ER-Alpha-cDNA As Part of a Bicistronic Transcript Gives Rise to High Frequency, Long Term, Receptor Expressing Cell Clones

Within the large group of Estrogen Receptor alpha (ERα)-negative breast cancer patients, there is a subgroup carrying the phenotype ERα−, PR−, and Her2−, named accordingly “Triple-Negative” (TN). Using cell lines derived from this TN group, we wished to establish cell clones, in which ERα is ectopically expressed, forming part of a synthetic lethality screening system. Initially, we generated cell transfectants expressing a mono-cistronic ERα transcription unit, adjacent to a separate dominant selectable marker transcription unit. However, the yield of ERα expressing colonies was rather low (5–12.5%), and only about half of these displayed stable ectopic ERα expression over time. Generation and maintenance of such cell clones under minimal exposure to the ERα ligand, did not improve yield or expression stability. Indeed, other groups have also reported grave difficulties in obtaining ectopic expression of ERα in ERα-deficient breast carcinoma cells. We therefore switched to transfecting these cell lines with pERα-IRES, a plasmid vector encoding a bicistronic translation mRNA template: ERα Open Reading Frame (ORF) being upstream followed by a dominant-positive selectable marker (hygroR) ORF, directed for translation from an Internal Ribosome Entry Site (IRES). Through usage of this bicistronic vector linkage system, it was possible to generate a very high yield of ERα expressing cell clones (50–100%). The stability over time of these clones was also somewhat improved, though variations between individual cell clones were evident. Our successful experience with ERα in this system may serve as a paradigm for other genes where ectopic expression meets similar hardships.


Introduction
Tumor expression of estrogen receptor alpha (ERa) plays an important role in the clinical care of breast cancer patients both as a prognostic factor and as a therapeutic target. Unfortunately, about two-thirds of breast cancer patients have an estrogen receptor alpha-negative disease. Within this large group of ERa 2 negative/ endocrine therapy-resistant breast cancer patients, the Triple-Negative (TN) subgroup has bad prognosis, as it tends to develop metastases. So far, this group is being treated by surgery/irradiation and for the most part nonspecific chemotherapy [1].
Genes, whose activity, expression or dependence is considered to have increased in cancer, are prime candidates for therapeutic intervention. Cancer cells may depend upon such changes in gene expression, not only during tumor initiation, but also during malignancy progression (i.e. ''oncogene addiction''). This is exemplified by the choice of the oncogene ERBB2/HER2 as drug target in ERBB2/HER2-positive breast cancer [2]. Alternatively, using the concept of synthetic lethality [3], efforts have been directed towards identification of chemicals/drugs or target genes whose activation or ablation, respectively, synergizes with mutations in either oncogenes or tumor suppressor genes [4,5]. The availability of large-scale synthetic low-molecular-weight chemical libraries has allowed high-throughput-screening (HTS) for compounds that are synergistically lethal with defined human cancer aberrations in activated oncogenes or tumor suppressor genes; the so called ''chemical synthetic lethality screens''. The generation of human/mouse genome-wide siRNAs and shRNAexpressing libraries has significantly advanced the complementing approach of ''genetic synthetic lethality screen''. The latter is being performed either at the single gene level, in an array format, or primarily by retroviral/lentiviral-pools carrying shRNA expression cassettes that are used to infect target cells at low multiplicity of infection [6,7]. In the case of the Triple-Negative derived BRCA1/BRCA2-deficient breast cancers, poly(ADP-ribose) polymerase (PARP), with or without DNA damaging agents, is synthetic lethal with BRCA1-or BRCA2-deficiency [8,9]. Likewise, the frequent inactivation of the PTPN12 tyrosine phosphatase tumor suppressor gene in TN derived tumors renders them sensitive to inhibitors of multiple tyrosine kinases [10].
The first system toward which our groups have decided to apply the synthetic lethality screening approach entails ERa-negative breast carcinoma TN-derived cultured cells. In order to do so, one needs to test the specificity of the identified targets in an in vitro cell culture system. A compulsory control ingredient of the synthetic lethality screening in the ERa-deficient TN breast carcinoma cell lines is stable transfectants expressing the human ERa cDNA. In view of the heterogeneity observed in the TN breast cancer group, it is essential to generate such complemented systems in several different TN-derived cell lines.
In light of the difficulty in creating stable expression of ERa (see below), this manuscript offers an alternative methodology [11,12] of doing so with greater success and fidelity. The generated ERa expressing clones can serve for the long term study of a variety of ERa associated topics.

A. Plasmids and constructs
pCDNA3-ERa, was constructed by the late Dr. Arnold Simons by first subcloning a 1820 bp SalI fragment encoding the complete coding sequence of wild type hERa from the GAL4 DB-hER plasmid [13] into the pBluescript II SK 2 plasmid. Then the XhoI -HindIII fragment encoding the ERa sequence from BlueScript was cloned into the pCDNA3.3 a neo expression vector from Invitrogen. pCDNA3 by itself was named pCDNA3-empty, and used to construct the G418 resistant ERa non-complemented cell clones. The neo R coding region is driven by the SV40 early genes promoter.
The pIRES-ERa plasmid (Fig. 1) is a derivative of the pIREShyg3 bicistronic vector (Clontech). The expression cassette of this vector contains the human cytomegalovirus (CMV) major immediate early gene promoter, followed by multiple cloning sites for cDNA/coding region insertion. A synthetic intron, is included downstream of the multiple cloning site. The encephalomyocarditis virus (EMCV) Internal Ribosome Entry Site (IRES) is followed by the bacterial hygromycin B resistance gene (Hygro R ) and the SV40 polyadenylation signal. The coding sequence of the human ERa-cDNA was cloned downstream of the CMV promoter into the EcoRV site of PIREShyg3, as a blunted EcoRV-XhoI fragment.
The pCMV-Bam-ERa-Hygro was constructed first by deleting the BamHI fragment encoding CD20 from pCMV-CD20 and religation of the vector. Next, an XbaI-HindIII fragment encoding TK-neo from the pCMV-Bam-neo was replaced with an NruI-SalI fragment encoding for TK-hygro R cassette from pCEP4 (Invitrogen). The coding sequence of human cDNA ERa was then cloned into pCMV Bam-Hygro by cloning an EcoRV-XhoI blunt-ended fragment encoding human ERa from pCDNA3-ERa (see above), into the BamHI site of pCMV-Bam-Hygro.

B. Cells growth
MDA-MB-231 [15] and GILM2 [16] were a kind gift from Prof. J. Price, MD Anderson. MDA-MB-435 and BT549 breast carcinoma cell lines were purchased from ATCC. MCF7 (ATCC) was a kind gift from R. Pinkas-Kramarski. Cell lines were routinely cultured at 37uC, 5% CO 2 , in DMEM supplemented with 5% fetal bovine serum (FBS), 4 mM L-glutamine, and penicillin/ streptomycin; these five medium ingredients were purchased from Biological Industries (Israel). ERa transfected cell clones were maintained in phenol red-free DMEM medium (Biological Industries, Israel) supplemented with 5% dextran coated Charcoal Stripped fetal calf Serum (CSS, manufactured by Hyclone, US) to prevent ERa activation (see below).

C. Cell transfection and clonal selection
MDA-MB-231, MDA-MB-435, and GILM2 transient and stable transfections were carried out using jetPEI reagent (PolyPlus Transfection, France) according to the manufacturer's instructions. In order to produce stable clones, a 1:5-1:20 dilutions of 5610 6 transfected cells was performed into 100 mm Petri dishes 48 hours post transfection. Selection was commenced the day after. Selective media consisted of DMEM without phenol-red, supplemented with 5% dextran charcoal fetal bovine serum (FBS), 4 mM L-glutamine, antibiotics (10 units/ml of penicillin and 50 mg/ml streptomycin) and the selective drug. Selection of stable clones was performed at 0.4 mg/ml G418 (Calbiochem) for pCDNA3-neo based clones, or at 0.2 mg/ml Hygromycin B (A.G. Scientific) for pIREShyg3 and pCMV-Bam-ERa-Hygro R based clones. Selective media was refreshed every 3 days thereafter. When colonies were big enough and interspaced, they were transferred to 48-well cell culture plates. For long term maintenance, 0.2 mg/ml G418, or 0.1 mg/ml Hygromycin B were used.

D. Western blot analysis
MCF7, BT549, MDA-MB-231, MDA-MB-435, GIML2 and their clonal derivatives were washed twice with cold Hanks buffer (Biological Industries, Israel), scraped with a rubber policeman and collected to a new tube. The cells were then centrifuged at 2000 rpm, 4uC for 5 minutes and pellets were lysed in ice-cold modified RIPA buffer (1% NP-40, 50 mM Tris pH 8, 0.15 M NaCl, 5 mM EDTA, 0.5% DOC and 1 mM PMSF, without SDS). Lysates were incubated on ice for 10 minutes, then cleared by centrifugation and stored in 270uC until use. For the Western blot analysis, the protein of each cell lysate was quantified by using the Bradford assay. 50 mg of each lysate was diluted 1:2 with a 46 SDS-PAGE sample buffer to a final concentration of 26 SDS-PAGE sample buffer (0.12 M Tris-Cl pH 6.8, 4% SDS, 20% glycerol, 0.2 M DTT, 0.008% bromophenol blue). These lysates were denaturated and separated on 10% polyacrylamide gel at 100v for 90 minutes at room temperature. Proteins were transferred to nitrocellulose membranes (BioScience, Germany) by electroblotting for 120 minutes at 12-20v or 120-150 mA. Membranes were blocked with blocking solution-1% nonfat dry milk in PBS-T (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 containing 0.1% tween) 20 for 1 hour at room temperature. The membranes were then probed with hERa mouse monoclonal primary antibody (NCL-ER-6F11; Novocastra Labs Ltd, England) at 1:1000 dilution in blocking solution overnight at 4uC, followed by 365 min washes in PBS-T 0.1%. Next, the membranes were incubated for 1 hour at room temperature with Goat anti mouse IgG HRP conjugated secondary antibody (Sigma, Israel), at 1:5000 dilution in blocking solution. Next, 365 min washes were preformed and the membranes were incubated with a home-made chemiluminescence solution (ECL solution-100 mM Tris, pH 8.5, 1.25 mM luminol, 0.2 mM p-cumaric acid, 0.01% H 2 O 2 ) for 1 minute. Blots were then exposed to film (Kodak) and developed. Signal quantization was performed by densitomentric analysis using a GE ImageQuant 350 scanner. After antibody stripping, a-tubulin was probed with a mouse monoclonal antibody (Sigma, Israel) at a dilution of 1:5000 and used as a cellular normalizing marker.

E. Dual-luciferase reporter assay
In order to assay ERa activity, cells were seeded in 24-well cell culture plates at 50-70% density, in DMEM supplemented with 5% FCS. The next day, the cells were transiently co-transfected with 0.5 mg p2xERE-PS2-luc (primary reporter vector containing the firefly luciferase gene under the ERa Response Element; i.e. ERE) and 0.3 mg pRNL-TK-luc (secondary reporter vector containing the Renilla luciferase gene under the constitutive HSV TK promoter), using the jetPEI transfection reagent. Forty-eight hours post-transfection, cells were washed twice with Hanks' (Biological industries), a balanced salts solution without phenol red, and cell lysates were prepared as described in the manufacturer's protocol for dual-luciferase reporter assay (Promega, USA). Briefly, cells were lysed with 45 mL/well of Passive Lysis Buffer for 10 minutes at room temperature. The firefly luciferase assay was initiated by adding 5-15 mL aliquot of cell lysate to 50 mL of Luciferase Assay Reagent II (LAR II). After recording the luminescence, 50 mL of Stop & Glo reagent was added to the same tube in order to quench the firefly luciferase reaction and simultaneously activate the Renilla luciferase reaction. Firefly and Renilla luciferase activities were measured using a LKB Wallac 1250 Luminometer. The firefly luciferase luminescence measured was proportional to the amount of active ERa protein present in the cells. The Renilla luciferase luminescence was proportional to the efficiency of the transfection. This internal control provides a convenient and reliable assay of efficiency. Normalized luciferase luminescence was calculated as followed: [(firefly luciferase activity/Renilla luciferase activity)6100]. These results, determined from lysates ERa complemented clones (as well as the positive and negative control), were then normalized again to MCF-7 positive control by dividing them to the same ratio obtained from the positive control: [(firefly/Renilla lumines-cence6100)/(MCF-7 firefly/Renilla luminescence6100)6100]. All experiments were performed several times in duplicates.

F. RT-PCR
For expression confirmation originating from the pIRES-ERa construct, RT-PCR was conducted. Two mg of total RNA extracted using EZ-RNA isolation kit (Biological Industries, Israel) were transcribed into first strand cDNA by hexamer priming, followed by PCR reactions as specified in the Long range RT-PCR kit (Qiagen). The PCR conditions included preincubation for 3 minutes at 93uC and 40 cycles comprised of 30 seconds at 93uC, 30 seconds at 54uC, 4.5 minutes at 68uC, and finishing up 10 minutes at 68uC.

G. Estimation of the cell growth doubling time
Each clone was seeded at a density of 2.5-3610 4 cells in 24-well tissue culture plates, and was incubated at 37uC in 5% CO 2 . The cells were counted every day for 5-6 days, using a cell counting chamber (Hemocytometer). The doubling time of each clone was calculated as following: [2624 hours/(Ave (no. of cells in day (X+1)/no. of cells in day X)].

Results
A. Generation & characterization of ERa-expressing MDA-MB-231 stable transfectants with the pCDNA3-Era expression vector A1. Transfection and selection of ERa expressing clones in MDA-MB-231 cells. In order to establish a supporting control system for synthetic lethality screening of ERa-negative breast cancer cells (of the TN subgroup), two human epithelial breast carcinoma cell lines, BT549 and MDA-MB-231, were initially utilized as recipients for the ERa-expressing constructs. The MDA-MB-231 cell line was particularly suitable for such preliminary studies since it is highly aggressive both in vitro as cell culture and in vivo upon grafting [15].
These two cell lines were initially transfected with pCDNA3-ERa. Simultaneously, these cell lines were also stably transfected with the pCDNA3 vector by itself, to serve as a negative control (pCDNA3-empty). Subsequent selection with G418 resulted in the establishment of the two groups of stable cell clones, ERacomplemented and ERa-empty (non-complemented). Initial studies performed in the presence of DMEM medium supplemented with 5% FCS showed that similar to other groups [17,18], the ERa complemented clones were much harder to establish than the empty vector control group. This observation was also reminiscent of studies demonstrating that ERa expression following long-term estrogen deprivation in ERa-positive breast cancer cells is thereafter manifested by an initial phase of estrogen hypersensitivity. This phase is characterized by apoptosis and rapid tumor regression at concentrations of estrogen (E 2 , Sigma Israel) below 10 213 M [19].
For these reasons, we decided to attempt generating the ERacomplemented MDA-MB-231 clones (which in some way are analogous to E 2 deprived ERa positive cells) in DMEM without phenol red, supplemented with 5% CSS. This way, exposure to residual ERa-receptor-activating agents was minimized, making the clones less sensitive to the ectopic expression of ERa.
In order to examine whether estrogen deprivation affects the stability and long term expression of functional ERa, newly emerging clones were grown simultaneously in e the regular phenol red-free DMEM supplemented with 5% CSS, as well as in DMEM supplemented with 5% FCS.
Forty established MDA-MB-231 cell clones selected for G418 resistance were then tested for ERa expression by Western blot analysis. Fig. 2 shows that only few clones (ERa-2, ERa-7a, ERa-8a and ERa-17a) of the established cell clones, express the protein. MCF7, a bona fide ERa-positive cell line, was used as positive control. ERa-16a, a cell clone which was established after transfection with pCDNA3-ERa integrating plasmid, turned out not to express the protein. It also served us as a negative control, when required. The four clones show various levels of ERa expression, as compared to the positive control (Fig. 2).
A2. Selecting for ERa-active clones. The next phase was to analyze whether the selected clones synthesizing the ERa, protein express a functional receptor. The quantification of ERa activity was performed by the dual luciferase reporter assay (see Methods).
MDA-MB-231 established clones were initially tested while grown in DMEM supplemented with 5% FCS, which naturally contains estrogen (E 2 ). and 231-ERa-20, expressed between 0% and 30% receptor activity, as compared to MCF7. Not only were these levels very low, but further ERa reporter assays showed that these five clones continued loosing activity over time. Additional tests performed on the former four ERa expressing clones, showed maintenance of appreciable levels of activity despite fluctuations over time (see below).
A3. Responsiveness to ligand. The next step was to determine whether the ectopically expressed ERa was under hormonal regulation. Several studies have shown that adding estrogen to serum starved ERa -positive cells, or to ERa ectopically expressed cells, can down-regulate expression of the receptor. The decrease requires a functional receptor and occurs at both the protein and mRNA levels [19,20]. This phenomenon has led us to systematically examine our ERa-complemented clones for responsiveness to regulation by estrogen. In order to do so, MCF-7 cells and four ERa complemented clones; 231-ERa-2, 231-ERa-7a, 231-ERa-8a and 231-ERa-17a, were seeded in 60 mm dishes under three different growth conditions: DMEM supplemented with 5% FCS, phenol red-free DMEM supplemented with 5% dextran coated charcoal filtered FCS (Dex), and phenol red-free DMEM supplemented with 5% dextran coated charcoal filtered FCS and 2610 28 M E 2 . After 24 hours, expression of ERa in these clones, under the three conditions was determined by Western immunoblot analysis. Fig. 4. reveals that all dishes treated with E 2 expressed a lower level of ERa compared with the parallel estrogen starved cells (Dex). As also expected, dishes treated with 5% FCS (FCS) expressed a lower level of estrogen receptor compared to the Dex cells, in accordance with estrogen saturating levels found in FCS.
Cell clones responding to the ligand regulation were also assayed for receptor activity under the different treatments. In order to do so, MDA-MB-231 established clones were seeded in 24-well tissue culture plates at the three different growth medium conditions, as mentioned above. Luciferase reporter plasmids were then transfected. After 24 hours, cell extracts were prepared and assayed. Fig. 5 summarizes the results obtained from these clones, comparing them to the positive control MCF7, grown in FCS.
When cell clones were seeded in DMEM supplemented with 5% FCS, they exhibited expression levels of 35% to 85% as compared to the expression of the positive control MCF7, which was assigned 100% relative activity. When cells were seeded in phenol red-free DMEM supplemented with 5% CSS, they behaved similarly to MCF7 and manifested an insignificant level of active ERa, in line with absence of the ligand (E 2 ). Naturally, the receptor was not activated, leading to its inability to bind to the ERE in p2xERE-pS2-luc reporter. However, when cells were seeded in phenol red-free DMEM supplemented with 5% CSS treated with 2610 28 M added E 2 , a significant increase in the activity level was exhibited.
Obviously, MDA-MB-231 parental cell-line did not display any significant expression level under all three conditions, since there is no ERa to be activated in the first place.
A4. Stability of the cell clones. In order to determine the clones' stability over time, ERa activity was assayed periodically   As mentioned above, clones established from the parental MDA-MB-231 cell line were maintained in culture with DMEM supplemented with 5% FCS, but also in phenol red-free DMEM supplemented with 5% CSS. We did not observe any difference in the cell clones' stability of receptor activity under the two growth conditions (data not shown).
Because drug administration efficiency is affected by the cultured cells' proliferation rate, we wished to compare the growth rate of MDA-MB-231 cell clones expressing ERa. As it turned out, the doubling time of two ERa complemented vs. four ERa noncomplemented (empty) clones was similar (around 21 hours), with almost identical growth curves (data not shown). . These inefficient attempts to recover ERa expressing cell clones had initiated the trial to establish an improved ectopic expression system using a bicistronic mRNA template for ERa translation (Fig. 1). The vector consists of a single transcription unit having the ERa ORF as the upstream cistron, and a dominant-positive selectable marker (Hygro R ), forming the downstream cistron, translated from an Internal Ribosome Entry Site (IRES).

B. Generation and characterization of MDA-MB-231 stable transfectants with ERa expressed from a bicistronic transcription unit
This configuration has the advantage that selection for the IRES-directed selectable marker gene expression may protect the transcription unit as a whole, including the upstream ERa ORF. Thus, this linkage may lead to a high yield of ERa-expressing clones. Accordingly, MDA-MB-231 parental cells were stably transfected with the ERa-IRES construct (Fig. 1). Transfection and selection were performed under minimal estrogen growth conditions, where phenol red-free DMEM medium supplemented with 5% CSS was used.
Screening of Hygromycin B resistant clones for ERa expression was initially performed by Western immunoblot analysis. Surprisingly, all hygromycin B resistant clones (8/8) showed some level of ERa expression. Upon testing these clones for ERa activity by the dual luciferase reporter assay, it became evident that the high frequency of ERa protein expression in the selected clones is accompanied by ERa activity (Fig. 8). Yet, as also observed by others, the relationship between Immunoblot quantification and activity is not always linear, for various potential reasons such as misfolding of the protein or proteolytic cleavage of terminal amino acids leading to loss of activity, etc.
Nevertheless, all nine IRES-ERa descendant clones of the MDA-MB-231 parental cell line showed high ERa-mediated activation of the reporter gene, amounting from 85% to 841% (!) of the level displayed by MCF7.  In order to evaluate stability of ERa expression over time, dualluciferase assays were performed intermittently over a relatively long time period (Fig. 9). Cell clones were kept under Hygromycin B selection, in phenol red-free DMEM medium supplemented with 5% CSS, in order to minimize potential expression suppression by the ligand. At each time point of assay, dual luciferase activity was normalized to the activity obtained in MCF7, transfected at the same time point, alongside the clones.
Examination of the stability of the ectopically expressed ERa-IRES clones in MDA-MB-231 over time (Fig. 9), revealed that three out of four clones still retain appreciable activity (as compared to that of the reference, MCF7 cells) 155 days after being initially monitored, amounting to over six months post transfection with the pERa-IRES vector.
B2. Characterization of the ERa-IRES-Hyg R containing transcript. After obtaining a high yield of ERa expressing cell clones among the hygromycin B resistant transfectants in MDA-MB-231 (9/9), we set to characterize the ERa-containing hybrid transcript. A 3.2 kb fused transcript, encoding both cistrons: the ERa and the hygromycin B resistance selectable marker gene, was anticipated. Noteworthy, the plasmid sequence contains an intron situated downstream of the ERa ORF in such a way that splicing of the intron would result in RT-PCR product shorter by 295 bp (3.2 kb), compared to the plasmid template PCR product (3.5 Kb).
Lane 8 of Fig. 10A and the second and third lanes from left of Fig. 10B show that in two representative clones: ERa-IRES-5 and ERa-IRES-3, the expected 3.2 kb RT-PCR product of the spliced fused transcript is amplified. In contrast, ERa-2 a cell clone, described above and created by the monocistronic construct pCDNA3-ERa and devoid of the Hygromycin selectable marker, expressed the 1.8 kb ERa RNA product (lane 3 of Fig. 10A). The ERa-2 cell clone was used as a positive control for the ERa RT-PCR reaction, since based on prior Northern blot analysis, we observed that it contained an intact mRNA of the active ERa protein (data not shown). Thus, we have demonstrated in our MDA-MB-231 cell clones the presence of a full-length fused ERa-IRES-Hygro R transcript, which is also correctly spliced; see in Fig. 10B for example the PCR product size from pERa-IRES (3.5 kb) vs. the RT-PCR product from RNA belonging to each of the two ERa-IRES cell clones (3.2 kb).
B3. Generating stable transfectants with ERa expressed from a bicistronic transcription unit in MDA-MB-435 & GILM2 cells. The high yields (9/9) of ERa expressing cell clones in the TN-derived ERa-deficient breast carcinoma cell line MDA-MB-231 stands in contrast to previous studies carried out in our laboratory (and described in part above), as well as those of others (mentioned before), in which the efficiency of generating stable ectopic ERa expressing breast carcinoma cells is a cumbersome and inefficient procedure, resulting in a mere ,5% valuable clones.
In view of the vast genetic heterogeneity within breast carcinomas, and the triple-negative breast cancer patients' group in particular [21][22][23][24], there is demand for generating additional ectopic ERa producers.
Accordingly, we decided to attempt the bicistronic vector approach in two more triple-negative breast cancer cell lines: MDA-MB-435 [15,25], and GILM2 [16]. Both cell lines were transfected with the pERa-IRES construct as before and selection with hygromycin B was carried out under minimal estrogen exposure conditions, as previously outlined.
Initially, we tested the clones for ERa expression by Western blot analysis. Fig. 11A & 11B show the results obtained with MDA-MB-435 clones. Importantly, similar to MDA-MB-231, most of the hygromycin B resistant clones of MDA-MB-435 (13 out of 13) and GILM2 (2 out of 4; data not shown) displayed some level of ERa expression. We next tested these clones for their ERa activity, by the dual luciferase reporter assay (Fig. 12). ERa activity was measured in 12 out of 13 MDA-MB-435 (Fig. 12A) and in two out of four of the GILM2 ERa-IRES clones (Fig. 12B).
In order to evaluate the stability of ectopic ERa expression over time of each cell clone, dual-luciferase assays were performed over a relatively long time period for both cell lines (Fig. 13). Cell clones were kept under hygromycin B selection and in phenol red-free DMEM supplemented with 5% CSS for the entire period. They were exposed to estrogen only at the time of assay.
As it turned out, in the case of MDA-MB-435 derived ERa-IRES stable transfectants, 11 clones out of 12 retained at least 50% of their initial ERa activity (data not shown). As compared to the activity of MCF-7 cells which were assayed alongside the clones, at each time point, five of the six cell clones had at least 75% of MCF-7 ERa activity (Fig. 13A). In the case of the two GILM2 derived ERa-IRES expressing cell clones, there was at least retention of their initial ERa activity (Fig. 13B).

Discussion
This project was aimed at establishing an efficient method for ERa complementation in various ERa-deficient cell lines. The generation of these complemented clones served as a counterpart control ingredient for synthetic lethality screening systems in ERadeficient TN breast carcinoma cell lines.
Studies performed by others [17,18], together with unpublished experiments performed in our laboratory in breast cancer BT549 cells and MDA-MB-231 cells (mentioned in Section B.1) have shown that the yield of cell transfectants expressing appreciable levels of ectopically mono-cistronic transduced ERa is very low (5-12.5%). Our attempts to generate such MDA-MB-231 stable transfectants under minimal exposure to the ERa ligand did not improve this low efficiency (Figs. 2-6). Moreover, stable ERa expression in the MDA-MB-231 cell line transfectants lasted for 130 days in only three out of the initial five ''stable'' clones monitored for prolonged periods (Fig. 6). We therefore decided to try setting up a modified system in which the fraction of ERa expressing cell clones would be higher and the expression perhaps more stable.
Based on the discovery of the EMCV IRES element by E. Wimmer's laboratory, researchers have started using IREScontaining bicistronic mammalian vectors to co-express multiple genes [11,12,26,27]. Following that line, we chose the pIREShyg3 mammalian bicistronic expression vector. This vector is equipped with multiple cloning sites downstream of the strong capdependent CMV immediate early promoter and upstream of an intron fused to the IRES element which directs the translation of the dominant selectable marker-Hygromycin B resistance. Despite weaker translation from the downstream IRES element, the Hygromycin B resistance gene can be easily selected for. We cloned the ERa ORF into this bicistronic vector and then transfected it into the MDA-MB-231 breast carcinoma cell line. Maintaining the cells under phenol red-free DMEM supplemented with 5% CSS while selecting for Hygromycin B resistance, led to the isolation of nine clones. The ERa producing clones were identified by Western immunoblot analysis. All Hygromycin B resistant cell clones expressed the correct size ERa protein (Fig. 7). When assayed, the ERa protein turned out to be functionally active (Fig. 8). Importantly, nine out of the nine clones had high levels of ERa expression. Evidently, the selection for expression of the downstream Hygro R gene had a protecting effect on the upstream ERa gene expression from the same (bicistronic) transcription unit.
The mRNAs of the ERa producing clones were tested in an RT-PCR assay, verifying the integrity of the bicistronic mRNA (Fig. 10). Yet, with regard to MDA-MB-231 parental cells, although we were able to obtain cell clones such as ERa-IRES-3, which retained significant activity over a period of at least 155 days, most of MDA-MB-231 IRES-ERa descendents had   (Fig. 9). Nevertheless, this intermediate level of expression, alongside plentiful clones was sufficient to complete any screen or study required.
Following MDA-MB-231 cells, we attempted usage of the pIRES-ERa vector in other breast cancer cell lines: MDA-MB-435, and GILM2. We received high initial yields of ERa expression; 12 out of 13 clones for MDA-MB-435, and 2 out of 4 for GILM2 (Figs. 11 & 12). Most of these cell clones retained high activity for at least 98 days (Figs. 13a & 13b). Noteworthy, usage of the Hygro R selectable marker gene driven by the relatively strong promoter HSV TK, such as in the monocistronic ERa expressing vector pCMV-Bam-ERa-Hygro, led to only 3-4 stable MDA-MB-231 cell clones out of forty eight which express the ERa receptor (Lilach Wallerstein-Shomrony M.Sc. Thesis Tel Aviv University 2006). So the difference in the yield of stable ERa expressing cell clones between the bicistronic vector and the mono-cistronic pCDNA3-ERa cannot be due to the usage of a different selectable marker (Hygromycin R vs. neo R ) or a weaker promoter driving the selectable gene (CMV vs. SV40 early), respectively.
The proven ability of the bicistronic vector to generate multiple ERa expressing clones at very high yields, which for the most part retain stable expression upon further propagation, is the major  point of this manuscript. We would like to suggest those who are encountering hardships in other ectopic gene expression systems, to adopt the usage of such bi-or multi-cistronic vectors.
Using the cell systems generated in MDA-MB-231, MDA-MB-435 and GIML2, we are now focusing our attempts on genetic synthetic lethality screenings [7]. These screenings entail a group of 100 human antiapoptotic/survival genes (known to be expressed in human breast cancers), and thereby promoting tumor growth and survival, as well as a lentiviral pool of shRNAs expressing clones targeted against all known human coding RNAs [7].