Reversal of P-glycoprotein–Mediated Multidrug Resistance by the Murine Double Minute 2 Antagonist Nutlin-3

Murine double minute 2 (MDM2) negatively regulates the activity of the tumor suppressor protein p53. Nutlin-3 is a MDM2 inhibitor under preclinical investigation as nongenotoxic activator of the p53 pathway for cancer therapy (1) that also exerts antiangiogenic effects (2). Although the anticancer activity of nutlin-3 initially seemed to be limited to cells harboring wild-type p53 (1), recent results revealed that nutlin-3 also exerted anticancer effects in p53-negative and p53-mutant human tumor cells through E2F1 binding to MDM2 (3, 4).

In this report, nutlin-3 was investigated in combination with cytotoxic drugs in the pediatric cancer entities neuroblastoma and rhabdomyosarcoma. In p53 wild-type neuroblastoma cells, nutlin-3 caused apoptosis and/or neuronal differentiation (5) and increased sensitivity to chemotherapy-induced apoptosis (6). Information about the activity of nutlin-3 in rhabdomyosarcoma is lacking. Because nutlin-3 was shown to also exert anticancer effects that are independent of MDM2 inhibition (3), p53-mutated cells were included. Our most striking finding is that nutlin-3 sensitizes p53-mutated multidrug-resistant cancer cells to chemotherapy by interference with the ATP binding cassette (ABC) transporters P-glycoprotein [P-gp; encoded by multidrug resistance gene 1 (MDR1/ABCB1)] and multidrug resistance protein 1 (MRP-1/ABCC1).

Cell lines. The neuroblastoma cell lines UKF-NB-4 and UKF-NB-3 as well as the UKF-NB-3 sublines adapted to doxorubicin (20 ng/mL; UKF-NB-3^rDOX²⁰) or vincristine (10 ng/mL; UKF-NB-3^rVCR¹⁰) were established as described before (7, 8). Be(2)-C cells were obtained from the American Type Culture Collection (ATCC). The alveolar rhabdomyosarcoma cell line UKF-Rhb-1 was established in our laboratory from a bone marrow metastasis. The alveolar rhabdomyosarcoma cell line Rh30 was kindly provided by Dr. P.J. Houghton (St. Jude's Children's Research Hospital, Memphis, TN). The vincristine-resistant rhabdomyosarcoma cell lines UKF-Rhb-1^rVCR¹⁰ and Rh30^rVCR¹⁰ were established by adaptation of UKF-Rhb-1 or Rh30 to growth in the presence of vincristine 10 ng/mL.

P53 status and P-gp expression status of the cell lines UKF-NB-3, UKF-NB-3^rDOX²⁰, UKF-NB-3^rVCR¹⁰, UKF-NB-4, Be(2)-C, UKF-Rhb-1, UKF-Rhb-1^rVCR¹⁰, Rh30, and Rh30^rVCR¹⁰ are shown in Table 1 (refs. 8–11; data for UKF-Rhb-1, UKF-Rhb-1^rVCR¹⁰ detected as described below).

The cell lines MDCKII MDR1 (12) and MDCKII MRP1 (12) and the nontransfected control cell line MDCKII were kind gifts of Drs. P. Borst, M. de Haas, and A.H. Schinkel (Nederlands Kanker Instituut, NKI, Amsterdam, the Netherlands). P53-negative, P-gp–negative PC-3 prostate cancer cells expressing low MRP-1 levels were received from ATCC. Adaptation of PC-3 to growth in the presence of vincristine 10 ng/mL resulted in a cell line that shows MRP-1 overexpression but no P-gp expression (PC-3^rVCR¹⁰; data not shown).

All cell lines were propagated in Iscove's modified Dulbecco's medium supplemented with 10% fetal bovine serum, 100 IU/mL penicillin, and 100 mg/mL streptomycin at 37°C.

Mutation analysis of p53. TP53 gene sequencing on cDNAs was performed using the following four pairs of primers: TP53 Ex2-3-f GTGACACGCTTCCCTGGAT and TP53 Ex2-3-r TCATCTGGACCTGGGTCTTC; TP53 Ex4-5-f CCCTTCCCAGAAAACCTACC and TP53 Ex4-5-r CTCCGTCATGTGCTGTGACT; TP53 EX6-7f GTGCAGCTGTGGGTTGATT and TP53 Ex6-7r GGTGGTACAGTCAGAGCCAAC; Tp53 Ex8-9-f CCTCACCATCATCACACTGG and TP53 Ex8-9-r GTCTGGTCCTGAAGGGTGAA. In addition, all cell lines were examined for TP53 mutations by sequence analysis of genomic DNA as described previously (13). PCR was performed as described before (8, 13). Each amplicon was sequenced bidirectionally.

Drugs. Vincristine was obtained from Sigma-Aldrich. Doxorubicin was from Cell Pharm. Nutlin-3, rhodamine 123, calcein-AM, and JC-1 were purchased from Merck Biosciences.

Viability assay. Cell viability was tested by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye reduction assay after 96-h incubation modified as described before (14).

Real-time PCR. Total RNA was isolated from cell cultures using the TRI reagent (Sigma-Aldrich). Reverse transcription and real-time PCRs were performed as described before (15). The following primer sequences were used: 18s rRNA forward 5′-GTGAAACTGCGAATGGCTCAT, 18s rRNA reverse 5′-CTGACCGGGTTGGTTTTGAT, 18s rRNA probe 5′-(VIC) TGGTCGCTCCTCTCCCAC-(TAMRA), MDM2 forward 5′-TGTTGGTGCACAAAAAGACA, MDM2 reverse 5′-CACGCCAAACAAATCTCCTA, p21 forward 5′-GCCCGTGAGCGATGGAA, p21 reverse 5′-ACGCTCCCAGGCGAAGTC, GADD45 forward 5′-GCACGCCGCGCTCTCT, GADD45 reverse CTTATCCATCCTTTCGGTCTTCTG. The results are presented as fold change.

Flow cytometry and investigation of efflux of ABC transporter substrates. Antibodies directed against P-gp (Alexis Biochemicals via AXXORA Deutschland), MRP-1, MRP-3, or breast cancer resistance protein (BCRP/ABCG2, all three from Kamiya Biomedical Company) were used to detect protein expression by flow cytometry.

To investigate ABC transporter–mediated substance efflux, cells were incubated with rhodamine 123 (0.1 μmol/L, P-gp, MRPs), JC-1 (0.1 μmol/L, P-gp), or calcein-AM (2 μmol/L, MRPs) for 60 min. Then, cells were washed with PBS and incubated for another 60 min to allow ABC transporter–mediated drug efflux. Subsequently, cellular fluorescence was analyzed by flow cytometry (FACSCalibur, Becton Dickinson). Rhodamine 123 was detected at the FL1 channel; JC-1 and calcein-AM were detected at the FL2 channel.

Determination of P-gp ATPase activity. The P-gp ATPase activity was determined using the Human PGP-ATPase Assay (BD Biosciences) following the manufacturer's instructions.

Viral transduction. Standard molecular cloning techniques were used to generate lentiviral vectors based on Lentiviral Gene Ontology (LeGO) vector technology⁵

Separation of enantiomers. Nutlin-3 was separated in its enantiomers nutlin-3a and nutlin-3b by chiral reversed-phase chromatography (Supplementary Fig. S2).

Influence of nutlin-3 on cancer cell viability. Nutlin-3 sensitivity markedly differed between p53 wild-type (<3 μmol/L) and p53 mutated cells (>17 μmol/L) after a 96-hour incubation (Table 1; Fig. 1A). Moreover, nutlin-3 (20 μmol/L) treatment induced expression of p53 target genes (p21, MDM2, GADD45) in p53 wild-type cells but not in p53-mutated cells as shown for UKF-NB-3, UKF-NB-3^rVCR¹⁰, UKF-Rhb-1, and UKF-Rhb-1^rVCR¹⁰ cells by real-time PCR (Fig. 1B and C).

Efficacy of nutlin-3 in combination with anticancer drugs. Nutlin-3 was already shown to increase the cytotoxic activity of anticancer drugs (1). In UKF-NB-3 (p53-wild-type, P-gp-negative) cells, nutlin-3 and vincristine induced additive cytotoxicity (Fig. 2A). In sharp contrast to the results obtained with UKF-NB-3, nutlin-3 concentrations that were nontoxic to UKF-NB-3^rVCR¹⁰ cells (p53-mutated, high P-gp expression) dramatically increased the vincristine sensitivity of these cells (Fig. 2A). The combination of nutlin-3 and vincristine in concentrations that alone did not affect UKF-NB-3^rVCR¹⁰ cell viability caused complete destruction of all UKF-NB-3^rVCR¹⁰ cells (Fig. 2A and C). Similar results were obtained for the comparison of UKF-Rhb-1 cells (p53 wild-type, P-gp-negative) and UKF-Rhb-1^rVCR¹⁰ (p53-mutated, high P-gp expression) cells (Fig. 2B and C). Investigation of further cell lines revealed that nontoxic or moderate toxic nutlin-3 concentrations only slightly decreased the IC₅₀ concentration of vincristine in p53 wild-type and P-gp–negative cells (<2-fold; UKF-NB-3, UKF-Rhb-1), but massively decreased the IC₅₀ of vincristine in p53-mutated and P-gp–positive cells [92- to 3,434-fold; UKF-NB-3^rVCR¹⁰, UKF-Rhb-1^rVCR¹⁰, UKF-NB-4; Be(2)-C; Table 1]. Moreover, nutlin-3 (10 μmol/L) induced a 3-fold reduction in the IC₅₀ for vincristine in p53-mutated, P-gp negative Rh30 cells but caused a 721-fold reduction in the IC₅₀ for vincristine in the P-gp–expressing vincristine-resistant subline Rh30^rVCR¹⁰ (Table 1). These findings suggest that nutlin-3 may interfere with P-gp–mediated vincristine efflux. In the p53 wild-type and P-gp–overexpressing cell line UKF-NB-3^rDOX²⁰, the IC₅₀ for vincristine was decreased by 28.9-fold by nutlin-3 (1 μmol/L; Table 1). The reduced effect on vincristine sensitivity compared with other P-gp–overexpressing cell lines may be explained by the relatively low nutlin-3 concentration that had to be applied due to the higher nutlin-3 sensitivity of this p53 wild-type cell line.

Moreover, UKF-NB-3^rVCR¹⁰ or UKF-Rhb-1^rVCR¹⁰ cells were treated with nutlin-3 in combination with cytotoxic drugs with different modes of action that are structurally unrelated to vincristine. Nutlin-3 (20 μmol/L), a concentration that did not significantly affect UKF-NB-3^rVCR¹⁰ viability (96 ± 8% cell viability relative to untreated control), strongly sensitized UKF-NB-3^rVCR¹⁰ cells to the P-gp substrates doxorubicin, paclitaxel, etoposide, mitomycin C, and actinomycin D but not to cisplatin that is known not to be a P-gp substrate (Fig. 2D). Nutlin-3 (20 μmol/L) substantially decreased the IC₅₀s of doxorubicin (IC₅₀: 110 ± 13 ng/mL; IC₅₀ + nutlin-3: 9 ± 1 ng/mL; fold decrease: 12.1), paclitaxel (IC₅₀: 51 ± 5 ng/mL; IC₅₀ + nutlin-3: 1.5 ± 0.2 ng/mL; fold decrease: 33.6), etoposide (IC₅₀: 320 ± 47 ng/mL; IC₅₀ + nutlin-3: 16 ± 2 ng/mL; fold decrease: 19.9), mitomycin C (IC₅₀: 34 ± 3 ng/mL; IC₅₀ + nutlin-3: 4.6 ± 0.7 ng/mL; fold decrease: 7.5), or actinomycin D (IC₅₀: 22.3 ± 2.1 ng/mL; IC₅₀ + nutlin-3: 0.34 ± 0.03 ng/mL; fold decrease: 65.7) in UKF-NB-3^rVCR¹⁰ cells but did not significantly affect sensitivity to cisplatin (IC₅₀: 178 ± 24 ng/mL; IC₅₀ + nutlin-3: 145 ± 22 ng/mL; fold decrease: 1.2). Similar results were obtained in UKF-Rhb-1^rVCR¹⁰ cells (data not shown).

To investigate the effects of nutlin-3 in P-gp–expressing and non–P-gp-expressing cells that share the same genetic p53 background, p53-mutated, P-gp–negative Rh30 cells were transfected with lentiviral MDR1 vectors (see Supplementary Fig. S1). Transfection with empty vectors did not change the sensitivity of Rh30 (Rh30^{empty vector}) cells to nutlin-3, vincristine, or the combination of vincristine and nutlin-3 (Table 1). Rh30 cells transfected with MDR1 (Rh30^MDR1) showed decreased sensitivity to vincristine that was reverted by addition of nutlin-3.

Influence of nutlin-3 on the ABC transporter–mediated efflux of fluorescence substrates. The fluorescent P-gp substrate rhodamine 123 was used to investigate the influence of nutlin-3 on P-gp–mediated substance efflux. Rhodamine 123 (0.1 μmol/L) fluorescence was detected in P-gp–negative UKF-NB-3 cells but not in P-gp–overexpressing UKF-NB-3^rVCR¹⁰ cells (Fig. 3A). Nutlin-3 (20 μmol/L) treatment induced strong accumulation of rhodamine 123 in UKF-NB-3^rVCR¹⁰ cells (25-fold increase; Fig. 3A). In contrast, nutlin-3 (20 μmol/L) increased rhodamine 123 accumulation in UKF-NB-3 cells to a much lesser extent (1.7-fold increase; Fig. 3A). Noteworthy, nutlin-3–induced effects on rhodamine 123 accumulation were similar to those induced by verapamil that is known to inhibit P-gp being a P-gp substrate. Moreover, rhodamine 123 efflux inhibition by nutlin-3 occurred in a concentration-dependent manner in UKF-NB-3^rVCR¹⁰ cells (Fig. 3B). Investigation of the P-gp substrate JC-1 that is structurally not related to rhodamine 123 resulted in similar results: accumulation in UKF-NB-3 cells but not in UKF-NB-3^rVCR¹⁰ cells and strong inhibition of JC-1 efflux by nutlin-3 in UKF-NB-3^rVCR¹⁰ cells (data not shown).

MRP-1 is another family member of the ABC transporters that was shown to be involved in cancer cell chemoresistance (17, 18). P-gp plays a dominant role between the ABC transporters in UKF-NB-3^rVCR¹⁰ cells. No significant overexpression of MRP-3 or BCRP could be detected. Moreover, MRP-specific (MK571) or BCRP-specific (fumitremorgin C) inhibitors did not significantly affect efflux of different fluorescence substrates from UKF-NB-3^rVCR¹⁰ cells (data not shown). Accumulation of rhodamine 123 that is a substrate of P-gp and MRP-1 was measured in MDCKII cells stably overexpressing P-gp (MDCKII MDR1) or MRP-1 (MDCKII MRP1) with or without nutlin-3 treatment. Results indicated decreased rhodamine 123 accumulation in MDCKII MDR1 and in MDCKII MRP-1 cells. Moreover, nutlin-3 caused strong accumulation of rhodamine 123 in both cell lines (Fig. 3C). The P-gp inhibitor verapamil (5 μg/mL) showed comparable effects like nutlin-3 in MDCKII MDR1 cells and the specific MRP inhibitor MK571 (20 μmol/L) showed comparable effects like nutlin-3 in MDCKII MRP1 cells (data not shown). Nutlin-3 inhibited rhodamine 123 efflux from MDCKII MDR1 or MDCKII MRP1 cells in a concentration-dependent manner, showing still weak effects in a concentration of 1.25 μmol/L but no effect in a concentration of 0.625 μmol/L anymore (data not shown). Similar results were obtained by the use of the fluorescence dye calcein-AM in MDCKII MRP1 cells (data not shown). To investigate the influence of nutlin-3 on MRP-1 in a second model, nutlin-3 was examined in PC-3^rVCR¹⁰, the vincristine-resistant subline of the p53-negative PC-3 cell line, showing high MRP-1 expression but no P-gp expression. In PC-3^rVCR¹⁰ cells, nutlin-3 (20 μmol/L) decreased the IC₅₀ value for vincristine from 45.78 ± 3.87 ng/mL to 15.45 ± 2.21 ng/mL. The MRP inhibitor MK571 (20 μmol/L) reduced it to 11.57 ± 0.45 ng/mL. Moreover, nutlin-3 caused calcein-AM accumulation similarly to the MRP inhibitor MK571 (20 μmol/L) in PC-3^rVCR¹⁰ cells (Supplementary Fig. S3).

Nutlin-3 is transported by P-gp. The question if nutlin-3 may act as direct inhibitor of P-gp function or as P-gp substrate was addressed by two different approaches. First, the time kinetics of nutlin-3–mediated inhibition of rhodamine 123 efflux was investigated in UKF-NB-3^rVCR¹⁰ and MDCKII MDR1 cells. Whereas P-gp inhibitors commonly show prolonged activity even after a washing-out period, the activity of P-gp substrates rapidly declines after substance removal (19, 20). To examine this, cells were incubated with rhodamine 123 in the presence of nutlin-3 (20 μmol/L) or verapamil (5 μg/mL), which is a P-gp substrate, for 60 minutes. Then, cells were washed and rhodamine 123 fluorescence was detected after 0, 15, 30, 60, and 120 minutes. Results indicated that rhodamine 123 fluorescence rapidly decreases after wash-out of nutlin-3 or verapamil (Fig. 3D).

To further investigate if a substance may inhibit P-gp or act as P-gp substrate, the P-gp ATPase activity can be determined (19, 20). The P-gp ATPase activity is often stimulated in the presence of transported substrates such as verapamil. Determination of the P-gp ATPase activity revealed that nutlin-3 (20 μmol/L) and verapamil (5 μg/mL) stimulated P-gp ATPase activity to a similar extent and no antagonism between verapamil and nutlin-3 in stimulation of P-gp ATPase activity (Fig. 3D). Taken together, these results suggest that nutlin-3 acts as a P-gp transport substrate.

Influence of the nutlin-3 enantiomers nutlin-3a and nutlin-3b on P-gp. To investigate if the nutlin-3 enantiomers nutlin-3a and nutlin-3b exert differential effects on P-gp, their influence was tested on UKF-NB-3^rVCR¹⁰ cell viability in combination with vincristine (Supplementary Table S1), on rhodamine 123 efflux in UKF-NB-3^rVCR¹⁰ cells, and on P-gp ATPase activity (Supplementary Fig. S4). Results revealed that both enantiomers showed similar effects on P-gp function.

In this report, the MDM2 inhibitor nutlin-3 was investigated alone or in combination with chemotherapeutic drugs in a panel of chemosensitive and chemoresistant neuroblastoma and rhabdomyosarcoma cell lines. In concordance with previous studies (5), nutlin-3 was much more effective in p53 wild-type than in p53-mutated neuroblastoma cells. Moreover, we could show that the anticancer activity of nutlin-3 also includes p53 wild-type rhabdomyosarcoma cells. A previous report had shown that nutlin-3 sensitizes p53 wild-type neuroblastoma cells to chemotherapy-induced cell death (6). We found that nutlin-3 interferes with P-gp function being a P-gp substrate. Nontoxic nutlin-3 concentrations strongly decreased the IC₅₀ for vincristine (and other P-gp substrates structurally nonrelated to vincristine) in different p53-mutated, P-gp–expressing cell lines (92- to 3,434-fold).

In addition to P-gp, other ABC transporters may play a role in cancer cell chemoresistance (17, 18). Our results indicate that nutlin-3 also interferes with MRP-1 activity. Systematic analysis of the influence of nutlin-3 on the activity of different ABC transporters is warranted. Notably, nutlin-3 is a racemate with nutlin-3a being the enantiomer that inhibits MDM2, whereas nutlin-3b does not (1). Both enantiomers similarly interfere with P-gp, indicating that the underlying structure-activity relationships differ from those defined for MDM2 inhibition.

In conclusion, our data show that nutlin-3 interferes with P-gp– or MRP-1–mediated drug efflux in multidrug-resistant cancer cells, representing a novel anticancer mechanism of this substance. This newly discovered activity requires further consideration during the development of this drug. Moreover, previous findings about synergisms between nutlin-3 and cytotoxic drugs may be reanalyzed concerning the role of ABC transporters. Because overexpression of ABC transporters and namely P-gp belongs to the most important chemoresistance mechanisms, nutlin-3 may be a new lead structure for the design of ABC transporter inhibitors.

J. Cinatl, H.W. Doerr, and M. Michaelis have filed a patent application for the use of nutlin-3 as an ABC-transporter inhibitor. The other authors declared no potential conflicts of interest.

Grant support: Hilfe für krebskranke Kinder Frankfurt e.V. and their foundation Frankfurter Stiftung für krebskranke Kinder (M. Michaelis, F. Rothweiler, D. Klassert, H.W. Doerr, and J. Cinatl). The work was in part supported by the EU-funded project SYNLET (contract No. 043312).

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 the friendly society “Hilfe für krebskranke Kinder Frankfurt e.V.” and its foundation “Frankfurter Stiftung für krebskranke Kinder” for support; Eva Bechtold, Janette Spitznagel, Lena Stegmann, and Jochen Meyer for their skillful technical assistance; Dr. U. Delabar, Institute of Pharmacology and Toxicology, Department of Clinical Pharmacology (University Hospital Tübingen), for her analytic advice and her assistance in sample collection of nutlin-3a and nutlin-3b; and Axel Schambach and Christopher Baum, MH Hannover, Germany, for providing the MDR1 cDNA.