A New Dual-Promoter System for Cardiomyocyte-Specific Conditional Induction of Apoptosis

Apoptosis is a key determinant of major pathological processes, including chronic cardiac failure. We developed and tested in vitro a novel system to induce cardiomyocyte-specific apoptosis by virus-mediated delivery of a conditional transgene. The entire system was packaged in a lentiviral vector. We used the cardiomyocyte-specific Na+-Ca2+ exchange promoter to control the transcription of the reverse tetracycline transactivator, while the transgene expression was driven by the tetracycline-responsive element. The proapoptotic transgene of choice was the short isoform of the apoptosis-inducing factor, known to quickly induce the caspase-independent apoptosis when overexpressed in cells. Transduction of cardiomyocyte cells with this vector caused a tetracycline-regulated induction of apoptosis, which was not observed in noncardiac cells. Therefore, our system proved a valuable molecular tool for inducing controlled apoptosis in selected cells. Furthermore, the vector we developed is suitable for “lentivirus transgenesis,” an interesting strategy recently proposed for the genetic manipulation of animals other than mice, including large mammals.


Introduction
Apoptosis is a key process involved in the pathogenesis of major diseases affecting the immune, nervous, gastroenteric, and cardiovascular system [1]. For instance, studies in humans and in dogs have documented the presence of approximately 3000 apoptotic cells per million in tissue harvested from failing hearts [2][3][4]. is number might seem very small and pathologically insigni�cant; however, studies in genetically modi�ed mice have shown that a few hundred apoptotic cells per million are sufficient to cause dilated cardiomyopathy [5,6], a malignant cardiac disease that, in humans, accounts for 50% of the heart transplants [7]. It should also be bear in mind that apoptosis, once triggered, is a rapid process (24-48 h) and consequently any histological evaluation provides a limited "snapshot" of a dynamic and rapid sequence of cell deaths.
Given the importance of apoptosis as a pathogenic mechanism, inducing this particular form of cell death in animal models can provide important insights. Ideally, apoptosis induction should be restricted to selected organs or systems and adequately controlled to avoid fatal massive systemic damage. While such requisites can be met in conditional transgenic mice, genetic manipulations are more challenging in clinically relevant large animal models. A potential strategy in an adult organism is the induction of apoptosis in selected organs by virus-mediated delivery of conditional transgenes. Alternatively, controllable transgenes could be stably incorporated in cells, via viral vectors, during the embryonic life for later activation [8]. However, no vector-carried transgene system, to date, has been developed to induce apoptosis in a controllable fashion. Nevertheless, a possible candidate gene for such a strategy is the apoptosis-inducing factor (AIF).
Two complementary routes leading to programmed cell death have been characterized, namely, the caspase-dependent and caspase-independent pathways, which differ in the events needed to trigger the apoptotic cascade [9]. AIF was the �rst identi�ed protein involved in caspase-independent cell death [10]. Recent studies have characterized a novel AIF isoform called AIFshort (AIFsh), which comprises 7 exons derived from exon 10 to 16 of AIF. e resulting cytosolic protein corresponds to the C-terminal part of AIF, lacking its N-terminal domain, and is able to trigger caspaseindependent cell death upon expression in recipient cells by means of inducing chromatin condensation and large scale (50 kb) DNA fragmentation [11].
Given our speci�c interest in lentivirus-mediated gene transfer [12,13] and its cardiovascular applications [14], here, we developed a lentiviral vector for cardiomyocyte-speci�c, conditional expression of AIFsh. is vector contains two transcriptional units in tandem, one coding for the inducible prokaryotic tetracycline (Tet) activator (rtTA), which was placed under the control of a cardiac-speci�c promoter, and the second the AIFsh gene regulated by the rtTA-responsive element.

Construction of Dual-Promoter Lentiviral Vectors.
Cardiac-speci�c, Tet-regulated lentiviral vectors were developed using an approach similar to the one described by Gascòn and colleagues [15] in neurons. In the Tet-on system, the rtTA transactivator mediates the transcription of a transgene controlled by the TRE inducible promoter. Transcriptional activation is positively regulated by the presence of Tet, which allows the binding of rtTA to the TRE promoter, thereby promoting transgene expression. erefore, regulated expression with the Tet-on system requires the delivery in the same target cell of both the rtTA transactivator and the transgene.
e lentiviral backbone used in this study was derived from pTY2-CMV-GFP, an optimized self-inactivating (SIN), integrase-defective human immunode�ciency virus type-1 (HIV-1) vector obtained by Dr. A. Cara, Istituto Superiore di Sanità, Rome, Italy [17]. For our purposes, the parental pTY2-CMVGFP vector was modi�ed in the multiple cloning site (MCS), which was completely redesigned between the KpnI and SalI restriction sites, and the CMV-GFP cassette that was replaced by an inducible, cardiac cell speci�c expression system. e inducible expression vector was obtained by inserting the NCX-1 promoter [18], into pTY2-CMV-GFP. is promoter drives the expression of the reverse repressor of the rtTA gene. Another expression cassette downstream contains the TRE sequence placed upstream the hAIFsh coding sequence [11] (Figure 1(c)).
AIFsh sequence was ampli�ed from p3xFLAG-AIFsh (a kind gi from Dr. Susin [11]) and cloned in the new MCS between the SacII and NheI sites; TRE element was subcloned from pTRE2 (Clontech Europe, France) by means of digestion with XhoI and SacII and subsequent ligation upstream AIFsh, thus obtaining the pTY2-TRE-AIFsh. ereaer, rtTA sequence was ampli�ed from pTetON (Clontech Europe, France) and cloned between AscI and EcoRV sites. Finally, the NCX1 promoter sequence was ampli�ed from pNCX1831 plasmid (a gi from Dr. Menick [18]) and cloned into the AscI site upstream of rtTA. e correct orientation and the absence of mutations in all fragments were ver-i�ed by sequencing. e resulting construct, named pTY2-NCX1-rtTA-TRE-AIFsh (pNRTA), was used for lentiviral vector production.

Lentivirus Production.
Recombinant lentiviruses were produced by a three-plasmid transfection procedure as described elsewhere [12,13,19]. Brie�y, HEK 293T cells were cotransfected with the pNRTA vector or the pTY2-CMV-GFP vector, the packaging vector pΔ8.91, and a vector encoding the G-protein of the vesicular stomatitis virus (VSV-G).
e virus supernatants were collected 48 hours aer transfection, pooled, and �ltered through 0.45 m �lters. Viral titers were calculated by measuring HIV capsid p24 titer in the viral supernatants by an enzyme-linked immunosorbent assay (ELISA) (Innogenetics, Gent, Belgium).

Lentivirus-Mediated
Transduction. Transduction of cell lines was performed in 12-well plate, 1 mL volume per well, according to Bonci et al. [20]. Lentivirus supernatants were added to 10 5 target cells (400 ng p24/well, in Opti-MEM 2% FBS) and incubated for 6 hours. en cells were washed and the medium changed to Claycomb medium 10% FBS. 48 hours aer transduction, cells were assayed for inducible apoptosis upon the administration of 1 g/mL Tet (see below). e same protocol was used for HeLa cells with the exception of postinfection medium, which was switched to DMEM 10% FBS.
Efficiency of transduction was evaluated using the same amount of pTY2-CMV-GFP construct and testing the percentage of green cells aer 48 hours by �ow cytometry.
Goat anti-AIF antibody and donkey anti-goat FITC conjugated were obtained from Santa Cruz Biotechnologies.

Statistical
Analysis. e statistical signi�cance was assessed by calculating values from nonparametric twotailed Student's -test or ANOVA when 3 or more groups were compared. e analyses were conducted using GRAPHPAD Prism 5.00 soware.

Dual-Promoter Lentiviral Vectors for Inducible Expression in Cardiomyocytes.
Dual-promoter lentiviral vectors for tissue speci�c expression of selected transgenes have already been described for several highly specialized tissues [21,22]. Here, we applied and further improved this system to obtain the conditional expression of a transgene in cardiomyocytes. is was achieved by placing AIFsh under control the of the feline Na + -Ca 2+ exchanger (NCX1) cardiac-speci�c promoter [18] and Tet-inducible activator rtTA. To this purpose, we engineered TY2-CMV-GFP, an HIV-1 vector expressing the green �uorescent protein (GFP) under the control of cyto- megalovirus (CMV) promoter. rough the intermediates described in Section 2, the CMV promoter and GFP were replaced by two expression cassettes. e CMV promoter was replaced by the cardiac-speci�c promoter (NCX1) and rtTA� GFP was replaced by the Tet-responsive element cassette (TRE) and AIFsh. e resulting construct (pTY2-NCX1-rtTA-TRE-AIFsh) has been named pNRTA for shortness ( Figure 1). Cell-speci�c and Tet-dependent expression of AIFsh was tested by transducing TY2-TRE-AIFsh in mouse cardiac  . Similarly, the number of AIFsh-expressing cells in induced versus noninduced cells increased and was a high statistical signi�cance ( ; the number of AIFsh expressing cells was calculated as the ratio between green �uorescent versus non�uorescent cells in each �eld counted. e efficiency of transduction was inferred with HL-1 cells transduced with pTY2-CM�-�F� and examined for �uorescence by �ow cytometry; the percentage of transduced cells was estimated to be between 60% and 70% in all transduction experiments (data not shown). HeLa cells transduced in parallel showed no transgene expression even aer induction (data not shown), thus con�rming that the expression was, in fact, cardiomyocyte speci�c.

Tetracycline-Regulated Transgene Expression.
To determine the time-course kinetic of the transgene expression, HL-1 cells were transduced with the inducible vector and cell survival was evaluated aer the exposure to Tet. First, the optimal, nontoxic concentration of Tet was estimated by treating cells with the increasing concentrations of the drug and evaluating cell survival aer 24 hours. e optimal concentration, that is the best induction with minimal the toxicity versus the similarly treated nontransduced cells was 1 g/mL (Figures 3(a), 3(b), and 3(c)). is concentration was, therefore, used in the following time-course experiment in which cell survival was measured aer 2, 4, 6, 8, and 24 hours of Tet addition and consequent AIFsh induction. Cell counts showed a marked reduction (∼50%) in cell survival over time in transduced HL-1 cells treated with Tet, compared to untreated cells (Figure 3(d)). ese results indicated that the apoptotic process occurred predominantly between 4 and 24 hours aer AIFsh induction, while no further decrease in cell survival was observed at later time points (data not shown). A small percentage of cell death could still be observed in untreated cells and can probably be ascribed to TRE promoter leakage.
To con�rm the cell speci�city of the observed effect, we performed the same experiments in HeLa cells. Despite good efficiency of transfection (65%), survival of this noncardiac cell line was not affected upon Tet stimulation of gene expression (Figure 3(f)), thus proving that the effect on cell survival is lineage speci�c and driven by the cardiomyocyte-speci�c promoter.
To rule out the possibility that the observed reduction of cell survival was either due to Tet treatment or lentiviral transduction, we assessed cell survival in nontransduced HL-1 and HeLa cells, treated or untreated with Tet (Figure 3(e)), and in HL-1 and HeLa cells transduced with the lentiviral backbone lacking the inducible system and the pro-apoptotic transgene (Figure 3(g)). e results clearly showed that the reduction in cell survival was not due to Tet treatment or vector transduction, thus con�rming the speci�city and inducibility of the system.

Tet Treatment Triggers Apoptosis in Transduced HL-1 Cells.
To provide evidence that Tet induction of AIFsh expression was effectively leading to cell death by apoptosis, we performed a TUNEL, an immuno�uorescence assays on transduced HL-1 cells (Figure 4(a)). is analysis showed a strong �uorescent signal, corresponding to DNA fragmentation, 4 hours aer Tet addition in approximately 50% cells (Figure 4(b)) (consistent with the decrease in cell survival found in the cell counts). In contrast, no signal was observed in either nontransduced or transduced but Tet-untreated cells. Statistical analysis showed that the difference in �uorescence was highly signi�cant ( 0.001) when induced cells were compared to noninduced cells or the negative control, while no signi�cance was found when comparing the �uorescence of the negative control with that of uninduced cells.
ese results showed that pTY2-TRE-AIFsh is an efficient and reliable in vitro system for inducing apoptosis of cardiomyocytes in a tightly regulated manner. e system consists of a lentiviral vector carrying a dual-promoter system that allows speci�c induction and restricted the expression of genes in cardiomyocytes. Induction was achieved by means of rtTA-regulated transcription of a cell type-speci�c promoter, whereas the transgene expression was driven by the TRE. In comparison to other gene delivery systems, in which the transfected genome is delivered in two separate plasmids, dual-promoter lentiviral vectors with modular design boast several advantages as they allow easy replacement of the transgenes and permit stable and regulatable expression of both transcriptional units in the recipient cells.
e evidence that neither latent expression from TRE in the absence of Tet (in HL-1 cells) nor the administration of Tet in transduced HeLa cells signi�cantly altered the cells viability strengthens the validity of our system as a potential molecular tool for further in vivo studies.
Lentiviral vectors are well known as tools to deliver and stably express transgenes in a variety of cell types, including cells of cardiac lineage [14], and have been successfully used in a number of gene therapy studies showing their ability to sustain long-term transgene expression in vivo (reviewed in [23]). Moreover, the type of vector developed here is suitable for "lentivirus trangenesis, " an interesting strategy recently proposed as an alternative to direct microinjection of DNA into pronuclei [8]. Lentiviruses can efficiently transfer and integrate cDNAs into the host cell genome of oocytes and early embryos, thus making possible the genetic manipulation of animals other than mice, including large mammals. e transgene can then be activated or deactivated at any time of postnatal life simply by administering or withdrawing Tet. In agreement with this idea, Tet-regulated transgene activation has been successfully used in several transgenesis approaches in vitro and in vivo. ese studies also demonstrated long-term responsiveness of the inducible system to Tet (reviewed in [24,25]).
Wencker et al. have shown, in genetically modi�ed mice, that very low levels of myocyte apoptosis are sufficient to cause a lethal, dilated cardiomyopathy in otherwise normal hearts [5], supporting the hypothesis that programmed cell death plays a major role in clinical heart failure [2,3]. ere-fore, a potential application of our construct would be the generation of clinically relevant, large animal models of dilated cardiomyopathy, which are currently unavailable and urgently required to study pathophysiological alterations and test new therapeutic interventions [26]. Alternatively, the dual-promoter system can be delivered to somatic cells during the adult life. In such a case, it might be necessary to produce an ad hoc vector to enhance the efficiency of the transduction of adult cardiomyocytes, which have proven almost completely refractory to conventional vector systems [27].

Conclusions
Our results offer a new and interesting tool that can be exploited to study apoptosis-related pathological conditions. Given the potential development of lentivirus transgenesis to develop novel and engineer existing animal models in the near future, the lentiviral construct described here may pave the way to produce clinically relevant models for further dilated cardiomyopathy research.

Con�ict of �nterests
e authors declare that they have no con�ict of interests.