Overexpression of ABCG2 confers resistance to pevonedistat, an NAE inhibitor
Liu-Ya Weia,b,1, Zhuo-Xun Wub,1, Yang Yanga,1, Min Zhaoa, Xiang-Yu Maa, Jin-Sui Lic,
Dong-Hua Yangb, Zhe-Sheng Chenb,∗, Ying-Fang Fanb,c,∗∗
a School of Pharmacy, Weifang Medical University, Weifang, 261053, China
b Department of Pharmaceutical Sciences, College of Pharmacy and Health Sciences, St. John’s University, Queens, NY 11439, USA
c Department of Hepatobiliary Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou 510282, China

A R T I C L E I N F O

Keywords:
ABCG2
Multidrug resistance MLN4924
Pevonedistat
NEDD8 activating enzyme inhibitor

A B S T R A C T

Pevonedistat is a potent, selective, first-in-class NEDD8 activating enzyme inhibitor. It is now under multiple clinical trials that investigate its anticancer effect against solid tumors and leukemia. ATP-binding cassette (ABC) transporters are membrane proteins that are involved in mediating multidrug resistance (MDR). In this article, we reveal that pevonedistat is a substrate of ABCG2 which decreases the therapeutic effect of pevonedistat. The cytotoXicity of pevonedistat was significantly weakened in ABCG2-overexpressing cells, and the drug resistance can be reversed by ABCG2 inhibitors. The ATPase assay suggested that pevonedistat can stimulate ABCG2 ATPase activity in a concentration-dependent manner. Pevonedistat showed little effect on the expression level or subcellular localization of ABCG2 after 72 h treatment. Furthermore, a pevonedistat resistance cell line S1-PR was established and overexpressed ABCG2. Generally, our study provides evidence that ABCG2 can be a pro- minent factor leading to pevonedistat-resistance. Furthermore, ABCG2 may also be utilized as a biomarker to monitor the development of pevonedistat resistance during cancer treatment.

1. Introduction

Protein homeostasis is critical for many cellular processes such as cell division, signal transduction, as well as cell death [1]. The ubiquitin proteasome system (UPS) is a key regulator of the protein homeostasis, which governs the turnover of most short-lived proteins [2,3]. Neural precursor cell expressed, developmentally downregulated 8 (NEDD8) is a 9-kDa protein encoded by the NEDD8 gene [4]. NEDD8 is a member of the ubiquitin-like proteins (UBLs) that modulate the activity of ubi- quitin [5–8]. Recently, evidences are emerging that overactivated neddylation is associated to progression and prognosis of various types of human cancers [9]. Therefore, the inhibition of neddylation pathway is being characterized as a promising target for cancer treatment [10,11]. Pevonedistat, also known as MLN4924, is characterized as a potent, selective inhibitor of NEDD8-activating enzyme (NAE). It se- lectively inhibits neddylation modification by binding to the ATP binding site of NAE, thereby preventing the activation of NEDD8 pro- tein [7,12–14]. Preclinical studies of pevonedistat have shown potent anticancer effect against a wide range of human cancers [15,16]. To

date, multiple clinical trials have been initiated to investigate the clinical effect of pevonedistat. Ongoing clinical trials include combi- nation of pevonedistat with chemotherapy for liver cancer (NCT04175912) and non-small-cell lung cancer (NSCLC, NCT03965689, NCT03228186), pevonedistat alone or combined with azacitidine for acute myeloid leukemia (AML, NCT03862157, NCT03330821, NCT04090736). Taken together, pevonedistat may be a novel anticancer drug that has a promising therapeutic effect.
Drug resistance, which leads to decreased or diminished response in cancer patients, remains a major obstacle for cancer treatment. It has been reported that mutations in the UBA3 gene (Y352H or I310 N) confer resistance to pevonedistat in leukemia cell lines [17]. Another study reported that A171 mutation of UBA3 confers resistance to pe- vonedistat in colon cancer cell line HCT116 [18]. Although the in vivo data is lacking, it remains important to investigate possible mechanisms that may lead to pevonedistat resistance. In the past three decades, numerous studies have demonstrated that ABC transporters, especially ABCB1, ABCG2 and ABCC1, are significantly associated with clinical chemoresistance [19]. These members of the ABC transporter family

∗ Corresponding author.
∗∗ Corresponding author. Department of Hepatobiliary Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou 510282, China.
E-mail addresses: [email protected] (Z.-S. Chen), [email protected] (Y.-F. Fan).
1 These authors contributed equally to this study.

https://doi.org/10.1016/j.yexcr.2020.111858

Received 11 December 2019; Received in revised form 16 January 2020; Accepted 19 January 2020
0014-4827/©2020ElsevierInc.Allrightsreserved.

are membrane transporters that are able to extrude drugs from cancer cells and cause multidrug resistance (MDR) [20]. Some clinically uti- lized chemotherapeutic agents or small molecular tyrosine kinase in- hibitors (TKIs) were identified as substrates of ABC transporters, sug- gesting that ABC transporters may limit the efficacy of these drugs. For instance, vincristine and etoposide have been reported as overlapping substrates of ABCB1 and ABCC1 [21,22]. MitoXantrone, doXorubicin, and topotecan were identified as substrates of ABCG2 [23]. Apart from the conventional chemotherapeutic drugs, several TKIs such as im- atinib, nilotinib, gefitinib are also substrates of ABCG2 [24,25]. Of note, other than association with MDR, ABCG2 may also be a marker of chemoresistance of non-substrate drugs [26].
In this study, we demonstrated that overexpression of ABCG2 may significantly confer resistance to pevonedistat. Therefore, ABCG2 may be used as biomarker of drug resistance in cancer patients receiving pevonedistat treatment.

2. Materials and methods

2.1. Chemicals

Pevonedistat was purchased from ChemieTek (Indianapolis, IN). DMEM, penicillin/streptomycin, FBS, trypsin EDTA were purchased from Corning Incorporated (Corning, NY). Ko143 was purchased from Enzo Life Sciences (Farmingdale, NY). PBS and BSA were purchased from VWR chemicals, LLC (Solon, OH). The anti-BCRP antibody, clone BXP-21 (catalog number MAB4146) was bought from Millipore (Billerica, MA). GAPDH monoclonal antibody (GA1R) (catalog number MA5-15738), Alexa Fluor 488 conjugated goat anti-mouse IgG sec- ondary antibody were obtained from Thermo Fisher Scientific Inc (Rockford, IL). HRP-conjugated secondary antibody (catalog number 7076S) was purchased from Cell Signaling Technology Inc (Danvers, MA). Cisplatin, mitoXantrone, MTT dye, DMSO, formaldehyde, Triton X-100, and DAPI and all other chemicals were purchased from Sigma Chemical Co (St. Louis, MO).

2.2. Cell lines and cell culture

Human non-small-cell lung cancer (NSCLC) NCI-H460 cells and its mitoXantrone-selected ABCG2-overexpressing subline NCI-H460/ MX20, human colon cancer S1 cells and mitoXantrone-selected ABCG2 overexpressing subline S1-M1-80 cells were cultured in DMEM medium containing 10% FBS and 1% penicillin/streptomycin. S1-M1-80 cells and NCI-H460/MX20 cells were maintained in medium with 80 μM or

without pevonedistat for designated concentrations or time periods. Protein concentrations were determined using Pierce™ BCA Protein Assay Kit (Thermo Scientific, Rockford, IL). Equal amount of lysates was subjected to 10% SDS-PAGE and transferred onto PVDF mem- branes. Primary antibodies against ABCG2 (1:1000) or GAPDH (1:1000) were used to detect the corresponding proteins. HRP-con- jugated anti-mouse antibody was used as secondary antibody. Signals were detected using Pierce™ ECL Western Blotting Substrate (Thermo Scientific, Rockford, IL), and protein bands were quantified using ImageJ software.

2.5. Immunofluorescence assay

The IFA protocol was as previously described[31]. To carry out the assay, cells (1 × 105) were seeded into 24-well plates and incubated for 24 h. Pevonedistat was added and further incubated for the designated incubation period. FiXation was done with 4% formaldehyde for 15 min, follow by permeabilization with 0.25% Triton X-100 for 15 min then blocked with 6% BSA for 1 h. The cells were then stained with primary antibody against ABCG2 followed by Alexa Fluor 488 con- jugated anti-mouse IgG secondary antibody. The cells were further in- cubated with DAPI for 15 min to counterstain the nuclei. Microscopy was performed with a Nikon TE-2000S microscope (Nikon Instruments Inc., Melville, NY, USA).

2.6. Fluorescent drug accumulation assay

The accumulation assay was performed using a FACSort flow cyt- ometer and analyzed by FlowJo software[32]. Cells were trypsinized, centrifuged and resuspended in complete medium, then separated into different group with 3 × 105 cells per group. The fluorescent substrate mitoXantrone (5 μM) was added into each group with or without Ko143
or pevonedistat. The cells were further incubated for 2 h then pelleted
at 500 × g, followed by resuspension in 300 μl PBS prior to FACS analysis.

2.7. ATPase assay

The ABCG2-related ATPase activities were determined using PRE- DEASY ATPase Kits (TEBU-BIOnv, Boechout, Belgium) with modifica- tions as previous described[33]. Briefly, different concentrations of pevonedistat, topotecan, or topotecan plus pevonedistat, with or without Na3VO4- were added to the ABCG2 membrane suspension. The

20 nM of mitoXantrone[27,28], respectively. HEK293/pcDNA3.1,
HEK293/ABCG2 cells were cultured in medium with G418 (2 mg/ml). All aforementioned cell lines were maintained at 37 °C incubator with 5% CO2 humidified air. Cells were cultured in drug-free medium at least 2 weeks before assay.

2.3. Cytotoxicity assay

MTT assay was performed to determine the cytotoXicity of tested compounds as previous described[29,30]. Generally, 5000 to 6000 cells per well were seeded into 96-well plates. After 24 h incubation, drugs were diluted with culture medium and added into the designated wells. The cells were further incubated with drugs for 72 h. After treatment,
20 μl of MTT solution were added to each well in the plate which was incubated for 4 h at 37 °C. After incubation, the medium was discarded
and DMSO was added to dissolve the formazan. The absorbance was measured at 570 nm in an accuSkanTM GO UV/Vis Microplate Spec- trophotometer (Fisher Sci., Fair Lawn, NJ).

2.4. Western blot analysis

Total proteins were harvested after cells were treated with or

miXtures were incubated at 37 °C for 5 min and the reaction was in- itiated by the addition of 5 mM Mg2+ATP. After a 40-min incubation at 37 °C, the inorganic phosphate (Pi) released was determined color- imetrically. The changes of relative light units were determined by comparing Na3VO4- treated group with the corresponding pevonedistat
-treated groups.

2.8. Establishment of pevonedistat-resistant cell line

The resistant cell line, S1-PR, was created by continuously culturing S1 cells in medium containing increased concentrations of pevonedi- stat. The initial concentration of pevonedistat was 5 μM and stepwisely
increased to final concentration of 30 μM. The established S1-PR cells
were maintained in drug-free medium for at least 2 months before ex- periments.

2.9. Statistical analysis

Data are presented as mean ± S.D from at least three independent experiments. One-way ANOVA was performed, and results were con- sidered statistically significant at P < 0.05.

Fig. 1. CytotoXicity of pevonedistat in parental and ABCG2-, ABCB1-overexpressing cells. (A) Chemical structure of pevonedistat (B) Concentration-viability curves for NCI-H460 and NCI-H460/MX20 cells (C) Concentration-viability curves for S1 and S1-M1-80 cells (D) Concentration-viability curves for HEK293/pcDNA3.1, HEK293/ABCG2-R482, HEK293/ABCG2-G482 and HEK293/ABCG2-T482 cells. Data are expressed as mean ± SD, representative of three independent experiments.

3. Results

3.1. Pevonedistat showed attenuated cytotoxicity in cancer cells overexpressing ABCG2

The chemical structure of pevonedistat is shown in Fig. 1 A. To evaluate the cytotoXicity of pevonedistat, MTT assay was performed in NSCLC and colon cancer cell lines. As shown in Fig. 1B and C, both NCI- H460/MX20 cells (which overexpress wild-type ABCG2) and S1-M1- 80 cells (which overexpress G482 mutant ABCG2) were significantly resistant to pevonedistat compared to the corresponding parental cancer cell lines. The IC50 values and resistance-fold (RF) are listed in Table 1. Reversal experiments were carried out to confirm that ABCG2 transporter is the major contributor to pevonedistat resistance. By blocking the activity of ABCG2 transporter using Ko143, a potent ABCG2 inhibitor, we examined whether the drug resistance to pevo- nedistat could be reversed. As shown in Table 1, 5 μM of Ko143 was
sufficient to reverse the drug resistance and sensitize the ABCG2-
overexpressing cells to pevonedistat to the similar level as the

corresponding parental cell lines. The resistance of pevonedistat de- creased from 6.58-fold to 1.38-fold in NCI-H460/MX20 cells and from 11.49-fold to 3.59-fold in S1-M1-80 cells. Therefore, the cytotoXicity results suggested pevonedistat may be a potential substrate of ABCG2.

3.2. Pevonedistat showed attenuated cytotoxicity in transfected cells overexpressing ABCG2

To further confirm that ABCG2-overexpression cells are resistant to pevonedistat, HEK293 transfected cell lines were used to perform the MTT assay. The same trend was observed in HEK293 cells transfected with either empty vector, wild-type ABCG2 (R482) or mutant ABCG2 (G482 and T482) as shown in Fig. 1 D. The ABCG2-overexpressing cells showed significant resistance to pevonedistat as shown in Table 2, with resistance ranging from 33- to 45-fold. Similarly, co-treatment with ABCG2 inhibitor Ko143 was able to completely reverse the pevonedi- stat resistance. These data indicate that overexpression of ABCG2 can confer resistance to pevonedistat.

Table 1
The cytotoXicity of pevonedistat in cancer cells overexpressing ABCG2.
Cell line IC50 value ± SDa (μM, Resistance-fold)

MitoXantrone MitoXantrone + Ko143 5 μM Pevonedistat Pevonedistat + Ko143 5 μM
NCI-H460 0.069 ± 0.014 (1.00) 0.064 ± 0.011 (0.93) 8.075 ± 1.357 (1.00) 7.889 ± 2.308 (0.98)
NCI-H460/MX20 4.894 ± 0.990b (71.27)
0.172 ± 0.028 (2.50) 53.122 ± 9.196b (6.58)
11.157 ± 1.269 (1.38)
S1 0.046 ± 0.009 (1.00) 0.048 ± 0.03 (1.05) 10.727 ± 1.656 (1.00) 14.943 ± 5.872 (1.39)
S1-M1-80 31.11 ± 2.195b (675.29)
0.087 ± 0.010 (1.89) 123.30 ± 34.12b (11.49)
33.993 ± 4.520b (3.59)

a IC50 values are represented as mean ± SD of at least three independent experiments performed in triplicate.
b P < 0.05 versus the control group.

Table 2
The cytotoXicity of pevonedistat in transfected cells overexpressing ABCG2.
Cell line IC50 value ± SDa (μM, Resistance-fold)

MitoXantrone MitoXantrone + Ko143 5 μM Pevonedistat Pevonedistat + Ko143 5 μM
HEK293/pcDNA3.1 0.068 ± 0.009 (1.00) 0.082 ± 0.021 (1.20) 0.751 ± 0.203 (1.00) 0.770 ± 0.392 (1.02)
HEK293/ABCG2-R482 2.281 ± 0.216b (33.62) 0.061 ± 0.018 (0.91)
26.028 ± 8.821b (34.66)
1.208 ± 0.143 (1.61)
HEK293/ABCG2-G482 2.307 ± 0.784b (34.00) 0.064 ± 0.008 (0.94)
33.903 ± 4.158b (45.14)
0.695 ± 0.285 (0.93)
HEK293/ABCG2-T482 2.491 ± 0.717b (36.72) 0.144 ± 0.017 (2.13)
25.022 ± 5.741b (33.32)
2.385 ± 0.714 (3.18)
a IC50 values are represented as mean ± SD of at least three independent experiments performed in triplicate.
b P < 0.05 versus the control group.

Fig. 2. Effect of pevonedistat on the ABCG2 ATPase activity. (A) The ABCG2 ATPase activity after membranes were incubated with 0–40 μM of pevonedistat
(B) The ABCG2 ATPase activity after membranes were incubated with
0–160 μM of topotecan alone (black), or topotecan with 1 μM of pevonedistat (orange). Data are expressed as mean ± SD, representative of three in- dependent experiments. (For interpretation of the references to color in this
figure legend, the reader is referred to the Web version of this article.)

3.3. Pevonedistat stimulated ABCG2 ATPase activity

Because ABCG2 transporter utilizes the energy derived from ATP hydrolysis to pump out substrates, we examined whether pevonedistat can stimulate the ATPase activity of ABCG2. The ATP hydrolysis was measured in the presence of pevonedistat (0–40 μM). According to the result (Fig. 2A), pevonedistat showed a concentration-dependent

stimulation of ABCG2 ATPase activity. The stimulation reached a maximum of 2.43-fold stimulation compared to the basal ABCG2 AT- Pase activity level and the EC50 value was 2.01 μM. The stimulation suggested that pevonedistat can interact with the drug-binding site of
the transporter. Since pevonedistat can interact with the ATP-binding pocket of NAE, we investigated whether pevonedistat can bind to the ATP-binding site of ABCG2 and hinder the function of ABCG2 ATPase. To minimize the stimulation effect of pevonedistat, 1 μM of pevonedi- stat was selected to conduct the combination study. As shown in
Fig. 2B, topotecan, a well-established ABCG2 substrate, was able to stimulate the ATPase activity. Meanwhile, the addition of pevonedistat did not significantly alter the topotecan-stimulated ATPase activity. Therefore, the interaction between pevonedistat and ABCG2 may in- volve only on the drug-binding cavity of transmembrane, but not an ATP-competitive inhibitor to target ABCG2.

3.4. Pevonedistat did not affect ABCG2 protein expression level and subcellular localization

Since pevonedistat is being identified as a substrate of ABCG2, it is possible that pevonedistat can stimulate the protein expression of ABCG2 after incubation with cells for a certain period of time. Therefore, Western blotting was performed to investigate this potential effect. As shown in Fig. 3, both parental NCI-H460 and resistant NCI- H460/MX20 cells were incubated with various concentrations of pe- vonedistat for designated time points (0, 72 h). No significant difference was observed after the cells were treated with pevonedistat for 24 h. Besides, although pevonedistat did not significantly increase the ex- pression level of ABCG2 after 72 h treatment, a trend of upregulated expression was observed. Since ABCG2 is a membrane-bound trans- porter, immunofluorescence assay was carried out to explore whether pevonedistat can alter the localization of ABCG2. As shown in Fig. 4 A, after treatment with pevonedistat, the subcellular localization of ABCG2 was not affected.

3.5. The effect of pevonedistat on intracellular accumulation of mitoxantrone in ABCG2-overexpressing cells

To further study the interaction between pevonedistat and ABCG2 as well as between pevonedistat and other ABCG2 substrates, mitoX- antrone accumulation assay was done to evaluate the short-term effect of pevonedistat on ABCG2-mediated drug effluX activity. These ex- periments were performed in the presence or absence of increased concentrations of pevonedistat or the positive ABCG2 inhibitor Ko143. As shown in Fig. 4B and C, without altering the drug accumulation in parental NCI-H460 cells, pevonedistat showed no effect towards mi- toXantrone accumulation in the resistance cells. The reversal effect of pevonedistat was also determined by MTT assay. Due to the significant high toXicity in parental cells, the reversal experiments were done in only ABCG2-overexpressing cells to avoid overlapping effect. Based on the results (Table 3), pevonedistat showed limited ability to reverse ABCG2-mediated mitoXantrone resistance. In NCI-H460/MX20 and

HEK293/ABCG2 cells, the resistance decreased from 71.27- to 47.17-

Fig. 3. The effect of pevonedistat on protein expression level and subcellular localization in cells overexpressing ABCG2. (A) The effect of pevonedistat on the protein expression of ABCG2 after the cells were treated with pevonedistat
(0, 0.1, 1, 10 μM) for 24 h. (B) The effect of pevonedistat on the protein ex- pression of ABCG2 after the cells were treated with pevonedistat (0, 0.1, 0.3, 1,
3 μM) for 72 h. Data are expressed as mean ± SD, representative of three independent experiments. *p < 0.05, compared with control group.

fold and from 20.58- to 13.31-fold, respectively.

3.6. S1-PR cells showed significant cross-resistance to pevonedistat and mitoxantrone

The above results strongly suggested pevonedistat is a substrate of ABCG2, we moved on to establish a pevonedistat-resistant cell line. By treating S1 cells with pevonedistat continuously, we successfully es- tablished a pevonedistat-resistant cell line (S1-PR). After the cells were cultured in drug-free medium for 2 months, MTT assays were per- formed to evaluate the drug resistance profile. Western blotting analysis was done to confirm S1-PR cells overexpress ABCG2. The results (Fig. 5A) showed that the S1-PR cells overexpress ABCG2 and the ex- pression level was stable after 2 months without pevonedistat treat- ment. As shown in Fig. 5 B and C, the viability curve of S1-PR showed significant right-shift compared to S1 cells. S1-PR cells showed sig-
nificant resistance to pevonedistat with IC50 over 100 μM. The resistant S1-PR cells also showed 21-fold resistance to mitoXantrone compared to
parental S1 cells. Meanwhile, the viability curves of both cell lines showed no difference with cisplatin treatment (Fig. 5D).

4. Discussion

Pevonedistat was identified as a first-in-class, selective inhibitor of NEDD8 activating enzyme [16]. By inhibiting cullin neddylation, it shows potent anticancer effect both in vitro and in vivo [34]. Multiple ongoing clinical trials include combination of pevonedistat with che- motherapy for liver cancer and NSCLC, pevonedistat alone or combined with azacitidine for AML. Despite the promising anticancer effect, drug resistance remains a major obstacle in cancer treatment. It is well- documented that ABCG2 is related to clinical MDR [35]. Some well- known conventional chemotherapeutic agents such as irinotecan, mi- toXantrone, doXorubicin were identified as substrates of ABCG2. Also, several clinically relevant TKIs like gefitinib and lapatinib were char- acterized as substrates of ABCG2 [36]. Besides, overexpression of ABCG2 was reported to be a biomarker for chemoresistance. In view of the critical role ABCG2 plays in MDR, it is essential to investigate whether pevonedistat is a substrate of ABCG2 transporter.
In this study, we reported that pevonedistat may be a potential substrate of ABCG2; Therefore ABCG2 overexpression may confer re- sistance to pevonedistat. Firstly, the cytotoXicity of pevonedistat was determined in both cancer cell lines and transfected cell lines. Since pevonedistat is in clinical trials for NSCLC patients, the NSCLC parental cells NCI-H460 and the ABCG2-overexpressing resistance cells NCI- H460/MX20 were selected. The colon cancer cells S1 and ABCG2- overexpressing subline S1-M1-80 were also used to validate the results. It showed that both ABCG2-overexpressing cancer cells were sig- nificantly resistant to pevonedistat. Besides, the addition of ABCG2 inhibitor Ko143 restored the cytotoXic effect of pevonedistat to a si- milar level as in parental cells. These results suggested that pevonedi- stat may be a substrate of ABCG2. It was reported that mutation at position 482 of ABCG2 may lead to a distinct drug-binding and effluX profile [37]. Subsequently, the HEK293 transfected cells, in which ABCG2 overexpression is the sole contributor to MDR, were used to carry out the MTT assay. Identically, the HEK293/ABCG2 cells, both wild-type and mutants, showed resistance to pevonedistat compared to parental HEK293/pcDNA3.1 cells. Moreover, utilizing Ko143 was able to completely reverse the drug resistance. The cytotoXicity results strongly suggested pevonedistat is a substrate of both wild-type and mutant ABCG2.
ABCG2 is characterized as a “half-transporter”, comprising a
transmembrane domain (TMD) where substrate binds and a nucleotide- binding domain (NBD) where ATP binds and hydrolyzes. Since ABC transporters utilize energy derived from ATP hydrolysis to extrude its substrates, the ABCG2 ATPase activity in the presence of pevonedistat was measured. Consistently, the results showed that pevonedistat can stimulate the ATPase activity of ABCG2 in a concentration-dependent manner. Therefore, pevonedistat may bind to the substrate binding site of ABCG2, which activates ABCG2 ATPase to provide energy for drug effluX. It has been reported that pevonedistat can interact with the ATP- binding pocket of NAE [38], therefore we further investigated whether pevonedistat can inhibit ABCG2 ATPase function. At 1 μM, which pe-
vonedistat showed modest stimulation effect, it failed to inhibit topo-
tecan-stimulated ATPase activity. These data suggested the interaction of pevonedistat with ABCG2 transporter may occur at substrate-binding site only, and the result indicated that pevonedistat may not be an ATP- competitive inhibitor to target ABCG2. Then Western blotting analysis and immunofluorescence assay were performed to investigate whether pevonedistat affects to the protein expression level or subcellular lo- calization of ABCG2. After the parental and resistant cells were treated
with pevonedistat up to 10 μM for 24 h, the ABCG2 protein level re- mained unchanged. Furthermore, the resistant cells were treated with
pevonedistat for 72 h with no significant increase of protein expression. However, after 72 h treatment, the ABCG2 protein expression did showed a trend of upregulation. Therefore, the effect of long-term treatment with pevonedistat should be further investigated. Immuno- fluorescence assay was carried out to visualize the ABCG2 localization

Fig. 4. Effect of pevonedistat on ABCG2 subcellular localization and mitoXantrone accumulation (A) Subcellular localization of ABCG2 expression in NCI-H460/ MX20 after the cells were incubated with 3 μM of pevonedistat for 0, 24, 48, 72 h. (B) The effect of pevonedistat on intracellular accumulation of mitoXantrone in NCI-H460 cells after co-treatment for 2 h. (C) The effect of pevonedistat on intracellular accumulation of mitoXantrone in NCI-H460/MX20 cells after co-treatment for 2 h.

after pevonedistat treatment. The results showed that pevonedistat did not affect the subcellular localization of ABCG2 transporter.
It is suggested that some ABC transporter inhibitors are capable of reversing MDR by competitive inhibition of substrate effluX [39], downregulation or localization of these membrane transporters [40]. We have demonstrated that pevonedistat showed no activity towards the protein expression or localization. Next, we explored whether pe- vonedistat can serve as a competitive inhibitor and reverse drug re- sistance. Therefore, the mitoXantrone accumulation experiment was performed. It was postulated that pevonedistat may be able to compete with mitoXantrone, another ABCG2 substrate, for the transporter and thereby decrease the effluX activity of mitoXantrone in ABCG2-over- expressing cells. The accumulation of mitoXantrone was determined by

mitoXantrone binding site. We also tested whether pevonedistat can reverse ABCG2-mediated drug resistance at non-toXic concentration after 72 h treatment. The results showed that pevonedistat failed to significantly restore the sensitivity of resistance cells to mitoXantrone, suggesting pevonedistat may not be an effective ABCG2 reversal agent. Most of the ABCG2-overexpressing cells were established by se- lecting the cells with ABCG2 substrates such as mitoXantrone, flavo- piridol [39,43]. Based on the 72 h Western blotting results, the pevo- nedistat-treated resistant cells showed slightly increase of ABCG2 expression. To further validate ABCG2 is a prominent factor that may lead to pevonedistat-resistance, we established a resistant cell line S1- PR by continuously exposed S1 cells to pevonedistat. The cytotoXicity and Western blotting results suggested S1-PR cells overexpress ABCG2

measuring the intracellular fluorescent

mitoXantrone.

The results

which lead to cross-resistance to

mitoXantrone

and pevonedistat.

showed that, at concentration up to 20 μM, pevonedistat did not in- crease the intracellular accumulation of mitoXantrone in resistant cells. Since it was reported that ABCG2 have multiple substrate binding sites
[41,42], pevonedistat may bind to a distinct site other than

Therefore, it provides strong evidence to show pevonedistat is a sub- strate of ABCG2 and long-term pevonedistat treatment may lead to the upregulation of ABCG2.
Pevonedistat is structurally related to adenosine 5′-monophosphate

Table 3
The reversal effect of pevonedistat in cancer and transfected cells overexpressing ABCG2.
Cell line IC50 value ± SDa (μM, Resistance-fold)

NCI-H460 NCI-H460/MX20 HEK293/pcDNA3.1 HEK293/ABCG2-R482
MitoXantrone 0.069 ± 0.014 (1.00) 4.894 ± 0.990 (71.27) 0.077 ± 0.015 (1.00) 1.586 ± 0.086 (20.52)
+ Ko143 (5 μM) 0.064 ± 0.011 (0.93) 0.172 ± 0.028b (2.50)
0.090 ± 0.019 (1.16) 0.072 ± 0.021b (0.94)

+ pevonedistat (5 μM) – 3.239 ± 0.517 (47.17) – 1.029 ± 0.487 (13.31)
a IC50 values are represented as mean ± SD of at least three independent experiments performed in triplicate.
b P < 0.05 versus the control group.

Fig. 5. CytotoXicity of different anticancer drugs in S1 and S1-PR cells. (A) The Western blotting of ABCG2 expression level in S1, S1-PR, and S1-M1-80 cells. (B) CytotoXicity of pevonedistat, mitoXantrone, and cisplatin in S1 and S1-PR cells. (C) CytotoXicity of mitoXantrone in S1 and S1-PR cells. (D) CytotoXicity of cisplatin in S1 and S1-PR cells. Data are expressed as mean ± SD, representative of three independent experiments. *p < 0.05, compared with control group.

(AMP), allowing the interaction with the nucleotide-binding site of NAE [15]. It remains attractive to investigate whether other NAE inhibitors such as TAS 4464 [44], ZM223 [45], and some structurally related compounds [46,47] are substrates of ABCG2. These studies may not only provide indications for clinical usage, but also allow more specific and effective identification of ABCG2 substrates.
In conclusion, our study reveals that overexpression of ABCG2 transporter can confer resistance to pevonedistat. Therefore, the drug response of pevonedistat may be attenuated in patients with high ABCG2 expression. Future study is required to validate ABCG2-medi- ated pevonedistat resistance in vivo.

CRediT authorship contribution statement

Liu-Ya Wei: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. Zhuo-Xun Wu: Conceptualization, Methodology, Writing – original draft. Yang Yang: Methodology, Writing – original draft. Min Zhao: Methodology. Xiang-Yu Ma: Methodology. Jin-Sui Li: Methodology. Dong-Hua Yang: Writing – review & editing. Zhe-Sheng Chen: Conceptualization, Supervision. Ying-Fang Fan: Conceptualization, Writing – review & editing, Supervision.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgement

This work was supported by the National Natural Science Foundation of China (81700167), Natural Science Foundation of Shandong Province (ZR2016HM47) for Liu-Ya Wei. We would like to thank Drs. Susan E. Bates and Robert W. Robey (NCI, NIH, Bethesda, MD) for providing the cell lines. We thank Ms. Yanglu Chen (Columbia

University College of Physicians and Surgeons) for editing the article.

References

[1] E. Reinstein, A. Ciechanover, Narrative review: protein degradation and human diseases: the ubiquitin connection, Ann. Intern. Med. 145 (2006) 676–684.
[2] A. Hershko, The ubiquitin system for protein degradation and some of its roles in the control of the cell-division cycle (Nobel lecture), Angew Chem. Int. Ed. Engl. 44 (2005) 5932–5943.
[3] A.L. Schwartz, A. Ciechanover, Targeting proteins for destruction by the ubiquitin system: implications for human pathobiology, Annu. Rev. Pharmacol. ToXicol. 49 (2009) 73–96.
[4] T. Kamitani, K. Kito, H.P. Nguyen, E.T. Yeh, Characterization of NEDD8, a devel- opmentally down-regulated ubiquitin-like protein, J. Biol. Chem. 272 (1997) 28557–28562.
[5] Z.-Q. Pan, A. Kentsis, D.C. Dias, K. Yamoah, K. Wu, Nedd8 on cullin: building an expressway to protein destruction, Oncogene 23 (2004) 1985–1997.
[6] F.G. Whitby, G. Xia, C.M. Pickart, C.P. Hill, Crystal structure of the human ubi- quitin-like protein NEDD8 and interactions with ubiquitin pathway enzymes, J. Biol. Chem. 273 (1998) 34983–34991.
[7] R.N. Bohnsack, A.L. Haas, Conservation in the mechanism of Nedd8 activation by the human AppBp1-Uba3 heterodimer, J. Biol. Chem. 278 (2003) 26823–26830.
[8] J. Jones, K. Wu, Y. Yang, C. Guerrero, N. Nillegoda, Z.-Q. Pan, L. Huang, A targeted proteomic analysis of the ubiquitin-like modifier nedd8 and associated proteins, J. Proteome Res. 7 (2008) 1274–1287.
[9] S. Ni, X. Chen, Q. Yu, Y. Xu, Z. Hu, J. Zhang, W. Zhang, B. Li, X. Yang, F. Mao,
J. Huang, Y. Sun, J. Li, L. Jia, Discovery of candesartan cilexetic as a novel ned- dylation inhibitor for suppressing tumor growth, Eur. J. Med. Chem. (2019) 111848.
[10] R.Z. Orlowski, D.J. Kuhn, Proteasome inhibitors in cancer therapy: lessons from the
first decade, Clin. Cancer Res. 14 (2008) 1649–1657.
[11] V.N. Podust, J.E. Brownell, T.B. Gladysheva, R.S. Luo, C. Wang, M.B. Coggins,
J.W. Pierce, E.S. Lightcap, V. Chau, A Nedd8 conjugation pathway is essential for proteolytic targeting of p27Kip1 by ubiquitination, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 4579–4584.
[12] D.T. Huang, B.A. Schulman, EXpression, purification, and characterization of the E1 for human NEDD8, the heterodimeric APPBP1-UBA3 complex, Methods Enzymol. 398 (2005) 9–20.
[13] H. Walden, M.S. Podgorski, D.T. Huang, D.W. Miller, R.J. Howard, D.L. Minor Jr.,
J.M. Holton, B.A. Schulman, The structure of the APPBP1-UBA3-NEDD8-ATP complex reveals the basis for selective ubiquitin-like protein activation by an E1, Mol. Cell 12 (2003) 1427–1437.
[14] Q. Xu, G. Lin, H. Xu, L. Hu, Y. Wang, S. Du, W. Deng, W. Hu, W. Cheng, K. Jiang,

MLN4924 neddylation inhibitor promotes cell death in paclitaxel-resistant human lung adenocarcinoma cells, Oncol Lett 15 (2018) 515–521.
[15] T.A. Soucy, P.G. Smith, M.A. Milhollen, A.J. Berger, J.M. Gavin, S. Adhikari,
J.E. Brownell, K.E. Burke, D.P. Cardin, S. Critchley, C.A. Cullis, A. Doucette,
J.J. Garnsey, J.L. Gaulin, R.E. Gershman, A.R. Lublinsky, A. McDonald, H. Mizutani,
U. Narayanan, E.J. Olhava, S. Peluso, M. Rezaei, M.D. Sintchak, T. Talreja,
M.P. Thomas, T. Traore, S. Vyskocil, G.S. Weatherhead, J. Yu, J. Zhang, L.R. Dick,
C.F. Claiborne, M. Rolfe, J.B. Bolen, S.P. Langston, An inhibitor of NEDD8-acti- vating enzyme as a new approach to treat cancer, Nature 458 (2009) 732–736.
[16] S.T. Nawrocki, P. Griffin, K.R. Kelly, J.S. Carew, MLN4924: a novel first-in-class inhibitor of NEDD8-activating enzyme for cancer therapy, EXpert Opin. Investig. Drugs 21 (2012) 1563–1573.
[17] G.W. Xu, J.I. Toth, S.R. da Silva, S.-L. Paiva, J.L. Lukkarila, R. Hurren, N. Maclean,
M.A. Sukhai, R.N. Bhattacharjee, C.A. Goard, B. Medeiros, P.T. Gunning, S. Dhe- Paganon, M.D. Petroski, A.D. Schimmer, Mutations in UBA3 confer resistance to the NEDD8-activating enzyme inhibitor MLN4924 in human leukemic cells, PLoS One 9 (2014) e93530-e93530.
[18] J.I. Toth, L. Yang, R. Dahl, M.D. Petroski, A gatekeeper residue for NEDD8-acti- vating enzyme inhibition by MLN4924, Cell Rep. 1 (2012) 309–316.
[19] J.I. Fletcher, R.T. Williams, M.J. Henderson, M.D. Norris, M. Haber, ABC trans- porters as mediators of drug resistance and contributors to cancer cell biology, Drug Resist. Updates 26 (2016) 1–9.
[20] A. Domenichini, A. Adamska, M. Falasca, ABC transporters as cancer drivers: po- tential functions in cancer development, Biochim. Biophys. Acta Gen. Subj. 1863 (2019) 52–60.
[21] S.V. Ambudkar, C. Kimchi-Sarfaty, Z.E. Sauna, M.M. Gottesman, P-glycoprotein: from genomics to mechanism, Oncogene 22 (2003) 7468–7485.
[22] R.G. Deeley, S.P.C. Cole, Substrate recognition and transport by multidrug re- sistance protein 1 (ABCC1), FEBS Lett. 580 (2006) 1103–1111.
[23] Q. Mao, J.D. Unadkat, Role of the breast cancer resistance protein (BCRP/ABCG2) in drug transport–an update, AAPS J. 17 (2015) 65–82.
[24] M. Dohse, C. Scharenberg, S. Shukla, R.W. Robey, T. Volkmann, J.F. Deeken,
C. Brendel, S.V. Ambudkar, A. Neubauer, S.E. Bates, Comparison of ATP-binding cassette transporter interactions with the tyrosine kinase inhibitors imatinib, nilo- tinib, and dasatinib, Drug Metab. Dispos. 38 (2010) 1371–1380.
[25] C. Hegedüs, K. Truta-Feles, G. Antalffy, G. Várady, K. Német, C. Ozvegy-Laczka,
G. Kéri, L. Orfi, G. Szakács, J. Settleman, A. Váradi, B. Sarkadi, Interaction of the EGFR inhibitors gefitinib, vandetanib, pelitinib and neratinib with the ABCG2 multidrug transporter: implications for the emergence and reversal of cancer drug resistance, Biochem. Pharmacol. 84 (2012) 260–267.
[26] K. Natarajan, Y. Xie, M.R. Baer, D.D. Ross, Role of breast cancer resistance protein (BCRP/ABCG2) in cancer drug resistance, Biochem. Pharmacol. 83 (2012) 1084–1103.
[27] R.W. Robey, Y. Honjo, A. van de Laar, K. Miyake, J.T. Regis, T. Litman, S.E. Bates, A functional assay for detection of the mitoXantrone resistance protein, MXR (ABCG2), Biochim. Biophys. Acta 1512 (2001) 171–182.
[28] K. Miyake, L. Mickley, T. Litman, Z. Zhan, R. Robey, B. Cristensen, M. Brangi,
L. Greenberger, M. Dean, T. Fojo, S.E. Bates, Molecular cloning of cDNAs which are highly overexpressed in mitoXantrone-resistant cells: demonstration of homology to ABC transport genes, Cancer Res. 59 (1999) 8–13.
[29] N. Ji, Y. Yang, C.-Y. Cai, Z.-N. Lei, J.-Q. Wang, P. Gupta, S. Shukla, S.V. Ambudkar,
D. Kong, Z.-S. Chen, Selonsertib (GS-4997), an ASK1 inhibitor, antagonizes multi- drug resistance in ABCB1- and ABCG2-overexpressing cancer cells, Cancer Lett. 440–441 (2019) 82–93.
[30] N. Ji, Y. Yang, Z.-N. Lei, C.-Y. Cai, J.-Q. Wang, P. Gupta, X. Xian, D.-H. Yang,
D. Kong, Z.-S. Chen, UliXertinib (BVD-523) antagonizes ABCB1- and ABCG2- mediated chemotherapeutic drug resistance, Biochem. Pharmacol. 158 (2018)

274–285.
[31] J.-Q. Wang, B. Wang, Z.-N. Lei, Q.-X. Teng, J.Y. Li, W. Zhang, N. Ji, C.-Y. Cai, L.-
Y. Ma, H.-M. Liu, Z.-S. Chen, Derivative of 5-cyano-6-phenylpyrimidin antagonizes ABCB1- and ABCG2-mediated multidrug resistance, Eur. J. Pharmacol. 863 (2019) 172611-172611.
[32] J.J. Gribar, M. Ramachandra, C.A. Hrycyna, S. Dey, S.V. Ambudkar, Functional characterization of glycosylation-deficient human P-glycoprotein using a vaccinia virus expression system, J. Membrane Biology 173 (2000) 203–214.
[33] Z.-X. Wu, Q.-X. Teng, C.-Y. Cai, J.-Q. Wang, Z.-N. Lei, Y. Yang, Y.-F. Fan, J.-
Y. Zhang, J. Li, Z.-S. Chen, Tepotinib reverses ABCB1-mediated multidrug resistance in cancer cells, Biochem. Pharmacol. 166 (2019) 120–127.
[34] M. Oladghaffari, J.P. Islamian, B. Baradaran, A.S. Monfared, MLN4924 therapy as a novel approach in cancer treatment modalities, J. Chemother. 28 (2016) 74–82.
[35] H. Amawi, H.-M. Sim, A.K. Tiwari, S.V. Ambudkar, S. Shukla, ABC transporter- mediated multidrug-resistant cancer, Adv. EXp. Med. Biol. 1141 (2019) 549–580.
[36] G.L. Beretta, G. Cassinelli, M. Pennati, V. Zuco, L. Gatti, Overcoming ABC trans- porter-mediated multidrug resistance: the dual role of tyrosine kinase inhibitors as multitargeting agents, Eur. J. Med. Chem. 142 (2017) 271–289.
[37] A. Pozza, J.M. Perez-Victoria, A. Sardo, A. Ahmed-Belkacem, A. Di Pietro, Purification of breast cancer resistance protein ABCG2 and role of arginine-482, Cell. Mol. Life Sci. 63 (2006) 1912–1922.
[38] S. Verma, A. Singh, A. Mishra, Molecular dynamics investigation on the poor sen- sitivity of A171T mutant NEDD8-activating enzyme (NAE) for MLN4924, J. Biomol. Struct. Dyn. 32 (2014) 1064–1073.
[39] R.W. Robey, W.Y. Medina-Pérez, K. Nishiyama, T. Lahusen, K. Miyake, T. Litman,
A.M. Senderowicz, D.D. Ross, S.E. Bates, Overexpression of the ATP-binding cas- sette half-transporter, ABCG2 (MXr/BCrp/ABCP1), in flavopiridol-resistant human breast cancer cells, Clin. Cancer Res. 7 (2001) 145–152.
[40] K. Natarajan, J. Bhullar, S. Shukla, M. Burcu, Z.-S. Chen, S.V. Ambudkar, M.R. Baer, The Pim kinase inhibitor SGI-1776 decreases cell surface expression of P-glyco- protein (ABCB1) and breast cancer resistance protein (ABCG2) and drug transport by Pim-1-dependent and -independent mechanisms, Biochem. Pharmacol. 85 (2013) 514–524.
[41] R. Clark, I.D. Kerr, R. Callaghan, Multiple drugbinding sites on the R482G isoform of the ABCG2 transporter, Br. J. Pharmacol. 149 (2006) 506–515.
[42] R.J. Ferreira, C.A. Bonito, M.N.D.S. Cordeiro, M.-J.U. Ferreira, D.J.V.A. dos Santos, Structure-function relationships in ABCG2: insights from molecular dynamics si- mulations and molecular docking studies, Sci. Rep. 7 (2017) 15534.
[43] R.W. Robey, K.M. Pluchino, M.D. Hall, A.T. Fojo, S.E. Bates, M.M. Gottesman, Revisiting the role of ABC transporters in multidrug-resistant cancer, Nat. Rev. Cancer 18 (2018) 452–464.
[44] C. Yoshimura, H. Muraoka, H. Ochiiwa, S. Tsuji, A. Hashimoto, H. Kazuno,
F. Nakagawa, Y. Komiya, S. Suzuki, T. Takenaka, M. Kumazaki, N. Fujita,
T. Mizutani, S. Ohkubo, TAS4464, a Highly Potent and Selective Inhibitor of NEDD8 Activating Enzyme, Suppresses Neddylation and Shows Antitumor Activity in Diverse Cancer Models, Molecular Cancer Therapeutics, 2019 mol- canther.0644.2018.
[45] H. Ma, C. Zhuang, X. Xu, J. Li, J. Wang, X. Min, W. Zhang, H. Zhang, Z. Miao, Discovery of benzothiazole derivatives as novel non-sulfamide NEDD8 activating enzyme inhibitors by target-based virtual screening, Eur. J. Med. Chem. 133 (2017) 174–183.
[46] H.-J. Zhong, K.-H. Leung, S. Lin, D.S.-H. Chan, Q.-B. Han, S.L.-F. Chan, D.-L. Ma, C.-
H. Leung, Discovery of deoXyvasicinone derivatives as inhibitors of NEDD8-acti- vating enzyme, Methods 71 (2015) 71–76.
[47] K.-J. Wu, H.-J. Zhong, G. Li, C. Liu, H.-M.D. Wang, D.-L. Ma, C.-H. Leung, Structure- based identification of a NEDD8-activating enzyme inhibitor via drug repurposing, Eur. J. Med. Chem. 143 (2018) 1021–1027.