The PI3K subunits, P110α and P110β are potential targets for overcoming P-gp and BCRP-mediated MDR in cancer

Human epidermoid carcinoma KB-3-1 cell line was used as the parental drug-sensitive cell line. The MDR cell line, KB-C2, over-expressing P-gp, was induced from KB-3-1 with colchicine. Both cell lines were kindly provided by Dr. Shinichi Akiyama (Kagoshima University, Japan). Non-small cell lung cancer (NSCLC) NCl-H460 cell line and its resistant subline H460/MX20 were kindly provided by Drs. Susan E. Bates and Robert W. Robey (NIH, Bethesda, MD). H460/MX80 cells with enhanced drug-resistant ability were generated by inducing H460/MX20 cells with mitoxantrone at increasing concentrations up to 80 μM. KB-3-1 and its derivative cell subline KB-C2 and its knocked-out sublines without the PI3K 110α and 110β subunits were cultured with DMEM supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified incubator containing 5% CO₂ at 37 °C. H460, H460/MX20, H460/MX80 and the derivative gene deficient cells were cultured with RPMI1640 medium under the same conditions. CRISPR/Cas9 all-in-one plasmids encoding SgRNA and Cas9 for knock-out were purchased from GeneCopoeia Inc. (Rockville, MD). KB-C2 and H460/MX80 showing high MDR were used for evaluation of the intensive reversal of MDR after P110α or P110β subunit was knocked out in vitro by using the CRISPR/Cas9 technique.

BAY-1082439, the PI3K 110α/β inhibitor, and GSK-2110183 (afuresertib), the inhibitor of AKT (AKT1, AKT2 or AKT3), were purchased from ChemieTek (Indianapolis, IN). Paclitaxel, mitoxantrone, doxorubicin and 3-(4,5-Dimethylthiazol-yl)-2,5-diphenyltetrazolium bromide (MTT), dimethyl sulfoxide (DMSO) were purchased form Sigma Chemical Co. (St. Louis, MO, USA). Mouse originated anti-P-gp and anti-BCRP, HRP ligated or fluorescent secondary rabbit or goat-anti mouse antibodies were purchased from Invitrogen, Thermo Fisher (Carlsbad, CA). Mouse-anti-PIK3CA antibodies were purchased from BD Biosciences (San Jose, CA) and anti-PIK3CB and anti-AKT Pan antibodies were purchased from R&D systems (Minneapolis, MN). Rabbit-anti-GSK3β, mouse-anti-P53, rabbit-anti-phospho-FOXO3a (Ser253) and rabbit-anti-Caspase-9 were purchased form Beyotime Institute of Biotechnology (Shanghai, China). Other reagents were purchased from VWR International (West Chester, PA, USA).

MTT assay was used to determine the non-toxic concentrations of the inhibitors (generating a viability of over 80%). Exponentially growing cells were treated with trypsin and seeded into 96 well plates at 5 × 10³ cells/well. After 72 h of cell culture, 20 μL of MTT (5 mg/mL) was added to each well and incubated for 4 h. The medium containing MTT was removed carefully and 150 μL of DMSO was added to each well. The plates were agitated until the dark blue crystal dissolved completely. The absorbance was measured using an ELx 800 Universal Microplate Reader (Bio-Tek, Inc. Winooski, Vermont) at wavelength of 490 nm. All MTT assays were performed with three repeated treatments and independently repeated three times. The survival rate was statistically analyzed by SPSS 20.0 (SPSS, Chicago, IL, USA), and the curves were generated by Origin 9.0. The data are expressed as the mean ± standard error of the mean (SEM). Student’s t-test was used for analyzing the differences between groups. Differences were statistically significant at P < 0.05. IC₅₀ was analyzed according to the survival rate-concentration curve.

Cells were seeded into 96-well plates at a density of 5 × 10³ cells per well and cultured for 8 h. The cell cultures were treated with BAY-1082439 or GSK-2110183 for 1 or 2 h, followed by gradient concentrations of the anti-cancer drugs, starting from 0.001 μM as a non-toxic concentration. Decrease in the viability of MDR cells and in the IC₅₀ of the anti-cancer drugs, together indicating MDR reversal efficacy, were evaluated by the MTT assay. Two specific drug substrates of P-gp and BCRP, colchicine and mitoxantrone, were used as antitumor drugs to induce high level of P-gp in KB-C2 or BCRP in H460/MX20 and H460/MX80, respectively.

Parental drug-sensitive cells and drug-induced MDR cells were treated with inhibitor at the same conditions set for MTT assay. After 24 to 72 h of culture, Western blot and immune-fluorescence (IF) were conducted to determine the levels of P-gp and BCRP protein expression.

For the Western blot, the cells were lysed with SDS lysate reagents and boiled with 5× SDS-PAGE sample loading buffer. The proteins were then separated by electrophoresis, followed by regular protocols for the Western blot (refer to product instructions of the antibodies).

For IF analysis, KB-C2 and H460/MX80 cells were washed with PBS buffer, fixed with formaldehyde for 10 min at RT, followed by standard protocols for binding with primary and fluorescent antibodies provided by the manufacturers, and observed with Thorlabs (Newton, New Jersey) fluorescent microscope systems. The experiment was repeated independently for three times.

Because doxorubicin is an fluorescent substrate of both P-gp and BCRP and can be pumped out from the cells by both ABC transporters, it was used at low concentration that can keep cells healthy and report whether BAY-1082439 can interact directly with these two transporters and inhibit the accumulation of doxorubicin within the cells. The parental and MDR cells were seeded into 96-well plates at a density of 5 × 10³ cells per well and cultured for 8 h. The cells were co-incubated with 10 μM of BAY-1082439 for 1 h. Doxorubicin (0.2 μM) was then used to supplement the cell cultures and co-incubated with the cells for 2 h. This short period of time and relatively low concentration of drug were applied to keep cells alive. After gently washing with PBS buffer thrice, the cells were lysed with 100 μL lysis buffer, quantified with BCA Protein Assay Kit and diluted with lysis buffer to constant concentration. Fluorescent intensity indicating doxorubicin accumulation within the attached cells was determined in a Synergy™ 4 Multi-Mode Microplate Reader (Bio Tek Instruments, Inc. Vermont, USA) (Ex/Em: 450/550 nm. Plate type: Corning 96 Flat Bottom black, clear bottom Polystyrene). The experiment was repeated three times, and the results were analyzed with SPSS 20.0 using Student’s t-test.

For the long-term consequence of the weakened efflux ability, cells were seeded at 3 × 10³ cells per well, cultured for 8 h, and treated with BAY-1082439 (10 μM) and doxorubicin (1 μM for the MDR cells and 0.1 μM for drug-sensitive cells to provide drug pressure and typical cell viability not less than 60%) for 72 h. Because of the inaccuracy of quantification of doxorubicin due to membrane injury of the dead or severely apoptotic cells, only fluorescent images showing distribution of doxorubicin were recorded as supplementary figures. This experiment was independently repeated three times to have statistical significance.

The ATPase activity of P-gp and BCRP in crude membranes of High-Five insect cells was measured in the presence of BAY-1082439 (0 to 40 μM) by PREDEASY ATPase Kits with modified protocols, as previously described [23, 24].

Structure information of the interactions between BAY-1082439 and target proteins was calculated on PatchDock Server based on molecular docking and simulation of stable structures of the complexes. PyMOL (version 1.8.x) was used to analyze the data and present the results that indicate the most stable structures, binding forces and positions, and residues or chemical groups.

The cells were seeded into a 6-well plate at 1.2 × 10⁶ cells per mL and cultured with respective anti-tumor drugs and the P110α/P110β inhibitor BAY-1082439 at non-cytotoxic concentration (10 μM) for 48 h. Untreated cells were used as a control. Cell apoptosis and necrosis were determined with an Annexin V-FITC Apoptosis Detection Kit (Beijing Solarbio Science & Technology Co., Ltd., Beijing) in a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA). To slow down the growth rate of the cells set as control (MDR cells treated with drug only, or the drug-sensitive cells that were not influenced by BAY-1082439), we used 5% FBS as substitution for the 10% FBS in the media. And before the samples were tested in the flow cytometer, trypsin was used to treat the cells for better dispersity. The cells were dispersed with 2 mL of detection buffer provided with the detection Kit. The experiments were independently repeated three times. Results from a representative experiment were presented.

For the study of the regulatory functions of PI3K-110α or PI3K-110β on ABC transporters (P-gp and BCRP), the CRISPR/Cas9 knockout system was introduced for specific deletion of Pik3ca or Pik3cb gene. KB-C2 or H460/MX80 cells were detached with trypsin, washed with PBS and seeded to 48-well plates with serum-free DMEM or RPMI1640 media, respectively. To avoid the potential risk of un-specified chromosomal recombination during transfection mediated by viral vectors, polymer-based GenePORTER transfection reagents for gene delivery were mixed with plasmids HCP213150-CG01–1-10 and HCP213151-CG01–1-10, developed as sgRNA/Cas9 all-in-one expression clones respectively targeting PIK3CA (NM_006218.2) and PIK3CB (NM_006219.1) (GeneCopoeia Inc., Rockville, MD). This was then incubated at RT for 20 min, and added to the cell cultures. After incubation for 4 h, FBS was added to the cells at 10% of the volume. The cells were transfected again using the same method after 48 h of culture to achieve high transfection efficiency. During transfection and cell stabilization, anti-tumor drugs were not applied in order to maintain survival of the cells with Pik3ca or Pik3cb deletion that may attenuate the expression of ABC transporters and reverse MDR capability. Transcription and expression of target genes were determined by RT-PCR and Western blot after culturing the gene deficient cells for 1 month. Healthy and stable cell populations with deficiency of P110α or P110β were obtained by adequate culture.

To determine transcription of PI3K 110α subunits, Primers (5′) CCTCCACGACCATCAT (3′) and (5′) TGCCTACTGGTTCAAT (3′) were used for RT-PCR, and for PI3K 110β subunits, primers (5′) TGCTTCAGTTTCATAA (3′) and (5′) GAAGAAAAGGTCTGAC (3′) were used.

In addition to the MTT assay, a JC-1 mitochondrial membrane potential analysis was performed to compare drug sensitivity of the cells before and after knock-out of the target genes Pik3ca and Pik3cb [25]. The protein expression of P-gp and BCRP in the knockout cell lines were analyzed with Western blot.

MTT assay of the drug-sensitive and the drug-resistant cells was performed as mentioned previously, to evaluate whether GSK-2110183, an inhibitor specific for AKT, can reverse MDR of the cancer cells over-expressing P-gp and BCRP.

Fluorescent immunoblotting analysis for protein levels was performed with Cytell™ Image Cytometer (GE Healthcare, Washington) based on 9–11 fields per well and at least three repeated experiments. Briefly, the cells were seeded into 96-well plates at a density of 5 × 10³ cells per well and cultured for required time with corresponding treatment in each experiment. Before analyzing, the cells were fixed with formaldehyde (4%), followed by a 10-min treatment with 0.1% of Triton-100 and then 6% of BSA. The cells were co-incubated with target-specific primary antibodies for 1 h at 37 ̊C, followed by adequate wash with PBS (pH 8.0) and a 30-min co-incubation with the Cy3 labeled secondary antibodies. After four times of wash with PBS, 0.5 μg mL^− 1 of DAPi in PBS was applied to staining of the living cells and characterization. Average fluorescent intensity of the markers (Cy3) per cell was automatically calculated by dividing total Cy3 fluorescent intensity by total nuclei area (indicated by DAPi) in correspondence with the cell counts.

BAY-1082439 is a small-molecule that specifically inhibits PI3K class I P110α, P110β, and mutated forms of P110α (Fig. 1a). The survival rate of the KB-C2 and H460/MX20 MDR cells was greatly reduced in the presence of mid-to-high concentrations of the respective antitumor drugs and increasing concentrations of BAY-1082439, indicating greatly enhanced sensitivity of the cells to these anticancer drugs (Fig. 1a). The cell viability of the parental KB-3-1 and H460 cells was not apparently changed upon treatment with BAY-1082439 and the antitumor drugs. Compared to that of the parental cells, the cell viability of the KB-C2 and H460/MX20 cells was decreased, which indicated that inhibiting the PI3K 110α and 110β subunits with nontoxic or minimally toxic concentrations of BAY-1082439 can potentially reverse the P-gp- or BCRP- mediated MDR in these cells, which express the drug-efflux transporters P-gp or BCRP, a primary characteristic that distinguishes them from their parent cells.

P-gp was downregulated in KB-C2 cells cotreated with colchicine and BAY-1082439 for only 24 h. After the dual treatment, the amount of P-gp did not exceed 18% of the P-gp in the cells treated with only colchicine (Fig. 1b). In the H460/MX20 cells, the BCRP was downregulated to approximately 20% that of the cells cotreated with mitoxantrone and BAY-1082439 after 48 h of coculture (Fig. 1b). Immunofluorescence microscopy showing the variations in the P-gp and BCRP levels and the distributions of each on the cell surface and within the cells also indicated that both P-gp and BCRP expressed by the drug-treated MDR cancer cells were downregulated by BAY-1082439 (Fig. 1c). Therefore, inhibition of PI3K 110α and/or 110β by BAY-1082439 may directly reduce the MDR protection induced by ABC transporters by downregulating their expression. These results indicated that P-gp and BCRP may have at least one common regulation factor, i.e., the P110α and/or P110β subunits in the PI3K signaling pathway.

Model complex structure showed the interactions stabilizing BAY1082439/PIK3CA and BAY-1082439/PIK3CB complexes (Additional file 1: Figure S1). The binding information of BAY-1082439 with P-gp and of BAY-1082439 with BCRP was also delineated through structural analysis in a statistical prediction framework, which was used because no information has yet been reported. The results revealed interactions between BAY-1082439 and human P-gp and BCRP, as shown in Fig. 2.

P-gp comprises two symmetrical transmembrane domains (TMDs) and two cytoplasmic nucleotide-binding domains (NBDs). The NBDs form a dimer that occludes two ATPs [26]. A docking analysis generated the top 10 models showing the likely interaction between BAY-1082439 and P-gp and the localization of BAY-1082439 within P-gp. The number 1 model indicated that BAY-1082439 has the highest likelihood of binding to the central space of the TM cavity in the human P-gp that is in an outward-facing state (Fig. 2). A polar interaction was revealed between the BAY-1082439 molecule and the Val-835 residue of P-gp. These results indicated that BAY-1082439 has the ability to bind to human P-gp and may influence its ATPase activity.

The BAY-1082439/BCRP complex was predicted to have a stable structure when BAY-1082439 was bound to the internal membrane side of the BCRP dimer (Fig. 2). The BCRP dimer occluded BAY-1082439 at the near-surface cavity. In addition to surface adaptability, BAY-1082439 can dock at BCRP through the unsaturated nitrogen in the 2-pyridine interacting with Arg-193; the oxygen on the carbonyl group and the aldimine interacting with Glu-317; and -CH (adjacent to -O-C-) at 1-benzene interacting with ILE-318. Compared with a recently reported series of drug/BCRP crystal structures, this predicted binding location for BAY-1082439/BCRP was not typical. The predicted complex structure of BAY-1082439/BCRP indicated that BAY-1082439 has the ability to bind to human BCRP and may also influence its ATPase activity.

Although ATP binding to P-gp, but not ATP hydrolysis, has recently been reported to promote the release of various substrates [26], ATPase alterations can theoretically provide evidence for the induced interactions between small inhibitors/activators and ABC transporters. The results showed that the vanadate-sensitive ATPase activity of P-gp or BCRP was positively correlated with the amount of BAY-1082439 at the tested concentrations; in general, they had a fold increase of 3.3- or 1.9-fold, respectively, compared to the basal activity levels (Additional file 2: Figure S2A), indicating a stimulation of ATPase activity by BAY-1082439 in a concentration-dependent manner. The stimulation of ATPase activity may accelerate the efflux of substances under specific circumstances; however, binding of BAY-1082439 to P-gp or BCRP might also lead to P-gp or BCRP transporter mediated efflux of BAY-1082439 such that it is in competition with the antitumor drugs and results in the increase in drug accumulation within the cells. In general, our studies indicated that, upon binding to P-gp and BCRP, BAY-1082439 enhanced the ATPase activity of both P-gp and BCRP, and this finding implies the potential of BAY-1082439 to increase drug accumulation by blocking the efflux activity of both P-gp and BCRP.

As indicated by the data analyzed in Additional file 2: Figure S2B, the fluorescent intensity of KB-C2 treated with Dox (0.2 μM) was 45,407 URL mg^− 1 protein. Co-culture with BAY-1082439 (10 μM) elevated the fluorescent intensity to 76,036 URL mg^− 1 protein. The fluorescent intensity of KB-3-1 treated and not treated with BAY-1082439 was 211,462.7 URL mg^− 1 and 196,077 URL mg^− 1 protein, respectively. The fluorescent intensity of the H460/MX20 cells was elevated from 110,496.8 to 177,531.1 URL mg^− 1 protein when the cells were treated with BAY-1082439. The fluorescent intensity of H460 with and without treatment of BAY-1082439 was 242,648.8 and 208,913.3 URL mg^− 1 protein, respectively. The results indicated that BAY-1082439 had less inhibitory impact on the P-gp and BCRP transporters to prompt an increase of doxorubicin accumulation.

Amazingly, downregulation of P-gp and BCRP that causes long-term reversal of drug efflux behavior may have an impact on the increased drug sensitivity of KB-C2 and H460/MX80 cells in the presence of BAY-1082439, as BAY-1082439 is highly specific only for the PI3K 110α and 110β subunits (Additional file 2: Figures S2C, Additional file 3: Figure S3). Although doxorubicin is a potent intercalator of DNA, no apparent accumulation of doxorubicin appeared in the DOX (+)/BAY-1082439 (−) treatment groups even though 1 μM doxorubicin (which is a physiological concentration) was used. This outcome could have been caused by the highly efficient efflux of the doxorubicin by the KB-C2 and H460/MX80 cells with highly expressed levels of P-gp and BCRP, respectively, which endowed them with remarkably strong MDR ability. Interestingly, the florescence intensity of the sensitive cell lines was several-fold higher than that of the DOX (+)/BAY-1082439 (-)-treated MDR cells; this could be caused by the rapid increase in the accumulation of doxorubicin due to the lack of expressed P-gp or BCRP, which can efflux doxorubicin from MDR cells. According to the concentrations of doxorubicin, the drug-sensitive cells treated with 0.1 μM doxorubicin internalized lower amounts of drugs than the MDR cells treated with 1 μM doxorubicin and BAY-1082439, which inhibited P-gp and BCRP expression and reversed drug resistance.

Both the KB-C2 and H460/MX80 cells showed a remarkable decrease in apoptosis resistance in the presence of BAY-1082439 in addition to the antitumor drugs specific to P-gp or BCRP, as indicated by flow cytometry. The ratio of apoptotic to necrotic KB-C2 cells cotreated with BAY-1082439 and colchicine was increased to approximately 42%, compared with 33% in the cells treated with colchicine only, with the major contribution made by the change in cell apoptosis, which was increased to 22% from 14% (Fig. 3a). The H460/MX80 cells treated with additional BAY-1082439 at a noncytotoxic concentration showed a two-fold rate of apoptosis compared with the control cells, of which 16% underwent apoptosis (Fig. 3b). The KB-3-1 and H460 cells did not undergo reversed antiapoptosis when treated with the same dose of BAY-1082439 in addition to the antitumor drugs.

The inhibition of P110α or P110β by BAY-1082439, which in turn downregulates the P-gp and BCRP transporters, requires validation, since some PI3K inhibitors, such as CUDC-907 and afatinib, can nonspecifically inhibit other targets, such as histone deacetylase and P-gp [27, 28]. In addition, P-gp or BCRP may not be completely downregulated because of inadequate BAY-1082439 inhibition of the P110α or P110β subunit. Two cycles of gene transfection led to the successful delivery of CRISPR/Cas9 plasmids to nearly 100% of the receptor cells (Fig. 4a-c). The deletion of target genes (~ 66 kDa P110α with mutation or 110 kDa P110β in both the KB-C2 and H460/MX80 cells) in the vast majority of derivative cells and the truncation of the N′-terminus of PIK3CB in the 110β subunit-k.o. MX80 cells via gene recombination were detected at both the translational level (Fig. 4b) and the transcriptional level (Fig. 4c). Immunofluorescence also showed the knockout of target genes in most 110α subunit-k.o. and 110β subunit-k.o. KB-C2 cells and showed weak signals in the 110β subunit-k.o. H460/MX80 cells due to the N′-terminal truncation of P110β (Additional file 4: Figure S4). In comparison, nontargeted P110β in the P110α-deficient cell (110α subunit-k.o. KB-C2 or 110α subunit-k.o. H460/MX80) populations and P110α in the P110β-deficient cell (110β subunit-k.o. KB-C2 or 110β subunit-k.o. H460/MX80) populations were detected with positive fluorescence signals, indicating successful knockout of the targeted P110α or P110β subunits only. The P110α- and P110β-deficient KB-C2 cells were named KB-C2-k.o.110α and KB-C2-k.o.110β; the P110α- and P110β-deficient H460/MX80 were named MX80-k.o.110α and MX80-k.o.110β, respectively.

KB-C2 cells with either Pik3ca or Pik3cb gene deficiency were sensitive to both colchicine and paclitaxel. The IC₅₀ values indicated that 110α subunit-k.o. KB-C2 cells were only 2.36-fold more sensitive to colchicine than the drug-sensitive KB-3-1 cells (Fig. 4d). Approximately 26.3% of the IC₅₀ level of paclitaxel remained in the 110α subunit-k.o. KB-C2 cells compared with that of the parental cell line, KB-C2. 110β subunit-k.o. KB-C2 cells treated with colchicine and paclitaxel maintained 13.7% and 19.4% of the respective IC₅₀ level in the KB-C2 cells (Fig. 4e).

The drug resistance of the 110α subunit-k.o. H460/MX80 cells was reduced to a low level, approaching that of the drug-sensitive parental H460 cells, and the IC₅₀ value of the remaining 110α subunit-k.o. H460/MX80 cells was only 5–6-fold greater than that of the parental H460 cells (Fig. 4f). The MDR ability of 110β subunit-k.o. H460/MX80 cells compared with that of 110α subunit-k.o. H460/MX80 cells was relatively unchanged. Deletion of P110α in the KB-C2 cells also led to a more substantial reversal of P-gp-mediated drug resistance compared to that caused by the deletion of P110β. These results suggest that PI3K 110α could have more important function in the regulation of MDR mediated by ABC transporters, including P-gp and BCRP.

P110β deficiency in the KB-C2 and H460/MX80 cells generated differences in cell viability in the presence of respective antitumor drugs. Results from the JC-1 analysis showed severe depolarization of the mitochondrial membrane, indicating cell apoptosis of 110β subunit-k.o. KB-C2 cells after 1 μM paclitaxel treatment (Additional file 5: Figure S5). Although the MTT assay results indicated no apparent evidence for increased apoptosis of the 110β subunit-k.o. H460/MX80 cells, compared with the number of the H460/MX80 cells undergoing apoptosis, when both were treated with mitoxantrone at 10 μM, the weak green fluorescence of the JC-1 monomers indicated slightly elevated apoptosis of the 110β subunit-k.o. H460/MX80 cells.

Deficiency in P110α and P110β was associated with the downregulation of the two main MDR-relevant transporters, P-gp and BCRP, in both the KB-C2 and H460/MX80 cells, as indicated by Western blot and IF-Cytell analyses (Fig. 5a-c). As indicted by the fluorescent marker intensity value, determined by IF in combination with statistical analysis completed with a Cytell™ image cytometer, the representative expressed marker intensity was 649, 647.5, and 635.5 in the KB-C2, 110α subunit-k.o. KB-C2 and 110β subunit-k.o. KB-C2 cell populations, respectively, and was 770, 749 and 726 in the H460/MX80 and the P110α- or P110β-knockout cell populations, respectively. The expression of AKT was not altered significantly in the P110-knockout cell populations. This finding challenges a formerly presented supposition that AKT expression might be directly associated with mutated P110α subunits, a premise based on a tumor sample analysis instead of gene knockouts [29]. This phenomenon could further support the hypothesis that MDR reversal caused by P110-knockout may not be dependent on AKT expression levels.

As some researchers have previously suggested, AKT activated by PI3K might be associated with the MDR of cancer cells [30, 31]. Our studies proposed that the AKT-mediated MDR-enhancing pathway might be independent of the regulation of ABC transporter-mediated MDR and may be activated during crisis conditions, such as when the ABC transporters fail to efficiently pump toxins out of the cells.

Although several studies have sought to determine whether PI3K catalytic subunits are related to ABC transporter-mediated MDR through the PI3K/AKT pathways, a direct regulatory target has not been confirmed. In this part of our study, we found contradictory results when we cotreated MDR cells with GSK-2110183 (an ATP-competitive inhibitor specific for AKT) and antitumor drugs specifically pumped out by each overexpressed ABC transporter.

None of the examined cancer cells exhibited changes in sensitivity to the antitumor drugs after being cultured with GSK-2110183; however, the cells not treated with antitumor drugs showed the inhibitory effects of GSK-2110183 treatment (Fig. 5d and e). Considering the evidence from the studies described above, we suggest that the inhibition of AKT activity does not have a regulatory effect on P-gp- or BCRP-mediated MDR in these cells.

GSK-2110183 administered alone to inhibit AKT (without antitumor drugs) inhibited the cell viability of the tested cell lines, except for the KB-C2 cells, in a dose-dependent manner (Fig. 5f). This finding may indicate that inhibition of AKT in the H460/MX80 cells generates a greater influence on cell viability than it does in the KB-C2 cells. On the basis of the combined observation that the AKT levels did not significantly change in the P110-deficient KB-C2 or H460/MX80 cells, we deemed that AKT may have a greater influence on enhancing H460/MX80 cell survival and proliferation and/or antiapoptosis, which could be independent of the MDR mediated by P-gp or BCRP.

To demonstrate that the downregulation of P-gp or BCRP mediated by the inhibition of P110α and P110β with BAY-1082439 was independent of AKT expression, Western blot analysis was performed, and the results showed minimally enhanced expression of AKT in the cells treated with BAY-1082439 (Fig. 6). We then studied the effect of BAY-1082439 on the MDR and the parental cells on the basis of the expression levels of AKT and five pivotal proteins downstream of AKT in the PI3K/AKT pathways. In addition to Western blot analysis, a Cytell™ image cytometer was used for the quantitative analysis and comparison of the Cy3 signal intensity, which indicated the amounts of the protein targets that had bound to the antibodies (Cy3-labeled). The five other proteins are 1. glycogen synthase kinase 3 beta (GSK3β), which has multiple functions in the regulation of multiple downstream pathways (such as GYS metabolism, Myc cell cycle progression and CCND1 cell cycle progression); 2. P53 (also known as a proapoptotic protein), which regulates cell survival; 3. phosphorylated Forkhead box O3 (phospho-FOXO3a or phospho-FOXO3), which has multiple functions in the regulation of multiple factors and pathways (such as DNA-PEPCK/G6Pace metabolism, DNA-CCNDY cell cycle progression, DNA-P27KIP1/RBL2 cell cycle progression, and DNA-FasL/Bim cell survival); 4. caspase-9 (also an important apoptosis factor that regulates multiple apoptosis pathways) regulates cell survival; and 5. Bcl-xL, which is intimately related to cell survival. The results showed that, compared with the cells treated with antitumor drug alone, the MDR KB-C2 cells treated with BAY-1082439 in addition to colchicine expressed increased levels of AKT, GSK3β, phospho-FOXO3a and caspase-9 and equal levels of P53 and Bcl-xL, indicating there was no negative influence from the inhibition that P110α or P110β induced on the expression level of AKT or that of its downstream factors, indicating the ability of BAY-1082439 to induce magnified caspase-9-induced apoptosis signals and possible cell adaptation induced by other factors/pathways in addition to P110α or P110β (Fig. 6). The H460/MX80 or H460/MX20 cells treated with BAY-1082439 in addition to mitoxantrone expressed increased levels of AKT and caspase-9 and unchanged levels of the GSK3β, phospho-FOXO3a and P53 proteins, showing little negative influence from the inhibition of P110α or P110β on the level of AKT expression and that of its pivotal downstream factors, but it indicated severe apoptotic effects caused by BAY-1082439.

The drug-sensitive parental KB-3-1 cells were unaltered when the expression of AKT was changed, and all the examined factors, including caspase-9, showed little influence of BAY-1082439 on the PI3K/AKT pathways or on the growth and survival of the KB-3-1 cells (Fig. 6) Combined with the data from the BAY-1082439-induced downregulation of P-gp in the KB-C2 cells, but not in the KB-3-1 cells, the findings indicate that BAY-1082439 has a more severe effect on KB-C2 cells because of the downregulation of P-gp, which affects the efflux of drugs from the cells.

The H460 cells showed a minimal increase in AKT and P53; unchanged Bcl-xL levels; a slight decrease in GSK3β, phospho-FOXO3a and caspase-9; and unchanged levels of phospho-FOXO3a to BAY-1082439 treatment, showing good cell survivability and weak influence of BAY-1082439 on PI3K/AKT or its downstream pathways (Fig. 6). These results also explained the mechanisms inducing the increased apoptosis of the BAY-1082439-treated MDR H460/MX80 cells, but not the drug-sensitive H460 cells, as was also shown by the results from the flow cytometry analysis (Fig. 3). The major difference between the H460/MX80 and H460 cells is the degree to which downregulated BCRP in the H460/MX80 cells induced this phenomenon. Similar to the KB-C2 cells, BAY-1082439 treatment led to greater accumulation of the antitumor drugs in the cells, because of inhibited P-gp or BCRP expression, inducing increased apoptosis of these MDR cells, as suggested by caspase-9 levels and other results from the flow cytometry analysis (Fig. 3, Fig. 6). These phenomena also support the conclusion that downregulated P-gp or BCRP expression mediated by inhibition of P110α and P110β may be due to the action of pathways that are not regulated by AKT.