Cellular folate status modulates the expression of BCRP and MRP multidrug transporters in cancer cell lines from different origins

Multidrug resistance (MDR) is characterized by the development of resistance to a broad spectrum of structurally and functionally unrelated drugs. This is usually associated with the active and ATP-dependent transport of drugs out of the cell by efflux pumps belonging to the ATP-binding cassette family of transporters (1). Although for >30 years P-glycoprotein (ABCB1) has been the most studied MDR transporter (2), it is now clear that other transporters such as the breast cancer resistance protein (BCRP/ABCG2; refs. 3, 4) and the family of MDR proteins (MRP1-9/ABCC1-6 and 10-12; refs. 5, 6) are also involved in conveying a MDR phenotype. These MDR transporters can confer resistance to different groups of anticancer agents including anthracyclines (e.g., doxorubicin and daunorubicin), Vinca alkaloids (e.g., vincristine and vinblastine), podophyllotoxins (e.g., etoposide), camptothecins (e.g., irinotecan and SN-38), antimetabolites (e.g., methotrexate), and new generations of targeted therapeutic agents (e.g., imatinib and gefitinib). Additionally, they have the ability to transport several physiologic substrates such as glutathione conjugates (e.g., leukotriene C₄), glucuronide conjugates (e.g., 17β-estradiol 17β-d-glucuronide), oxidized glutathione, cyclic nucleotides (cAMP and cGMP), and folates (3, 5).

Folates play a key role in one-carbon metabolism, acting as coenzymes in many important reactions including the synthesis of DNA and RNA precursors, synthesis of several amino acids, and DNA methylation (7). Folate homeostasis is regulated by several enzymes in the folate metabolic pathway and also different transporters that mediate the uptake and efflux of folates from cells (8). Cellular uptake of folates may proceed via three different routes, including the reduced folate carrier (RFC) (9), folate receptors (10, 11), and the recently cloned proton-coupled folate transporter (12). Folylpolyglutamate synthetase (FPGS) catalyzes folate polyglutamylation process, which is responsible for increasing the cellular retention of this vitamin. Folate efflux from cells is facilitated by several MDR transporters (MRP1-5, MRP8, and BCRP; refs. 8). Consequently, it has been shown that these transporters play a role in the modulation of the intracellular levels of folates. Consistently, human ovarian carcinoma 2008 cells transfected with MRP1-3 displayed 32% to 38% diminished total folate pools relative to wild-type 2008 cells (13). Additionally, it has been reported that expression and functional activity of MRP1 and BCRP are severely affected by cellular folate status. MRP1-overexpressing cells (2008/MRP1) cultured for 2 days in folate-free medium showed a decrease in daunorubicin efflux, which is indicative of a decrease in MRP1 activity. Of note, folate replenishment of the folate-deprived cells resulted in a rapid resumption of MRP1-dependent daunorubicin efflux activity (14). Similarly, gradual deprivation of folates induced a nearly complete loss of BCRP expression along with 5-fold down-regulation of MRP1 levels in MCF-7/MR cells (15). In contrast, however, we have recently reported that Caco-2 cells gradually deprived of folates showed a significant increase in BCRP protein and mRNA levels, associated with an increased resistance to mitoxantrone (16).

Because cellular folate levels seem to have a different effect on cancer cells from different origins, the relationship between folate status and MDR transporters warrants further clarification. This relationship is of particular interest and relevance given the notion that folates are often included in chemotherapeutic regimens (17, 18) and that cancer patients take vitamin supplements during treatments on a regular basis. Of note, the physiologic levels of folates in the human plasma are usually in the low nanomolar range (5-30 nmol/L; refs. 19–21) and in tumor tissues as well.⁵

Folic acid, leucovorin, mitoxantrone dihydrochloride, propidium iodide, and sulforhodamine B were obtained from Sigma. Methotrexate was purchased from Pharmachemie. Protease inhibitor cocktail was obtained from Roche Diagnostics. Ko143 was generously provided by Prof. G.J. Koomen (University of Amsterdam).

The human cancer cell lines KB (nasopharyngeal), OVCAR-3 and IGROV-1 (ovarian), ZR75-1/R/MTX (breast), SCC-11B and SCC-22B (human head and neck squamous cell carcinoma; refs. 23, 24), and WiDr (colon; ref. 25) were either grown in standard RPMI 1640 (Cambrex BioScience) containing 2.3 μmol/L supraphysiologic folic acid (HF) supplemented with 10% FCS (Greiner Bio-One) and 20 mmol/L HEPES (Cambrex BioScience) or adapted to more physiologic concentrations of folates (1-5 nmol/L folic acid or leucovorin; LF) by gradually decreasing the folate concentrations in the medium over a period of up to 4 months, depending on the cell line, followed by monitoring these cells under these conditions for at least 1 year. To this end, cells were cultured in a folic acid-free RPMI 1640 (Invitrogen) supplemented with 10% dialyzed FCS (Invitrogen) and 20 mmol/L HEPES to which gradually decreasing folic acid or leucovorin concentrations were added. The final concentration of folates in the LF cell lines were as follows: 2 nmol/L folic acid in KB, IGROV-1, and SCC-22B LF; 2 nmol/L leucovorin in OVCAR-3 LF; 2.5 nmol/L leucovorin in WiDr LF; and 5 nmol/L folic acid in ZR75-1/R/MTX and SCC-11B LF.

Growth inhibition by mitoxantrone, methotrexate, gemcitabine, and 5-fluorouracil (5-FU) was determined with the sulforhodamine B assay (26). Briefly, cells were seeded in 100 μL medium, in triplicate, in 96-well flat-bottomed plates at a density of 5,000 per well. Cells were allowed to attach for 24 h at 37°C. Attached cells were then exposed to various concentrations of drug in 100 μL medium. The BCRP-specific inhibitor Ko143 (27) was added 15 min before the drugs and was present during the next 72 h at a concentration of 200 nmol/L. After a drug exposure time of 72 h, cells were fixed with TCA and stained with sulforhodamine B protein dye. Absorbance was measured at 492 nm. Control wells containing cells in drug-free medium were cultured for 1 day (day 0) followed an additional 72 h and were used to determine the control cell growth at 72 h compared with the initial number of cells (day 0 value); wells with culture medium only were used as blanks. To distinguish between cell growth inhibition and cell kill, the absorbance after 72 h was corrected for the mean absorbance observed for the control wells at the day of drug addition (day 0 value). The IC₅₀, defined as the drug concentration that corresponds to a reduction of cellular growth by 50% when compared with values of untreated control cells, was used as a measure of resistance.

Total lysates were prepared in buffer containing 50 mmol/L Tris (pH 7.6), 20% (v/v) glycerol, 5 mmol/L DTT, 0.5% (v/v) NP-40, and 4.0% (v/v) of a protease inhibitor cocktail. Lysates were sonicated for 3 × 5 s with 30 s interval (MSE Soniprep 150, 4°C, amplitude 6-7) and centrifuged at 13,000 rpm for 10 min at 4°C. The protein-containing supernatant was collected and protein content was determined using the Bio-Rad protein assay. In each lane of a Bio-Rad minigel system (Bio-Rad), 30 μg protein was loaded. For detection of BCRP and MRP1-5, the following monoclonal antibodies were used: rat anti-MRP1 (MRPr1; 1:500; 0.5 μg/mL; ref. 28), anti-MRP4 (M₄I-10; 1:200; 1.25 μg/mL; ref. 29), anti-MRP5 (M₅I-1; 1:250; 1.0 μg/mL; ref. 30), and anti-BCRP (BXP-53; 1:200; 1.25 μg/mL; ref. 31) as well as mouse anti-MRP2 (M₂III-6; 1:500; 0.5 μg/mL) and anti-MRP3 (M₃II-21; 1:500; 0.5 μg/mL; ref. 30). As secondary antibodies, horseradish peroxidase-conjugated rabbit anti-rat or anti-mouse (DakoCytomation; 1:2,000) were used. As a loading control, expression of β-actin was determined using an antibody against β-actin (clone C4 from Chemicon International; 1:10,000; 0.01 μg/mL). Densitometric analysis of the images captured on the VersaDoc3000 instrument (Bio-Rad) was done with the Quantity One software.

Total RNA was isolated from the different cells using the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer's instructions. RNA was reverse transcribed to cDNA using random hexamers as described previously (32).

BCRP, MRP1, and MRP3-5 mRNA expression (MRP2 was not studied due to technical limitations) was studied by real-time PCR using the LightCycler instrument (Roche Diagnostics) and hybridization probes, essentially as described earlier (33). All samples were tested by the LightCycler FastStart DNA Master^PLUS HybProbe kit (Roche Diagnostics) according to the manufacturer's recommendations. PCRs were done in duplicate using 5 μL cDNA, which was added to 15 μL reaction mixture containing 4 μL LightCycler FastStart DNA Master^PLUS HybProbe “Master Mix” (prepared previously by the addition of 10 μL “Enzyme” to one vial of “Reaction Mix”), 4 μL (BCRP and MRPs), or 0.4 μL (β-actin) of a mix containing forward and reverse primers and probe, and water (PCR grade), in a final volume of 20 μL. The primers and probes used were designed using LightCycler probe design software (Roche Diagnostics); their sequences as well as their final concentrations are depicted in Table 1.

The PCR program for all targets consisted of an initial denaturation step at 95°C for 10 min and 45 cycles of warming up until 95°C immediately followed by 15 s at 60°C. After the final cycle, the capillaries were cooled for 30 s at 40°C. Fluorescence curves were analyzed with the LightCycler software (version LCS4 4.0.5.415). This software uses the second derivative maximum method to calculate the fractional cycle numbers where the fluorescence increases above background (crossing point), which is the point at which the rate of change of fluorescence is fastest.

The expression of BCRP and MRPs was quantified relative to β-actin. For this purpose, standard calibration curves were made for all targets by amplifying 5 μL of various dilutions of cDNA from a pool of three cell lines (CEM/7A, MCF-7/MR, and SW1573 2R/120) as described above. For the standard curves, crossing points were plotted versus log concentration for the standards. These standard curves were used to estimate the concentration of each sample.

Analysis of RFC, FPGS, DHFR, and TS mRNA expression was done by real-time PCR with the Applied Biosystems 7900HT sequence detection system (Applied Biosystems). PCRs were done in triplicate using 5 μL cDNA, 12.5 μL TaqMan Universal PCR Master Mix, 2.5 μL probe, and 2.5 μL forward and reverse primers in a final volume of 25 μL. Samples were amplified using the following thermal profile: an initial incubation at 50°C for 5 min, to prevent the reamplification of carryover PCR products by AmpErase uracil-N-glycosylase, followed by incubation at 95°C for 10 min, to suppress AmpErase uracil-N-glycosylase activity and denature the DNA, and 40 cycles of denaturation at 95°C for 15 s followed by annealing and extension at 60°C for 1 min.

Forward and reverse primers and probes were designed with Primer Express version 2.0 (Applied Biosystems) based on TS, DHFR, and FPGS gene sequence (34), whereas primers and probes for RFC were obtained from Applied Biosystems Assay-on-Demand Gene expression products (Hs00228858_m1). Amplification data were normalized to β-actin, and quantification of gene expression was done using standard curves obtained with dilutions of cDNA from Quantitative PCR Human Reference Total RNA (Stratagene).

Preliminary experiments were carried out with dilutions of cDNA obtained from Quantitative PCR Human Reference Total RNA (Stratagene) to determine the primer concentrations that give the minimum SD between CT values and to show that the efficiencies of amplification of all targets and reference (β-actin) genes are approximately equal.

WiDr HF and LF-adapted cells were seeded onto glass Lab-Tek Chamber Slides (8 wells; 0.8 cm²/well) at a density of 20 × 10⁴ per well (400 μL/well medium) and incubated for 1 to 2 days at 37°C. Then, the growth medium was removed, and cell monolayers were washed twice with HBSS solution and fixed with 3.7% formaldehyde in PBS for 15 min at room temperature. Cells were washed twice with PBS (similar PBS washes were included between each subsequent step) and permeabilized by saponin [2% (w/v) in PBS, 2 min at room temperature]. Nonspecific binding sites were blocked for 30 min at room temperature with PBS containing 5% rabbit serum (DakoCytomation). After serum removal, and without further washing, cells were incubated with a mouse anti-BCRP monoclonal antibody BXP-21 (35), diluted 1:50 in PBS with 4% bovine serum albumin, for 60 min at room temperature. BCRP staining was revealed by incubation with FITC-conjugated rabbit anti-mouse antibody (1:30; DakoCytomation) for 60 min at room temperature. Cell nuclei were stained with propidium iodide (5 μg/mL) for 5 min at room temperature. After a final wash with water instead of PBS, the slides were mounted in Vectashield (Vector Laboratories) and examined using a Leica TCS SP confocal laser scanning microscope equipped with ×20 objective with numerical aperture of 0.70. FITC was excited using the 488 laserline and propidium iodide using the 568 laserline. All images were recorded sequentially to reduce bleeding through. The settings were first optimized for the WiDr LF cells. These same settings were used for all pictures (for both WiDr HF and LF cells) to allow comparisons between the different cell lines. Control for nonspecific staining was the replacement of BXP-21 with a nonspecific antibody from the same class, mouse IgG2a (DakoCytomation).

Arithmetic means are given with SE. Statistical significance of the difference between two groups was evaluated by Student's t test. Differences were considered to be significant when P < 0.05.

The levels of MRP1-5 and BCRP protein expression were determined in the panel of HF and LF-adapted cell lines by Western blot. As shown in Fig. 1, the expression of these transporters is quite variable between the different cell lines. Consistent with the ubiquitous distribution of MRP1 in human tissues (36), this transporter was detected in all cell lines, with the head and neck cancer cells (SCC-11B and SCC-22B) displaying the highest expression. Whereas, for most of the cell lines, we did not observe differences in the MRP1 protein expression between HF and LF cells, the OVCAR-3 and ZR75-1/R/MTX LF-adapted cell lines showed a decreased expression of MRP1 compared with their HF counterparts. MRP2 was only detected in KB and ZR75-1/R/MTX cells, but no differences were observed between HF and LF cell lines. A very high expression of MRP3 was found in WiDr cells, both HF and LF, but this transporter was absent in the majority of the other cell lines. Interestingly, KB LF cells, but not the HF counterpart, expressed MRP3. MRP4 was highly expressed in the ovarian cancer cell line IGROV-1 but not in cells from other origins. Of note, in another ovarian cancer cell line, OVCAR-3, no detectable MRP4 expression was observed in HF cells, but OVCAR-3 LF cells displayed markedly increased MRP4 levels. MRP5 was poorly expressed in the panel of HF and LF cell lines. The highest level of this transporter was detected in ZR75-1/R/MTX HF cells, whereas it was not present in the LF counterpart. BCRP protein was only strongly detected in WiDr cells and the expression of this transporter was markedly (2.2-fold) higher in the LF-adapted cell line than in the HF cells.

The levels of MRP1-5 and BCRP mRNA were determined in the panel of HF and LF-adapted cell lines by real-time LightCycler PCR (Fig. 2). In general, the gene expression data were highly consistent with the protein expression profile in all cells. Most cell lines showed a comparable expression of MRP1 mRNA with a clearly higher expression in the SCC-22B cells; KB LF cells showed a 2-fold increase in the expression of this transporter compared with the HF cells (Fig. 2A). MRP3 mRNA was only highly expressed in WiDr cells and low levels were found in the remaining cell lines; MRP3 expression was 16.1- and 10.7-fold increased in KB and SCC-11B LF cells, respectively, but 12.5-fold decreased in IGROV-1 LF cells compared with the HF counterparts (Fig. 2B). The highest levels of MRP4 mRNA were detected in two ovarian cancer cell lines IGROV-1 and OVCAR-3; a 1.9 and 4.8-fold increase in MRP4 mRNA expression was found in KB and OVCAR-3 LF cells, respectively, and a 1.7-fold decrease in SCC-11B LF cells compared with the HF cell lines (Fig. 2C). High levels of MRP5 mRNA were found in SCC-22B followed by ZR75-1/R/MTX; a slight increase in MRP5 expression was observed in KB and OVCAR-3 LF cells (Fig. 2D). BCRP mRNA was highly expressed in WiDr cells and almost absent in OVCAR-3, IGROV-1, and ZR75-1/R/MTX cells. LF conditions induced 2.4- and 1.6-fold increase in BCRP expression in KB and WiDr LF cells, respectively, and 2.8-fold decrease in SCC-11B LF cells (Fig. 2E).

The effect of folate deprivation on the expression of other important genes involved in folate uptake and metabolism was determined by real-time PCR. RFC expression was 6-fold increased in ZR75-1/R/MTX LF and ∼2-fold increased in KB, SCC-11B, and SCC-22B LF compared with the HF cell lines. In contrast, the expression of this transporter was markedly decreased in the ovarian OVCAR-3 LF (4-fold) and almost unchanged in IGROV-1 and WiDr LF cells (Table 2). Regarding the enzymes FPGS, DHFR, and TS, less variation was observed between HF and LF cells. Of note, ZR75-1/R/MTX LF cells showed an increased gene expression of all these enzymes compared with the HF counterpart, whereas, in SCC-22B LF, a decreased expression of all enzymes was observed (Table 2).

In two cell lines, we determined the uptake of methotrexate to assess the relative importance of the influx and efflux transporters on folate accumulation. In line with the higher RFC activity (24), WiDr and SCC-11B LF cells showed a 15- and 8-fold increase in radioactive methotrexate uptake, respectively. These data indicate that folate accumulation is the result of action of different transporters and folate-metabolizing enzymes.

Following the observation that WiDr LF cells harbored an increased expression of BCRP, we assessed whether these cells would have acquired resistance to the BCRP substrate mitoxantrone. Indeed, WiDr LF cells were ∼2-fold resistant to mitoxantrone in comparison with the HF counterpart. BCRP inhibition by its potent and specific inhibitor Ko143 could restore the sensitivity of the LF cells, suggesting that BCRP is involved in the observed resistance. Additionally, Ko143 also partially sensitized the HF cells to mitoxantrone, underscoring that BCRP in WiDr cells, both HF and LF, is functionally active and capable of facilitating resistance to mitoxantrone (Fig. 3A).

The antifolate methotrexate is also a BCRP substrate (8). However, despite the higher BCRP expression in WiDr LF cells, they were ∼17-fold more sensitive to this drug (Fig. 3B) possibly because of the higher RFC and FPGS activity in these cells (24, 25).

The growth-inhibitory effect of 5-FU and gemcitabine was also tested in WiDr HF and LF cells, but no differences in sensitivity were observed for these drugs, consistent with the notion that they are non-BCRP substrates (Fig. 3C and D).

To investigate the subcellular localization of BCRP in the WiDr cell lines, these cells were stained with the BCRP monoclonal antibody BXP-21 followed by a FITC-conjugated secondary antibody. Confocal laser scanning microscopy showed positive staining of BCRP in WiDr cells, both HF and LF (Fig. 4). Interestingly, BCRP staining was mainly found in the intracellular compartment of HF and LF-adapted WiDr cells rather than at the plasma membrane (Fig. 4A and B). A more detailed view of WiDr LF cells after three-dimensional reconstruction showing X-Z and Y-Z sections confirmed the predominant intracellular BCRP staining in these cells (Fig. 4C). Furthermore, consistent with the protein and gene expression data, we found a more intense staining for BCRP in the LF-adapted cells (Fig. 4A and B).

In the present study, we show that changes in the cellular folate status in a panel of cancer cell lines were accompanied by alterations in the expression of not only BCRP and MRP1 but also some other MRPs. Notably, a marked increase in the gene and protein expression of MRP3 and MRP4 was observed in the KB and OVCAR-3 LF-adapted cell lines, respectively, in comparison with their HF counterparts, whereas a marked up-regulation of BCRP was noted in the human colon cancer cell line WiDr cultured under LF conditions. These results show that a mechanism of adaptation to LF conditions involving an up-regulation of MRPs/BCRP, as it was described previously for Caco-2 cells (16), can also be found in cell lines from different origins.

To our knowledge, this study constitutes the first report on the ability of folates to modulate MRP3 and MRP4 expression. Previous studies have shown that folate deprivation induces a decrease in MRP1 expression and functional activity (14, 37). Similarly, we have found that MRP1 expression was decreased in OVCAR-3 and ZR75-1/R/MTX LF-adapted cell lines. In contrast, MRP3 and MRP4 expression was markedly up-regulated in KB and OVCAR-3 LF-adapted cells, respectively. This result is in line with the reported up-regulation of BCRP in Caco-2 LF-adapted cells (16), showing that, similar to BCRP, MRPs can also be up-regulated under LF conditions. Taking into account the intracellular localization of BCRP in Caco-2 cells, one can speculate that MRP3 and MRP4 might also have intracellular rather than plasma membrane localization in KB and OVCAR-3 cells.

Regarding BCRP expression and functionality, WiDr cells share several features with Caco-2 cells: (a) BCRP was up-regulated in the WiDr LF-adapted cell line, (b) BCRP up-regulation conferred a modest level (2-fold) of resistance to mitoxantrone, and (c) Ko143, a specific BCRP inhibitor, established a reversal of mitoxantrone resistance. Interestingly, essentially the same phenotype of BCRP induction as described herein for WiDr LF cells was also observed for human CEM leukemia cells in which folate deficiency was induced by blocking folate uptake via the RFC using the anti-arthritic drug sulfasalazine (38, 39). Consistent with the observation for Caco-2 cells, WiDr LF cells were also more sensitive to methotrexate than the HF counterpart (16). In contrast, WiDr HF and LF cells were equally sensitive to the non-BCRP substrates 5-FU and gemcitabine, which further underscores the specific involvement of BCRP in the observed resistance to mitoxantrone in WiDr LF cells. Also in the Caco-2 panel, no difference in sensitivity for gemcitabine was observed.⁵

It is well documented that BCRP in many drug-resistant cell lines is predominantly localized in the plasma membrane where it facilitates the extrusion of drugs to the external milieu, thereby decreasing the intracellular concentration (40, 41). Moreover, plasma membrane expression of BCRP can also be reminiscent of a physiologic role that BCRP may play in specific cells/tissues (e.g., breast) where BCRP mediates the secretion of bioactive substances in milk (42). Notwithstanding these facts, an alternative mechanism of MDR was postulated by Rajagopal and Simon (43), who showed that, for some MDR transporters, including MRP1, P-glycoprotein, and BCRP, their expression was not exclusively at the plasma membrane but also in subcellular compartments. These compartments were tentatively identified as the lysosomes within which drugs such as doxorubicin were accumulated and sequestered, thereby conferring drug resistance. Consistently, we have shown that, in Caco-2 cells, BCRP is mainly present in the intracellular compartment. We hypothesized that BCRP might be located in the membrane of an intracellular organelle where it can pump and accumulate its substrates (e.g., folates and mitoxantrone). This notion would explain the physiologic importance of increased BCRP expression in the retention of folates in cells cultured under LF conditions and, concomitantly, an increased level of mitoxantrone resistance in the same cells (16). To understand whether a similar mechanism could be functioning in WiDr cells, we investigated the subcellular localization of BCRP in these cells. Confocal laser scanning microscopy revealed that BCRP is also located in the intracellular compartment of WiDr cells, both HF and LF-adapted cell lines, suggesting that the functionality of BCRP in both colon cancer cells (Caco-2 and WiDr) is quite similar. Although the up-regulation of BCRP in both colon cancer cells cultured under LF conditions is quite consistent, the exact mechanism responsible for this up-regulation remains unclear. It has recently been shown that BCRP expression may be regulated by promoter methylation. Turner at al. (44) and To et al. (45) showed that the BCRP promoter of cells with low BCRP expression was almost completely methylated, whereas the promoter of cells with high and medium BCRP expression was completely or almost completely unmethylated. Additionally, both studies showed that treatment of low BCRP-expressing cells with the demethylating agent 5-aza-2′-deoxycytidine induced a significant increase in BCRP mRNA and protein levels (44, 45). Folates play an important role in DNA methylation, with low levels of the vitamin being highly correlated with decreased DNA methylation (46, 47). Thus, one could speculate that decreased intracellular folate levels, either due to LF adaptation (current study) or by blocking RFC-mediated uptake (38, 39), might be associated with a decreased methylation status of the ABCG2 promoter region, which could then be responsible for the increased expression of this transporter.

Folate homeostasis is regulated not only by efflux transporters but also by some uptake transporters, such as RFC, and some folate-metabolizing enzymes, such as FPGS (8). Therefore, we determined the levels of RFC, FPGS, DHFR, and TS in the panel of HF and LF-adapted cell lines. Folate deprivation induced an up-regulation of RFC in the majority of the cells, which is consistent with several previous reports (16, 24, 48, 49). FPGS expression and/or activity is also often up-regulated in response to folate deprivation to enhance cellular retention of folates (15, 16, 50). However, in this study, only KB and ZR75-1/R/MTX LF-adapted cells showed an increased FPGS gene expression. Similar findings have been reported by Backus et al. (24) showing no significant changes in FPGS activity between HF and LF colon and head and neck cancer cells. Of note, although RFC and FPGS gene expression was not up-regulated in the WiDr LF-adapted cell line, it has been shown previously that the activity of the transporter and enzyme is increased in this cell line compared with the HF counterpart (24, 25). Additionally, for DHFR and TS gene expression, we did not observe major differences between HF and LF cells.

The present study further underscores the important role of cellular folate status in the regulation of BCRP/MRPs expression, with potential implications for MDR phenotypes. However, this study also emphasizes that the relationship between folates and MDR transporters is critically dependent on the natural localization of these transporters (plasma membrane or subcellular organelles).

In conclusion, the present study shows that up-regulation of intracellularly localized BCRP in response to adaptation to LF conditions may be a common feature as shown in the panel of colon cancer cell lines. Additionally, we can rationalize that, under these circumstances, folate supplementation might improve the efficacy of chemotherapeutic drugs by decreasing BCRP expression.

No potential conflicts of interest were disclosed.

We thank Prof. G.J. Koomen for the generous gift of Ko143.