Abstract
Background: Reversion of chemoresistance by inhibition of P-glycoprotein (P-gp) expression may overcome the chemoresistance observed in many cancer types and may allow for improved therapeutic ratio. We investigated whether siRNA specific for ABCB1 (MDR1) mRNA might restore sensitivity to chemotherapy in tumor cell lines known to overexpress the MDR1 gene. Materials and Methods: MDR1-expressing tumor cell lines were transiently transfected with anti-MDR1 silencing RNA (siRNA) before exposure to doxorubicin or methotrexate. The capacity of siRNA to reduce cell proliferation and increase the IC50 of the tested chemotherapies was investigated. Results: siRNA down-regulated MDR1 mRNA expression by 50% in breast carcinoma and osteosarcoma cell lines, and significantly inhibited tumor cell proliferation up to 90% (p<0.01), when co-administered with doxorubicin or methotrexate, despite the known chemoresistance of the cell lines. siRNAs reduced the IC50 of doxorubicin and methotrexate by more than 10-fold (p<0.01). Conclusion:These results suggest the potential clinical application of anti-MDR1 siRNA to restore chemosensitivity and possibly improve the therapeutic ratio of these cytotoxic drugs.
- P-Glycoprotein
- chemotherapy
- inhibition of multidrug resistance
- osteosarcoma cell lines
- carcinoma cell lines
- siRNA
Despite being very effective in most tumor types, cytotoxic agents have non-negligible side-effects mainly due to their lack of tissue specificity (1). These side-effects are usually dose-dependent. Whether constitutive or acquired during the course of chemotherapy (2), resistance towards these agents leads to a lack of therapeutic efficiency and usually requires increasing doses of the drugs or cessation of their administration, thus impairing the chance of survival for patients.
The multidrug-resistance phenotype (MDR) involves various protein families, from membrane transporters such as the ABC superfamily transporters (ATP binding cassette transporters) (3) to detoxification enzymes such as glutathione-S-transferase Pi (GST Pi) (4). Within the ABC transporters, ATP-binding cassette, sub-family B (MDR/TAP), member 1 (ABCB1, MDR1), ATP-binding cassette, sub-family C (CFTR/MRP), member 1 (ABCC1) and breast cancer resistant protein (BCRP) are the best characterized (3). P-Glycoprotein (P-gp) encoded by the MDR1 gene (5) acts as an energy-dependent efflux pump that conveys cytotoxic drugs (vinca alkaloids, anthracyclines, and taxanes) and radioactive tracers such as 99mTc Sestamibi (99mTc MIBI) out of tumor cells (2, 5, 6). P-gp was found to be overexpressed in many chemoresistant hematopoietic (7), as well as solid, malignancies including sarcoma (8). Overexpression of MDR1 has proven to be of prognostic value in leukemia, neuroblastoma and sarcoma (9), and given its implications in many carcinomas, it was hypothesized that inhibition of P-gp might restore chemosensitivity. Tsuro et al. were the first to report that inhibiting P-gp expression with verapamil was able to restore both in vitro and in vivo chemosensitivity in leukemia (10).
Inhibitors of P-gp have been developed; they can prevent the interaction of chemotherapeutic agents (e.g. cyclosporine) or ATP (e.g. via verapamil) with their binding site on P-gp, thus preventing the efflux of these molecules (11, 12). While different in their mechanisms of action, both types of agents prevented efflux of chemotherapies and were able to reverse chemoresistance in vitro and in vivo in leukemia and myeloma cell lines (10, 13). Despite good preclinical efficacy, clinical applications of these agents are limited because of their poor therapeutic ratio (12, 14). Thus, molecular approaches specifically targeting the MDR1 gene represent an elegant alternative. Antisense RNA specific for MDR1 mRNA prevented its protein translation (15) and hybridization of triple helix-forming oligonucleotides (TFO) to MDR1 DNA inhibited its transcription (16). Both TFO and antisense strategies were shown to restore chemosensitivity in vitro to leukemia cells overexpression MDR1 (15, 17). MDR1 interference restored chemosensitivity in epidermoid carcinoma in vitro and in vivo in xenografts models established in nude mice (18-20).
Based on these promising results, we evaluated the capacity of MDR1-specific siRNA to restore chemosensitivity in vitro in chemoresistant sarcoma and carcinoma cell lines overexpressing MDR1.
Materials and Methods
Cell lines and reagents. Human MNNG/HOS (osteosarcoma) and 293T (embryonic kidney) cells were cultured in Dulbecco's Modified Eagle Medium and were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA) (DMEM, Invitrogen, Cergy Pontoise, France) supplemented with 10% heat inactivated fetal bovine serum (BioWhittaker, Verviers, Belgium), 1% penicillin-streptomycin, 1% L-glutamine (Invitrogen) (hereafter referred to as complete medium). The human adenocarcinoma cell lines MDA-MB549 provided by Alain Puisieux (Centre Leon Berard) and Cal51 (German Collection of Microorganisms and Cell Culture, DSMZ, Braunschweig, Germany) and the neuroblastoma cell line Sk N-Be (ATCC) were maintained in complete medium supplemented with 1% non-essential amino acids (Invitrogen). All cell lines were maintained at 37°C in a humidified atmosphere under 5% of CO2. All cell lines used in these experiments, except 293T used as sensitive cells control, cells are known to be chemoresistant.
Chemotherapeutic agents were chosen based on the dependence of their chemosensitivity profile to MDR1. Cell resistance to doxorubicin (Dox; Baxter, Deerfield, IL, USA) depends on P-gp, whereas that to methotrexate (MTX; MERCK, Darmstadt, Germany) depends on P-gp (21) and dihydrofolate reductase (DHFR) (22).
siRNA sequences and transfection. Two 21-mer siRNAs (herafter referred to as d1 and d3) specific for the MDR1 sequence were purchased from Dharmacon (Chicago, IL, USA). The anti-MDR1 siRNA d1 sequence (5’-GAAACTGCCTCATAAATTTTT-3’) is localized 1960 bp downstream of the start codon of human MDR1 mRNA. The anti-MDR1 siRNA d3 sequence (5’-GCTGATCTATG CATCTTATTT-3’) is localized 1330 bp downstream of the initiation codon of the MDR1 mRNA. A non-targeted irrelevant siRNA (Dharmacon) herafter referred to as siIrr was used as the negative control.
Tumor cells were plated on tissue culture dishes and grown for 24 h before transfection was performed using a lipofectant (Dreamfect, Oz Biosciences, Marseille, France) according to the manufacturer's instructions. Anti-MDR1 siRNAs d1 or d3 were transfected at the optimal concentration previously established for each cell line and defined as the concentration of siRNA enabling the best inhibition of MDR1 expression without showing cell toxicity (160 nM for MNNG/HOS, 293T, Mda-Mb549 and 200 nM for Cal51 and Sk-N-Be). siRNA toxicity was assayed 24, 48 and 72 h after transfection with the corresponding siRNA and compared to that induced by transfection agent alone.
Detection of MDR1 expression by quantitative reverse transcription PCR. MDR1 mRNA expression was evaluated by quantitative reverse transcription PCR (Q-RT-PCR). Cell lines were seeded in 6-well plates (1×105 cells/well) 24 h before transfection with the anti-MDR1 siRNA d1 or d3. Chemotherapeutic agent was added (Dox or MTX) to cell cultures 24 h after transfection at their determined IC50. Tumor cells were harvested 48 or 72 h after the addition of chemotherapy to evaluate the effects of anti-MDR1 siRNAs on MDR1 mRNA expression. All molecular biology regents and primers were from Qiagen (Courtaboeuf, France). Total RNA was extracted using the RNeasy Plus Kit and cDNA was synthesized using the reverse transcription kit according to the manufacturer's instructions. Q-RT-PCR was performed on an ABPrism 7000 instrument (Applied Biosystems, Foster City, CA, USA) with QuantiFast SYBRGreen and primers specific for MDR1 and hypoxanthine phosphoribosyl transferase (HPRT). Amplicons were validated by the specificity of the melting curve and by electrophoresis on Qiaxcel (Qiagen).
Cell growth assays. To evaluate whether anti-MDR1 siRNAs might restore chemosensitivity and induce a decrease in cell proliferation, cells were seeded in 96-well plates, transfected and treated with chemotherapeutic agent 24 h after transfection. Chemotherapy lasted for 72 h. To assess whether the efficacy of siRNA could be increased, multiple transfection assays were performed on the osteosarcoma cell line (MNNG/HOS) and the neuroblastoma cell line (Sk-N-Be). For this particular assay, cells were transfected cells with siRNA d3 every 3 days. Chemotherapeutic agent was added (at its IC50 concentration) after each transfection as described above. The proliferation assay was conducted after a period of 12 days and as described below.
For all assays, cell proliferation was assessed on a daily basis using a tetrazolium salt based-assay (Cell Counting Kit-8, Dojindo, Gaithersburg, MD, USA) according to the manufacturer's instructions. Cell absorbance was read at 450 nm on an Infinite F500 TECAN reader (TECAN, Männedorf, Switzerland) after 3 h of incubation at 37°C. Inhibition of cell proliferation under each condition was normalized against the proliferation of control cells (i.e. cells not transfected but treated with the corresponding chemotherapeutic agent).
Statistical analysis. All treatments were performed in duplicate and three time. All results were expressed as the mean±standard deviation (SD). Statistical significance was tested by Student's t-test using SigmaPlot software (Systat Software Inc., Chicago, IL, USA). A p-value less than 0.01 was considered significant.
Results
Tumor cell lines express different levels of MDR1 mRNA and P-gp. Basal MDR1 mRNA expression determined by Q-RT-PCR ranged from weak to high, with the highest expression level being found in the MNNG/HOS osteosarcoma and in the Cal51 mammary carcinoma cell lines, with relative expression values of 0.96 and 0.99, respectively (Figure 1B). These results were supported by Western blot analysis showing a higher expression of the protein in these two cell lines (data not shown). In contrast, the neuroblastoma cell line (Sk-N-Be) showed a weak expression of MDR1 mRNA (relative expression value of 0.75) (Figure 1B) and expression of P-gp was barely detected by Western blot.
Anti-MDR1 siRNAs inhibit P-gp expression in chemoresistant sarcoma and carcinoma cell lines. Whereas no inhibition of the MDR1 gene expression was observed when exposing cells to siIrr (Figure 1B), exposing tumor cells to either siRNA d1 or d3 reduced the level of MDR1 mRNA (Figure 1A, 1B). Overall, the siRNA efficacy was similar for d1 and d3 at inhibiting MDR1 mRNA expression. Anti-MDR1 siRNA d3 induced a 50% inhibition of the MDR1 mRNA expression as assessed by Q-RT-PCR 48 h after transfection in the MNNG/HOS osteosarcoma cell line (Figure 1B). This inhibition was maintained 72 h after transfection and was confirmed by electrophoresis (Figure 1A). In the Cal51 mammary carcinoma cell line, anti-MDR1 siRNA d3 caused a 65% inhibition in MDR1 mRNA, 48 h after transfection (Figure 1B). Conversely, both anti-MDR1 siRNAs failed to induce any decrease in MDR1 mRNA levels in the Sk-N-Be neuroblastoma cell line.
Similarly to situations observed in vivo in patients, we saw acquired drug resistance in all tumor cell lines with an increase in their MDR1 mRNA content upon exposure to Dox or MTX (23, 24). The highest MDR1 mRNA level upon exposure to chemotherapy was observed in the MNNG/HOS osteosarcoma cell line in which Dox increased MDR1 mRNA levels by 30% (Figure 1C). In this particular condition, anti-MDR1 siRNAs had to compete with the increased concentration of mRNA and, as a consequence, the effect of anti-MDR1 siRNA was lowered and a basal level of mRNA expression was maintained (Figure 1C). Despite this chemotherapy-induced increase in MDR1 mRNA concentrations, anti-MDR1 siRNA d3 was more efficient than d1; d3 lowered MDR1 mRNA by 30% compared to control tumor cells treated with Dox only and d1 reduced MDR1 mRNA by 20% compared to control tumor cells treated with Dox (Figure 1C).
Anti-MDR1 siRNAs inhibit cell proliferation in chemoresistant sarcoma and carcinoma cell lines. Both siRNAs were able to restore chemosensitivity in the osteosarcoma and carcinoma cell lines. Significant inhibition of cell proliferation was observed after exposure to Dox or MTX when osteosarcoma cells were transfected with anti-MDR1 siRNA d1 or d3 (Figure 2 and Table I). The most significant inhibitions obtained for all cell lines are shown in Table I. siRNA d3 showed a higher capacity for inhibiting cell proliferation, even in the presence of chemotherapeutic agents, as compared to anti-MDR1 siRNA d1 (Table I). The maximal inhibition was observed in the MNNG/HOS osteosarcoma cell line (Table I). Inhibition of cell proliferation was dependent upon the concentration of siRNA, ranging from 20% inhibition with 40 nM anti-MDR1 siRNA d3 to 90% (p<0.01) with 160 nM (Figure 2). Dox was more efficient than MTX at inducing tumor cell death when combined with anti-MDR1 siRNA: 84% (p<0.01) inhibition was obtained with Dox compared to 70% with MTX (p<0.01) (data not shown). This higher efficacy of siRNA in inhibiting cell proliferation in the presence of Dox rather than in the presence of MTX could be due to the implication of other proteins in addition to MDR1 in resistance to MTX (25). All tumor cell lines tested showed dependence upon doses and schedule of siRNA transfection and chemotherapy exposure.
Anti-MDR1 siRNAs reduce the IC50 of Dox and MTX. When exposed to anti-MDR1 siRNAs, chemoresistant osteosarcoma and carcinoma cell lines became significantly more sensitive to chemotherapeutic agents as early as 48 h after siRNA transfection. In the MNNG/HOS osteosarcoma cell line, transfection with the anti-MDR1 siRNA d3 resulted in a 10-fold decrease of Dox IC50 value (from 0.5 μM to 0.05 μM) (Figure 3) and a 15-fold reduction in MTX IC50 (from 150 μM to 10 μM) (p<0.01) (Figure 3). Similar results were observed for the carcinoma cell lines Mda-Mb-549 and Cal51. IC50 values after siRNA exposure in other tumor cell lines tested are shown in Figure 4. As expected by its lower MDR1 gene expression, no improvement in the IC50 value of chemotherapeutic agents was observed for the neuroblastoma cell line Sk-N-Be after transfection with anti-MDR1 siRNAs (Figure 4).
Single vs. multiple transfections. A single transfection by anti-MDR1 siRNA was able to significantly reduce the IC50 chemotherapeutic agents and inhibit cell proliferation but only over a short period of time. To investigate whether multiple treatments with anti-MDR1 siRNA would completely inhibit MDR1 gene expression and cell proliferation, or if there were anti-MDR1 siRNA-resistant clones, multiple transfections were performed. MNNG/HOS was transfected three times over a period of 12 days by siRNA d1 or d3 (transfections were performed three days apart from each other). Chemotherapeutic agent was administered after each transfection as presented. After three cycles of transfection/chemotherapy exposure, cell proliferation was similar to the basal level. After the second and third transfections, cell proliferation showed the same profile as after the first transfection: proliferation decreased over 96 h and then returned to the basal level.
Cell proliferation decreased until day 5 of 80% after one treatment and it increased slowly until day 11 (Figure 4B). The same observation was made after two or three transfections: five days after treatment the cell proliferation increased slowly to be inhibited by only 20% at day 11. After each new cycle of transfection/chemotherapy treatment the inhibition of cell proliferation reached the same level.
For the neuroblastoma cell line, Sk-N-Be, no difference in cell proliferation was observed between day 1 and day 12 (i.e. between the first and third transfection/chemotherapy cycles; cell proliferation was always inhibited by about 10% irrespective of the number of transfections and the siRNA with which cells were transfected (siRNA d3 or siIrr). These results show that the P-gp efflux pump was not inhibited by multiple transfections of MDR1 siRNA in the neuroblastoma cell line.
MDR1 expression level was analyzed in the osteosarcoma cell line (MNNG/HOS) and the neuroblastoma cell line after three cycles of transfection/chemotherapy. For a single transfection, we observed a 40% decrease in MDR1 mRNA expression two days after transfection. This inhibition lasted until day 4 then the MDR1 mRNA expression returned to its basal level by day 6. The same decrease in MDR1 mRNA was obtained after three cycles of transfections; each transfection cycle induced an inhibition of 40% for three days and the mRNA level returned to its basal level six days after transfection (Figure 4A). For the neuroblastoma cell line, a decrease in MDR1 mRNA expression was not observed. These results demonstrated the existence of anti-MDR1 siRNA-resistant clones in spite of multiple transfections.
These results showed that in vitro, the efficacy of siRNA is limited in time: four days after transfection cell proliferation and gene expression returned to their basal levels. The efficacy of anti-MDR1 siRNA is also limited in dose; the multiple transfections did not decrease the level of MDR1 gene expression, or increase the effect of siRNA and drugs on cell proliferation.
Discussion
Inherent or acquired overexpression of P-gp has been observed in different types of cancer and has been linked in part to chemoresistance towards conventional therapeutic approaches used in the treatment of cancer. In this study, we evaluated whether two siRNAs specific for MDR1 mRNA could reverse the chemoresistance phenotype in sarcoma and carcinoma cell lines known to overexpress MDR1. Our results showed that anti-MDR1 siRNA significantly inhibited cell proliferation in the presence of Dox or MTX, two cytotoxic agents whose efficiency is known to be dependent on MDR1 expression. This decrease in cell proliferation was correlated with a decrease of MDR1 mRNA level. Furthermore, when tumor cells were treated by chemotherapy after being transfected by anti-MDR1 siRNA, an increased efficacy of the considered drugs was obtained as shown by their lower IC50 values. These results showed that anti-MDR1 siRNA reduced innate chemoresistance and prevented acquired chemoresistance.
siRNA targeting MDR1 mRNA was more efficient at restoring chemosensitivity to Dox than that to MTX. This data is consistent with the respective mechanisms involved in chemoresistance patterns of these drugs. Dox chemoresistance is only mediated by P-gp, while the mechanism of chemoresistance to MTX is more complex and involves P-gp as well as other transporters, among which a folate carrier, DHFR, is of major importance. Tumor cells deficient in this enzyme or with mutated DHFR use P-gp to transport MTX (22). Thus the implication of P-gp in chemoresistance to MTX is not as important as in that to Dox and even appears controversial (8, 25). In various osteosarcoma cell lines (including MNNG/HOS), cyclosporine, an inhibitor of P-gp, proved inefficient at changing the IC50 of MTX (8). Using anti-MDR1 siRNA to transfect MNNG/HOS cells, we showed a 10-fold decrease in the IC50 of MTX. Norris et al. demonstrated in human leukemia cell lines that the inhibition of P-gp using a monoclonal antibody was able to partially restore the sensitivity of the cells to MTX (25). These observations and our results are in favor of the significant role of MDR1 in MTX chemoresistance even though inhibition of one of the two transporter pathways may clearly not be sufficient to fully restore cell chemosensitivity toward this drug. Partial reversion of chemoresistance by inhibition of MDR1 may be compensated by overexpression of alternative transporter pathways such as DHFR (48). Alternatively, inhibitors with low specificity towards P-gp, such as cyclosporine, may have similar efficacy to molecular ones (e.g. siRNA) or immune inhibitors (e.g. monoclonal antibodies) which are more specific to P gp.
Factors influencing siRNA efficacy include the secondary structure of the mRNA target and the subsequent stability of the complex formed by siRNA and mRNA (26, 27). The better efficacy of the anti-MDR1 siRNA d3 in restoring chemosensitivity over siRNA d1 may be due to its enriched content in GC motifs compared to d1, which may have led to more stable interactions between the siRNA and the MDR1 mRNA. The transient transfection used in our experiments resulted in a transitory restoration of chemosensitivity, with a maximum effect observed 48 h post-transfection. It is well established that short hairpin RNAs (shRNA) are more stable than siRNAs and slightly more efficient than transient transfection in inhibiting gene expression. Thus, they may be a better option than siRNA (28). Unless targeted delivery to tumor cells can be achieved in vivo, the capacity of shRNA to integrate into the host's genome may represent a serious drawback. Non-specific integration of anti-MDR1 shRNA in liver or kidney cells might lead to permanent inactivation of P-gp and accumulation of toxins. Thus, in our view, transient blockage of MDR1 activity via siRNA is preferable for future clinical applications.
The improvement in overall survival in many cancer types requires reducing long-term side-effects of cytotoxic treatment, particularly, but not only, in children. Reducing the IC50 of the cytotoxic compound is essential to improve its therapeutic ratio and thus its immediate and long-term tolerance. Vinca alkaloids are essential to the treatment of B-cell malignancies but carry significant side-effects, including neurotoxicity, which can limit their clinical application. Blocking the ABC transporter pathways using verapamil led to restoration of chemosensitivity to vincristine and vinblastine and a 30-fold reduction in the IC50 of vincristine in a leukemia model (10). Unfortunately, clinical trials using verapamil to restore chemosensitivity to doxorubicin led to unacceptable toxicity in non-specific organs (29). Valspodar, a non-immunosuppressive cyclosporine derivative targeting the ABC transporter pathway, can sensitize canine osteosarcoma to doxorubicin in vivo and reduce doses infused by 30% (30). Although they do not generate the risks of immunosuppression, chemosensitizers interact with P-gp in normal tissue. But P-gp is expressed more in cancer cells than in normal tissue (9), except in liver or kidney, hence inhibition of P-gp in these organs can be toxic (31).
The multiple transfections showed that the addition of siRNA does not lead to an increase of efficacy of the siRNA in vitro. In vivo application of siRNAs has been hampered by their rapid degradation by nucleases. An anti-MDR1 siRNA cloned into an expression plasmid showed an inhibitory effect for only three days after intravenous administration in normal mice and in a xenograft model of prostate cancer (32, 33). Increasing the molecular stability of siRNA is the key to any future clinical application and this may be facilitated by encapsulation into particles such as liposomes (34, 35). Subsequent engineering of these liposomes may also prove essential for specific tumor targeting (36). Cancer cells can also be targeted with Regioselectively Addressable Functionalized Template-arginine-glycine-aspartic acid (RAFT-RGD) owing to the RGD motif which allows the recognition of the αvβ3 integrins (37) overexpressed in the neovascular endothelium.
Conclusion
Inhibition of MDR1 is an interesting approach to reverse inherent or acquired resistance to critical cytotoxic compounds for the treatment of cancer. Beyond restoration of chemosensitivity, targeting MDR1, thus allowing lowering of the doses of chemotherapy while maintaining tumor control, represents a desirable goal. This is particularly relevant for pediatric tumors, such as osteosarcoma, where chemotherapy cures should be associated with minimum if no impact at all on long-term quality of life. We demonstrate here that targeting the MDR1 resistance pathway using siRNA is efficient at restoring chemosensitivity towards Dox and MTX, two key chemotherapeutic agents of multiregimen osteosarcoma treatments. Prior to any clinical application, experiments are ongoing to assess the feasibility of linking these siRNAs to vectors specifically targeting the tumor cells.
Acknowledgements
Jennifer Perez was supported by a Cible grant from the region Rhône Alpes, France.
- Received April 22, 2011.
- Revision received June 16, 2011.
- Accepted June 17, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved