Vitamin D(3)-up-regulated protein-1 is a stress-responsive gene that regulates cardiomyocyte viability through interaction with thioredoxin.

The protein-disulfide reductase thioredoxin is critical for redox signaling during apoptosis and growth. In this study, we demonstrate that vitamin D(3)-up-regulated protein-1 regulates thioredoxin in conditions of biomechanical or oxidative stress and critically regulates cardiomyocyte viability. Expression of vitamin D(3)-up-regulated protein-1 but not of thioredoxin in rat cardiomyocytes was rapidly suppressed by biomechanical strain or hydrogen peroxide at both mRNA and protein levels. Mechanical suppression of vitamin D(3)-up-regulated protein-1 gene expression was blocked by N-acetylcysteine. The half-life of vitamin D(3)-up-regulated protein-1 transcripts in cardiomyocytes was only 1.1 h and remained unchanged after mechanical stimulation, suggesting that rapid responses in vitamin D(3)-up-regulated protein-1 gene expression occur through transcriptional control. Vitamin D(3)-up-regulated protein-1 down-regulation by strain or hydrogen peroxide led to increased thioredoxin activity, whereas adenovirus-mediated overexpression of vitamin D(3)-up-regulated protein-1 suppressed thioredoxin activity. Overexpression of vitamin D(3)-up-regulated protein-1 but not of thioredoxin induced cardiomyocyte apoptosis. Furthermore, overexpression of vitamin D(3)-up-regulated protein-1 sensitized cells to hydrogen peroxide-induced apoptosis, whereas overexpression of thioredoxin protected against injury. These data identify vitamin D(3)-up-regulated protein-1 as a key stress-responsive inhibitory switch of thioredoxin activity in cardiomyocytes and demonstrate that the vitamin D(3)-up-regulated protein-1/thioredoxin axis has an important role in the preservation of cellular viability.

Biomechanical strain, hypoxia, and other types of stress induce hypertrophy, apoptosis, contractile failure, and other myocardial changes that directly or indirectly predispose to cardiac failure. The molecular pathways responsible for these changes are only partly understood, but oxidative reactions, either by injurious levels of reactive oxygen species (ROS) 1 generated during reperfusion and inflammation or by ROS as mediators of organized signal transduction, are important events in the initiation and progression of cardiac disease (1)(2)(3)(4). These pathways are at least partly counterbalanced by adaptive responses that, when impaired, enhance susceptibility to environmental stress and induce premature cardiomyocyte dysfunction and heart failure (5,6).
In the present study we hypothesized that thioredoxin (TRX), a major protein-disulfide reductase and determinant of cellular redox state (7), participates in cardiomyocyte responses to environmental stress. Thioredoxin has functions in defense against oxidative stress and control of growth and apoptosis (7,8). Recently, several investigators independently described binding of TRX to vitamin D 3 -up-regulated protein-1 (VDUP1) (9 -11), which acts as an endogenous inhibitor of TRX by interacting with the catalytic active center of TRX. In this study, we demonstrate that VDUP1 is a biomechanical and oxidative stress-responsive gene and that the VDUP1 gene product controls cardiomyocyte survival through regulation of TRX activity. We describe the following: (i) biomechanical and oxidative stress down-regulate VDUP1 mRNA and protein expression while increasing TRX activity, (ii) overexpression of VDUP1 inhibits TRX activity, and (iii) overexpression of VDUP1 induces apoptosis and sensitizes cells to oxidative stress-mediated apoptosis. Based on these data, we conclude that VDUP1 acts as a crucial, environmental stress-mediated regulator of cardiomyocyte viability through TRX.

EXPERIMENTAL PROCEDURES
Culture of Cardiac Myocytes-Neonatal rat ventricular myocytes (NRVM) from 1-day-old Harlan Sprague-Dawley rats (Charles River, Boston, MA) were isolated by previously described methods (12). Mechanical deformation was applied to a thin and transparent membrane on which cells were cultured, an approach that produces controlled biaxially uniform cellular strain as well as visualization of cells. After 24 h of plating, NRVM were washed twice with phosphate-buffered saline and made quiescent by incubating with Dulbecco's modified Eagle's medium containing 1% insulin-transferrin-sodium selenite supplement (Sigma) for approximately 48 h. To eliminate the variable of time-dependent changes due to cell age or effects of adhesion, in each experiment all cells were cultured for an identical time period, and cells and media from all samples were harvested at the same time. For example, in a time course experiment with strain, the time point represents the time prior to harvest that strain was initiated such that the strain sample and control sample were harvested at the same time.
Northern and Western Analyses-Northern analysis was performed and analyzed as described previously (12). For Northern blotting, 5 g of total RNA was loaded on a 1.0% agarose formaldehyde gel, trans-* This work was supported by Grants HL64858 (to R. T. L.) and HL67554 (to Y. W.) from the NHLBI, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  ferred to a nylon membrane (Stratagene), and UV cross-linked with a UV Stratalinker (Stratagene). The membrane was hybridized with a 600-base pair VDUP1 cDNA radiolabeled by the random priming method with [␣-32 P]dCTP and the Klenow fragment of DNA polymerase (Stratagene). For Western analyses, NRVM were subjected to 0 or 8% mechanical strain at 1 Hz for varied periods of times and lysed in radioimmune precipitation assay buffer. An equal amount of protein was separated on SDS-PAGE and electrotransferred to a polyvinylidene difluoride membrane. Western analysis was performed using polyclonal antisera against TRX (91-105) or VDUP1 (339 -358) .
Recombinant Adenoviral Construction-The recombinant adenovirus expressing green fluorescent protein alone (GFP-Ad) and recombinant adenoviruses for VDUP1 (VDUP-Ad) and TRX (TRX-Ad) were generated with the Ad-Easy system and the pAdTrack-CMV vector as described previously (13). Full-length VDUP1 was generated by PCR with the primer set containing the 5Ј-GAAGATCTCAATCATGGTGA-TGTTCAAG-3Ј and 5Ј-GCTCTAGAGCTTCACTGCACGTTGTTG-3Ј oligonucleotides, which have BglII and XbaI restriction sites, respectively. Full-length TRX was generated by PCR with the primer set containing the 5Ј-GAAGATCTCAACAGCCAAAATGGTGAAGCTG-3Ј and 5Ј-GCTCTAGAGGTTTTAAACAGCTG-3Ј oligonucleotides, which have BglII and XbaI restriction sites, respectively. cDNAs were cloned in the GFP-containing pAdTrack-CMV shuttle vector (kindly provided by Dr. Bert Vogelstein, Johns Hopkins University) and selected with kanamycin. After cloning, sequences were verified by DNA sequencing. Subsequently, clones were linearized, cotransformed with the adenoviral plasmid PadEasy-1 in electrocompetent BJ5183 cells, and selected with ampicillin. Recombinant plasmids were linearized and propagated in 293 cells. Stock titers were 10 9 plaque-forming units/ml for each vector.
TRX Activity Assay-The insulin disulfide reduction assay was performed to measure TRX activity as described (8). NRVM were infected with GFP-Ad, VDUP-Ad, or TRX-Ad at a multiplicity of infection (MOI) of 10 for 48 h. Total cellular proteins were extracted using a buffer containing 20 mmol/liter HEPES (pH 7.9), 300 mmol/liter NaCl, 100 mmol/liter KCl, 10 mmol/liter EDTA, 0.1% Nonidet P-40, plus protease inhibitors. Equal amounts of protein (90 g) were preincubated at 37°C for 15 min with 1 l of dithiothreitol activation buffer with 50 mmol/ liter HEPES (pH 7.6), 1 mmol/liter EDTA, 1 mg/ml bovine serum albumin, and 2 mmol/liter dithiothreitol in a total volume of 35 l to reduce TRX. Then 20 l of reaction mixture containing 200 l of 1 M HEPES (pH 7.6), 40 l of 0.2 mol/liter EDTA, 40 l of NADPH (40 mg/ml), and 500 l of insulin (10 mg/ml) were added. The reaction was started by addition of 5 l of bovine TRX reductase (American Diagnostica Inc., Greenwich, CT) and an equal volume of water to control samples. The samples were incubated at 37°C for 20 min. The reaction was stopped by the addition of 0.25 ml of 6 M guanidine hydrochloride and 0.4 mg/ml 5,5Ј-dithiobis(nitrobenzoic acid) in 0.2 M Tris-HCl, pH 8.0, and absorbance at 412 nm was measured.
Terminal Deoxynucleotidyltransferase-mediated dUTP Nick End-labeling (TUNEL) Assay-TUNEL assay was performed using the in situ cell death detection kit (Roche Molecular Biochemicals) according to the manufacturer's protocol. To quantify the number of apoptotic cells nuclei were counterstained with 4Ј,6-diamidino-2-phenylindole, and the total numbers of nuclei and TUNEL-positive nuclei were counted in 10 low power fields in three independent experiments. For each experimental condition, contiguous visual fields were counted to accumulate data on at least 1000 myocytes per condition per experiment.
FACS Analysis-NRVM were detached with trypsin, washed in phosphate-buffered saline (PBS), and fixed in 70% ethanol/PBS. Cells were suspended in PBS with propidium iodide (20 g/ml, Sigma) and RNase A (5 Kunitz units/ml, Sigma). Ten thousand cells from each sample were counted using a Becton Dickinson FACScan (excitation, 488; emission, 590). There was some variability in the absolute number of apoptotic myocytes under control and treated conditions, and for this reason each set of experiments was compared with its own internal positive and negative controls for statistical analysis.
Statistical Analysis-All data are presented as mean Ϯ S.E. Comparison of groups was done using analysis of variance followed by post-hoc analysis with the Tukey-Kramer test. Statistical significance was achieved when p Ͻ 0.05.

RESULTS
Biomechanical Strain or Hydrogen Peroxide Suppresses Expression of VDUP1-Exposure of NRVM to mechanical strain or hydrogen peroxide reduced the expression of VDUP1 at both mRNA and protein levels (Fig. 1, A and B). Cyclic mechanical strain (8%, 1 Hz) rapidly suppressed VDUP1 mRNA expression. Down-regulation of VDUP1 occurred as early as 30 min after the initiation of strain and was sustained for at least 24 h (mean down-regulation at 3 h of strain, 2.3 Ϯ 0.1-fold, p Ͻ 0.05, n ϭ 8). Biomechanical strain also reduced expression of VDUP1 protein (mean decrease 1.7 Ϯ 0.1-fold at 24 h, n ϭ 3, Fig. 1A, lower panel). Importantly, levels of TRX, the interacting partner of VDUP1, remained unaltered during the same duration of strain (Fig. 1A, lower panel).
To explore the biochemical mechanism for biomechanical down-regulation of VDUP1, we tested whether or not angiotensin II or endothelin-1 that were released upon the stretch of cardiomyocytes (14,15) participated. NRVM were pretreated with AT 1 antagonists losartan (1 mol/liter) or CP191,166 (0.1 mol/liter) 30 -60 min prior to mechanical strain and exposed to strain for 3 h. VDUP1 mRNA down-regulation, however, remained unaltered in the presence of AT 1 antagonism (data not shown). Exogenously added angiotensin II (100 nmol/liter, 3 h) did not affect VDUP1 mRNA in the absence of strain. Similarly, pretreatment with the endothelin-A receptor blocker BQ123 (1 mol/liter) or the nonspecific endothelin receptor blocker PD145065 (1 mol/liter) did not affect VDUP1 downregulation by strain, and exogenous endothelin-1 (100 nmol/ liter, 3 h) did not affect VDUP1 mRNA in the absence of strain (not shown). Furthermore, cyclosporin (1 mol/liter), a calcineurin inhibitor; SB203580 (10 mol/liter), a p38 mitogenactivated protein kinase inhibitor; PD98059 (20 mol/liter), a mitogen-activated protein kinase kinase inhibitor; or genistein (50 mol/liter), a tyrosine kinase inhibitor (Fig. 1C), did not inhibit the suppressive effect of mechanical strain on VDUP1 mRNA.
Several previous reports have demonstrated that mechanical strain induces the generation of reactive oxygen species in cardiomyocytes and that concomitant changes in cell redox state participate in downstream signaling events evoked by strain that can be blocked by antioxidants such as N-acetylcysteine (12, 16). Fig. 1C demonstrates that in NRVM pretreated with N-acetylcysteine (10 mmol/liter), down-regulation of VDUP1 mRNA by strain (2.5 h, 8%, 1 Hz) was completely inhibited.
To further characterize stress-induced VDUP1 mRNA downregulation in cardiomyocytes, we measured VDUP1 mRNA half-life. NRVM were exposed to 0 or 8% strain for 3 h and then incubated with actinomycin D (5 g/ml) to inhibit transcriptional activity. The half-life of VDUP1 mRNA was only 1.1 Ϯ 0.1 h and remained unaltered after strain (0.9 Ϯ 0.2 h). These experiments suggest that biomechanical strain regulates VDUP1 by suppressing synthesis of VDUP1 mRNA and that inhibitory transcriptional events induced by strain participate in the regulation of this redox-regulatory switch.
Effect of Biomechanical Strain and Hydrogen Peroxide on TRX Activity in Cardiomyocytes-We demonstrated that either biomechanical strain or oxidative stress suppressed expression of the endogenous TRX inhibitor, VDUP1. After exposures to biomechanical strain or hydrogen peroxide, TRX activity was consistently increased. Using insulin-reducing assays, mechanical strain (1 Hz, 8%, 24 h) increased TRX activity by 37 Ϯ 4% and hydrogen peroxide (50 M, 24 h) by 56 Ϯ 6% (Fig. 2, p Ͻ  0.05, n ϭ 3 independent experiments). Effects of VDUP1 on TRX-reducing Activity-To explore further the functional interaction between VDUP1 and thioredoxin, we prepared replication-defective adenovirus constructs expressing wild type VDUP1 or TRX. Transduction of cardiomyocytes with VDUP-Ad resulted in a 7-fold increase in VDUP1 protein (compared with control GFP-Ad-infected cells) but did not change TRX protein expression (Fig. 3A). On the other hand, transduction of cardiomyocytes with TRX-Ad resulted in a 2.5-fold increase in TRX protein expression (compared with control GFP-Ad-infected cells) but did not change VDUP1 expression (Fig. 3A).
To determine whether VDUP1 affects TRX activity, TRXspecific insulin-reducing assays were performed. As shown in Fig. 3B, overexpression of VDUP1 suppressed TRX activity by 43% (data are representative of three separate experiments, p Ͻ 0.01), whereas overexpression of TRX increased TRX activity by 144% (data are representative of three separate experiments, p Ͻ 0.01). These data confirmed that gene transfer of VDUP1 or TRX affects TRX activity as anticipated.
When cells were exposed to hydrogen peroxide for 24 h, the pro-apoptotic effects of VDUP1 were even more pronounced, and protective effects of TRX became evident. TUNEL assay showed that hydrogen peroxide dose-dependently increased the proportion of apoptotic cells transduced with the GFP-Ad from 11.5 Ϯ 1.4 to 26.7 Ϯ 1.3 and 47.8 Ϯ 1.6% and from 18.5 Ϯ 1.8 to 45.1 Ϯ 3.3 and 60.7 Ϯ 2.8% in cells transduced with VDUP-Ad (data are representative of three separate experiments that were performed in triplicate, p Ͻ 0.05), whereas overexpression of TRX inhibited hydrogen peroxide-induced cardiomyocyte apoptosis (Fig. 4B). Similar results were obtained when apoptosis was quantified with FACS analysis (Fig. 5). DISCUSSION In the present study, we focused on VDUP1, an intrinsic TRX inhibitory protein, as a stress-responsive gene in cardiomyocytes that regulates cell viability. We found that biomechanical strain and oxidative stress rapidly down-regulated VDUP1 expression at mRNA and protein levels. Given the known interactions between VDUP1 and TRX (9 -11), these observations argued for a role of VDUP1 as a stress-responsive regulatory switch of TRX in cardiomyocytes. Suppression of VDUP1 expression allowed strain and oxidative stress to induce TRX activity at unchanged levels of TRX protein. Importantly, overexpression of VDUP1 inhibited TRX activity and induced spontaneous cardiomyocyte apoptosis. Furthermore, in conditions of oxidative stress, forced expression of VDUP1 induced apoptosis. These findings demonstrate that VDUP1 is a pivotal gene in the control of a TRX-dependent survival pathway in cardiomyocytes.
Expression of VDUP1 was suppressed by hydrogen peroxide and its mechanical down-regulation inhibited by antioxidants. These observations suggest that VDUP1, in addition to regulating cellular redox state through TRX, acts as a sensor of reactive oxygen species and as a critical mediator between environmental stress and adaptive cellular responses. Consist-ent with previous reports (9 -11), overexpression of VDUP1 inhibited TRX activity. TRX is a major ubiquitous disulfide reductase that controls rapid thiol-disulfide exchange reactions of structural or catalytic SH groups of many proteins. TRX has many biological activities, including protection against oxidative stress and inhibition of apoptosis (7,8). In this study, inhibition of endogenous TRX by VDUP1 induced apoptosis even in the absence of imposed oxidative stress. TRX is thus more than a defense system in cardiomyocytes against exogenous oxidative stress and may be a critical regulator of cell viability in non-oxidative conditions. These conditions likely include biomechanical overload, as suggested by our observation that mechanical deformation induced VDUP1 down-regulation and activation of TRX. Intriguingly, mechanical deformation of cardiomyocytes has been reported previously to induce functionally relevant changes in the cardiac muscle redox state (12,16).
The mechanisms through which TRX acts as an anti-apoptotic regulator are still under investigation, but recent reports have shown that TRX inhibits ASK-1, a mammalian mitogenactivated protein kinase kinase kinase that delivers apoptotic signals by activating c-Jun NH 2 -terminal kinase and p38 pathways (11,17). VDUP1 competes with ASK-1 to bind to TRX (11). TRX may, however, promote survival by other mechanisms, including the modulation of nuclear factor B activity, a stress-related survival-related transcription factor in cardiomyocytes and other cell types (18,19).
These data suggest that the regulation of VDUP1 may participate in protection against apoptosis from mechanical overload or oxidative stress. It is important to consider this protective mechanism within the context of potential pro-apoptotic effects of mechanical strain and oxidative stress. We and others have found that very large deformations (25% or greater) can promote apoptosis (16,20). However, we have found this proapoptotic effect only in association with obvious partial cell detachment from the culture substrate; we have not observed an increase in cardiac myocyte apoptosis by any measure with the smaller deformations that regulate gene expression that were used in this study. In addition, our data suggest that overexpression of TRX and an increase in TRX activity can partially protect against oxidative stress, but this protection is not complete. Thus, we would anticipate that even with complete absence of VDUP1, extreme conditions of oxidative stress could exceed the beneficial effects of increased TRX activity.
This study argues for a role of the VDUP1/TRX pathway in myocardial remodeling, and previous reports further suggest a role of TRX in cardiac disease (21)(22)(23). Serum levels or local expression of TRX are enhanced in heart failure and myocarditis (22,23), and TRX prevents reperfusion-induced arrhythmias (21). The present study demonstrates that the earliest regulatory event in the recruitment of TRX may be stressinduced down-regulation of its inhibitory partner VDUP1 rather than increased expression of TRX. Delayed or deficient down-regulation of VDUP1 may result in incomplete adaptation in the early phases of cardiac disease and further alter disease progression.