In vitro and in vivo studies have demonstrated that UCB (unconjugated bilirubin) is neurotoxic. Although previous studies suggested that both MRP1 (multidrug resistance-associated protein 1) and MDR1 (multidrug resistance protein 1) may protect cells against accumulation of UCB, direct comparison of their role in UCB transport was never performed. To this end, we used an inducible siRNA (small interfering RNA) expression system to silence the expression of MRP1 and MDR1 in human neuroblastoma SH-SY5Y cells. The effects of in vitro exposure to clinically-relevant levels of unbound UCB were compared between unsilenced (control) cells and cells with similar reductions in the expression of MRP1 or MDR1, documented by RT–PCR (reverse transcription–PCR) (mRNA), immunoblotting (protein), and for MDR1, the enhanced net uptake of a specific fluorescent substrate. Cytotoxicity was assessed by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] test. MRP1-deficient cells accumulated significantly more UCB and suffered greater cytotoxicity than controls. By contrast, MDR1-deficient cells exhibited UCB uptake and cytotoxicity comparable with controls. At intermediate levels of silencing, the increased susceptibility to UCB toxicity closely correlated with the decrease in the expression of MRP1, but not of MDR1. These data support the concept that limitation of cellular UCB accumulation, due to UCB export mediated by MRP1, but not MDR1, plays an important role in preventing bilirubin encephalopathy in the newborn.
In humans and rodents, UCB (unconjugated bilirubin) may be neurotoxic. In some neonates with severe hyperbilirubinaemia, UCB accumulates in neurons and astroglial cells in specific brain regions, causing encephalopathy and kernicterus . Initial in vitro studies of the effects of UCB on mouse neuroblastoma cell lines revealed a higher sensitivity to UCB toxicity of mitotically active N2AB-1 cells than of ‘mature’ neurons , and a progressive and irreversible toxicity in N-115 cells exposed to UCB for one or more hours .
Toxic effects of UCB are greatly decreased by tight binding to plasma albumin, since only the unbound diacid species of UCB, which is dominant at physiological pH [4,5], can diffuse rapidly across cell membranes . In vitro exposure of neurons or astrocytes to UCB has shown neuroprotection at low, physiological concentrations of Bf (unbound UCB), but neurotoxicity at modestly higher Bf values [5,7]. Others reported that UCB activated the mitochondrial pathway of apoptosis in rat brain neurons, inducing mitochondrial depolarization and Bax translocation via physical interaction with membranes . Studies performed on human NT2-N neurons showed that UCB induces rapid necrosis at high and moderate concentrations and delayed apoptosis at low and moderate concentrations . In human neuroblastoma SH-SY5Y cells, clinically relevant UCB concentrations cause early disruption of the mitochondrial membrane potential and subsequently induce apoptosis .
Collectively these findings indicate that intracellular accumulation of UCB may cause serious damage to nerve cells. This prompted us to explore the mechanisms that may limit toxic intracellular accumulation of UCB. The multidrug resistance-associated protein, MRP1 (ABCC1), and the multidrug resistance protein, MDR1 (P-Glycoprotein, ABCB1), belong to the superfamily of the ABC (ATP-binding cassette) transporters . MRP1 transports amphipathic anionic, glucuronide and sulphate conjugates of steroid hormones and bile salts. In the presence of physiological concentrations of reduced glutathione, MRP1 also transports unconjugated xenobiotics and uncharged drugs [11–13]. MDR1 transports hydrophobic, either uncharged or weakly positively charged compounds in their unmodified forms and, in particular conditions, even highly charged, anionic compounds . Both MRP1 and MDR1 protect nerve cells from toxic substances by promoting their active export from the cells [14–18].
Studies using various experimental models demonstrated that MRP1 is directly involved in the transport of UCB [19–21]. In plasma membrane vesicles from MDCKII cells (Madin-Darby canine kidney II cells), the stable expression of human MRP1 significantly increased the export of UCB with high affinity, Km=10±3 nM (Bf) . In cultured mouse astrocytes and primary cultures of rat neurons and astrocytes, functional inhibition of Mrp1 with MK571 engendered an increase in UCB-induced toxicity , even at a normally non-toxic Bf of 40 nM . Collectively, these findings show that MRP1 exports UCB at relatively low, physiologically relevant Bf, and thus may protect the cell from accumulation of toxic levels of UCB.
Other studies have suggested that Mdr1/MDR1 might also transport UCB. UCB may be a weak substrate for Mdr1a  and competitively inhibits the labelling of brain capillary MDR1 with a photoaffinity substrate . A greater proportion of UCB administered intravenously is taken up by the brains of mdr1a (−/−) knockout mice compared with their mdr1a (+/+) controls . Although these studies suggest that MDR1 may be involved in the transport of UCB, they were all performed with high doses of UCB that yielded bilirubin concentrations vastly higher than clinically relevant levels [27,28]. However, drugs known to inhibit MDR1 function may increase the risk of bilirubin encephalopathy in jaundiced neonates .
The aim of the present work was to directly compare the role of MRP1 and MDR1 in the prevention of UCB-related cytotoxicity in the human neuroblastoma SH-SY5Y cell line, which expresses both proteins [30,31] and is highly susceptible to UCB toxicity . The expression of either MRP1 or MDR1 in these cells was modulated by using siRNAs (small interfering RNAs) produced by an inducible vector, and net accumulation and cytotoxicity of UCB were assessed. Our experiments are the first to compare directly the protective role of MRP1 and MDR1 against UCB toxicity in a single cell line under identical experimental conditions. We demonstrate that knockdown of MRP1, but not MDR1: a) decreases the cellular export of UCB and increases the cytotoxicity of UCB over a range of clinically relevant Bf levels (40–140 nM); and b) renders the cells sensitive to UCB toxicity at low Bf levels that are non-toxic to control cells.
DMEM (Dulbecco's modified Eagle's medium), L-glutamine, penicillin streptomycin and trypsin (0.05%)/EDTA (0.02%) in PBS without Phenol Red, Ca2+/Mg2+ were purchased from Euroclone (Pero, Milano, Italy). Fetal bovine serum tetracycline-free (Tet System Approved FBS, US-Sourced) was purchased from Clontech. MEM (modified Eagle's medium), non-essential amino acids, MEM vitamin solution, Opti-MEM I reduced serum medium and Lipofectamine™ 2000 were purchased from Invitrogen. Blasticidin S hydrochloride, puromycin dihydrochloride, doxycycline hyclate, PBS (Dulbecco's Phosphate Buffered Saline) without Ca2+/Mg2+, TRI REAGENT™, bicinchoninc acid protein assay kit, Ponceau S-solution (0.1% w/v) in 5% acetic acid (v/v), MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide], DMSO and Hoechst 33342 were purchased from Sigma. CPRG (chlorophenol red-β-D-galactopyranoside) was purchased from Boehringer Mannheim (Indianapolis, IN, U.S.A.). iScript™ cDNA Synthesis Kit and iQ™ SYBR Green Supermix were purchased from Bio-Rad Laboratories. Cell lysis buffer was purchased from Cell Signaling. ECL® (enhanced chemiluminescence)-plus Western blotting detection system solutions were purchased from Amersham-Pharmacia Biotech. UCB was purchased from Sigma (St. Louis, MO, U.S.A.) and purified as described by McDonagh and Assisi . [3H]-Bilirubin (29.25 mCi/mmol) was biosynthetically labelled in vivo and then highly purified from the bilirubin conjugates in bile as described . Liquid-scintillation cocktail for radioassay (Filter counter N° 6013149) was purchased from Packard Bioscience. MRP1-A23 rabbit antibody was prepared at our laboratory, as described in ; anti-mouse IgG (Fc specific)-peroxidase antibody produced in goat affinity isolated antibody, rabbit anti-actin affinity isolated antibody and peroxidase conjugated-goat anti-rabbit IgG-whole molecule affinity isolated antigen specific antibody were purchased from Sigma. Anti-C219 (MDR1) monoclonal antibody was purchased from Signet Laboratories (Dedham, MA, U.S.A.).
Construction of pSUPERIOR-MRP1 and pSUPERIOR-MDR1 clones
We used pSUPERIOR.puro vector (Oligoengine, Seattle, WA, U.S.A.), a tetracycline (or doxycycline)-regulated vector, for inducible siRNA expression. The siRNA sequence against MDR1 was obtained by Wu et al. , whereas that against MRP1 was the silencer validated MRP1-siRNA (#51321-1651; Ambion, Austin, TX, U.S.A.). The pSUPERIOR.puro vector was digested by BglII and HindIII and the annealed oligonucleotides were ligated and cloned into the vector. The constructs were confirmed by sequencing. SH-SY5Y cells were transfected (Lipofectamine 2000) with the pcDNA6/TR vector (Invitrogen) expressing high levels of the TetR (tetracycline repressor) gene and cells were grown in medium containing 5 μg/ml blasticidin. After transient transfection of the pcDNA4/TO/lacZ vector (Invitrogen), single colonies were isolated and expanded. They were screened for expression of the highest levels of tetracycline repressor from pcDNA6/TR by assaying β-galactosidase expression upon induction with 3.9 μM doxycycline, used as an alternative inducing agent to tetracycline (half-life: 48 compared with 24 h, respectively), according to the pSUPERIOR manufacturer's suggestions. The clone showing the highest ratio of β-galactosidase activity between induced and non-induced samples was selected for transfecting the pSUPERIOR vector containing the siRNA-expressing sequence targeting MRP1 or MDR1. The empty pSUPERIOR.puro vector was also transfected and used as control. Puromycin (0.8 μg/ml)-resistant colonies were expanded for two additional months. The clones showing the highest knockdown of MRP1 (pSUPERIOR-MRP1) or MDR1 (pSUPERIOR-MDR1) upon induction with 3.9 μM doxycycline were selected by gene, protein and functional analysis.
SH-SY5Y cells were kindly supplied from the Mario Negri Institute for Pharmacological Research (Laboratory of Molecular Pharmacology) in Milan, Italy. Stable SH-SY5Y clones constitutively expressed the tetracycline repressor from pcDNA6/TR vector and inducibly expressed the siRNAs from pSUPERIOR vector containing the siRNA-expressing sequence targeting MRP1 (pSUPERIOR-MRP1 clone) or MDR1 (pSUPERIOR-MDR1 clone). The cells were grown in DMEM supplemented with 2 mM L-glutamine, 10% (v/v) tetracycline-free fetal bovine serum, 1% (v/v) MEM non-essential amino acids, 1% (v/v) MEM vitamin solution and 1% (v/v) penicillin/streptomycin (100 units/ml penicillin and 100 μg/ml streptomycin). Blasticidin (5 μg/ml) and puromycin (0.8 μg/ml) were added as selection antibiotics for pcDNA6/TR and pSUPERIOR.puro respectively. Clones were maintained at 37 °C in humidified air with 5% CO2. On the day prior to the experiment, confluent cells were trypsinized and plated in either: 1) 35 mm plates at a density of 2.5×106 cells/well for RT (reverse transcription)-real time PCR and Western blot analyses; 2) in 12-well plates at a density of 1×106 cells/well for the [3H]UCB-uptake assay; or 3) in 96-well plates at a density of 0.2×106 cells/well for MTT assay and for Hoechst 33342 dye accumulation assay. Doxycycline (3.9 μM) was used to induce the siRNA expression from pSUPERIOR.puro vector. Before the harvest or treatments, clones were incubated with and without doxycycline (induced and non-induced samples respectively) at 37 °C for various time intervals.
RNA isolation, reverse transcription and RT-real time PCR
The pSUPERIOR-MRP1 clone was incubated with doxycycline (3.9 μM) for 24, 48 or 72 h, or for 48 h followed by an additional 48 h without doxycycline (induced samples). pSUPERIOR-MDR1 was incubated with doxycycline (3.9 μM) for 24 or 48 h, or for 24 h followed by additional 24 h without doxycycline (induced samples). The control clone, stably transfected with pcDNA6/TR vector and with an empty pSUPERIOR.puro vector, was treated with doxycycline (3.9 μM) as described for pSUPERIOR-MRP1 and pSUPERIOR-MDR1 clones. For each induced sample, a non-induced sample was tested. Total RNA was isolated from each clone and from its control using TRI REAGENT™, according to the manufacturer's recommendations. One μg of total RNA from cell culture was reverse transcribed using the iScript cDNA Synthesis Kit, according to manufacturer's instructions, on a thermal cycler (Gene Amp PCR System 2400, PerkinElmer, Boston, MA, USA). The relative levels of MRP1 mRNA with sense (GCCAAGAAGGAGGAGACC) and anti-sense (AGGAAGATGCTGAGGAAGG) primers and the MDR1 mRNA with sense (TGCTCAGACAGGATGTGAGTTG) and anti-sense (AATTACAGCAAGCCTGGAACC) primers were assayed by RT-real time PCR using iQ SYBR Green Supermix, as described previously . Amplification, data acquisition and data analysis were carried out using the iCycler IQ (Bio-Rad Laboratories, Hercules, CA). Each sample was analysed in triplicate. For each amplification, a blank (nuclease-free water) was added in place of cDNA. The thermal cycler conditions were: 95 °C for 3 min, 40 cycles at 95 °C for 20 min, 60 °C for 20 min and 72 °C for 30 min. In each reaction, a standard curve was generated by serial dilution of isolated and quantified cDNA (chosen among the cDNA samples) and a melting curve was generated to verify the specificity of the amplification. The results were normalized to each housekeeping gene and the initial amount of the template of each sample was determined as relative expression against the non-induced control sample chosen as reference. β-ACTIN with sense (CGCCGCCAGCTCACCATG) and anti-sense (CACGATGGAGGGGAAGACGG) primers and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) with sense (CCCATGTTCGTCATGGGTGT) and anti-sense (TGGTCATGAGTCCTTCCACGATA) primers were used as endogenous controls to normalize the expression level of target genes.
pSUPERIOR-MRP1 clone was incubated with doxycycline (3.9 μM) for 24, 48 or 72 h, or for 72 h followed by 72 h without doxycycline (induced samples). pSUPERIOR-MDR1 clone was incubated with doxycycline (3.9 μM) for 24 or 48 h, or for 48 h followed by 48 h without doxycycline (induced samples). The control clone, stably transfected with pcDNA6/TR vector and with an empty pSUPERIOR.puro vector, was treated with doxycycline as described for pSUPERIOR-MRP1 and pSUPERIOR-MDR1 clones. For each induced sample, a non-induced sample was tested. Proteins of each sample were subjected to SDS/PAGE with 10% acrylamide and the resolved protein bands transferred onto a nitrocellulose membrane (0.2 lm Protran BA 83, Schleicher and Schuell, Dassel, Germany). Membranes were blocked for 1 h at room temperature (22 °C) in milk-TTBS [4% (w/v) non-fat dried skimmed milk, 0.2% Tween 20, 20 mM Tris/HCl and 500 mM NaCl, pH 7.5] and then incubated overnight at 4 °C with a 1:600 dilution of the MRP1-A23 rabbit antibody, 1:50 dilution of the anti-C219 (MDR1) mouse antibody or 1:1500 dilution of rabbit anti-actin antibody. After washing three times with milk-TTBS, the membranes were incubated in blocking solution for 1 h at room temperature with secondary antibodies: peroxidase conjugated-goat anti-rabbit (1:6000) or anti-mouse IgG (Fc specific)-peroxidase (1:2000). Protein bands were detected by ECL®-Plus, transferred to Kodak film and visualized by Kodak EDAS 260 (Kodak Instruments, New Haven, CT, U.S.A.) using Kodak 1D image software. The intensities of protein bands were quantified by Scion Image (Frederick, MD, U.S.A.). To quantify the relative expression of the proteins (MRP1/ACTIN or MDR1/ACTIN), serial dilutions of a reference sample were loaded on each gel, optical density was recorded and a standard curve was generated from the serial dilutions by non-linear regression analysis (Curver Expert 1.3 software, Starkville, MS, U.S.A.). The correlation coefficient of the calibration curve was never lower than 0.995.
[3H]-UCB uptake in cultured cells
pSUPERIOR-MRP1 clone was treated with UCB 72 h after induction of the siRNA-expressing sequence transcription by doxycycline (3.9 μM), and pSUPERIOR-MDR1 clone was treated with UCB 48 h after removing the doxycycline from cells previously incubated with doxycycline for 48 h. These incubation conditions were predetermined to be the best for observing the silencing of MRP1 and MDR1 respectively (see the Results section). [3H]UCB (29.25 mCi/mmol) was dissolved in 0.3% (v/v) DMSO and diluted in culture medium supplemented with 15% (v/v) fetal bovine serum containing 54 μM albumin. Cells were incubated for 30 min in the presence of [3H]UCB (15 μM) at Bf=40 nM, determined as described previously . The cells were then carefully washed with ice-cold PBS to arrest UCB uptake and remove loosely-bound [3H]UCB. The quantity of [3H]UCB retained by cells, normalized for protein concentration, was measured as previously described .
MTT assay and Hoechst 33342 dye accumulation assay
MTT assay is used for measuring the activity of enzymes that reduce MTT to formazan. Yellow MTT is reduced to purple formazan in the mitochondria of living cells by active mitochondrial reductase enzymes; therefore conversion is used as a measure of viable cells.
pSUPERIOR-MRP1 and pSUPERIOR-MDR1 clones were incubated with doxycycline (3.9 μM) for the same durations selected to analyse protein content (see above). For each induced sample, a non-induced sample was included, and each treatment was performed in quadruplicate. After the appropriate doxycycline incubation time, cells were treated for 4 h with three different concentrations of UCB (15 μM, 20 μM or 30 μM). UCB was dissolved in 0.3%, 0.4% or 0.6% (v/v) of DMSO respectively, and diluted in 100 μl of complete medium supplemented with 15% fetal bovine serum containing 54 μM albumin. Bf concentrations were 40 nM, 70 nM and 140 nM respectively, measured as described previously . For each induced and non-induced sample treated with UCB, an induced and a non-induced sample treated with the same volume of DMSO, but not UCB, were included as controls. Induced and non-induced samples, not treated with either UCB or DMSO, were tested also. MTT assay was performed as described previously , with absorbance measured at 562 nm, using a microtiter plate reader (Beckman Coulter). MTT assay was also performed on pSUPERIOR-MDR1 clone exposed to ceftriaxone, a cephalosporin antibiotic transported by MDR1 but not MRP1 [29,38,39]. Cell viability, assessed by MTT reduction ability, of induced and non-induced samples treated with UCB plus DMSO, was expressed as a percentage of the respective samples treated with DMSO alone.
To assess MDR1 export function, accumulation of the Hoechst 33342 dye, a substrate for MDR1 , was measured in a microtiter plate fluorescence reader (Beckman Coulter) at 350 nm (excitation)/460 nm (emission), using 0.2×106 cells/well in 96 well black plates. After the appropriate incubation time with doxycycline, cells were washed twice with Hanks solution and exposed for 6 min to three different concentrations of Hoechst 33342 (1.0, 2.5 and 5.0 μM), dissolved in Hanks solution. Cellular dye uptake was assessed by the increase in fluorescence. At the end of each experiment, cell viability was assessed by the MTT test, as described above.
Results are given as means±S.D. of at least three separate experiments. Statistical analysis was performed by Student's t test and a P value of less than 0.05 was considered statistically significant.
Gene and protein expression changes according to the incubation time with doxycycline
MRP1/MRP1 expression of pSUPERIOR-MRP1 clone
In the samples induced with doxycycline for 72 h, MRP1 mRNA expression (Figure 1A) decreased by about 45% as compared with the non-induced samples (P<0.02); no decrease was observed with a shorter incubation time (48 h). A 40% decrease (P<0.03) was also observed in samples incubated with doxycycline for 48 h and then without doxycycline for an additional 48 h.
Time course of the relative gene and protein expression of MRP1/MRP1 and MDR1/MDR1 in SH-SY5Y clones
Protein analysis confirmed the variations observed in gene expression. As shown in Figure 1(B), no change in the MRP1 level was observed in both induced and non-induced samples after 48 h of incubation with doxycycline. On the contrary, a 60% reduction (P<0.03) was found after 72 h of incubation. A similar decrease in MRP1 content (approx. 50%) occurred in samples induced by doxycycline for 72 h and then incubated without doxycycline for an additional 72 h (P<0.04). The control clone, stably transfected with pcDNA6/TR vector and with an empty pSUPERIOR.puro vector, showed a 10% upregulation of MRP1/MRP1 after 72 h of incubation with doxycycline (results not shown).
MDR1/MDR1 expression of pSUPERIOR-MDR1 clone
After a 24 h incubation with doxycycline, the expression of MDR1 mRNA (Figure 1C) decreased by 30% (P<0.05), but partially recovered by 48 h. The reduction in gene expression was greater than 50% (P<0.02) when the clone was incubated for 24 h in the presence of doxycycline followed by an additional 24 h in its absence. The expression of MDR1 protein (Figure 1D) did not show any decrease after 24 or 48 h of incubation with doxycycline, but was reduced by approx. 60% (P<0.02) 48 h after the removal of doxycycline from cells previously incubated with doxycycline for 48 h. Taking into account the limited specificity of the anti-C219 antibody, these results are compatible with the observation that the silencing by siRNA was hidden by the induction of MDR1/MDR1 expression caused by doxycycline, as previously reported . Doxycycline-related induction of 25% was also observed in the control clone stably transfected with pcDNA6/TR vector and with an empty pSUPERIOR.puro vector, after a 24 h (for gene) or 48 h (for protein) incubation in the presence of the antibiotic followed by a 24 h (for gene) or 48 h (for protein) incubation in its absence (results not shown). This finding explains why the silencing effect could be observed mainly only after removing the doxycycline. The small reduction in the gene but not in the protein expression observed after a 24 h incubation with doxycycline (Figures 1C and 1D), as well as the different incubation times needed to achieve maximum silencing in gene or protein expression, is related to the shorter half life of mRNA than the protein .
These findings defined the optimal incubation times to achieve comparable decreases in the expression levels of MRP1 or MDR1 to correlate with UCB accumulation and cytotoxicity.
UCB accumulates in MRP1-deficient cells, not in MDR1-deficient cells
Figure 2 shows the effects of decreased expression of MRP1 or MDR1 on the intracellular accumulation of UCB after incubation for 30 min with [3H]UCB (Bf=40 nM). Accumulation of [3H]UCB was markedly higher (2.6-fold) in the doxycycline-induced samples of the pSUPERIOR-MRP1 than in the non-induced ones (P<0.0006). By contrast, the intracellular accumulation of [3H]UCB was similar in the induced and non-induced pSUPERIOR-MDR1. Since the decreased expression of both proteins was comparable in the two clones (see Figures 1B and 1D), these results confirm that MRP1, but not MDR1, can regulate the cellular content of UCB at the low Bf of 40 nM.
3[H]UCB accumulation by MRP1- or MDR1-deficient SH-SY5Y clones
Impairment of transport function and viability in MDR1-deficient cells
As shown in Figure 3, compared with non-induced cells, induced pSUPERIOR-MDR1 samples showed a dramatic increase in the cellular accumulation of three different concentrations of the fluorescent dye Hoechst 33324. These results demonstrate that the knockdown of MDR1 expression resulted in functional impairment of the export of the dye, a known substrate for MDR1 .
Cellular accumulation of a fluorescent MDR1 substrate by MDR1-deficient SH-SY5Y clone
When the induced clone with its MDR1 level decreased by 60% was exposed for 24 h to 15 mM ceftriaxone, cell viability was decreased by two-thirds compared with the non-induced clone (35±6% compared with 100±4%, P<0.002) (results not shown). This confirmed that the decreased export function of MDR1 in the induced cells engendered greater susceptibility to cytotoxicity from a known MDR1 substrate.
UCB-induced cytotoxicity occurs in cells deficient in MRP1 but not MDR1
Figure 4 shows the cytotoxic effects of 4 h of exposure to UCB at Bf=40, 70 or 140 nM, on pSUPERIOR-MRP1 and pSUPERIOR-MDR1 clones with comparable silencing of MRP1 and MDR1 respectively. Cell viability was assessed by the MTT assay. All pSUPERIOR-MRP1 samples, whether or not induced with doxycycline for 72 h, showed progressive reduction in cell viability by increasing Bf. However, the decrease in viability was significantly greater in the induced cells with a 60% reduction of MRP1 (see Figure 1B) at each Bf level (Figure 4A), and this effect was UCB dose-dependent (Figure 4C).
Cell viability of SH-SY5Y MRP1- or MDR1-deficient clones after exposure to UCB
In contrast to the pSUPERIOR-MRP1, no difference in cytotoxicity was found between induced and non-induced samples of the pSUPERIOR-MDR1 (Figures 4B and 4C), even at the highest Bf (140 nM), although the induced clone reached a 60% reduction of MDR1 expression by incubation with doxycycline for 48 h followed by additional 48 h in its absence (see Figure 1D). In spite of the substantial reduction in MDR1 expression, the susceptibility of the induced clone to UCB was comparable with that of the non-induced clone, in keeping with the lack of difference in UCB accumulation (Figure 2). This is in sharp contrast to the clear increase, in induced pSUPERIOR-MDR1 cells, of accumulation of Hoechst 33342 (Figure 3), and cytotoxicity of ceftriaxone, two known MDR1 substrates, which confirmed the decreased export function of MDR1 in the induced cells.
In summary, greater accumulation of UCB markedly affected the cell viability of MRP1-deficient cells and this effect was dose-dependent. On the other hand, the MDR1-deficient cells showed UCB accumulation and cell viability identical to their non-induced control, in spite of a demonstrated impairment in transport activity of the protein for a known substrate (Figure 3). Collectively these results strongly point to a role of MRP1 but not of MDR1 in the protection from UCB-induced cytotoxicity. This conclusion was further supported by the strong linear correlation between the relative expression of MRP1 and the cytotoxic effect of UCB at all Bf concentrations tested (Figure 5).
Correlation between UCB-induced cytotoxicity and MPR1 protein expression in SH-SY5Y clones
Previous studies have indicated that nerve cells are particularly sensitive to UCB toxicity [2,3,7–10,27]. MRP1 and MDR1 have each been proposed to protect cells against UCB toxicity by limiting the intracellular accumulation of the pigment. We had previously demonstrated that MRP1 transports UCB with high affinity  and protects MEF (mouse embryonic fibroblast) cells from UCB toxicity by exporting the pigment from the cells . Both Gennuso et al. , using differentiated mouse astrocytes, and, more recently, Falcão et al.  using mature rat astrocytes and neurons, have shown dramatic increases in UCB-induced cytotoxicity (impaired MTT reduction) following blockade of Mrp1 by MK571, an inhibitor of Mrp1 function . Gennuso et al.  showed further that MK571 increased vulnerability of astrocytes to UCB-induced membrane, mitochondrial and nuclear damage, even at a normally non-toxic Bf of 40 nM. Others have suggested that UCB is also a substrate for MDR1  and that this transporter may protect neural cells from toxicity by limiting accumulation of UCB [25,26,29]. However, those studies were performed at vastly high and clinically irrelevant concentrations of UCB.
The present study is the first to compare directly the functional roles of MRP1 and MDR1 in protecting against UCB toxicity in identical cells exposed to identical, increasing concentrations of UCB at clinically relevant Bf levels. The comparable reduction we obtained in the expression levels of MRP1 and MDR1, utilizing RNA interference technology, also allowed a direct comparison of their relative roles in the transport of UCB and in the protection against UCB neurotoxicity in cells differing only in the reduced expression of either transporter.
Our experiments demonstrated that, at clinically relevant Bf levels, protection from UCB cytotoxicity was correlated with the level of expression of MRP1 but not MDR1. No significant UCB-induced impairment in cell viability was observed, even when maximal MDR1 silencing was achieved and its impaired transport function confirmed by the significantly greater cellular accumulation, in the cells with induced siRNA expression, of Hoechst 33342 dye, a substrate for MDR1 . In contrast, the cytotoxic effect caused by ceftriaxone, another known substrate for MDR1 [29,38,39], was significantly greater in the induced cells than in non-induced cells. These findings clearly indicate that decreased expression and function of MDR1 is not associated with greater UCB cytotoxicity at low Bf values.
The [3H]UCB uptake results confirm further the involvement of MRP1 in UCB export, previously demonstrated by using MEF cells isolated from Mrp1 knockout mice , and exclude participation of MDR1 in the transport of UCB at low Bf levels. They indicate also that the protective effects of MRP1 against UCB cytotoxicity are largely related to the ability of this transporter to export UCB from the cell, thus limiting intracellular accumulation of the pigment [19,20,22]. This is in agreement with the chemical characteristics of the substrates of the two proteins; MDR1 ‘prefers’ uncharged or slightly positively-charged compounds, while MRP1 primarily transports amphipathic, anionic conjugates, as well as unconjugated xenobiotics and uncharged drugs [11–13]. Although our results show no role for MDR1 in the transport of UCB at Bf=40 nM, they do not exclude the possibility, suggested by the work of others [24–26,29], that MDR1 transports UCB at much higher Bf levels. Many compounds are substrates for both high-affinity (low Km) transporters that operate at low concentrations, and low-affinity (high Km) transporters that are operative at much higher concentrations of substrate [15,44,45]. In the case of UCB, MRP1 clearly is the high-affinity transporter, whereas MDR1 may become operative only at Bf levels that are far above clinically-relevant concentrations.
Our MRP1-knockdown model validly mimics the increases in mRNA and protein expression of Mrp1 [15,46], and the decreased susceptibility to UCB toxicity , that occur during maturation of rodent neurons and astrocytes. Similar changes in the expression of MRP1 might explain the higher vulnerability of premature infants to UCB-induced neurological dysfunction during moderate to severe neonatal jaundice [48–50].
Thanks are due to Drs. Sebastian D. Calligaris and Pablo Giraudi for their help in performing the bilirubin uptake experiments.
unbound unconjugated bilirubin
Dulbecco's modified Eagle's medium
multidrug resistance protein
mouse embryonic fibroblast
modified Eagle's medium
multidrug resistance-associated protein
small interfering RNA
This study was supported in part by a grant from Telethon [grant number GGP05062]. A. A. is a Ph.D. student sponsored by a fellowship from the Ministry of Foreign Affairs, Rome, Italy.
In Memoriam: This paper is dedicated to the late Professor Rudi Schmid for his outstanding contributions in the bilirubin field.