TTNPB

The modified firefly luciferase reporter gene (luc+) but not Renilla luciferase is induced by all-trans retinoic acid in MCF-7 breast cancer cells

Key words: all-trans retinoic acid, MCF-7 cells, modified firefly luciferase, Renilla luciferase, reporter assay

Summary

Luciferase genes are widely used as reporters to analyze promoter and regulatory elements. We found that a luciferase reporter gene vector with a modified firefly luciferase gene (luc+), but not Renilla luciferase (Rluc), was induced by all-trans retinoic acid (tRA) in the MCF-7 breast cancer cell line. tRA (5 × 10—6 M) increased luciferase activity of the pGL3 promoter vector (containing luc+) up to ~3.8-fold in MCF-7 cells, but not in LNCaP prostate cancer cells or JEG-3 choriocarcinoma cells. Chimeric plasmids were constructed and showed that tRA-induction required the luc+ gene, but not any specific promoter or vector sequence. Time course and dose-response studies of tRA-induction indicated that longer treatment (>24 h) and higher tRA dose (>10—6 M) were required for luc+ induction compared with those for a positive retinoic acid response element (maximum induction at 6 h and 10—8 M tRA). Studies with the translation inhibitor, cycloheximide, indicated the half-life of the luc+ protein was increased from 9.7 ± 1.5 to 22.1 ± 3.1 h with tRA treatment. Other retinoids, TTNPB, a retinoic acid receptor β/γ-specific ligand, and a retinoid X receptor ligand, did not significantly increase luc+ expression. Caution is needed in analysis of retinoid responsive gene regulation with the luciferase reporter system in MCF-7 cells, especially at high retinoid concentrations.

Introduction

Transient expression of reporter genes in transfec- ted plasmids is an established method for analysis of promoters and upstream regulatory elements. Fire- fly luciferase assays are simple to perform, sen- sitive, and linear over a wide range of enzyme concentrations [1, 2]. The sequence of some re- porter vectors, however, contains cis-elements which modify gene expression in response to various hor- mones [1]. The herpes simplex virus thymidine kinase (HSV-TK) promoter, commonly used as a heterologous promoter to analyze upstream regulat- ory sequence, responds to triiodothyronine (T3) and contains a thyroid hormone receptor binding site [3, 4]. T3 has been reported to decrease the fire- fly luciferase expression through a negative T3 re- sponse element in the luciferase gene [5, 6]. A vitamin D derivative, 1,25-dihydroxycholecalciferol, reduces firefly luciferase activity in the pGL2 vector, up to 40%, and requires vitamin D receptor expression [7]. To avoid the effects of trans-acting factor(s) on lu- ciferase gene expression, a modified firefly luciferase gene [8] or luciferase genes from different species, such as sea pansy (Renilla reniformis) [9], have been developed as reporter genes.

The MCF-7 human breast cancer cell line is widely used in breast cancer research. All-trans retinoic acid (tRA) inhibits cell cycle progression and induces ap- optosis in many tumor cell line [10], including MCF-7 cells [11–14]. tRA also induces some differentiated functions in the MCF-7 cells, such as iodide accu- mulation [15]. tRA regulation of gene transcription is mediated through two families of nuclear receptors, retinoic acid receptors (RAR) and retinoid X recep- tors (RXR), both of which are expressed in MCF-7 cells [16]. The tRA–RAR complex recruits RXR and possibly other cofactors and binds to a cis-element (retinoic acid response element, RARE) in the regu- latory regions of the RA-responsive gene [17, 18].

During a study of retinoid responsive genes in breast cancer, we investigated the effects of tRA on re- porter gene expression in MCF-7 cells. We unexpect- edly found that MCF-7 cells transiently transfected with a pGL3 promoter vector (Promega) responded to
5 × 10—6 M of tRA without any additional sequence in the vector. We found that the luc+ gene, a modi- fied firefly luciferase gene in the pGL3 vector but not Renilla luciferase (Rluc), is induced by tRA, and the effect of tRA is due, in part, to a prolonged half-life of the luc+ protein.

Materials and methods

Chemicals, cells, and vectors

MCF-7 breast cancer cells (lot #F15100, and 205623), JEG-3 cells, and LNCaP cells were ob- tained from the American Type Culture Collec- tion (ATCC, Rockville, MD) and maintained as described [15]. Two lots of MCF-7 cells were utilized 3–8 passages after being obtained from ATCC. All-trans retinoic acid (lot #60K1776 and 84H0658), TTNPB (4-[E2-5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl-1-propenyl]benzoic acid), were purchased from Sigma (St. Louis, MO). Two retinoid compounds, AGN 190168 [19] and AGN 194204 [20] were provided by Allergan (Irvine, CA). These retinoids were dissolved in di-methyl sulfoxide (DMSO) to 10—2 M and stored at —20◦C. Luciferase reporter vectors, pGL3 pro- moter and pRL-TK, were purchased from Promega (Madison, WI). A Renilla luciferase vector containing four RAREs (pRL-DR5) was constructed by blunt- end ligation of annealed synthetic oligonucleotides derived from human RARβ2 gene [21, 22], 5r- GGGTAGGGTTCACCGAAAGTTCACTCGGATCC- 3r, and 5r-GGATCCGAGTGAACTTTCGGTTAACCCTACCC-3r, into the BglII site of the pRL-TK vector treated with T4 DNA polymerase (New England Biolabs, Beverly, MA). Sequence analysis by Laragen Co. (Santa Monica, CA) indicated four RAREs were inserted to the BglII site. Plasmids used for trans- fection analysis were purified with Qiafilter Plasmid Maxi Kit (Qiagen, Valencia, CA). All restriction enzymes and DNA modification enzymes were pur- chased from New England Biolabs (Beverly, MA).

The RARα expression vector [23] was provided by Dr R.M. Evans (Salk Institute, La Jolla, CA).

Transient expression analysis

Cells (3.0 × 104 per well) were seeded in 24-well dishes 24 h before transfection. Unless otherwise noted, 0.2 µg luciferase construct and 0.1 µg pSV- β-galactosidase control vector (Promega) were transfected to the cells by effectene transfection reagent (Qiagen, Valencia, CA), and the transfection medium was changed to growth media 24 h after the beginning of the transfection. Luciferase assays were performed 72 h after the beginning of transfection with com- mercial reporter lysis buffer and the luciferase assay substrate (both from Promega). β-Galactosidase assay was performed as described [24], and the transfec- tion efficiency of luciferase reporter constructs was normalized to β-galactosidase activity. In some exper- iment, DEAE-dextran (Promega) or LipofectAMINE reagent (Invitrogen, Carlsbad, CA) was used for trans- fection as described [25].

Statistical analysis

All numerical data are expressed as the mean ± SD. Statistical significance of differences was determined by paired Student’s t-test.

Results

All-trans retinoic acid increases the luciferase activity in the MCF-7 cells transfected with a luciferase reporter vector.In MCF-7 cells (lot #F15100) transfected with a pGL3 promoter vector, which contains a modified firefly lu- ciferase gene (luc+) under the control of the SV40 promoter, the addition of 5 × 10—6 M of tRA increased luciferase activity up to ~3.8-fold (Figure 1). tRA-induction was not observed in other cell lines, LNCaP human prostate cancer cells and JEG-3 hu- man choriocarcinoma cells (Figure 1), indicating that the response of the luciferase reporter vector was cell-selective. We modified a number of reagents to determine the influence on the RA-response. Equiva- lent results were obtained with a different lot of tRA. Dilution of tRA in ethanol rather than DMSO made no significant difference in tRA-induction (data not shown). Effectene transfection reagent (Promega), a non-liposomal lipid formulation, was utilized to in- troduce plasmid DNA into the cells. Transfection with other reagents, LipofectAMINE (Invitrogen), a liposome-based reagent, or DEAE-dextran (Promega) positively charged polymers, made no significant dif- ference in the response to tRA (data not shown).

The response to tRA is dependent on the modified firefly luciferase gene, luc+

The pGL3 promoter vector consists of a modified fire- fly luciferase (luc+) gene coding region, the SV40 promoter controlling the luc+ expression, and the other vector sequence, including a β-lactamase (ampicillin resistant) gene and the plasmid replication start site (Figure 2). To clarify which part of the reporter construct was conferring tRA-induction, we construc- ted chimeric plasmids from the pGL3 promoter vector and pRL-TK vector, which carries Renilla luciferase (Rluc) gene under the control of the HSV-TK promoter (Figure 2). To produce the luc+ constructs with the other vector and/or promoter sequences, the SV40 promoter and/or the pGL3 vector sequences were re- moved from the pGL3 promoter and the HSV-TK promoter and/or pRL vector sequences from the pRL- TK were fused to the luc+ gene (Figure 2). Chimeric constructs of the Rluc gene were made as well as those with luc+ (Figure 2). As shown in Table 1, among the various combination only the luc+ gene was associated with tRA-induced luciferase activity in MCF-7 cells (lot #F15100), up to 5.2-fold. The re- sponse was not dependent on the vector sequence or promoter. Clonal variation of MCF-7 cells has been reported [26], however, a different lot of MCF-7 cells (#205623 from ATCC) also responded to tRA after the transfection of the luc+ constructs, although with less magnitude (Table 1). On the other hand, the constructs with the Rluc gene did not significantly respond to tRA (Table 1). These results indicated that the luc+ gene is required and sufficient for the response to tRA in MCF-7 cells.

Comparison of the response of the luc+ with that of positive RARE

The effects of tRA on gene regulation are usually mediated by RAR, which acts as a transcriptional trans regulator by binding to cis-acting regulatory elements. The RARE typically consists of two half sites, 5r-[A/G]G[G/T][T/A]CA-3r, arranged as a di- rect repeat with two or five bases between two half sites (direct repeat DR-2 or DR-5). Although tRA increased the expression of the luc+ independently of the other sequence of the constructs, our inspec- tion of the sequence of the luc+ gene indicated no consensus RARE. To compare the induction of luc+ by tRA with the activation of positive RARE from another RA-responsive gene promoter, we made a construct, pRL-DR5, containing an Rluc gene con- trolled by an HSV-TK promoter and four positive RAREs (DR-5), whose sequence was derived from the human RARβ promoter [21, 22]. We compared the tRA response of the luc+ gene in the pGL3 promoter vector to the positive RARE. Luciferase activity in MCF-7 cells transfected with the luc+- containing pGL3 promoter vector was still increasing at 48 h (Figure 3(A)), while the response of the posit- ive RAREs reached a maximum at 6 h (Figure 3(A)), indicating the delayed response of the luc+ gene to tRA. The dose-response study showed that higher con- centrations of tRA (10—6 M or more) were required for the induction of the luc+, compared with that (10—8 M) for stimulation of the RARE (Figure 3(B)). tRA induces its own metabolism to polar derivatives and isomers by induction of converting enzymes [27], and a high metabolic rate has been reported in MCF-7 cells [28]. The higher dose of tRA required for luc+ induction compared to the RARE may be due, in part, to greater RA metabolism over the longer time frame of the luc+ response.

Half-life of the luc+ protein in MCF-7 cells treated with tRA

Our time course study of RA stimulation of the luc+ gene suggested an influence on post-transcriptional events. To investigate whether tRA affects the stabil- ity of the luc+ protein, we assessed the half-life of the protein in the MCF-7 cells treated with or without tRA. We utilized cycloheximide (CHX), which in- hibits translation and allows for determination of the decay of preexisting protein. Since luciferase does not require post-translational modification for activity [9], the half-life of the luciferase activity after the addition of CHX indicates the half-life of the luc+ protein. The half-life of luciferase activity in MCF-7 cells without tRA treatment was 9.7 ± 1.5 h, longer than the previous data (~3 h) in HepG2 human hepatoma cells [29], while treatment with tRA (5 × 10—6 M) stabilized the luc+ gene products, extending the half-life to 22.0 ± 3.1 h (Figure 4). tRA treatment did not significantly change the half-life of co- transfected β-galactosidase in MCF-7 cells (data not shown).

Response of the luc+ gene to retinoid compounds

The retinoid receptors in MCF-7 cells include RARα, RARγ, and RXRα [16]. To evaluate if RAR mediates the induction of the luc+ gene, we assayed luciferase activity in MCF-7 cells transfected with pGL3 pro- moter in response to several retinoid compounds. The RAR-specific ligand, TTNPB [30], an RAR β/γ ligand AGN 190168 [19], and an RXR ligand AGN 194204 [20], however, did not significantly increase the lu- ciferase activity (Table 2). On the other hand, the pRL-DR5, an Rluc construct with positive RAREs, responded to all retinoids tested (Table 2).

Exogenous RAR expression in MCF-7 cells transfected with the luc+ gene

Previous studies of the response of reporter genes to T3 or vitamin D3, demonstrated increased hormone- response by addition of exogenous receptors for those ligands. MCF-7 cells express RAR and RXR, while JEG-3 choriocarcinoma cells, in which no response of the luc+ to tRA was observed, express RXRα, but not the other RARs or RXRs [31, 32]. To investigate whether additional RAR expression induces the re- sponse of the luc+ gene in these cell lines, we carried out co-transfection studies of the RARα expression vector with the pGL3 promoter vector. Co-transfection of RARα, however, did not increase the response to tRA in MCF-7 cells or JEG-3 cells (data not shown).

Discussion

We found that the expression of a modified firefly lu- ciferase gene, luc+, is increased by tRA in MCF-7 cells due, at least in part, to stabilization of its gene product. The magnitude of tRA stimulation varied with the lot of MCF-7 cells, but was not dependent on the lot of tRA, transfection method, or tRA solvent.

The firefly luciferase reporter gene responds to T3 [5, 6], and a vitamin D3 derivative [7]. Co-transfection of nuclear receptors, thyroid hormone receptor [5], or vitamin D3 receptor [7] in each case, increases the effects of the ligands, indicating the participation of the nuclear receptors in the response of the luciferase gene. Our sequence inspection of the luc+ gene, however, demonstrated no typical RARE sequence, DR-2 or DR-5. In addition, our co-transfection stud- ies showed no significant effect of the addition of RARα on the response of the luc+. We also found differences of time course and dose-response to tRA between the luc+ gene and the positive RARE ex- pression. These findings suggest less influence of tRA on transcriptional regulation of the luc+ response. We demonstrated post-translational regulation of the luc+ by tRA with a prolonged half-life of the luc+ gene product in tRA-treated MCF-7 cells. Some proteins involved in cell cycle regulation, such as p53 [33] and p27Kip1 [34], are post-translationally stabilized by retinoic acid. Elongation of a half-life of p27Kip1 in neuroblastoma cells by retinoic acid is associated with down-regulation of ubiquitin–proteasome degrading pathway [34].

DMSO interacts with the lipid components of cell membranes and has effects on the stability and dynamics of biomembranes, therefore, it is commonly applied to cell biology as a cryoprotectant of cells, cell fusogen, and a permeability enhancing agent [35]. A non-toxic concentration (1%) of DMSO enhances up- take of DNA delivered with liposomes in MCF-7 cells [36]. The expression of apolipoprotein A-I mRNA and its promoter activity can be enhanced by 1% DMSO [37]. In our study, we first used DMSO to dissolve tRA for stock solution. The final concentration of DMSO in the medium, however, was 0.1%, and our additional study with ethanol as another tRA solvent indicated
that the effects of tRA on the luc+ expression were independent of DMSO.

No response of the luc+ to an RAR-specific ag- onist, TTNPB, an RAR β/γ ligand, or an RXR ligand was observed in the present study, while tRA, the other RAR-specific ligand, increased the luc+ expression. Although TTNPB binds to RARα, β and γ and is able to express RA-responsible genes as well as tRA, its affinity to those receptors is ~10-fold less than those of tRA [38]. Since the RARβ/γ ligand did not induce the luc+, high affinity binding of ligand to RARα might be required for the luc+ induction. The other possible explanation for the differential response of the luc+ gene to those ligands is a different conform- ation of the ligand/receptor complex. Analogous to a recent report of distinct conformation changes in ERs by different ER ligands [39], the ligand/receptor conformation may have an important influence on the post-translational response observed in our study. On the other hand, additional RARα did not significantly increase the luc+ expression, indicating the possibil- ity that RARs or RXRs are not directly involved in the response. Further study is required to clarify the mechanism of the luc+ induction by retinoids.

The MCF-7 cell line is one of the most commonly used cell line in breast cancer studies, and the fire- fly luciferase gene is also one of the most common reporter gene. Our results indicates the need to exer- cise caution in the study of retinoid response in breast cancer cell lines with luciferase reporter constructs to avoid overestimating the response of promoter activity to retinoids, especially at high ligand concentrations.