ARS-1620 was a free sample from ChemieTek (Indianapolis, IN). The monoclonal antibodies against ABCB1 (clone F4, Cat # SAB4200775) and other chemicals were purchased from Sigma Chemical Co (St. Louis, MO) except otherwise specified. Antibiotics (penicillin/streptomycin), trypsin-EDTA, and fetal bovine serum (FBS) were obtained from Hyclone (Waltham, MA, United States). The Horseradish peroxidase (HRP)-conjugated rabbit anti-mouse IgG secondary antibody (Cat # 7076S, lot #: 32) was the product of Cell Signaling Technology Inc. (Danvers, MA, United States). The GAPDH loading control monoclonal antibody (GA1R) (Cat # MA5-15738, lot #: SA247966), Alexa Fluor 488 conjugated goat anti-mouse IgG cross-adsorbed secondary antibody (Cat # A32723) were acquired from Thermo Fisher Scientific Inc. (Rockford, IL, United States). [³H]-paclitaxel (15 Ci/mmol) was obtained from Moravek Biochemicals, Inc. (Brea, CA).

All the cells were cultured as previously described (Zhang et al., 2020). In brief, the human epidermoid carcinoma cell line KB-3-1 and its colchicine-induced ABCB1-overexpressing KB-C2 cell line, human colon carcinoma cell line SW620, and its doxorubicin-induced ABCB1-overexpressing SW620/Ad300 cell line were used (Lyall et al., 1987; Lai et al., 1991). The KB-3-1 and KB-C2 were not KRAS-G12C mutation cell lines, the SW620 and SW620/Ad300 were KRAS-G12V mutation but not KRAS-G12C mutation cell lines. HEK293/pcDNA3.1, HEK293/ABCB1, HEK293/ABCG2, and HEK293/ABCC1 were human embryonic kidney HEK293 cells transfected with empty vector pcDNA3.1, full-length ABCB1, full-length ABCG2, and full-length ABCC1, respectively (Fung et al., 2014). The moist incubator (37°C, 5% CO₂) was used to culture the above cell lines.

Cell survival percentage after treatment with ARS-1620 and other substrates was determined by MTT assay as described earlier (Dong et al., 2020). In short, the cells were seeded into 96-well plates (5×10³/well). After cell attachment, the cells were treated with serial diluted ARS-1620 or other chemotherapeutic agents in the presence or absence of ABCB1 inhibitor verapamil for 68 h. Then, MTT solution was added for another 4 h incubation. Later, the supernatant was discarded and DMSO was added to dissolve the formazan. The optical density was obtained by reading the plate at 570 nm with a microplate analyzer.

We performed Western blotting as described earlier (Feng et al., 2020). Briefly, cells were treated with ARS-1620 (3 μM) at 37°C for 0, 24, 48, or 72 h, and then lysed. Protein concentration in each group was determined by BCA protein quantitative method. An equivalent amount (20 μg) of protein was subjected to SDS-PAGE followed by transfer to the PVDF membranes. Next, blocking of the membranes were processed in 5% skimmed milk at room temperature for 2 h. Primary antibodies (ABCB1 or GAPDH) (1:1000) were applied to membranes at 4°C overnight. After washing the membranes with TBST 3 times, the HRP-conjugated anti-mouse antibody (1:1000) was incubated with the membranes at room temperature for 2 h. ECL chemiluminescence kit was used and the signal was captured by X-ray film. The analysis of protein expression levels was performed by ImageJ (NIH, MD).

Immunofluorescence was carried out as previously described (Wang JQ. et al., 2020). Cells (1×10⁵) were incubated in 24-well plates with ARS-1620 (3 μM) for 0, 24, 48, or 72 h. The cells were fixed and permeated, then blocked with 6% BSA and incubated overnight with ABCB1 antibody (1:1000). Later, Alexa Fluor 488 IgG secondary antibody (1:1000) was applied to the cells for 2 h, and nuclei were re-stained with DAPI. The image was captured by EvoS FL fluorescence microscope (Life Technologies, MD).

KB-3-1 and KB-C2 cells (3×10⁵) were seeded into 6-well plates with 2 ml complete DMEM. After 2 days incubation, the medium was replaced with plain DMEM with or without 3 μM verapamil and incubated for 2 h, followed by adding 30 μM ARS-1620 to the wells for another 2 h. Afterwards, the cells were washed with PBS twice then lysed with 0.5% sodium dodecyl sulfate and acetonitrile. Samples were harvested and centrifuged at 14,000 rpm for 10 min. The supernatant was collected for intracellular concentration analysis using HPLC.

24-well plates were used for the culture of the cells (1×10⁵). On the second day, the cells were treated for 2 h with or without ARS-1620 or verapamil. Afterward, [³H] -paclitaxel was added to the specified wells and incubated for 2 h. After washing with PBS, the cells were gathered and shifted to scintillation vials and the radioactivity of different treatments were measured using Packard Tricarb 1900CA liquid scintillation analyzer (Packard Instrument, Dners Grove, IL).

As previously described, the vanadate-sensitive ATPase activity of ABCB1 was measured in terms of the amount of inorganic phosphate (P_i) generated by hydrolysis of ATP (Sarkadi et al., 1992). The quantity of P_i was determined using the improved colorimetric method by Murphy and Riley (Ojida et al., 2004).

Docking analysis was conducted with software Maestro 11.5 (Schrödinger, LLC, New York, NY, 2018) (Cai et al., 2020). Ligand preparation was essentially performed with the default protocol. Human ABCB1 (PDB ID: 6QEX) (Alam et al., 2019) protein preparation was conducted to optimize the structure, remove waters, and minimize the energy. Subsequently, a grid of 20 Å at the binding pocket of ABCB1 protein was generated. Glide XP docking was performed, followed by induced-fit docking (IFD) was carried out using the default protocol.

All data are expressed as mean ± SD and were acquired from a minimum of three independently repetitions. Statistical analysis was based on one-way ANOVA followed by the Dunnett’s test. p-value below 0.05 represents significant differences.

Firstly, we conducted the MTT assay to examine the cytotoxicity of ARS-1620 (the chemical structure is presented in Figure 1A) in KB-3-1, KB-C2, SW620, and SW620/Ad300 cell lines. As Figures 1B,C showed, the parental KB-3-1 and SW620 cells were significantly more sensitive to ARS-1620 than their drug-selected KB-C2 and SW620/Ad300 cells that overexpressing ABCB1. To further certify our observation, the cytotoxicity of ARS-1620 was also tested in gene-transfected cells HEK293/pcDNA3.1 and HEK293/ABCB1 (Figure 1D). Consistent with our previous results, HEK293/ABCB1 was more resistant and less sensitive to ARS-1620 compared with its parental cells HEK293/pcDNA3.1. Meanwhile, the cell survival curve of wild-type HEK293/ABCG2 and HEK293/ABCC1 cells displayed no significant increased sensitivity or reduced sensitivity. Additionally, the IC₅₀ values (concentrations for 50% inhibition) of ARS-1620 and resistance fold (RF) were summarized in Table 1. The KB-C2 and SW620/Ad300 cells, compared with their parental KB-3-1 and SW620 cells, possessed 6.30- and 2.95-fold resistance to ARS-1620, respectively. HEK293/ABCB1 cells showed 3.07-fold resistance to ARS-1620 compared with HEK293/pcDNA3.1. These results indicated the possibility that the overexpression of ABCB1 decreased the cytotoxicity of ARS-1620, which renders the ability of ABCB1-overexpressing cells to survive with high concentrations of ARS-1620.

Towards to get more supporting evidence that the overexpression of ABCB1 transporter leads to ARS-1620 resistance in ABCB1-overexpressing cells, verapamil, the pervasive used first-generation inhibitor of ABCB1, was used to co-treat with ARS-1620 in KB-C2, SW620/Ad300, and HEK293/ABCB1 cells. As shown in Figures 1B,C and Table 1, in the presence of 3 μM verapamil, the resistance to ARS-1620 was significantly reversed in KB-C2, SW620/Ad300, and HEK293/ABCB1 cells, the resistance fold dropped to 2.69-, 0.99-, and 1.09-fold, respectively. These results reinforced our assumption that the overexpression of ABCB1 exists as one predominant contributor to the resistance of ARS-1620.

Next, we tested the accumulation of intracellular [³H]-paclitaxel when the cells were co-treated with distinct concentrations of ARS-1620 (3, 10, 30, 100 μM) over a short period. As observed in Figure 2, ARS-1620 at the concentration of 3 μM or even the toxic concentration of 10 μM did not influence the accumulation of [³H]-paclitaxel in KB-C2 cells. However, when the concentrations of ARS-1620 reached 30 and 100 μM, the [³H]-paclitaxel level was significantly increased in KB-C2 cells. In addition, the accumulation of [³H]-paclitaxel in parental KB-3-1 cells was not perturbed by ARS-1620. Notably, the toxic concentrations chosen in this assay were unlikely to affect the viability and function of the parental and resistant cells in this short treatment period. The restored accumulation of [³H]-paclitaxel in resistant cell treating with high concentrations of ARS-1620 implied that high concentrations of ARS-1620 might compete with paclitaxel for binding to substrate binding site, providing potential and indirect evidence that ARS-1620 is a substrate of ABCB1.

The ATPase activity was considered a non-negligible factor affecting the function of ABCB1, since the energy supplied to pump out the substrate drugs was generated through the hydrolysis of ATP. Thus, we measured the effect of ARS-1620 (0–160 μM) on the ATP hydrolysis in ABCB1 membrane vesicles. As displayed in Figure 3A, ARS-1620 showed a stimulatory effect on ABCB1-associated ATPase activity in a dose-depend manner. The stimulatory effect of ARS-1620 reached 50% maximal (EC₅₀) at 14.8 μM. Besides, a maximum of 4.23-fold of the basal activity was achieved by ARS-1620. Understanding that stimulation of ATPase activity is commonly relevant to ABC-mediated substrates transport, the above results illustrated that ARS-1620 may act as a substrate of ABCB1 transporter which was in accord with the cytotoxicity results.

For further verification of ARS-1620 is a substrate of ABCB1 transporter, an HPLC assay was performed to examine the concentration of ARS-1620 with or without adding verapamil in KB-3-1 and KB-C2 cells. In Figure 3B, the amount of ARS-1620 was 5-times higher in KB-3-1 than KB-C2. After co-incubation with verapamil, the concentration of ARS-1620 was significantly increased in KB-C2 cells compared with KB-3-1 cells. These results provided direct and strong evidence that ARS-1620 resistance was conferred by overexpression of ABCB1 and efflux by ABCB1.

It has been reported that some substrates could act as reversal reagents that competed with other conventional substrates like paclitaxel or doxorubicin at the substrate-binding site of ABCB1 transporter (Wang J. et al., 2020; Wu et al., 2020), resulting in a higher concentration of paclitaxel or doxorubicin remained in cells. Therefore, reversal experiments were carried out in three pairs of cell lines to examine this potentiality. Based on the results of the cytotoxicity assay (Figure 1), the non-toxic concentration of ARS-1620 (0.3, 1, 3 μM) was selected. As shown in Figure 4, the IC₅₀ values of doxorubicin and paclitaxel were slightly but not significantly increased in drug resistant cells by the addition of ARS-1620 in a concentration-dependent manner which signifies that ARS-1620 could not behave as an inhibitor of ABCB1 at non-toxic concentrations. The verapamil that significantly re-sensitized the resistant cells was used as a positive inhibitor of ABCB1.

Knowing that substrate drugs may interact with ABCB1 protein which brings about increasing the protein expression of ABCB1 or altering the localization of ABCB1 on the cell membrane (Wu et al., 2020), we conducted immunoblotting and immunofluorescence to investigate these potentialities. According to Figure 5A, the non-toxic concentration of ARS-1620 (3 μM) did not affect the expression level of ABCB1 in KB-C2 cells after comparing the relative intensity of groups with different treating periods (0, 24, 48, 72 h). Furthermore, the results from Figure 5B showed that ABCB1 protein still located on the cell membrane without alterations during the 24–72 h treating periods with ARS-1620. The parental KB-3-1 cells in these experiments were set as the negative control.

The results showed that ARS-1620, doxorubicin, and paclitaxel have high binding affinities with human ABCB1 (Figure 6) with docking scores -12.807, -12.241, and -15.527 kcal/mol, respectively. As exhibited in Figures 6C,D, the interactions between ARS-1620 and ABCB1 include hydrogen bonds and π-π interaction. The three aromatic rings of ARS-1620 are involved in the π-π interactions with the residues Phe239, Phe770, and Phe994. The carbonyl group and a nitrogen atom on quinazoline ring ARS-1620 formed hydrogen bonds with Gln725 and Gln838 as hydrogen-bond acceptors, while the hydroxyl group interacted with Asn296 as hydrogen bond donor. Distinct from ARS-1620, paclitaxel and doxorubicin with more oxygen atoms in their structures, tended to form hydrogen bonds rather than π-π interactions with ABCB1. Residues Gln347, Glu 875, Gln946, Tyr953, and Gln990 formed hydrogen bonds with doxorubicin. Gln946 interacted with a carbonyl group and a hydroxyl group as hydrogen bond donor and acceptor, respectively. Interestingly, the amino group became a cation and interacted with Gln990 with a hydrogen bond. Due to the acidic cancer microenvironment (Boedtkjer and Pedersen, 2020), the amino group can become protonated to a cation, and bind with the residues. Paclitaxel formed hydrogen bonds with residues Tyr310, Gln725, and Gln990, where Tyr310 interacted with both carbonyl and hydroxyl groups.