A Novel Hypoxia-Inducible Spliced Variant of Mitochondrial Death Gene Bnip3 Promotes Survival of Ventricular MyocytesNovelty and Significance
Rationale: Alternative splicing provides a versatile mechanism by which cells generate proteins with different or even antagonistic properties. Previously, we established hypoxia-inducible death factor Bnip3 as a critical component of the intrinsic death pathway.
Objective: To investigate alternative splicing of Bnip3 pre-mRNA in postnatal ventricular myocytes during hypoxia.
Methods and Results: We identify a novel previously unrecognized spliced variant of Bnip3 (Bnip3Δex3) generated by alternative splicing of exon3 exclusively in cardiac myocytes subjected to hypoxia. Sequencing of Bnip3Δex3 revealed a frame shift mutation that terminated transcription up-stream of exon5 and exon6 ablating translation of the BH3-like domain and critical carboxyl-terminal transmembrane domain crucial for mitochondrial localization and cell death. Notably, although the 26-kDa Bnip3 protein (Bnip3FL) encoded by full-length mRNA was localized to mitochondria and provoked cell death, the 8.2-kDa Bnip3Δex3 protein encoded by the truncated spliced mRNA was defective for mitochondrial targeting but interacted with Bnip3FL resulting in less association of Bnip3FL with mitochondria and diminished apoptotic and necrotic cell death. Forced expression of Bnip3FL in cardiac myocytes or Bnip3−/− mouse embryonic fibroblasts triggered widespread cell death that was inhibited by coexpression of Bnip3Δex3. Conversely, RNA interference targeted against sequences encompassing the unique exon2-exon4 junction of the Bnip3Δex3 sensitized cardiac myocytes to mitochondrial perturbations and cell death induced by Bnip3FL.
Conclusions: Given the otherwise lethal consequences of deregulated Bnip3FL expression in postmitotic cells, our findings reveal a novel intrinsic defense mechanism that opposes the mitochondrial defects and cell death of ventricular myocytes that is obligatorily linked and mutually dependent on alternative splicing of Bnip3FL during hypoxia or ischemic stress.
Loss of cardiac cells by programmed cell death has been posited as a central underlying cause of ventricular remodeling and the decline in cardiac performance following myocardial ischemic injury.1 For this reason, there has been considerable effort over the last decade in deciphering the signaling pathways and cellular targets that govern cardiac cell death under normal and pathological conditions. Although the molecular signaling events remain poorly defined, there is considerable evidence that the mitochondrion is a conduit for integrating signals for apoptosis, necrosis, and autophagy.2 Several lines of investigation have implicated certain members of the Bcl-2 gene family as critical regulators of the permeability transition pore formation, cytochrome c release, and cell death.3,–,7 Notably, the proteins Bnip3 (Bcl-2/adenovirus E1B 19-kDa interacting protein 3) and Nix/Bnip3L (Bcl-2/adenovirus E1B interacting protein 3-like) are a subclass of evolutionary conserved BH3 domain–like members of the Bcl-2 gene family that provoke mitochondrial perturbations and cell death in response to distinct biological stresses.8,–,11 Previously, we established that the carboxyl-terminal transmembrane domain of Bnip3 and Nix are crucial for insertion into mitochondrial membranes and cell death.8 Despite the ability of Bnip3 and Nix proteins to trigger mitochondrial perturbations, they are transcriptionally activated under different physiological conditions. Indeed, we previously established that Bnip3 is induced in postnatal ventricular myocytes during hypoxia,8,12 whereas Nix is selectively activated by Gq signaling in response to pathological hypertrophy.9,13 Despite their overlapping ability to disrupt mitochondrial function in different cardiac pathologies, the biological significance of Bnip3 and Nix proteins in regulating cell death more generally is undetermined. Given our interest in hypoxic and ischemic cell injury, we focused our attention on inducible Bnip3 transcription.
Previously, we established that Bnip3 promoter activity is strongly repressed under basal normoxic conditions but is readily induced during hypoxia. We attributed the induction of Bnip3 gene transcription and cell death during hypoxia to the displacement of inhibitory NF-κB-HDAC complexes, which relieves the steric hindrance on the Bnip3 promoter.14,15 Nevertheless, despite the tight regulation of Bnip3 transcription, certain cancer cells are resistant to hypoxic injury.16,–,21 Although the underlying mechanisms for this apparent resistance to hypoxic stress are unknown, it likely reflects an adaptive survival mechanism to oppose the otherwise lethal actions of Bnip3 on apoptosis. We therefore speculated that a putative endogenous inhibitor may oppose Bnip3. In this regard, alternative gene splicing provides a versatile mechanism by which cells generate proteins with different or even antagonistic properties. Notably, spliced variants of the Bcl-2 gene family members exist, although none to date are known to be transcriptionally regulated or alternatively spliced during hypoxia or ischemic stress.
In this report, we identify a novel, previously unrecognized Bnip3 isoform that is solely generated during hypoxia. We specifically show that the alternatively spliced Bnip3 isoform, Bnip3Δex3, encodes a novel truncated protein that promotes survival by antagonizing the mitochondrial defects induced by Bnip3. We further show that inhibition of Bnip3Δex3 predisposes cardiac myocytes to cell death. Our findings reveal a novel survival mechanism to curtail mitochondrial injury and cell death that is obligatorily linked and mutually dependent on the alternative splicing of Bnip3. Thus, therapeutic interventions designed to selectively promote Bnip3Δex3 activity in the heart may prove beneficial in suppressing apoptotic and necrotic cell death during hypoxia.
Postnatal ventricular myocytes from 1 to 2 day old Sprague–Dawley rats were subjected to hypoxia for 24 hours in an air-tight chamber under serum-free culture conditions continually gassed with 95% N2, 5% CO2, and po2 ≤5 mm Hg, as previously reported.8,15,22 Bnip3−/− MEFS were cultured as previously reported.13
Quantitative Real-Time RT-PCR and Radioactive Semiquantitative RT-PCR
Total RNA (1 μg) from postnatal cardiac myocytes or adult heart tissue was reverse-transcribed with oligo dT20 (Invitrogen), and one-twentieth of the reaction was then amplified with gene specific primers for Bnip3, Bnip3Δex3, or house-keeping control gene L32, respectively. Real-time RT-PCR was performed using iQ5 multicolor real-time PCR detection system (Bio-Rad). For radioactive RT-PCR, 1 μg of the total RNA from postnatal cardiac myocytes was reverse-transcribed with oligo dT,18 and one-tenth of the reaction was then amplified in 25 cycles of PCR with Bnip3 primers, and the reverse primer was 32P-labeled. The PCR products were resolved on 8% polyacrylamide/8 mol/L urea denaturing gels. The gel was dried, exposed, and scanned in a PhosphorImager (Fuji Medical Systems, Roselle, IL).
Cloning and Constructs
Bnip3 gene was amplified from rat genomic DNA, whereas Bnip3Δex3 PCR product was purified from radioactive gel. Both PCR products were cloned into pcDNA4/HisMax TOPO TA expression vector or pcDNA6.2/EmGFP vector (Invitrogen) to generate expression plasmids encoding either Bnip3 or Bnip3Δex3 His-tag or GFP-fusion constructs. Short hairpin interference (sh)RNA-Bnip3FL was designed to directly target Bnip3 exon3 to knock down full-length Bnip3. Bnip3Δex3 small interfering RNA (siRNA-Bnip3Δex3) was designed to target Bnip3 exon2-exon4 junction to selectively knock down Bnip3Δex3. All the constructs were confirmed by DNA sequencing.
Hypoxia-Induced Alternatively Spliced Variant Bnip3ΔEx3 in Ventricular Myocytes
We previously established hypoxia-induced activation of the Bnip3 promoter, resulting in a Bnip3FL transcript of 564 base pairs. Sequence analysis of the PCR product verified the Bnip3FL mRNA was comprised of exons 1 through 6 and encoded for a Bnip3 protein of approximately 26 kDa. Interestingly, during the course of our studies we discovered a faster migrating mRNA transcript of 482 base pairs by radioactive PCR that was only detectable in ventricular myocytes subjected to hypoxia. Notably, baseline mRNA levels of this smaller mRNA were virtually undetectable relative to Bnip3FL in normoxic cells. Sequence analysis revealed the smaller mRNA band in hypoxia cells was an isoform of Bnip3FL that was missing 82 nucleotides encompassing exon3 (Figure 1A and 1B). Further inspection of the spliced isoform revealed that the exclusion of exon3, and subsequent fusion of exon2 and exon4, introduced a premature stop codon terminating transcription up-stream of exon5 and exon6. Hence, the truncated Bnip3 protein encoded by the alternatively spliced Bnip3 mRNA that we designate Bnip3Δex3 has a predicted mass of 8.2 kDa and retains the N terminus of full-length Bnip3, but lacks the putative BH3 and carboxyl-terminal transmembrane domain crucial for integrating into mitochondrial membranes (Figure 1A through 1C).
As a first step toward determining the physiological significance of Bnip3Δex3 in ventricular myocytes, we assessed the relative temporal expression of Bnip3Δex3 isoform in ventricular myocytes under physiological conditions. For these studies, we designed quantitative PCR primers that specifically amplified the Bnip3Δex3 isoform in cardiac cells under normoxic and hypoxic conditions. As shown in Figure 1D and 1E, our findings indicate that full-length Bnip3 and the Bnip3 splice variant are both induced in a manner dependent on the degree and severity of hypoxia. We determined that moderate to severe hypoxia of 0.1% to 5% not only activates Bnip3 gene transcription, but promotes Bnip3Δex3 splicing. This finding is in full agreement with our previous hypoxia studies demonstrating that a minimum threshold of oxygen deprivation is required to activate the Bnip3 promoter in ventricular myocytes.8
To verify that this alternative splicing was not restricted to neonatal myocytes, we next tested whether Bnip3 alternative splicing occurs in adult myocytes under relevant physiological conditions in vivo. For these studies, we assessed the presence of Bnip3FL and Bnip3Δex3 isoforms in adult myocardium following myocardial infarction in vivo. As shown in Figure 1F, in contrast to sham operated control hearts, a marked increase in Bnip3FL and Bnip3Δex3 isoforms were detected in myocytes following myocardial infarction. These findings corroborate our in vitro data and verify that alternative splicing of Bnip3FL occurs in adult ventricular myocytes under relevant ischemic stress conditions imposed by myocardial infarction. Because we previously established that the induction of Bnip3FL under hypoxia is attributable to the transcriptional derepression of the Bnip3 promoter by the displacement of NF-κB inhibitory complexes,14 we were compelled to test whether the Bnip3Δex3 splice variant is generated by a hypoxia-regulated mechanism or simply related to increased cellular processing of Bnip3 premRNA transcript levels. Therefore, we next assessed whether Bnip3 pre-mRNA would undergo alternative splicing under basal conditions in the absence of hypoxic signal. To test this possibility, we rendered the ventricular myocytes defective for NF-κB signaling with a nonphosphorylatable mutant of IκBα that we had previously shown to derepress the Bnip3 promoter and increase Bnip3FL gene transcription.14 As shown by real time PCR in Figure 1G, endogenous full-length Bnip3 mRNA was significantly increased in ventricular myocytes defective for NF-κB signaling, a finding consistent with our earlier work for the regulation of Bnip3 transcription by NF-κB. In contrast, we did not detect the alternative Bnip3Δex3 isoform in these cells, despite the increase in full-length Bnip3 transcript levels. These findings confirm that alternative splicing of Bnip3 is a hypoxia-regulated process.
Bnip3ΔEx3 Suppresses ROS, Mitochondrial ΔΨm, and Cell Death in Ventricular Myocytes
To assess the physiological significance of the Bnip3Δex3 in cardiac myocytes, we designed eukaryotic expression vectors encoding GFP-fusion proteins of Bnip3FL and Bnip3Δex3. As shown by epifluorescence microscopy in Figure 2A, expression of Bnip3FL in cells was punctuate and colocalized with MitoTracker Red to mitochondrial membranes, consistent with our earlier work demonstrating that Bnip3FL associates with mitochondria. However, in contrast to Bnip3FL, expression of Bnip3Δex3 in cells was not localized to mitochondria. Cell fractionation experiments confirmed that Bnip3FL, but not Bnip3Δex3, was associated with mitochondria (Figure 2E), verifying our epifluorescence data for the localization of Bnip3FL to mitochondria. Given early work establishing that localization of Bnip3FL to mitochondrial membranes is crucial for provoking permeability transition pore opening and cell death,8 we reasoned that an absence of Bnip3Δex3 localized at mitochondria may indicate its use as an intrinsic dominant-negative inhibitor of Bnip3FL to curtail mitochondrial injury and cell death during hypoxia.
Therefore, to test this possibility we assessed whether Bnip3Δex3 influences mitochondrial ΔΨm changes induced by Bnip3FL. As shown in Figure 2B, in contrast to control cells or cells expressing Bnip3Δex3, a marked reduction in mitochondrial TMRM staining was only observed in cells expressing Bnip3FL, a finding concordant with our epi-fluorescence data for localization of Bnip3FL to mitochondria. Moreover, a significant increase in reactive oxygen species (ROS) was observed in the presence of Bnip3FL, but not the Bnip3Δex3 (Figure 2C; Figure III in the Online Data Supplement, available at http://circres.ahajournals.org). Importantly, the loss of mitochondrial ΔΨm and ROS production induced by Bnip3FL in ventricular myocytes was suppressed by forced expression of Bnip3Δex3 (Figure 2B and 2C). Because earlier work by our laboratory established mitochondrial membrane integration of Bnip3FL as critical for disrupting mitochondrial function, we reasoned that Bnip3Δex3 may interact with Bnip3FL, preventing its ability to integrate into mitochondrial membranes. As shown by Western blot analysis in Figure 2D, Bnip3Δex3 immunoprecipitated with Bnip3FL, indicating that Bnip3Δex3 forms protein interactions with Bnip3FL. Importantly, cell fractionation studies revealed that less Bnip3FL was associated with mitochondrial membranes in the presence of Bnip3Δex3, supporting the notion that Bnip3Δex3 suppresses mitochondrial perturbations by disrupting integration of Bnip3FL into mitochondrial membranes.
Based on these findings, we tested the role of Bnip3Δex3 on cell survival. For these studies, ventricular myocytes were infected with adenoviruses encoding either Bnip3FL or Bnip3Δex3 and stained with vital dyes calcein-AM and ethidium homodimer-1 to mark the number of living (green) and dead (red) cells, respectively. As shown in Figure 2F, a significant increase in myocyte death was observed in the cells expressing Bnip3FL compared to vector control cells. Importantly, myocytes expressing Bnip3Δex3 were indistinguishable from vector control cells, with respect to cell viability, indicating that unlike Bnip3FL, the alternative spliced variant Bnip3Δex3 was not cytotoxic and did not provoke cell death. Furthermore, coexpression of Bnip3Δex3 rescued cell death induced by Bnip3FL in a dose dependent manner, to comparable levels of vector control cells (Figure 2G). These findings support our notion that Bnip3Δex3 promotes survival by acting as an endogenous inhibitor of Bnip3FL.
Bnip3ΔEx3 Abrogates Hypoxia-Induced Loss of Mitochondrial ΔΨm and Cell Death
Collectively, our data strongly suggest that Bnip3Δex3 promotes cell survival in ventricular myocytes by opposing the actions of Bnip3FL under physiological conditions. To test this notion we used three independent, but complementary strategies, to assess the impact of Bnip3Δex3 on mitochondrial membrane potential (ΔΨm) and cell viability during hypoxia. First, we tested whether forced expression of Bnip3Δex3 would be sufficient to suppress hypoxia-induced loss of mitochondrial ΔΨm in ventricular myocytes. As shown in Figure 3E, in contrast to normoxic control cells a significant reduction in ΔΨm was observed in ventricular myocytes during hypoxia, a finding concordant with our earlier work. Notably, the observed loss of mitochondrial ΔΨm during hypoxia was prevented in cells expressing Bnip3Δex3. Moreover, the preservation of mitochondrial ΔΨm by Bnip3Δex3 during hypoxia coincided with a marked reduction in cell death by apoptosis and necrosis and significantly increased cell viability compared to cells subjected to hypoxia alone (Figure 3F and 3G; Online Figure IV). Importantly, Bnip3Δex3 had no effect on the expression levels of Bcl-2 family members, Bax, Bak, or Beclin-1, excluding the possibility that Bnip3Δex3 promotes survival by influencing the expression levels of these factors (N.Y. and L.A.K., unpublished data, 2009).
To prove a survival role for endogenous Bnip3Δex3, we assessed whether selectively inhibiting full-length Bnip3 isoform would influence mitochondrial function and cell viability during hypoxia. For these studies we designed a shRNA (shRNA-Bnip3FL) specifically targeting sequences against Bnip3 exon3 as a means to selectively knockdown Bnip3FL. We reasoned that because exon3 sequences are only present in the Bnip3FL isoform and not the Bnip3Δex3 isoform, endogenous Bnip3FL and not Bnip3Δex3 would be affected by the shRNA knockdown (Figure 3A). As shown in Figure 3B and 3C, in contrast to normoxic control cells a significant increase in Bnip3FL gene and protein expression was observed in cells subjected to hypoxia, which is in agreement with our earlier work involving hypoxia-induced transcription of Bnip3.8
In the presence of shRNA-Bnip3FL, however, expression of Bnip3FL was markedly inhibited, whereas Bnip3Δex3 expression was unaffected, in cells during hypoxia (shown in Figure 3B and 3D, respectively). Notably, during hypoxia mitochondrial membrane ΔΨm and cell viability were indistinguishable from normoxic control cells following Bnip3FL knockdown, Figure 3E and 3F. Importantly, our preliminary and published studies verified the specificity and integrity of the shRNA against Bnip3FL,14 thereby excluding the possibility for off target effects of the Bnip3FL RNA interference on cell survival. Together these findings strongly suggest that Bnip3Δex3 protects cardiac myocytes against mitochondrial defects and cell death induced by Bnip3FL during hypoxia.
Secondly, to validate the notion that Bnip3Δex3 plays an important survival role in ventricular myocytes by antagonizing or opposing Bnip3FL, we reasoned that cells deficient or defective for Bnip3Δex3 isoform would be more susceptible to hypoxic injury. To test this possibility, we conducted reciprocal experiments in which we rendered ventricular myocytes defective for Bnip3Δex3 during hypoxia. For these studies, we designed siRNA targeted against the unique sequences encompassing the exon2-exon4 junction that is present only in the alternative spliced variant Bnip3Δex3 and not in the Bnip3FL (Figure 4A through 4C). Therefore, this would allow us to selectively knockdown Bnip3Δex3 isoform without influencing Bnip3FL. Remarkably, in contrast to control cells, viability was dramatically reduced in cells following Bnip3Δex3 knockdown during hypoxia (Figure 4D and 4E). These findings are consistent with our interaction data for Bnip3Δex3 and Bnip3FL (Figure 2D), and the ability of Bnip3Δex3 to suppress cell death by disrupting integration of Bnip3FL into mitochondria (Figure 2E).
Thirdly, to further prove that Bnip3Δex3 opposes the cytotoxic actions of Bnip3FL to promote survival, we next tested the impact of Bnip3 isoforms on cell viability in Bnip3−/− mouse embryonic fibroblasts (MEFs). Because these cells are deficient for generating both Bnip3 isoforms, we reasoned that we could assess the impact of one isoform in the presence and absence of the other, on cell viability in a Bnip3-null background. As shown in Figure 4F and 4G, Bnip3−/− MEF cells and Bnip3−/− MEFs expressing Bnip3Δex3 isoform were indistinguishable from each other with respect to cell viability. However, repletion of Bnip3FL into Bnip3−/− cells resulted in a significant increase in cell death. Importantly, cell death induced by Bnip3FL was rescued by coexpression of Bnip3Δex3. Collectively, our data strongly suggest that Bnip3Δex3 is an endogenous inhibitor that attenuates mitochondrial defects and cell death mediated by Bnip3FL during hypoxia.
To our knowledge, we provide the first demonstration that a member of the Bcl-2 gene family undergoes alternative splicing during hypoxia. Previously, we established that Bnip3 is crucial for provoking mitochondrial defects and cell death of ventricular myocytes during hypoxia.8,23,24 In this report, we provide new compelling evidence that alternative splicing of Bnip3 during hypoxia provides a molecular switch that determines whether Bnip3 triggers mitochondrial perturbations and cell death of cardiac myocytes. We specifically show that hypoxia not only drives Bnip3 transcription, but provides a molecular signal for alternative splicing of Bnip3 resulting in the exclusion of exon3. We determined that the Bnip3Δex3 mRNA resulting from the fusion of exon2 and exon4 introduces a frame shift mutation and stop codon that prematurely terminates transcription up-stream of exon5 and exon6, generating a truncated Bnip3 protein missing the putative BH3-like domain and carboxyl-terminal transmembrane domain required for mitochondrial targeting. Notably, in contrast to full-length Bnip3, which provokes cell death, Bnip3Δex3 promotes cell survival. The underlying mechanisms that account for alternative splicing of Bnip3 premRNA during hypoxia are unknown and are an active area of investigation; nevertheless, several salient and important findings arise from this work.
First, we found that Bnip3 gene transcription is not only activated in ventricular myocytes under relevant physiological conditions in vivo and in vitro, but hypoxia/ischemia provides a molecular signal that promotes alternative splicing of exon3 and synthesis of Bnip3Δex3. Second, and perhaps most compelling, were our findings demonstrating the ability of Bnip3Δex3 to interact and suppress the mitochondrial defects and cell death induced by Bnip3FL. Third, we show that cells defective for synthesizing Bnip3Δex3 displayed a marked increase in cell death during hypoxia.
Given that deregulated Bnip3 transcription would otherwise have catastrophic consequences in postmitotic cells, implies that Bnip3 must be highly regulated and under tight transcriptional control. Indeed, we have previously demonstrated that the Bnip3 promoter is subject to strong basal repression, but is highly induced during hypoxia. However, despite the increase in Bnip3 transcription during hypoxia, certain cells can reportedly avert death in response to Bnip3.17,18 The underlying mechanism for this resistance is undetermined, but likely involves a mechanism that antagonizes the actions of Bnip3. This notion is supported by the fact that at no time did we detect full-length Bnip3 transcripts in the absence of truncated Bnip3Δex3. Furthermore, activating endogenous Bnip3 gene transcription under basal normoxic conditions was sufficient to increase the full-length Bnip3 isoform, but not the Bnip3 splice variant. This implies that alternative splicing of Bnip3 mRNA during hypoxia is a selective regulated process and may represent a novel cellular defense mechanism that safe-guards against indiscriminant mitochondrial damage and cell death that would otherwise occur by Bnip3 gene activation alone, if unopposed during hypoxia. This view is supported by the fact that we showed by not one, but by three, independent approaches that the extent of mitochondrial damage and cell death mediated by Bnip3FL was significantly greater in cells deficient for Bnip3Δex3. When taken together, our data strongly suggest that the principle function of Bnip3Δex3, at least operationally, is to limit or curtail mitochondrial damage and cell death induced by Bnip3FL during hypoxia. Though the mode by which Bnip3Δex3 suppresses cell death was not determined, the fact that mitochondrial associated Bnip3FL was reduced in the presence of Bnip3Δex3 strongly suggests it behaves as a dominant-negative inhibitor interfering with the mitochondrial targeting of Bnip3FL and/or its ability to provoke mitochondrial perturbations, which if not curtailed would provoke cell death by apoptosis and/or necrosis pathways. This view is consistent with the loss of mitochondrial membrane potential and increased ROS production induced by Bnip3FL in cells deficient for Bnip3Δex3.
The closest homolog to Bnip3 is Nix/Bnip3L, which can reportedly undergo RNA splicing; however, unlike Bnip3, Nix is not induced in the heart during hypoxia or ischemia, but instead is transcriptionally activated by Gq signaling during pathological cardiac hypertrophy.9,25 Considering that the primary mode by which Bnip3 induces cell death is to disrupt mitochondrial function,8 it is possible that Bnip3FL may initially induce subtle mitochondrial changes still compatible with cell life, with more severely damaged or irreparable mitochondria removed from the cell by mitophagy. This would effectively postpone the induction of Bnip3-mediated apoptosis. This view is supported by a recent report documenting the dependency of Nix for efficient mitochondrial clearance in differentiating reticulocytes by ATG8/GABARAP.26,27 However, this caveat must be interpreted with caution, because it remains to be tested whether signaling events involved in mitochondrial clearance in reticulocytes, as part of a normal developmental process, are equivalently operational in the myocardium.28 It is equally undetermined whether Bnip3FL plays any role in clearing damaged mitochondria during hypoxia, which was not addressed here and is beyond the scope of the present study. Nevertheless, the fact that Bnip3FL-induced mitochondrial perturbations were dramatically reduced in the presence of Bnip3Δex3 is consistent with this theory and a cytoprotective role for Bnip3Δex3.
Therefore, based on the findings of the present study, we envision a model in which the synthesis of Bnip3Δex3 during hypoxia confers a survival role by opposing or dampening the mitochondrial defects induced by Bnip3FL; however, beyond a certain threshold Bnip3FL dominates and gives rise to irreversible mitochondrial injury and cell death (Online Figure I). This view is supported by the fact that ventricular myocytes rendered defective for Bnip3Δex3 were more sensitive to cell death induced by Bnip3FL and hypoxia. Based on our findings, we speculate the discordant and confounding reports on Bnip3FL's ability to provoke death in certain tumor cells may be explained in part by the unappreciated existence of the Bnip3Δex3 isoform, which we believe likely masks the ability of Bnip3FL to provoke cell death. This notion is supported by our findings demonstrating that knockdown of Bnip3Δex3 in pancreatic ductal carcinoma cells, which are resistant to hypoxic injury and which have elevated basal levels of the Bnip3Δex3 spliced variant, increased cell death in response to hypoxia (H.G. and L.A.K., unpublished data, 2010).
Thus, our findings provide the first direct evidence for the existence of a novel survival mechanism that is obligatorily linked and mutually dependent on hypoxia-induced alternative gene splicing of Bnip3. Hence, by curtailing the mitochondrial injury induced by Bnip3FL, Bnip3Δex3 variant may represent an adaptive mechanism to safe guard against excess Bnip3FL and cell death during hypoxic stress.
Sources of Funding
R.D. holds a postdoctoral fellowship from the Manitoba Health Research Council; J.W.G. holds a postdoctoral Canadian Institutes of Health Research/IMPACT fellowship. This work was supported by grants to L.A.K. from the Canadian Institutes of Health Research. L.A.K. is a Canada Research Chair in Molecular Cardiology.
We are grateful to the friendship of the late Dr Arnold H. Greenberg, to Drs Honey B. Golden and Harvey Weisman for critical comments on the manuscript, and to Pam Lowe for graphics and figures.
In February 2011, the average time from submission to first decision for all original research papers submitted to Circulation Research was 13.7 days.
Non-standard Abbreviations and Acronyms
- BCL2-associated X protein
- B-cell CLL/lymphoma 2
- Bcl-2/adenovirus E1B 19-kDa interacting protein 3
- Bnip3 full length
- Bnip3 with exon3 deletion
- histone deacetylase
- nuclear factor of κ light polypeptide gene enhancer in B cells inhibitor, α
- mouse embryonic fibroblast
- nuclear factor κB
- Bcl-2/adenovirus E1B interacting protein 3-like
- reactive oxygen species
- short hairpin interference RNA
- tetramethyl rhodamine methyl ester
- mitochondrial membrane potential
- Received December 14, 2010.
- Revision received March 4, 2011.
- Accepted March 8, 2011.
- © 2011 American Heart Association, Inc.
- Reed JC
- Adachi S,
- Cross AR,
- Babior BM,
- Gottlieb RA
- Crow MT,
- Mani K,
- Nam YJ,
- Kitsis RN
- Jacobson MD,
- Burne JF,
- Raff MC
- Regula KM,
- Ens K,
- Kirshenbaum LA
- Chen G,
- Cizeau J,
- Vande VC,
- Park JH,
- Bozek G,
- Bolton J,
- Shi L,
- Dubik D,
- Greenberg A
- Bruick RK
- Diwan A,
- Krenz M,
- Syed FM,
- Wansapura J,
- Ren X,
- Koesters AG,
- Li H,
- Kirshenbaum LA,
- Hahn HS,
- Robbins J,
- Jones WK,
- Dorn GW
- Shaw J,
- Yurkova N,
- Zhang T,
- Gang H,
- Aguilar F,
- Weidman D,
- Scramstad C,
- Weisman H,
- Kirshenbaum LA
- Baetz D,
- Regula KM,
- Ens K,
- Shaw J,
- Kothari S,
- Yurkova N,
- Kirshenbaum LA
- Green DR,
- Bissonnette RP,
- Cotter TG
- Bellot G,
- Garcia-Medina R,
- Gounon P,
- Chiche J,
- Roux D,
- Pouysségur J,
- Mazure NM
- Gurevich RM,
- Regula KM,
- Kirshenbaum LA
- Shaw J,
- Zhang T,
- Rzeszutek M,
- Yurkova N,
- Baetz D,
- Davie JR,
- Kirshenbaum LA
- Galvez AS,
- Brunskill EW,
- Marreez Y,
- Benner BJ,
- Regula KM,
- Kirschenbaum LA,
- Dorn GW
- Dorn GW II.
Novelty and Significance
What Is Known?
The hypoxia-inducible Bcl-2 family member Bnip3 provokes mitochondrial perturbations and cardiac cell death with features of both apoptosis and necrosis.
The biological role of Bnip3 in cardiac muscle and cancer is unclear.
What New Information Does This Article Contribute?
We identify a novel spliced variant of Bnip3, named Bnip3ΔEx3 that is exclusively generated in cardiac myocytes during hypoxia and myocardial infarction.
We demonstrate that Bnip3ΔEx3 acts as cell survival factor that antagonizes the cytotoxic effects of the native full-length Bnip3.
These findings demonstrate that Bnip3 exists in two isoforms, one that promotes cell death, whereas the other promotes survival.
Alternative gene splicing provides a versatile mechanism by which cells generate proteins with different or even antagonistic properties. The hypoxia-inducible protein Bnip3 is a critical component of death pathways. Understanding how Bnip3 functions in cardiac cells is of paramount importance toward the ultimate goal of developing novel therapeutics to mitigate cell death during heart failure. Here we identify a new spliced variant of Bnip3 that encodes a novel protein that opposes the cytotoxic actions of Bnip3. Discovery of the Bnip3 splice variant highlights a survival protein that is functionally linked to alternative gene splicing during hypoxia.