Regulation of Bcl-xL Expression by H 2 O 2 in Cardiac Myocytes* □ S

Oxidative stress promotes cardiac myocyte apoptosis through the mitochondrial death pathway. Since Bcl-2 family proteins are key regulators of apoptosis, we ex- amined the effects of H 2 O 2 on the expression of principal Bcl-2 family proteins (Bcl-2, Bcl-xL, Bax, Bad) in neona- tal rat cardiac myocytes. Protein expression was as-sessed by immunoblotting. Bcl-2, Bax, and Bad were all down-regulated in myocytes exposed to 0.2 m M H 2 O 2 , a concentration that induces apoptosis. In contrast, al- though Bcl-xL levels initially declined, the protein was re-expressed from 4–6 h. Bcl-xL mRNA was up-regulated from 2 to 4 h inneonatal rat or mouse cardiac myocytes exposed to H 2 O 2 , consistent with the re-expression of protein. Four different untranslated first exons have been identified for the Bcl-x gene (exons 1, 1B, 1C, and 1D, where exon 1 is the most proximal and exon 1D the most distal to the coding region). All were detected in mouse or rat neonatal cardiac myocytes, but exon 1D was not expressed in adult mouse hearts. In neonatal mouse or rat cardiac myocytes, H 2 O 2 induced the expression of exons 1B, 1C, and 1D, but not exon 1. These data demonstrate that the Bcl-x gene is selectively responsive to oxidative stress, and the response is medi- ated through distal promoter regions. Cardiac myocytes, the contractile cells of the heart, are ter-minally differentiated cells,

Cardiac myocytes, the contractile cells of the heart, are terminally differentiated cells, which withdraw from the cell cycle in the perinatal period. Consequently, myocyte cell death, as occurs (for example) following ischemia and/or reperfusion, significantly affects cardiac function. One of the principal insults that may promote cardiac myocyte death is oxidative stress, and the levels of oxidative stress are increased in ischemic hearts (1, 2) with higher levels being produced following reperfusion (3). H 2 O 2 (0.1-0.5 mM), as an example of oxidative stress, promotes cardiac myocyte apoptosis through the mitochondrial death pathway (4). However, lower concentrations are non-toxic and may have cytoprotective and/or growth promoting effects (5)(6)(7)(8).
Bcl-2 family proteins are key regulatory components of the mitochondrial cell death pathway (9,10). Some family members are cytoprotective (e.g. Bcl-2, Bcl-xL, Bcl-w, Mcl-1), whereas others promote apoptosis (e.g. Bad, Bak, Bax, Bid, Bim, Bmf). These proteins act at the mitochondria to regulate cytochrome c release. Bcl-2 family proteins act either as heterodimers or as oligomers and the dynamic equilibrium between such complexes appears to determine the predisposition to apoptosis. Translocation of Bax to the mitochondria and oligomerization of Bax and/or Bak promotes permeabilization of the outer mitochondrial membrane and release of cytochrome c and other apoptosis-inducing factors, possibly by forming a pore in the membrane. Heterodimerization with either Bcl-2 or Bcl-xL prevents oligomerization and protects from apoptosis. Other pro-apoptotic proteins, such as Bad, Bim, or Bmf compete for binding to Bcl-2/Bcl-xL causing release of Bax/Bak, which can then form oligomers and induce apoptosis. Thus, the balance between pro-and anti-apoptotic Bcl-2 family proteins influences the rate of apoptosis.
Bcl-2 and Bcl-xL are the principal Bcl-2 family proteins that protect cells from apoptosis. The regulation of the Bcl-x gene appears particularly complex. Initial studies of the mouse Bcl-x gene demonstrated that, like Bcl-2, the Bcl-x gene consisted of three separate exons (11). Exon 1 is untranslated and is separated from the first coding exon, exon 2, by a short facultative intron. Exon 3 codes for the C terminus of Bcl-xL and is separated from exon 2 by a large intron of at least 9 kb. At least four proteins potentially derive from alternatively spliced Bcl-x mRNAs. Exons 1, 2, and 3 are spliced together to produce Bcl-xL. Splicing from an alternative donor site within exon 2 to exon 3 results in a shorter form of the protein (Bcl-xS), which lacks a central region present in Bcl-xL, but the reading frame is maintained and the C terminus is identical (11). In contrast to Bcl-xL, Bcl-xS induces apoptosis. Read-through of the donor site at the end of exon 2 produces Bcl-x␤ (12), and splicing of exon 2 to a novel exon 4 generates Bcl-x␥ (13). Apart from alternative splicing of the coding exons of the Bcl-x gene, there is additional complexity in the regulation of the 5Ј non-coding region. Grillot et al. (11) reported the presence of at least two principal transcription initiation sites for exon 1 at Ϫ655 and Ϫ727, in addition to a cluster of initiation sites at the start of exon 2. Subsequently, an alternative first exon upstream of exon 1 (exon 1B) was identified in human lymphoma cells (14). A third study suggested that at least five different promoters (P1-P5) operate in the mouse Bcl-x gene, each of which is associated with a different first exon (exons A-E) (15). This terminology is confusing, since exons A and B correspond, respectively, to exons 2 and 1 identified by Grillot et al. (11), and exon C corresponds to the region identified as exon 1B (14). The different exons of the Bcl-x gene are summarized in Fig. 1. Despite the identification of potential promoters and transcriptional start sites further upstream in the Bcl-x gene (i.e. exons D and E (15)), expression of these exons was not detected in erythroid progenitor cells or differentiating erythroblasts in which Bcl-xL is up-regulated (16), and their significance in a biological system remains to be established.
A number of Bcl-2 family proteins are expressed in cardiac myocytes including Bcl-2, Bcl-xL, Bad, Bax, and Bid (4,17,18). As in other cells, overexpression of Bcl-2 in cardiac myocytes is cytoprotective (19 -21). H 2 O 2 -induced apoptosis in neonatal myocytes is associated with rapid translocation of Bad and Bax to the mitochondria (Ͻ5 min), an event which precedes cytochrome c release (4,17), suggesting that, in the context of oxidative stress, Bad/Bax translocation may be one of the initiating events. Bcl-2 family proteins are therefore key regulators of apoptosis in cardiac myocytes, as in other cells, and the expression of the various family members is likely to influence the rate of myocyte death. In this study, we examined the effects of H 2 O 2 on the expression of four key family members which are readily detected in cardiac myocytes (Bcl-2, Bcl-xL, Bax, and Bad).

Preparation of Neonatal Rat or Mouse Cardiac Myocytes-Myocytes
were dissociated from the ventricles of 1-3-day Sprague-Dawley rat hearts as described previously (22) and plated at a density of 1.4 ϫ 10 3 cells/mm 2 in 35-mm Primaria tissue culture dishes precoated with 1% (w/v) gelatin. Neonatal mouse myocytes were prepared from the ventricles of C57Bl6 mice using essentially the same procedure as for neonatal rat myocytes, and cells were plated at a density of 2.0 ϫ 10 3 cells/mm 2 . Myocytes were plated in Dulbecco's modified Eagle's medium/M199 (4:1) containing 5% (v/v) fetal calf serum and 10% (v/v) horse serum (16 h, 37°C), and then incubated in serum-free medium for 24 h prior to experimentation. Myocytes were exposed to H 2 O 2 (0.02 mM, 0.2 mM, 2 mM) for the times indicated.
Total RNA Isolation and cDNA Synthesis-Myocytes were homoge-nized in 0.5 ml of RNAzol B (AMS Biotech) and total RNA prepared according to the manufacturer's instructions. RNA pellets were dissolved in diethyl pyrocarbonate-treated water, and the concentrations were calculated from the A 260 . First strand cDNA synthesis was performed from total RNA (2 g) in a reaction mixture containing 500 oligo(dT) and 1 mM dNTP mix. Tubes were heated (65°C, 5 min). 5 ϫ first-strand buffer, 0.1 M dithiothreitol, recombinant ribonuclease inhibitor (40 units) and Superscript II RNase H reverse transcriptase (200 U) (Invitrogen) were added, and cDNA synthesis was performed (50 min, 42°C).
Preparation of Genomic DNA-Rat spleen (stored at Ϫ80°C) was powdered under liquid N 2 , homogenized in Buffer ATL (Qiagen DNeasy tissue DNA extraction kit), and passed through a 21-gauge needle 5-20 times. DNA was extracted according to the manufacturer's instructions, eluted in 100 l, and the concentration was calculated from the A 260 .
Ratiometric RT 1 -PCR Analysis-Ratiometric RT-PCR was performed as previously described, and in our hands, results from this method are comparable with those obtained by quantitative "real-time" PCR (24). Primers were designed for the coding regions of rat Bcl-2, Bad, and Bax and for mouse or rat Bcl-xL coding region. Primers for the Bcl-xL coding region (exon 2) were designed not to detect Bcl-xS. Primers were also designed for each of the potential first exons upstream of exon 2 of the Bcl-x gene and to amplify the region across exons 1 and 2 (Table I and Fig. 1). To avoid confusion, we have used the Grillot/MacCarthy-Morrogh (11,14) terminology referring to exons D and E of the Pecci et al. study (15) as exons 1C and 1D (Fig. 1). RT-PCR reactions were carried out in a 50-l volume containing 100 ng of cDNA or 20 ng of genomic DNA template, 100 M concentration of each primer, 50 mM KCl, 20 mM Tris-HCl (pH 8.4 at 25°C), 1.5 mM MgCl 2 , 0.01% (v/v) Tween 20, and a 0.2 mM concentration each of dATP, dCTP, dGTP, dTTP, using 1 unit of Taq polymerase (Invitrogen). The following conditions were used: 95°C, 3 min followed by 24 -31 cycles of denaturation (95°C, 50 s), annealing (61°C, 50 s), and extension (72°C, 50 s). Amplification products were analyzed by ethidium bromide-agarose gel electrophoresis using 2% (w/v) agarose gels and the bands captured under UV illumination. All primer sets generated clean products of the predicted sizes (Table I). Products were analyzed by scanning densitometry and normalized to GAPDH. Results are expressed relative to unstimulated controls.

H 2 O 2 Promotes Down-regulation of Bcl-2 Family Proteins in
Cardiac Myocytes-We examined the effects of different concentrations of H 2 O 2 (2, 0.2, and 0.02 mM) on the expression of four principal Bcl-2 family proteins, which are expressed in cardiac myocytes: Bcl-2, Bcl-xL, Bax, and Bad. 0.2 mM H 2 O 2 results in myocyte death associated with features of apoptosis (4), whereas higher concentrations induce non-apoptotic cell death. 0.02 mM H 2 O 2 is non-toxic, but has effects on gene expression. 2 Neonatal rat cardiac myocytes were exposed to H 2 O 2 for up to 24 h, and the levels of Bcl-2, Bcl-xL, Bax, and FIG. 1. Exon structure of the mouse Bcl-x gene. The exon numbering system has been adapted from the schemes by Grillot et al. (11) and MacCarthy-Morrogh et al. (14), with the system (exons A-E) used by Pecci et al. (15) shown below.
Bcl-x primer pairs that were used are also shown (for further details see Table I). bp numbering extends from the first coding ATG. Exons and introns are not shown to scale.
Bad were compared by immunoblotting. High concentrations of H 2 O 2 (2 mM) induced profound down-regulation of all four proteins (Fig. 2, A-D). Down-regulation of pro-apoptotic Bax and Bad was apparent within 1-2 h (Fig. 2, A and B), whereas down-regulation of anti-apoptotic Bcl-2 and Bcl-xL was not significant before ϳ4 h (Fig. 2, C and D). The levels of Bax and Bad were also decreased in myocytes exposed to 0.2 mM H 2 O 2 (Fig. 2, E and F), albeit to a lesser extent than in myocytes exposed to 2 mM H 2 O 2 (Fig. 2, A and B). Bcl-2 was downregulated to a similar degree and over a similar time course in response to both concentrations (Fig. 2, C and G). In contrast, although Bcl-xL protein initially decreased in myocytes exposed to 0.2 mM H 2 O 2 (ϳ70% over 2-4 h), it was re-expressed, and after 24 h the levels of Bcl-xL were similar to those of unstimulated control cells (Fig. 2H). The expression of all four Bcl-2 family proteins did not change in myocytes exposed to 0.02 mM H 2 O 2 (results not shown). These data indicate that, of the four Bcl-2 family proteins, only Bcl-xL is re-expressed following stimulation by oxidative stress at a concentration that promotes apoptosis.
H 2 O 2 Induces Bcl-xL mRNA Expression-To determine whether re-expression of Bcl-xL protein in response to 0.2 mM H 2 O 2 reflected an increase in mRNA expression (rather than an effect on protein synthesis or stability) and that this was a selective effect, neonatal rat cardiac myocytes were exposed to 0.2 mM H 2 O 2 , and ratiometric RT-PCR analysis of Bcl-xL, Bcl-2, Bax, and Bad was performed. Primers were designed to detect expression of the coding regions of the mRNAs for each gene (Table I). Consistent with the protein data (Fig. 2H), Bcl-xL (exon 2) mRNA was up-regulated from 2 to 4 h, and this was sustained over the 8-h period studied (Fig. 3A). In contrast, there was no change in expression of Bcl-2 or Bax mRNA (Fig. 3, B and C), and Bad was marginally down-regulated at the mRNA level (Fig. 3D). Since more detailed analysis of the regulation of Bcl-x has been performed for the mouse gene than for the rat gene (15), we confirmed that Bcl-xL mRNA was up-regulated in neonatal mouse cardiac myocytes in response to 0.2 mM H 2 O 2 . As in rat myocytes, Bcl-xL mRNA expression was increased at 2 and 4 h (Fig. 3F). Thus, 0.2 mM H 2 O 2 increases the expression of Bcl-xL mRNA in cardiac myocytes, and this probably accounts for the selective re-expression of Bcl-xL protein.
Up-regulation of Bcl-xL mRNA by H 2 O 2 Is Mediated by Selective Promoter Activation-As illustrated in Fig. 1, four putative untranslated first exons have been identified for the mouse Bcl-x gene (15). To avoid confusion we have extended the terminology used by MacCarthy-Morrogh et al. (14) and refer to exons 1 (the original first exon identified by Grillot et al. (11)), 1B, 1C, and 1D (which correlate to exons C, D, and E described by Pecci et al. (15)). Transcripts containing exons 1 and 1B had been previously detected in heart extracts, but whether exons 1C and 1D are also expressed had not been studied. Using specific primers to each of the untranslated exons of the mouse Bcl-x gene for ratiometric RT-PCR analysis, we confirmed that exons 1 and 1B were expressed in adult mouse heart and all other tissues studied (spleen, kidney, liver, and brain) (Fig.  4A). Exon 1C, but not exon 1D, was also detected in all adult mouse tissues ( Fig. 4A and results not shown). Bcl-x exons 1, 1B, and 1C were all readily detected in neonatal mouse cardiac myocyte cultures confirming expression in the myocytes themselves. However, although it was not detected in adult hearts, exon 1D was expressed in neonatal cardiac myocytes (Fig. 4B), suggesting that expression of this region is down-regulated during postnatal development of the heart. Since all primers were designed to amplify regions within individual predicted exons, we considered it necessary to confirm that there was no contamination of samples with genomic DNA and primers were designed for RT-PCR to amplify the region across the intron between exons 1 and 2. RT-PCR analysis of cDNA samples resulted in a single product of 433 bp, consistent with splicing of exon 1 to exon 2 generating a single mRNA species (Fig. 4C). This contrasts with a previous study in which differential splicing across this region produced three different mRNAs (15). Amplification of genomic DNA resulted in a product of 615 bp (Fig.  4C) encompassing the intron and confirming that there was no significant contamination of samples with genomic DNA.
To determine which of the untranslated exons are expressed in response to H 2 O 2 , neonatal mouse cardiac myocytes were exposed to 0.2 mM H 2 O 2 , and the different exons were analyzed by ratiometric RT-PCR. The expression of exon 1 did not change in response H 2 O 2 , whereas exons 1B, 1C, and 1D were all signif- icantly increased (Fig. 5A). The sequence for the region 5Ј to the rat Bcl-x gene has recently become available and comparison of the region encompassing all potential untranslated first exons with the mouse sequence indicates a high degree of conservation from some distance upstream of exon 1D through to the first coding ATG of the Bcl-x gene (see on-line Supplemental Material). Using primers specific for the rat sequence (Table I), we demonstrated that, consistent with the data for mouse myocytes (Fig. 5A), 0.2 mM H 2 O 2 selectively increased the expression of exons 1B, 1C, and 1D in neonatal rat cardiac myocytes with no change in expression of exon 1 (Fig. 5, B and C). These data suggest that oxidative stress results in specific expression of the distal untranslated exons of the Bcl-x gene in cardiac myocytes rather than the most proximal exon 1.

DISCUSSION
Down-regulation of Bcl-2 Family Proteins during Cardiac Myocyte Death-H 2 O 2 , as an example of oxidative stress, induces cardiac myocyte death. In response to 2 mM H 2 O 2 , which should induce a more necrotic phenotype, all four Bcl-2 family proteins studied were profoundly down-regulated (Fig. 2, A-D). In contrast, in response to 0.2 mM H 2 O 2 , which induces apoptosis, although the down-regulation of Bcl-2 was similar to that seen in response to 2 mM H 2 O 2 , the effect on Bax and Bad was less pronounced (Fig. 2, E-G). Furthermore, although Bcl-xL decreased initially, the protein was re-expressed (Fig.  2H). The effects of the two concentrations on Bcl-2 family proteins are therefore distinct and may be a reflection of the type of cell death. The mechanisms involved in the downregulation of these Bcl-2 family proteins were not investigated in this study. However, the down-regulation of Bcl-2, Bax, and Bad in response to 0.2 mM H 2 O 2 is likely to be primarily due to increased protein degradation and/or reduced protein synthesis, since the mRNA levels for Bcl-2 and Bax did not change and, although there was a decrease in Bad mRNA, this was relatively modest (Fig. 3, B-D). The initial decline in Bcl-xL protein also probably reflects increased degradation, since Bcl-xL mRNA expression did not change significantly before ϳ2 h (Fig. 3A). The subsequent increase in Bcl-xL mRNA presumably accounts for the re-expression of Bcl-xL protein from ϳ 4 -6 h (Fig. 2H). There was some variability in the time at which Bcl-xL protein was re-expressed, which may be due to a conflict between the rate of protein degradation and synthesis of new protein. The significance of Bcl-xL re-expression in FIG. 4. Expression of Bcl-x non-coding exons 1, 1B, 1C, and 1D in mouse tissues, and cardiac myocytes. RNA was extracted from adult mouse heart, spleen, kidney, liver, and brain (A and B) or from mouse neonatal cardiac myocytes (B and C), or genomic DNA was prepared (A and C). The expression of Bcl-x exons 1, 1B, 1C, and 1D, and the region encompassing the intron between exon 1 and exon 2 was examined by ratiometric RT-PCR analysis using the primers detailed in Table I. neonatal cardiac myocytes were exposed to 0.2 mM H 2 O 2 for the times indicated. RNA was extracted, and the expression of Bcl-x exons 1, 1B, 1C, and 1D was examined by ratiometric RT-PCR analysis using the primers detailed in Table I Regulation of Bcl-xL mRNA Expression in Cardiac Myocytes-Although Bcl-x transcripts have been reported to derive from at least five different promoters (i.e. the region between exons 1 and 2 and regions upstream of exons 1, 1B, 1C, and 1D), only the three most proximal (regions upstream of exons 2, 1, and 1B) were previously shown to be operative in the heart (15). In this study, we demonstrated that exon 1C was expressed in adult mouse heart and neonatal cardiac myocytes, and although it was not detected in adult hearts, exon 1D was expressed in neonatal cardiac myocytes (Fig. 4A). These data suggest that there is developmental regulation of the upstream promoters such that exon 1D is no longer expressed in adult hearts. Our preliminary data suggest that, in neonatal cardiac myocytes, exons 1D and 1C are expressed as a single exon that incorporates the intervening sequence (results not shown). This suggests that during development there is a switch in the transcriptional initiation site that is used. Exons 1C/1D do not appear to be continuous with exon 1B (results not shown). Consistent with this, comparison of the mouse and rat sequences indicates that there is a region of low homology between exon 1C and exon 1B (see on-line Supplemental Material).
Although all four untranslated exons were detected in neonatal cardiac myocytes, only the expression of exons 1B, 1C, and 1D was significantly increased by 0.2 mM H 2 O 2 , with no increase in exon 1 (Fig. 5). The reasons for multiple promoter usage and differential expression of untranslated exons are unclear, although previous studies suggest that the promoter may influence the isoform of Bcl-x which is produced (15). In cardiac myocytes, Bcl-xL (Fig. 2) and Bcl-xS (results not shown) were up-regulated concomitantly, and it is possible that, for example, exon 1B may preferentially induce Bcl-xL whereas exons 1C/1D may induce Bcl-xS. However, there is minimal evidence in support of such a scenario. Alternatively, switching of non-coding exons to produce different 5Ј-untranslated regions may influence the mode and rate of translation. In cardiac myocytes, H 2 O 2 (Ͼ0.1 mM) suppresses global protein synthesis by ϳ95% over at least 4 h, probably because of repression of cap-dependent initiation of translation (25). However, a small group of proteins continues to be synthesized, including p21 CIP1/WAF1 (26), presumably through cap-independent mechanisms. An increasing number of eukaryotic mRNAs have been identified which contain internal ribosome entry sites required for cap-independent protein synthesis (see ifr31w3.toulouse. inserm.fr/IRESdatabase), and many of these are associated with apoptosis (27). In the case of insulin-like growth factor 2, mRNAs are produced with three different 5Ј-untranslated regions, only one of which confers cap-independent protein synthesis (28). The specific up-regulation of exons 1B, 1C, and 1D, but not exon 1, by H 2 O 2 in cardiac myocytes may therefore be required for translation to occur in the context of global inhibition of protein synthesis. Most studies so far have focused on the regulation of expression of Bcl-xL from the promoter regions upstream of exons 1 and 2, and binding sites for a number of transcription factors have been identified in this region (11). It is now clear that further analysis of sequences upstream from exons 1B, 1C, and 1D is required to identify the regions that are responsive to H 2 O 2 and establish which transcription factors are involved.
It is of note that the homology between rat, mouse and human Bcl-x gene is extremely high across the region encompassing exons 1B through to exon 3. However, although the homology between the rat and mouse sequences across exons 1C and 1D is high (see on-line Supplemental Material), we have had problems aligning this region with the human sequence (42% identity at best). It is possible exons 1C and 1D represent a different gene from Bcl-x, but the homology between rat and mouse extends almost continuously from some distance 5Ј to exon 1D through the first coding ATG of Bcl-x, and the regulation of exons 1C and 1D in response to H 2 O 2 is similar to that of exons 1B and Bcl-xL exon 2. Furthermore, the homology between the rat and mouse sequences is lost abruptly some distance 5Ј to exon 1D (see on-line Supplemental Material), suggesting that this may be the extreme 5Ј end of the gene. The reasons for the lack of homology with the human sequence are not clear. If exons 1C and 1D are part of another gene, it is possible that this gene is elsewhere on the human genome, but we have been unable to find any human sequences with significant homology to the rat/mouse sequences. It seems more probable that evolutionary constraints on the primary sequence were not high. Motifs for internal ribosome entry sites are not conserved at the primary sequence level, and it is the secondary structure that appears to be of prime importance (29). If the upstream region of the Bcl-x gene produces a mRNA capable of promoting cap-independent translation, the constraints would therefore be expected to occur at the level of secondary structure rather than primary sequence. Further analysis is required to determine whether the upstream region is relevant to expression of human Bcl-x gene products and if this region does indeed promote translation through cap-independent mechanisms.