPMI 1640 containing 500 nmol/L paclitaxel plus 0.25 μCi [3H]paclitaxel (Moravek Biochemical, Brea, CA), or 500 nmol/L vinblastine plus 0.25 μCi [3H]vinblastine (Moravek biochemical). After removal of culture medium, cells were washed twice with 5% FBS RPMI 1640 and lysed with 0.2 mL of 5% FBS RPMI 1640 containing 0.02% SDS. Intracellular radioactivity was measured by liquid scintillation counting.
MDR1 induction by glycolipids. KB-V0.01 cells were seeded into 6.0-cm dishes in complete medium. After 24 hours, medium was removed and cells were incubated in 5% FBS DMEM medium containing either 30 μg/mL C8 β-d-glucosylceramide (Avanti Polar Lipids, Alabaster, AL), 0.5 μmol/L doxorubicin (LKT Laboratories, St Paul, MN) as positive control, or 10 μg/mL palmitic acid (Sigma) as negative control. Cells were treated for 48 hours, and RNA was analyzed by real-time RT-PCR.
Chemical inhibition of glucosylceramide synthase. d,l-Threo-1-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP) was from Biomol Research Laboratories (Plymouth Meeting, PA) and d,l-erythro-PPMP and d-threo-PPMP were from Matreya (Pleasant Gap, PA). These reagents were used as described in the figure legends.
Statistical analyses. Student's t test was used for statistical analysis.
Previously, we showed that GCS antisense transfection of multidrug-resistant MCF-7-AdrR cells enhanced cell sensitivity to doxorubicin, vinblastine, and paclitaxel ( 19). A doxorubicin-induced increase in ceramide levels and caspase activity is in keeping with ceramide-mediated cytotoxic responses to chemotherapy ( 16), but it is not clear whether ceramide is the only factor involved in the significantly (>100-fold) increased sensitivity of GCS antisense–transfected cells to Vinca alkaloids and paclitaxel.
Initial studies on ceramide production measured with [3H]palmitate showed that similar levels of tritiated ceramide were formed in both MCF-7-AdrR and MCF-7-AdrR/GCS antisense cells in response to vinblastine and paclitaxel challenge ( Fig. 1A ). Using mass analysis, however, we found that ceramide production in response to paclitaxel (1.0 μmol/L) was enhanced 3-fold in MCF-7-AdrR/GCS antisense cells compared with parental MCF-7-AdrR cells after 24 hours of treatment ( Fig. 1B). To further assess the influence of GCS antisense transfection on cell response to chemotherapy, we next measured uptake and efflux variables of vinblastine and paclitaxel. P-glycoprotein–mediated drug efflux is the most widely characterized drug resistance mechanism in cancer cells ( 28), and it is highly expressed in MCF-7-AdrR cells ( 13). We previously reported that levels of rhodamine-123, a substrate of P-glycoprotein, were ∼5-fold higher in MCF-7-AdrR/GCS antisense compared with MCF-7-AdrR cells ( 19). This suggests that GCS antisense transfection alters drug uptake and/or retention. Experiments with chemotherapy drugs showed that after 60 minutes, intracellular levels of vinblastine were 12-fold greater in MCF-7-AdrR/GCS antisense compared with MCF-7-AdrR cells ( Fig. 2 ). Similarly, uptake of paclitaxel increased 8.6-fold in GCS antisense transfectants compared with MCF-7-AdrR cells ( Fig. 2).
Because of the pronounced differences in drug levels in the two cell lines, we assessed P-glycoprotein expression, by mRNA and protein. As shown in Fig. 3 , the level of MDR1 mRNA, evaluated by reverse transcription-PCR (RT-PCR), was dramatically lower in MCF-7-AdrR/GCS antisense compared with MCF-7-AdrR cells. We confirmed this by Western blot; whereas MCF-7-AdrR cells contained characteristically elevated levels of P-glycoprotein, MCF-7-AdrR/GCS antisense cells were nearly devoid of P-glycoprotein. These data suggest that high drug levels attainable in MCF-7-AdrR/GCS antisense cells are a consequence of the dramatic decrease (∼80%) in P-glycoprotein expression. Thus, the stable 30% decrease in GCS activity of MCF-7-AdrR/GCS antisense cells ( 16) seems to have an important influence on intracellular drug levels and on P-glycoprotein expression. For this reason, we investigated whether partial inhibition of GCS would cause other alterations, in particular, in membrane lipid composition.
Steady-state [3H]palmitic acid radiolabeling (24 hours) of MCF-7-AdrR/GCS antisense and MCF-7-AdrR cells showed, in the former, a 30% decrease in sphingomyelin levels and a 44% decrease in the level of phosphatidylinositols (data not shown). There were no significant differences in cholesterol esters or other glycerophospholipids between transfected and parent cells. Because GCS is pivotal in the genesis of cerebrosides and gangliosides, we also looked for changes in glycosphingolipid content. Although MCF-7-AdrR and MCF-7-AdrR/GCS antisense cells expressed a similar ganglioside pattern (GM3, GM2, GD3, and GD1a; data not shown), sialic acid assay showed that the level of gangliosides was 4-fold lower in GCS antisense transfected cells (data not shown).
To determine whether depletion of gangliosides and reduced expression of P-glycoprotein were strictly a consequence of GCS down-regulation by antisense transfection, we evaluated the influence of d,l-threo-PPMP, a chemical inhibitor of GCS ( 29– 31), on ganglioside synthesis and P-glycoprotein expression in MCF-7-AdrR cells. A 4-day exposure to d,l-threo-PPMP produced a 34% decrease in ganglioside levels in MCF-7-AdrR cells ( Fig. 4A ). Moreover, PPMP greatly diminished the expression of MDR1 in MCF-7-AdrR cells, with shown stereospecificity ( Fig. 4B). Unlike d,l-threo-PPMP, d,l-erythro-PPMP is not a GCS inhibitor ( 32), and this stereoisomer had no influence on MDR1 expression. Real-time RT-PCR showed that MDR1 expression in MCF-7-AdrR cells treated with d,l-threo-PPMP and d,l-erythro-PPMP was reduced by 58% and 12%, respectively, compared with untreated MCF-7-AdrR cells ( Fig. 4B). To reinforce the results obtained with PPMP and to confirm that changes in MCF-7-AdrR/GCS antisense cellular MDR1 expression were not due to clonal artifacts, we used GCS siRNA to treat MCF-7-AdrR cells. As shown in Fig. 4C, after 48 hours, both GCS and MDR1 mRNA were dramatically decreased by GCS siRNA compared with LipofectAMINE only controls. The siRNA had no effect on expression levels of β-actin.
Whether chemical lowering of MDR1 expression affects cellular response to chemotherapy was next evaluated. Treatment of MCF-7-AdrR cells for 4 days with d,l-threo-PPMP enhanced vinblastine uptake by ∼3-fold at 30 and 60 minutes ( Fig. 4D), and as illustrated in Fig. 4E, vinblastine cytotoxicity, even at low concentrations (0.1 μmol/L), was enhanced ∼60% in cells that had been cultured with d,l-threo-PPMP. Thus, like GCS antisense transfection, inhibition of GCS by chemical means reversed resistance of MCF-7-AdrR cells to vinblastine. To examine the generality of this response, we used KB-V0.01 cells, a head/neck multidrug-resistant epidermoid carcinoma cell line that expresses both GCS and MDR1 ( 33). As shown in Fig. 5A , MDR1 expression in KB-V0.01 cells was lowered 70% by d-threo-PPMP (10 μmol/L) and 38% by d,l-threo-PPMP (15 μmol/L). Therefore, the d-threo isomer is the most effective inhibitor of GCS compared with the racemic mixture. KB-V0.01 cell treatment with d-threo-PPMP for a prolonged period (7 days) induced a dramatic decrease (84%) in MDR1 mRNA levels ( Fig. 5B). Moreover, P-glycoprotein protein levels in these cells diminished by 50%, compared with the untreated control ( Fig. 5C). We next investigated the effect of glycolipid supplementation on MDR1 expression. Growth of KB-V0.01 cells with cell-permeable C8-glucosylceramide (30 μg/mL) elicited a 2-fold increase in MDR1 mRNA levels ( Fig. 6 ), a response nearly comparable to the influence of Adriamycin (0.5 μmol/L) on MDR1 expression ( Fig. 6). Palmitic acid, used as a lipid control, had no influence on MDR1 expression.
The present study shows that GCS antisense transfection of multidrug-resistant human breast cancer cells modifies cellular lipid composition, reduces MDR1 expression, and enhances the cytotoxic effect of chemotherapeutic drugs. GCS antisense transfection decreased the levels of sphingomyelin in MCF-7-AdrR cells. Sphingomyelin is a major constituent of the external leaflet of the plasma membrane ( 34). Sphingomyelin, phosphatidylcholine, and proteins are laterally organized in biological membranes ( 35– 37). These organized domains have been suggested to participate in cellular processes, such as signal transduction, membrane trafficking, and protein sorting ( 38). Expression of the principal component of caveolae, caveolin-1, in MCF-7-AdrR and MCF-7-AdrR/GCS antisense cells, determined by Western blot, was not affected by antisense down-regulation of GCS (data not shown), although it should be mentioned that this assay is not a good estimate of the status of cellular lipids in rafts/caveolae. However, we found that transfected cells had lower levels of gangliosides, the sialic acid–containing glycosphingolipids. Gangliosides have been shown to influence lipid order and hydration of the lipid bilayer; such changes could play an important role in modulation of transmembrane molecular events ( 39). Moreover, gangliosides have been shown to influence membrane fluidity ( 40– 42). Cellular ganglioside levels decreased 4-fold in MCF-7-AdrR/GCS antisense cells compared with MCF-7-AdrR cells. Such a change could modify membrane permeability and facilitate entrance of natural-product chemotherapeutic agents such as vinblastine and paclitaxel.
In addition to their role as a structural component of the plasma membrane, gangliosides might regulate signaling events. In melanoma cells, transient ganglioside depletion by GCS inhibition reduced tumorigenic capacity ( 43). Gangliosides can also induce production of nitric oxide, tumor necrosis factor-α, and cyclooxygenase 2 and activate extracellular signal–regulated kinase and c-jun/stress-activated protein kinase kinase, p38, and nuclear factor κB (NF-κB; ref. 44). Our work showed that inhibiting the activity of GCS severely limited the expression of MDR1 and its product, P-glycoprotein. Studies have shown that the MDR1 promoter can be activated directly by anticancer agents such as vincristine, daunorubicin, doxorubicin, and colchicine ( 45); however, an association between glycolipids and the MDR1 promoter has not been clearly established. Some studies have suggested that glycolipids, in particular gangliosides, modulate multidrug resistance. For example, the up-regulation of GM3 biosynthesis in fenretinide-adapted A2780 ovarian cancer cells has been correlated with fenretinide resistance ( 46). In human leukemia cells, ganglioside depletion is believed to account for PDMP-mediated reversal of multidrug resistance, and GM3 and GD3 are thought directly involved via modulation of P-glycoprotein function through phosphorylation ( 47). Shabbits et al. ( 48) showed a relationship between drug transport and ceramide metabolism. Other support for a link between glycolipids and multidrug resistance may be found in the PPMP-modulated expression of MDR1 mRNA in SKOV3/AdrR human ovarian cancer cells ( 49), in KBV200 cells ( 50), and in the decreased efflux of [14C]paclitaxel and [3H]vincristine in a neuroblastoma cell model ( 51). Results of other studies show that verapamil, an antihypertensive formerly used clinically as a P-glycoprotein antagonist ( 52), limits the expression of MDR1 in human leukemia cells ( 53). Our group showed that verapamil, tamoxifen, and cyclosporine A block glucosylceramide formation and resultant downstream cerebroside and ganglioside biosynthesis in drug-resistant cancer cells ( 54).
Previously, we showed that doxorubicin treatment of MCF-7-AdrR/GCS antisense cells enhanced the production of ceramide ( 16). In the present study using radiolabeling, ceramide buildup was not evident in MCF-7-AdrR/GCS antisense cells challenged with either vinblastine or paclitaxel ( Fig. 1A); however, lipid mass analysis by TLC char clearly showed elevated ceramide levels in drug-challenged GCS antisense transfectants ( Fig. 1B). Failure of radiolabeling techniques to accurately portray mass is not uncommon. More importantly however is the apparent dual role that GCS antisense transfection and/or GCS blockade play in sensitizing multidrug-resistant cancer cells to chemotherapy. From our experiments, it is evident that GCS antisense (i) down-regulates expression of MDR1 and (ii) promotes ceramide buildup in cells that would otherwise scavenge ceramide via elevated GCS activity. This one-two punch could be of benefit in cancer treatment.
The present results suggest that glycolipids participate in MDR1 expression directly or via activation of a specific transcription factor. In a recent study, Bentires-Alj et al. ( 55) showed that NF-κB inhibition increased cellular uptake of daunorubicin and reduced expression of MDR1 mRNA and protein (P-glycoprotein) in colon cancer cells. NF-κB complexes can bind at a consensus NF-κB binding site in the first intron of the human MDR1 gene. Moreover, NF-κB can transactivate an MDR1 promoter luciferase construct ( 55).
In conclusion, our work shows that limiting GCS activity by either GCS antisense transfection, siRNA transfection, or PPMP treatment down-regulates the expression of P-glycoprotein. It should be noted, however, that drug resistance through enforced overexpression of GCS in wild-type MCF-7 cells, using a retroviral tetracycline-on expression system, did not rely on P-glycoprotein induction ( 13) but rather on ceramide scavenging. In addition, GCS antisense transfection retards clearance of ceramide generated in response to chemotherapeutics such as paclitaxel. Therefore, ceramide-signaled death cascades and depletion of cellular P-glycoprotein are likely contributory to heightened chemosensitivity in MCF-7-AdrR/GCS antisense. We propose that overexpression of GCS contributes to chemotherapy resistance by enhancing levels of cerebrosides and/or gangliosides that could promote the expression of MDR1. Doxorubicin-activated expression of GCS in MCF-7 drug-sensitive cells ( 56) lends support to this novel slant on the metabolism of ceramide.
Grant support: USPHS/National Cancer Institute grant CA95339, Department of Defense/U.S. Army postdoctoral award (V. Gouazé), the Susan G. Komen Breast Cancer Foundation, Associates for Breast and Prostate Cancer Studies (Santa Monica, CA), Sue and Larry Hochberg, and Fashion Footwear Association of New York Charitable Foundation (NY).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Jacques Portoukalian (Institut National de la Sante et de la Recherche Medicale U346, Lyon, France) for helpful discussions.
- Received June 30, 2004.
- Revision received February 8, 2005.
- Accepted February 24, 2005.
- ©2005 American Association for Cancer Research.