Decrease in Multidrug Resistance-associated Protein 2 Activities by Knockdown of Phosphatidylinositol 4-phosphate 5-kinase in Hepatocytes and Cancer Cells

- Purpose: The plasma membrane localization and transport activity of multidrug resistance-associated protein 2 (MRP2/ABCC2) and P-glycoprotein (P-gp/ABCB1) efflux transporters are governed by transporter-associated proteins. Phosphatidylinositol 4,5-bisphosphate (PIP 2 ) formed by phosphatidylinositol 4-phosphate 5-kinase type 1 (PIP5K1) activates the linker function of radixin for efflux transporters. Radixin is involved in the plasma membrane localization of efflux transporters. We examined whether PIP5K1 could be a target for the modulation of transporter activities in hepatocytes and cancer cells. Methods: The effects of PIP5K1 depletion by siRNA in mouse primary hepatocytes, PANC1 human pancreatic carcinoma cells, and HepG2 human hepatocellular carcinoma cells on the intracellular accumulation of MRP2 and P-gp substrates were examined. Results: PIP5K1A depletion resulted in increased intracellular accumulation of carboxydichlorofluorescein, a MRP2 fluorescent substrate, in mouse primary hepatocytes, PANC1 cells, and HepG2 cells. In PANC1 and HepG2 cells, the transport activities of MRP2 were significantly decreased by PIP5K1C depletion. However, the transport activities of P-gp were unchanged by PIP5K1 depletion. PIP 2 levels were unchanged between control and PIP5K1A- or PIP5K1C-depleted HepG2 cells. MRP2 mRNA levels showed few changes in HepG2 cells following PIP5K1A or PIP5K1C depletion. The expression of phosphorylated radixin was decreased by PIP5K1A and PIP5K1C depletion, although total radixin levels were unchanged. Conclusions: These data suggest that PIP5K1A and PIP5K1C could be target proteins for modulating MRP2 function, partly because of the resulting changes of the linker function of radixin.

ABSTRACT -Purpose: The plasma membrane localization and transport activity of multidrug resistanceassociated protein 2 (MRP2/ABCC2) and P-glycoprotein (P-gp/ABCB1) efflux transporters are governed by transporter-associated proteins. Phosphatidylinositol 4,5-bisphosphate (PIP2) formed by phosphatidylinositol 4-phosphate 5-kinase type 1 (PIP5K1) activates the linker function of radixin for efflux transporters. Radixin is involved in the plasma membrane localization of efflux transporters. We examined whether PIP5K1 could be a target for the modulation of transporter activities in hepatocytes and cancer cells. Methods: The effects of PIP5K1 depletion by siRNA in mouse primary hepatocytes, PANC1 human pancreatic carcinoma cells, and HepG2 human hepatocellular carcinoma cells on the intracellular accumulation of MRP2 and P-gp substrates were examined. Results: PIP5K1A depletion resulted in increased intracellular accumulation of carboxydichlorofluorescein, a MRP2 fluorescent substrate, in mouse primary hepatocytes, PANC1 cells, and HepG2 cells. In PANC1 and HepG2 cells, the transport activities of MRP2 were significantly decreased by PIP5K1C depletion. However, the transport activities of P-gp were unchanged by PIP5K1 depletion. PIP2 levels were unchanged between control and PIP5K1A-or PIP5K1C-depleted HepG2 cells. MRP2 mRNA levels showed few changes in HepG2 cells following PIP5K1A or PIP5K1C depletion. The expression of phosphorylated radixin was decreased by PIP5K1A and PIP5K1C depletion, although total radixin levels were unchanged. Conclusions: These data suggest that PIP5K1A and PIP5K1C could be target proteins for modulating MRP2 function, partly because of the resulting changes of the linker function of radixin. _______________________________________________________________________________________ PURPOSE ATP-binding cassette (ABC) transporters such as multidrug resistance-associated protein 2 (MRP2/ABCC2) and P-glycoprotein (P-gp/ABCB1), which function as efflux transporters, are determinants of the pharmacokinetics of some drugs (1)(2)(3). In the blood-brain barrier, P-gp and MRP2 transport intracellular drugs to the blood to protect the central nervous system (4). In the liver, MRP2 expressed on the canalicular membrane participates in the biliary excretion of bile acid and drug metabolites, and MRP2 dysfunction due to genetic diseases or drug inhibition causes intrahepatic cholestasis (5). The overexpression of MRPs and P-gp in cancer cells causes resistance to anti-cancer drugs such as paclitaxel, etoposide, and doxorubicin (1)(2)(3)6).
The plasma membrane localization and transport activity of efflux transporters is governed by transporter-associated proteins such as phosphatidylinositol 4,5-bisphosphate (PIP2), the ezrin/radixin/moesin (ERM) family proteins, and Na + /H + exchanger regulating factor 1/ERM binding phosphoprotein of 50 kDa (7)(8)(9). ERM proteins are activated by C-terminal phosphorylation by PIP2 and subsequently function as a linker between efflux transporters and actin filaments (8,(10)(11)(12)(13). Although the involvement of ERMs such as radixin in the plasma membrane localization and transport activity of MRP2 is well defined (14), few reports have examined the relation between PIP2 activity and MRP2.
The membrane phospholipid PIP2 is involved in a variety of cellular signaling pathways, such as cytoskeletal signaling, tissue remodeling, and vesicle trafficking (15), and regulates the membrane localization of ion channels and transporters (16,17). _________________________________________ PIP2 is produced by the phosphorylation of phosphatidylinositol 4-phosphate (PI(4)P) or phosphatidylinositol 5-phosphate (PI(5)P) and dephosphorylation of phosphatidylinositol (3,4,5)trisphosphate (18)(19)(20)(21). Phosphorylation of PI(4)P by the lipid kinase phosphatidylinositol 4-phosphate 5kinase type 1 (PIP5K1) is the predominant pathway for PIP2 biosynthesis (22). The intracellular levels of PI(4)P are approximately 10-fold higher than those of PI(5)P. PIP5K1 has three isozymes: PIP5K1A (human PIP5Kα and mouse PIP5Kβ), PIP5K1B (human PIP5Kβ and mouse PIP5Kα), and PIP5K1C (human and mouse PIP5Kγ). Previous studies showed that PIP5K1A overexpression is involved in the malignant transformation of prostate cancer cells (23). Cytosolic PIP5K1A mainly participates in the generation of PIP2 in humans (24). PIP5K1Bknockout mice exhibited anaphylactic responses caused by increased cytokine production by mast cells (25). PIP5K1C is highly expressed in the brain and is involved in the endocytosis-mediated recovery of synaptic vesicles (26). PIP5K1C binds to the scaffold protein talin to promote integrin expression on the plasma membrane (27). Previous studies showed that the activities of PIP5Ks are higher in cancer cells than in normal cells (28). However, the roles of PIP5K1A, PIP5K1B, and PIP5K1C in the transport activity of MRP2 remain unclear.
Determining the potential function of PIP5Ks in regulating MRP2 function may help elucidate MRP2 interacting proteins and other related pathways in hepatocytes and cancer cells. Therefore, here we investigated whether PIP5K1A, PIP5K1B, or PIP5K1C are involved in the transport activity of MRP2 in hepatocytes and cancer cells.

Cell culture
Hepatocytes were isolated via two-step collagenase perfusion as described previously (29). Hepatocytes were seeded at 1 × 10 5 cells/cm 2 in collagen-coated plates (Sumitomo Bakelite, Tokyo, Japan) in Williams' Medium E supplemented with 10% fetal bovine serum, 1 nM insulin, 1 nM dexamethasone, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM GlutaMAX. Cells were incubated at 37°C in the presence of 5% CO2 and 95% air. After cell attachment for 3 h, the medium was replaced with fresh serum-free Williams' Medium E containing the aforementioned supplements and 5 U/ml aprotinin.
Human pancreas cancer PANC1 and human liver cancer HepG2 cells (obtained from RIKEN Cell Bank, Ibaraki, Japan) were maintained in RPMI-1640 and DMEM, respectively, supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin. Cells were grown at 37°C in a humidified incubator equilibrated with 5% CO2 and were subcultured every 3-4 days using 0.25% trypsin and 1 mM EDTA. The medium was replaced every 2 days. PANC1 and HepG2 cells were seeded at 2.5 × 10 4 cells/cm 2 in 24-well plates.

Transfection of siRNA
Mouse primary hepatocytes were treated with 5 pmol/well siRNA targeting PIP5K1A (siPIP5K1A) or siNegative complexed with Lipofectamine RNAiMAX in Opti-MEM. PANC1 cells were transfected with siRNA (siPIP5K1A, siPIP5K1B, siPIP5K1C, siNegative; 5 pmol/well) complexed with Lipofectamine RNAiMAX in Opti-MEM. In HepG2 cells, reverse transfection (the addition of cells to siRNA complexes [5 pmol/well] on a plate) was performed. At 48 h after transfection, the cells were used for experiments.
Determination of expression of PIP5K1A, PIP5K1B and PIP5K1C by targeted proteomics The protein expression levels of PIP5K1A, PIP5K1B and PIP5K1C were determined by targeted proteomics as described previously (32,33). Briefly, cell lysates (50 g) diluted in 25 l phase transfer surfactant buffer were treated with 10 mM dithiothreitol and 42 mM iodoacetamide and digested by 1 g trypsin at 37°C for 18 h. Peptides were separated on a reversed-phase column (AdvanceBio Peptide Map C18, 2.1 × 150 mm, 2.7 m; Agilent Technologies, Richardson, TX, USA) and detected on a LC-MS/MS apparatus consisting of an LC system (UltiMate 3000 series, Thermo Fisher Scientific, Waltham, MA, USA) and a TSQ Endura Triple Quadrupole Mass Spectrometer with electrospray ionization (Thermo Fisher Scientific).
Functional studies of MRP2 and P-gp with fluorescent substrate Hepatocytes, HepG2 cells, and PANC1 cells were treated with siPIP5K1A, siPIP5K1B, siPIP5K1C, or siNegative for 48 h; the medium was removed and the cells were incubated in Hanks' balanced salt solution containing MK571 (50 µM) and verapamil (50 M) for 30 or 60 min at 37°C. CDFDA (10 µM) and R123 (50 M) was added to the cells, and cells were incubated for 60 or 120 min at 37°C. The medium was removed, and cells were washed twice with ice-cold phosphate-buffered saline. Next, 100 µl of 1 M NaOH were added, and the cell lysate was collected using a cell scraper. The fluorescence intensity (Ex/Em = 485 nm/530 nm) derived from carboxydichlorofluorescein (CDF) or R123 was measured using a SH-9000Lab fluorescent microplate reader (Corona Electric Co., Ibaraki, Japan).

Preparation of cell lysates and membrane proteins
Cell lysates were isolated from HepG2 cells at 48 h after siRNA transfection using RIPA lysis buffer. Membrane proteins were extracted using the Mem-PER eukaryotic membrane protein extraction reagent kit. The protein concentrations of cell lysates and membrane proteins were measured using a BCA protein assay kit.

Determination of PIP2 levels via dot blot analysis
HepG2 cell lysates (0.1 µg of protein in 2 µl) were spotted onto Hybond-P polyvinylidene difluoride membranes at the center of a grid drawn by pencil. The membrane was blocked with 5% ECL blocking reagent in Tris-buffered saline with 0.05% Tween 20 (TBS-T) for 60 min at room temperature. The membranes were incubated with primary antibodies for PIP2 (1:10,000 [v/v]) in TBS-T for 60 min at room temperature with gentle shaking. The membrane was washed three times for 10 min each with TBS-T. The membrane was subsequently incubated with HRP-conjugated anti-mouse secondary antibody (1:10,000 [v/v]) in TBS-T for 60 min at room temperature with gentle shaking and then washed three times for 10 min each. The immunoreactive dots were visualized using ECL Prime western blotting detection reagent with Ez-Capture (ATTO, Tokyo, Japan). The densitometric intensity of the dot blot bands was calculated using CS analyzer (ATTO).

Western blot analysis
Western blotting was performed as described previously (34). Membrane proteins (5 g) were diluted with loading buffer, denatured at 95°C for 3 min, and resolved using 6% sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE). The SDS-PAGE gel included 20 M Phos-tag for recognition of activated (phosphorylated) radixin (pradixin). Proteins were transferred to a polyvinylidene fluoride membrane and subjected to semidry blotting using a Transblot SD (Bio-Rad, Hercules, CA, USA). Radixin, p-radixin, and actin protein levels were detected using ECL Prime western blotting detection reagent with Ez-Capture. The densitometric intensity of each band was calculated using a CS analyzer.

STATISTICAL ANALYSIS
Data are shown as the mean ± standard error of the mean. Significance of differences between mean values was determined by Dunnett's test after analysis of variance using GraphPad Prism software (La Jolla, CA, USA). p < 0.05 was considered to indicate statistical significance.

RESULTS
We examined the involvement of PIP5K1A, the predominant PIP5K isoform in liver, in MRP2 activity in mouse primary hepatocytes using siRNA. To first confirm the efficiency of siPIP5K1A in depleting PIP5K1A in mouse primary hepatocytes, we evaluated the mRNA levels of PIP5K1A in cells transfected with siPIP5K1A for 48 h. The results confirmed that PIP5K1A mRNA levels were lower after treatment with both 5 and 50 pmol/well siPIP5K1A than the levels in control cells ( Figure   1A). We selected 5 pmol/well for use in subsequent studies since the two tested concentrations had similar depletion efficiencies.
We next evaluated MRP2 activity in mouse primary hepatocytes following PIP5K1A depletion by measuring the intracellular accumulation of CDF (a MRP2 substrate). An increased intracellular CDF accumulation was observed in siPIP5K1A-treated cells compared with controls ( Figure 1B).
We next examined the consequence of knocking down each isoform of PIP5K on MRP2 activity in PANC1 and HepG2 cancer cell lines. We previously confirmed that PANC1 and HepG2 cells have sufficient MRP2 activity to determine the effects of PIP5K knockdown (data not shown). In PANC1 cells, the mRNA levels of PIP5K1A and PIP5K1C, but not those of PIP5K1B, were reduced by siRNA ( Figure 2A). Intracellular CDF was increased in PANC1 cells transfected with siPIP5K1A and siPIP5K1C or MK571 (an MRP2 inhibitor) compared with controls ( Figure 2A). In HepG2 cells, the mRNA and protein levels of PIP5K1A, PIP5K1B, and PIP5K1C were decreased by siRNA treatment compared with controls ( Figure 2B). Intracellular CDF was also increased by siPIP5K1A and siPIP5K1C treatment in HepG2 cells.
The effects of PIP5K1 depletion on the activity of P-gp, another efflux transporter involved in multidrug resistance with chemotherapy, were also examined. PIP5K1 depletion had little effect on intracellular accumulation of R123 (a P-gp substrate) ( Figure 3A) and P-gp mRNA levels ( Figure 3B) in HepG2 cells.
We next examined PIP2 amounts in HepG2 cells depleted for PIP5K1A or PIP5K1C by dot blot analysis. PIP2 levels were unchanged between control and PIP5K1A-or PIP5K1C-depleted HepG2 cells (Figure 4). To clarify whether the levels of MRP2 were affected by PIP5K1A or PIP5K1C depletion in the absence of changes of PIP2 levels, the mRNA levels of MRP2 and protein levels of radixin and p-radixin were determined. Neither PIP5K1A nor PIP5K1C depletion had significant effects on MRP2 mRNA levels in HepG2 cells ( Figure 5A). Previous studies showed that p-radixin is involved in MRP2 activity (35). We found that pradixin expression was decreased by PIP5K1A and PIP5K1C depletion, although the total levels of radixin were unchanged ( Figure 5B).

DISCUSSION
Inactive ERM in the cytoplasm undergoes conformational changes upon binding with PIP2, and subsequent PIP2-mediated phosphorylation converts ERM to its active form. Active ERM proteins act as scaffold proteins for MRP2. PIP5K1 is the predominant enzyme for PIP2 biosynthesis. Here we examined whether PIP5K1A, PIP5K1B, and PIP5K1C may serve as target molecules for modulating MRP2 activity in hepatocytes and cancer cells.
In the liver, PIP5K1A is the predominant enzyme for converting PI(4)P to PIP2 (24). PIP5K1A depletion resulted in approximately 2-fold greater intracellular CDF accumulation in mouse primary hepatocytes compared with controls (Figure 1), suggesting that MRP2 activity was decreased by PIP5K1A depletion in mouse primary hepatocytes. This result suggests that this PIP2-mediated process could be a target for modulating MRP2 activity in liver.
Intracellular CDF accumulation was significantly increased by PIP5K1A or PIP5K1C depletion in HepG2 cells ( Figure 2B), indicating that reducing PIP5K1A and PIP5K1C levels led to decreased MRP2 activities in HepG2 cells. These results may reflect that radixin predominantly interacts with MRP2 in cancer cells as well as in hepatocytes (36,37). However, P-gp activity in HepG2 cells as estimated by intracellular R123 accumulation was not affected by PIP5K depletion ( Figure 3A). These results indicate that PIP5K1A and PIP5K1C are closely related to the transport activities of MRP2 but not P-gp. We also found that MRP2 activity was decreased by PIP5K1A and PIP5K1C depletion in PANC-1 and HepG2 cells (Figure 2A). These results suggest that PIP5K1A and PIP5K1C might participate in the modulation of MRP2 activity in cancer cells. Notably, PIP5K1B did not show similar results. The depletion efficiencies for siRNA targeting PIP5K1B were relatively low compared with those for PIP5K1A and PIP5K1C (Figure 2). Furthermore, PIP5K1B gene expression is extremely low in the liver, although PIP5K1B gene expression is detectable in some organs such as the brain and testes (38)(39)(40). Therefore, PIP5K1B may have less contribution to the modulation of MRP2 activity.
The interactions between radixin and MRP2 in HepG2 cells have been well defined. We thus examined PIP2 levels, MRP2 mRNA levels, and radixin and p-radixin levels in HepG2 cells. We also previously demonstrated that the interaction between MRP2 and radixin on the canalicular membrane in inflammation was decreased along with the ratio of p-radixin to total radixin (41). Our results showed that MRP2 activity in HepG2 cells was reduced by PIP5K1A and PIP5K1C depletion without alterations of PIP2 levels in the cell lysate. PIP is phosphorylated by PIP4K2 and PIP3 is dephosphorylated by PTEN in the PIP2 generation pathway (20,21). Following PIP5K1A depletion, PIP3 levels were decreased, and PIP2 levels were unchanged. A previous report showed that PIP5K1C depletion had no impact on both PIP3 and PIP2 levels (42). Our findings of unchanged PIP2 levels following PIP5K1A or PIP5K1C depletion are consistent with previous results. We also found that PIP5K1A or PIP5K1C depletion had no impact on MRP2 mRNA and radixin protein levels ( Figure 5). The decreased p-radixin expression could be partly involved in the reduction of MRP2 activities in HepG2 cells.
A previous report demonstrated that PIP5K1A overexpression resulted in malignant progression in cancer cells (23). Our findings indicate that the knockdown of PIP5K1A or PIP5K1C could be effective in the modulation of MRP2 activity in HepG2 cells, although further studies are needed to clarify the precise mechanism. Together these results suggest that PIP5K1 depletion in cancer cells could be advantageous because the resulting effects on MRP2 activity may prevent anti-cancer drug resistance and malignant progression.
In conclusion, PIP5K1A or PIP5K1C depletion resulted in reduced MRP2 activity in hepatocytes and cancer cells without effects on PIP2 levels, suggesting that PIP5K1A and PIP5K1C could be target proteins for modulating MRP2 function.