MRP1 gene expression level regulates the death and differentiation response of neuroblastoma cells

We have previously reported a strong correlation between poor prognosis in childhood neuroblastoma (NB) patients and high-level expression of the transmembrane efflux pump, Multidrug Resistance-associated Protein (MRP1), in NB tumour tissue. In this study, we inhibited the endogenous expression of MRP1 in 2 different NB tumour cell lines by stably transfecting an MRP1 antisense expression vector (MRP-AS). Compared with control cells, MRP-AS transfectant cells demonstrated a higher proportion of dead and morphologically apoptotic cells, spontaneous neuritogenesis, and, increased synaptophysin and neurofilament expression. Bcl-2 protein expression was markedly reduced in MRP-AS cells compared to controls. Conversely, we found that the same NB tumour cell line overexpressing the full-length MRP1 cDNA in sense orientation (MRP-S) demonstrated resistance to the neuritogenic effect of the differentiating agent, all-trans-retinoic acid. Taken together, the results suggest that the level of MRP1 expression in NB tumour cells may influence the capacity of NB cells for spontaneous regression in vivo through cell differentiation and death. © 2001 Cancer Research Campaign  http://www.bjcancer.com

using Trypan blue exclusion and morphologic evidence of neurite extension, as previously described (Marshall et al, 1995). Cells were treated with all-trans-retinoic acid (aRA) (Sigma Chemical Company, St Louis, MO, USA), 9-cis-retinoic acid and 13-cisretinoic acid (kind gift of Hoffman LaRoche, Basel, Switzerland) to final concentrations of 0.01, 0.1, 1 or 10 µM. The percentage of apoptotic cells was estimated by examination of cell morphology as previously described (Lock and Stribinskiene, 1996).
To examine cellular cytotoxicity, BE cells overexpressing MRP1 were seeded in 96-well plates and growth inhibition determined after 72 hours continuous exposure to various concentrations of sodium arsenate (Baker, Biolab Scientific, Sydney, Australia), using a microtitre-based assay with the Alamar Blue™ reagent (Astral Scientific, Sydney, Australia). Using a log scale for molar concentration, data points from at least 2 replicate assays were fitted to a spline curve, and ID 50 values determined from the curve as previously described (Haber et al, 1989).

Plasmid constructs and transfections
The pCEBV7-MRP1 (CEP MRP1) episomal expression vector containing the full length human MRP1 cDNA in sense orientation was the kind gift of Dr SPC Cole, Cancer Research Laboratories, Queen's University, Kingston, Ontario, Canada. CEP was derived from pREP7 (Invitrogen, San Diego, CA, USA) by substituting a cytomegalovirus promoter for the Rous sarcoma virus long terminal repeat . To create MRP1 antisense constructs, CEP MRP1 cDNA was digested with HindIII, to generate a 1.7 kb fragment including a short segment of the CEP polylinker and 0 -1633 bp of the 5′ end of the MRP1 cDNA, and a 1.3 kb fragment of the MRP1 cDNA from 1634 bp to 2962 bp. After purification, each fragment was cloned into the HindIII restriction enzyme site in the mammalian expression plasmid pMEP, where cloned fragments are under the control of the human metallothionein II A promoter. While the human metallothionein II A promoter is inducible by heavy metals, we have found that in NB cells a low level of constitutive expression of cloned sequences also occurs. As an empty vector control for comparison with SY MRP-AS clones, SY cells were stably transfected with the pREP plasmid. The pREP plasmid is identical to the pMEP plasmid, except that cloned cDNAs are under the control of the Rous sarcoma virus long terminal repeat, instead of the human metallothionein II A promoter. In prior experiments, we have shown no difference in the growth behaviour of untransfected BE or SY cells, or BE and SY cells transfected with either the pREP or pMEP empty vector. BE and SY cells were transfected by electroporation with empty vectors (CEP, REP, MEP), or plasmids containing MRP1 cDNA in sense (MRP-S), or antisense (MRP-AS) orientation, and transfectants were selected in Hygromycin B and perpetuated as previously described (Marshall et al, 1995).

RNA isolation and PCR
Total cellular RNA was isolated (Chomczynski and Sacchi, 1987) and used to make cDNA using Moloney murine leukaemia virus reverse transcriptase (Life Technologies) and random hexanucleotide primers. To evaluate the relative mRNA expression levels of MRP1 in MRP-S transfected cell lines, cDNA equivalent to 50 ng of mRNA from each line was subjected to co-amplification of target (MRP1) and control (β2-microglobulin) gene sequences. The gene-specific oligonucleotide primers for MRP1 and the β2-microglobulin gene, the PCR conditions and methods of estimating gene expression by reverse-transcriptase PCR have been previously reported Haber et al, 1994). To evaluate relative expression of Bcl-2 in MRP-AS cells, β2-microglobulin was the control gene. Gene-specific PCR primers for Bcl-2 were B11S (5′ACAA-CATCGCCCTGTGGATGAC3′), and B12AS (5′AGCCAGGAG AAATCAAACAGAGG3′), which amplified a 132 bp sequence from 1970 bp to 2102 bp of the Bcl-2 coding sequence.

Protein isolation and immunoblotting
To measure cellular MRP protein, crude membranes were prepared as previously described . For other immunoblotting experiments, whole cell lysates were prepared from harvested cell pellets by incubation in RIPA buffer (50 mM Tris.Cl (pH 7.5), 150 mM NaCl, 1% Nonidet P-40, 0.5% Na deoxycholate, 0.1% SDS) for 30 min on ice, with vortexing every 10 minutes. The lysate was clarified by centrifugation (14 g, 10 min, 4˚C) and the supernatant stored at -70˚C until further use. Protein content was determined using the BCA Protein Assay Kit (Pierce Chemical Co, Rockford, IL, USA) and samples adjusted to equal protein content before use.
Equal sample volumes were separated by SDS-PAGE, transferred and immunoblotted using monoclonal anti-MRP1 antibody, MRPr1 (Signet Laboratories, Inc, Dedham, MA, USA), or polyclonal anti-Bax, anti-Bcl-2, anti-Bcl-x L , and anti-actin antisera (PharMingen, San Diego, CA, USA). After incubation of membranes in horseradish peroxidase-conjugated secondary antibodies, immunostained bands were detected using chemiluminescent methods (ECL, Amersham International Plc, Little Chalfont, Bucks, UK; or SuperSignal Substrate, Western Blotting, Pierce Chemical Co) with image collection on X-ray film (ECL) or by phosphorimager (SuperSignal). Quantitation of immunostained bands was by densitometry (X-ray film), or by image analysis of phosphorimage data using Multi Analyst 1.02 software (Bio-Rad). Equal protein loadings were checked by immunostaining with an anti-β-actin rabbit polyclonal antibody, and, by visual inspection or densitometry of duplicate gels stained with Coomassie Blue.

Immunocytochemistry
Cells grown on collagen-coated cover slips were fixed in 4% paraformaldehyde in PBS, pH 7.4, rinsed 3 times in PBS, then further treated with ice cold methanol for 15 minutes, followed by 2 rinses with PBS. After blocking, the cells were immunostained with monoclonal antibodies directed against either synaptophysin (anti-SVP-38 antibody, Sigma Chemical Company), or high molecular weight neurofilament (anti-NF-200 antibody, Sigma Chemical Company). Stained coverslips were mounted on slides using the ProLong Antifade Kit (Molecular Probes, Inc). The slides were examined using an LSM GB 200 laser scanning confocal microscope (Olympus, Tokyo, Japan) with excitation wavelengths of 488 nm (Argon) and 543 nm (Helium). The images were collected electronically (1024 × 768 pixels) and analysed using NIH Image software (W. Richards, National Institutes of Health, Bethesda, MD, USA). To estimate relative cellular protein content, mean pixel density of 5 images each containing 20 individual cells was calculated by measuring the mean pixel density of 20 identical sized spots per image (excluding measurements made over the nucleus) and subtracting it from the mean non-cellular background measured in each image by the same method.

TUNEL labelling
NB cells were stained using the In situ cell death detection kit (Roche, Castle Hill, Australia). Briefly, cells were harvested, along with their respective supernatants, and washed once with cold PBS. A total of 5 × 10 8 cells were placed on a polylysine glass slide for cytospin preparation. Following cytospin, the slides were fixed in a 4% paraformaldehyde solution for 1 hour at room temperature. The slides were subsequently rinsed with PBS and incubated in a permeabilising solution (0.1% Triton X-100, 0.1% sodium citrate) for 2 minutes at 4˚C. The slides were then rinsed twice with PBS and dried. 50 µl of the TUNEL reaction mixture was added onto the slides, which were incubated at 37˚C for one hour. The slides were subsequently rinsed 3 times in PBS before analysis under a fluorescence microscope. 5 microscopic fields (> 100 cells), in duplicate experiments, were counted in duplicate slides to determine the proportion of TUNEL positive or apoptotic cells.

Down-regulation of endogenous MRP1 expression in NB cells
The NB tumour cell lines, BE and SY, were transfected with vectors constitutively expressing, in antisense, either a 1.3 kb (1634-2962 bp) or a 1.7 kb (0-1633 bp) fragment of the human MRP1 cDNA. Following 2 separate transfection experiments for each cell line, individual clones (MRP-AS) were isolated in Hygromycin B and characterised for MRP1 expression ( Figure 1A and 1B). There was a significant difference between the growth rate of BE MRP-AS clones, and control MEP cells (Figure 2). Compared with control cells, BE MRP-AS clones were markedly growth inhibited, with day 7 cell counts ranging from 15% to 39% of controls. Similarly, SY MRP-AS clones were markedly growth inhibited when compared with controls. Morphologic analysis of BE MRP-AS clones demonstrated marked spontaneous neurite formation (Figure 3). There was a strong inverse correlation between MRP1 expression level and the proportion of BE MRP-AS cells with neurites after 7 days in culture when individual clones were compared (Figures 1 and 3).
Increased expression of synaptophysin and high-molecularweight neurofilament are features of neuronally differentiated cells (Jahn et al, 1985). Confocal microscopy and image analysis untreated BE MRP-AS cells were equivalent to that seen in control cells undergoing morphologic differentiation following retinoid treatment (Figure 4).

MRP-AS clones demonstrate enhanced spontaneous cell death
We used trypan blue staining to determine the proportion of dead cells in culture one week after plating. BE and SY MRP-AS clones all had a significantly higher proportion (0.5-2.5 fold) of dead cells than control cells (data not shown). Cells grown on slides in 2-well chambers were stained with Wright's stain and examined for apoptotic morphology. A proportion of MRP-AS cells exhibited typical apoptotic morphology ( Figure 5A), and, in comparison with control MEP cells, MRP-AS clones had a higher proportion of morphologically apoptotic cells ( Figure 5B). To confirm the presence of apoptosis in MRP-AS cells we evaluated TUNEL positivity and Caspase 3 activation ( Figure 6). MRP-AS clones 1.3 and 22 exhibited a higher proportion of TUNELpositive cells compared to controls during log phase growth. The increase in TUNEL positivity was proportionately similar to the increase in apoptotic bodies shown for these clones in Figure  5B. Caspase 3 activity was similarly increased in both clones. Thus, in the MRP-AS transfectants, growth data, morphologic and immunocytochemical evidence showed that down-regulated MRP1 expression was associated with growth inhibition, neuritic differentiation, and apoptosis.

Bcl-2 protein expression is decreased in MRP-AS cells
Differentiation and apoptosis have been linked to the level of Bcl-2 expression in NB primary tumour tissue and cell lines (Hoehner et al, 1995). We therefore examined Bcl-2 mRNA and protein expression levels in the MRP-AS and control cells. Bcl-2 mRNA expression levels were similar for control and MRP-AS clones ( Figure 7A). However, immunoblots showed that the MRP-AS clones had markedly reduced Bcl-2 protein expression compared with control MEP and parental BE cells ( Figure 7B). Antisera directed against Bcl-x L and Bax demonstrated no significant difference in the expression levels of these proteins for MRP-AS and control cells ( Figure 7B). These results suggested that in the MRP-AS clones, reduced Bcl-2 protein expression was due to a translational or post-translational mechanism.

Ectopic MRP1 overexpression blocks the neuritogenic effects of retinoic acid
The BE cell line was transfected with a vector constitutively overexpressing the full-length human MRP1 cDNA (BE MRP-S). 2 clones  with an MRP1 expression level 2-fold higher than controls ( Figure 1C), were selected for further study. Control (BE CEP) and MRP-S cells showed similar morphology and growth characteristics to parental BE cells. To test the function of MRP1 in the MRP-S cells, we performed cytotoxicity assays using the MRP1-specific substrate, sodium arsenate . Compared with controls, MRP-S transfectants had a 2.3-fold (P < 0.05) higher resistance to the cytotoxic effects of sodium arsenate as measured by ID 50 .
To investigate the effect of increased MRP1 expression levels on the capacity of NB cells for neuritic differentiation, control and MRP-S cells were treated with aRA. We found that aRA induced a similar level of concentration-dependent growth inhibition in MRP-S, BE CEP and BE cells (Figure 8). However, the MRP-S clones were markedly resistant to neurite formation ( Figure 8). Similar results were observed when the cells were treated with 13-cis-retinoic acid or 9-cis-retinoic acid (data not shown).

DISCUSSION
In this study, we have shown that the level of MRP1 protein expression in NB tumour cells has unique effects on cell survival, and on the capacity of NB cells to undergo neuritic differentiation in vitro. Reduced MRP1 expression in MRP-AS cells correlated with reduced Bcl-2 protein levels through a post-transcriptional effect on Bcl-2 expression. Taken together, our results suggest the hypothesis that high-level MRP1 expression, may be an acquired characteristic of neuroblasts undergoing malignant transformation in vivo, which contributes to the malignant phenotype by blocking the developmentally regulated neuronal deletion which occurs during embryogenesis of the neural crest. A link between MRP1 expression level, and protection against cell death caused by substances other than chemotherapeutic cytotoxic drugs, has not been reported. However, recent studies indicate that another membrane-associated efflux pump, Pgp, may protect malignant cells against caspase-dependent apoptosis induced by activation of Fas, serum deprivation, and other death-inducing stimuli (Robinson et al, 1997;Smyth et al, 1998;Johnstone et al, 1999). Growth factors and other morphogens involved in neural crest development are possible MRP1 export substrates. Embryonal neuronal deletion is regulated by multiple neurotrophins in an ageand cell type-specific manner (Davies, 1997). Thus far, basic fibroblast growth factor, acting as a survival factor for Kaposi's and osteogenic sarcoma cells, is the only protein defined as an MRP1 export substrate Gupta et al, 1998).
Our studies are the first to link MRP1 expression level with a process of differentiation and cell death. Our previous findings linking MRP1 expression levels in NB primary tumour tissue and patient prognosis, combined with the data presented here, indicate that MRP1 may play a role in the resistance of some NB tumour cells to spontaneous regression. Further in vivo studies are required to substantiate this hypothesis and to define the mechanism of action of MRP1 in NB tumour cells.