Effects of NRF-1 and PGC-1α Cooperation on HIF-1α and Rat Cardiomyocyte Apoptosis Under Hypoxia

Hypoxia is a primary inducer of cardiomyocyte injury, its signicant marker being hypoxia-induced cardiomyocyte apoptosis. Nuclear respiratory factor-1 (NRF-1) and hypoxia-inducible factor (HIF)-1α are transcriptional regulatory elements implicated in multiple biological functions, including oxidative stress response. However, their roles in hypoxia-induced cardiomyocyte apoptosis remain unknown. The effect HIF-α, together with NRF-1, exerts on cardiomyocyte apoptosis also remains unclear. We established a myocardial hypoxia model and investigated the effects of these proteins on the proliferation and apoptosis of rat cardiomyocytes (H9C2) under hypoxia. Further, we examined the association between NRF-1 and HIF-1α to improve the current understanding of NRF-1 anti-apoptotic mechanisms. The results showed that NRF-1 and HIF-1α are important anti-apoptotic molecules in H9C2 cells under hypoxia, although their regulatory mechanisms differ. NRF-1 could bind to the promoter region of Hif-1α and negatively regulate its expression. Additionally, HIF-1β exhibited competitive binding with NRF-1 and HIF-1α, demonstrating a synergism between NRF-1 and the peroxisome proliferator-activated receptor-gamma coactivator-1α. These results indicate that cardiomyocytes can regulate different molecular patterns to tolerate hypoxia, providing a novel methodological framework for studying cardiomyocyte apoptosis under hypoxia. results that NRF-1 exerts an anti-apoptotic effect in hypoxia-induced cardiomyocyte injury. The specic mechanism or the molecular control method adopted by NRF-1 for regulating the process remains to be studied. The central element in hypoxia is the reduction in oxygen concentration. specic genes related to oxidative stress and the regulation of cardiomyocyte apoptosis by proteins encoded by these genes may help understand mechanism. HIF family proteins are extremely sensitive to oxygen; they are degraded by proteases under normoxia, remain stable hypoxia. under hypoxia in various ways rst NRF-1 HIF-1α. the inhibition of both molecules reduced cardiomyocyte proliferation; however, compared with that in the HIF-1α inhibition group, the reduction in cardiomyocyte proliferation level was alleviated in the NRF-1 inhibition group, and caspase-3 activity was suppressed. Further analysis showed that HIF-1α inhibition increased BCL-2 and BCL-xL expression, and BCL-2/BCL-xL levels were relatively lower in the NRF-1 and HIF-1α co-inhibition group than that in the HIF-1α inhibition group. Interestingly, HIF-1α inhibition led to an increase in NRF-1 expression, which has not been previously reported. This novel nding could also explain the simultaneous increase in BCL-2/BCL-xL expression with HIF-1α and NRF-1 inhibition. Supercially, HIF-1α inhibition leads to cardiomyocyte injury and apoptosis under hypoxia, indicating that HIF-1α serves as an anti-apoptotic molecule under hypoxia. However, under the simultaneous inhibition of NRF-1, the cell morphology appeared qualitatively better than that in the HIF-1α inhibition group; this may be related to the fact that NRF-1 inhibition can partially restore HIF-1α levels and alleviate, to a certain extent, cardiomyocyte hypoxia-induced injury. The results are concordant with the protective effect of HIF-1α on cardiomyocytes previously reported and suggest that HIF-1α plays a more signicant anti-apoptotic role than NRF1. The specic inhibition of HIF-1α an increase in BCL-2/BCL-xL expression, after the simultaneous NRF-1, BCL-2/BCL-xL HIF-1α ndings Choi al. al. Zhao al. indicated that BCL-2/BCL-xL primarily affected NRF-1 rather than HIF-1α. results suggest that


Establishment of the hypoxia model
The cells were subcultured in three T-75 cell culture bottles, and the experiment was initiated when the cells reached 90% con uence. A day before hypoxia induction, 1 × 10 7 cells were subcultured in a 10-cm culture dish and placed in an incubator at 37 ℃ overnight for subsequent use. On day 2 under anoxia, the medium was removed and 10 mL cell culture medium was added to each dish. The dishes were placed in a three-gas incubator, and the timer was set to start when the oxygen concentration reached 1%. After completion, the nitrogen valve in the anoxic operation table was opened, and the oxygen concentration detection table was used to reduce the oxygen concentration to 1%. After treatment under hypoxia, the cells were rapidly placed in the anoxic operation table to extract total protein or RNA.

Immunoprecipitation
Lysate preparation and total protein extraction were performed as described in 2.4. For western blotting, 100 μL supernatant was used, and the remaining was used for immunoprecipitation. The samples were stored on ice until further use. Each 900 μL protein lysate sample was mixed with 1 μg immunoprecipitation antibodies and placed in the microcentrifuge tube rack of a rotary mixer. The instrument was placed in a refrigerator at 4 °C overnight. The next day, 1 μL premixed magnetic beads (Thermo Fisher Scienti c) was added to the sample, mixed gently, and incubated in a refrigerator at 4 °C for 1 h. The liquid was then discarded, and 900 μL pre-cooled lysis buffer containing a protease inhibitor was added, gently mixed and washed thrice. Next, 50 μL 1× protein loading buffer (Solarbio Life Science) was added, mixed well, boiled at 100 °C for 10 min, incubated on ice, and stored. The subsequent steps were the same as described in 2.4.

ChIP assay
The cells were spread on 10-cm Petri dishes with a convergence degree of approximately 90%. To every culture dish, 10 mL DMEM and 275 μL fresh 37% formaldehyde (Sigma-Aldrich) were added, and cross-linking was allowed for 10 min. Next, 10× glycine was added to quench excess formaldehyde, and the suspension was mixed and incubated for 5 min. The residual medium was discarded, 10 mL pre-cooled PBS was added, and the cells were washed. Then, 1 mL pre-cooled PBS containing a protease inhibitor was added. The cells were then scraped into a microcentrifuge tube and centrifuged at 700 ´ g at 4 °C for 3 min. The supernatant was discarded, and the cells were resuspended in SDS lysis buffer containing a protease inhibitor and incubated on ice before ultrasonication (Xinzhi Biotechnology Co., Ltd, Ningbo, China). The ultrasonication conditions were: power, 50 W; ultrasonication for 5 s, stop for 55 s, repeated eight times. Next, protein-G agarose beads (Thermo Fisher, Norway) were added to the ultrasonicated products and mixed for 1 h on the rotary mixer at 4 °C. The supernatant was collected by centrifugation and distributed into two microcentrifuge tubes. Next, 1 μg of normal rat IgG (Sangon Biotech, Shanghai, China) and 1 μg antibody were added to the two tubes, and the mixture was incubated overnight at 4 °C. The next day, agarose beads were added to trap the protein-DNA complexes, followed by washing, complex elution using an eluent buffer, and DNA puri cation after cross-linking. Primer sequences for Hif-1α promoter region used in this study are listed in Table 1. preservation at 4 °C. The samples were separated by electrophoresis using a 0.5% agarose gel.

RNA extraction
RNA was extracted using TRIzol reagent (Thermo Fisher Scienti c). Brie y, 2-5 × 10 6 cells were subcultured in six-well plates. TRIzol reagent (1 mL) was added to each well, and the suspension was distributed in enzyme-free microcentrifuge tubes. Next, 200 μL chloroform (Sigma-Aldrich) was added; the mixture was shaken for 15 s, incubated on ice for 15 min, and centrifuged at 12,000 ´ g for 15 min. To the collected supernatant, an equal volume of isopropanol was added, mixed well, incubated on ice for 10 min, and centrifuged at 12,000 ´ g for 10 min. The supernatant was discarded; 75% ethanol was added to the pellet and the solution was centrifuged at 7,500 ´ g for 5 min. After air-drying, 50 μL enzyme-free ddH 2  curves. The target gene mRNA levels were calculated using the 2 -△△CT method; rat β-actin was used as an internal reference control. The reaction system and primer sequences are listed in Supplementary Table S4 and Supplementary Table S5, respectively.
2.11. Real-time evaluation of cell proliferation Before evaluating cell proliferation, the cells were counted and the concentration was adjusted to 15 × 10 5 cells/mL/well. Real-time cell analysis (ACEA Biosciences, CA, USA) was used. The detection time was adjusted to measure the cell proliferation level every 10 min for 24-48 h. The cells in each well were labeled, and 150 μL complete medium was added into the special plate (8-well) to deduct the base value. Next, 300 μL cell suspension was added into each well, and the cell state was assessed microscopically (EVOS™ XL Core Imaging System, Thermo Fisher). After evenly spreading out the suspension, the plates were placed on an ultra-clean workbench for 30 min and then transferred to the instrument, with the entire detector placed in the cell incubator. After adherence overnight, the cells were observed. For hypoxic cultures, after overnight adherence, the entire instrument was placed in the three-gas incubator, and cells were observed after the oxygen concentration was adjusted to 1%.

Evaluation of caspase-3 activity
Cells (5-10 × 10 6 cells) were seeded on a 10-cm cell culture dish 1 day in advance. The next day, 10 μL DTT was added per mL of lysis buffer and detection buffer. After hypoxic or normoxic culturing, the medium was discarded, and 500 μL of the aforementioned cold cell lysate was added. The cells were scraped into a microcentrifuge tube, mixed by vibration at high speed for 15 s, placed on ice for 15 min with shaking for 15 s every 5 min, and centrifuged at 12,000 ´ g for 15 min. The supernatant was collected into a fresh tube, and protein levels were quanti ed using the Bradford assay. A 96-well plate was prepared for each group, and three repeat wells were assigned. Next, 10 μL protein supernatant (containing 30-50 μg of total protein) was collected from each well, the lysate was added, and the suspension was mixed well. Thereafter, 10 μL Ac-DEVD-pNA (BestBio, Nanjing, China) was added, and the culture was incubated in the dark at 37 °C for 4 h until the solution turned yellow. Detection was performed by measuring absorbance at 405 nm. Caspase-3 activity was measured based on the ratio of the treated group absorbance to that of the blank control group.

Real-time observation of cell apoptosis
The experiment was performed in a Cytation 5 Imaging Reader (BioTek, VT, USA). Brie y, the plate bottom height was adjusted, the six-well plate mode was selected, and 1 × 10 6 cells were seeded in the six-well plate. After overnight adherence at 37 °C, the plate was placed on a test bench. The focus, exposure time, channel, and other parameters were adjusted, and red and blue uorescence were detected simultaneously for each well, once every 10 min. The temperature was set at 37 °C, and the oxygen concentration was xed at 1%. Automatic imaging and videography were performed. The cell proliferation curve was generated using the Gen5 software (BioTek).
2.14. Bisul te sequencing PCR DNA was extracted according to the method described in 2.7. Based on the sequence predicted by the MethPrimer tool [29][30][31], primers were designed for DNA extraction. Details of the reaction system and primer sequences are outlined in Supplementary Tables S6 and S7,  The gene sequence is as follows (5′-3′): GAGAGCAACGTGGGCTGGGGTGGGGCCTGGCCGCCTGCGTCCTTTCCCATTGGCTCTCGGGGAACCCGCCTCCGCTCAGGTGAGGCGGGCCCGCGGGTGCGCGCGTC After bisul te treatment, the cytosine residues were changed to thymine, with the CG sites retained as "C," as shown in the following sequence (5′-3′): GAGAGTAACGTGGGTTGGGGTGGGGTTTGGTCGTTTGCGTTTTTTTTTATTGGTTTTCGGGGAATTCGTTTTCGTTTAGGTGAGGCGGGTTCGCGGGTGCGCGCGTCG 2.15. Statistical analysis SPSS 17.0 (IBM Corporation, NY, USA) was used for statistical analysis. The results are expressed as mean ± standard deviation. Student's t-test was used to compare the means of two samples. One-way analysis of variance was used for comparing multiple samples. Student-Newman-Keuls test was used for pairwise comparison between groups. P < 0.05 was considered statistically signi cant.
To further verify whether the protective effect of NRF-1 is mediated by its anti-apoptotic effect, the different levels of cardiomyocyte apoptosis were evaluated. Caspase-3 is crucial for the execution of apoptosis. Activated caspase-3 can cleave DNA or induce its degradation to RNA, inhibit cytoskeletal protein synthesis, and ultimately induce nuclear pyknosis, fragmentation, and cell membrane disintegration [32,33]. Under hypoxia, caspase-3 activity increased gradually, and NRF-1 suppressed this increase ( Fig. 2A). Furthermore, with the prolongation in hypoxia duration and caspase 3 activation, the state of myocardial cells in each group gradually deteriorated, and the cells began to disintegrate and undergo pyknosis and nuclear fragmentation. As cell membrane integrity is lost during cell apoptosis, propidium iodide can enter the damaged nucleus and stain the DNA. Apoptotic and necrotic cells showed bright red uorescence, with peak uorescence at approximately 6 h; however, NRF-1 delayed this process to approximately 10 h (Fig. S1). Next, we compared the expression of the apoptosis regulatory molecules and found that with an increase in hypoxia, the expression of NRF-1 and anti-apoptotic proteins BCL-2 and BCL-xL decreased, whereas that of pro-apoptotic protein Bax marginally increased, suggesting that NRF-1 promoted BCL-2/BCL-xL expression and inhibited Bax expression (Fig. 2B, C). These results suggest that in hypoxia-induced myocardiocyte injury, NRF-1 offers protection from apoptosis.

NRF-1 binds to Hif-1α promoter region and inhibit its expression
In 293T cells, NRF-1 reportedly binds to Hif-1α promoter region and negatively regulates its expression [34]. The regulatory effects of NRF-1 on HIF-1α were therefore investigated. The primers were designed from the sequence of the rat Hif-1α promoter region available on the Ensembl genome browser (https://asia.ensembl.org/index.html). After evaluating each primer annealing temperature, ChIP was performed to con rm the NRF-1 binding site in Hif-1α promoter region. Our data showed that the promoter region between -1992--1511 and -1147--455 upstream of Hif-1α represented the NRF-1 binding site (Fig.  3A). To better elucidate the effect of NRF-1 on the binding site, qPCR was performed. We observed that HIF-1α expression decreased signi cantly with the increase in NRF-1 expression (Fig. 3B). Similar results were obtained in the western blotting experiment, wherein the HIF-1α protein levels increased gradually with the prolongation of the duration of hypoxia; however, its expression could be inhibited by NRF-1 (Fig. 3C). The above data suggest that the binding of NRF-1 to Hif-1α promoter region can inhibit its expression.
Additionally, NRF-1 reportedly activates Dnmt-1 promoter region, in uence its expression, regulate methylation levels, and further in uence the expression of other genes [34,35]. Since NRF-1 can inhibit the HIF-1α expression at the transcriptional level, we investigated whether this regulation involves its methylation. First, the PROMO online analysis tool (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3) was used to predict NRF-1 binding sites in Dnmt1 promoter region. Since PROMO does not have a rat genome resource, the human database was referred for the binding sequence and it was compared with the rat Dnmt1 promoter sequence to identify the corresponding sequence. The data showed that NRF-1 has two binding sites in the promoter region of human DNMT1, and comparison with rat Dnmt1 promoter sequence revealed two similar sequences, suggesting that the binding sequence is conserved. Subsequently, primers for the sequence were designed (Fig. S2 A, B). After optimizing the annealing temperature, the ChIP assay was performed and revealed that NRF-1 could bind to the sequence (Fig. S2C). Next, the effect of NRF-1 overexpression on the DNMT-1 protein level was evaluated using western blotting. The results indicate that under hypoxia (for 0, 1, 2, 3, 6, 12, and 24 h), Dnmt1 expression gradually decreased; however, it signi cantly increased following NRF-1 overexpression (Fig. S2D). Hence, we next determined whether NRF-1 modulates HIF-1α expression by regulating its methylation through DNMT-1, leading to subsequent regulation of Hif-1α transcription. Based on the information provided on the MethPrimer website (https://www.urogene.org/methprimer/), we concluded that a CpG island was present in Hif-1α promoter region. The methylation status of Hif-1α promoter before and after hypoxia was compared. The bisul te sequencing PCR results revealed no signi cant change in the methylation status of the predicted sequences (Fig. S2E, F). Thus, NRF-1-mediated regulation of DNMT-1 expression does not affect HIF-1α; NRF-1 directly inhibits HIF-1α.

HIF-1α inhibition by NRF-1 affects cardiomyocyte apoptosis
To evaluate the effect of HIF-1α inhibition by NRF-1 on cardiomyocyte apoptosis, HIF-1α and NRF-1 were simultaneously inhibited under arti cial conditions. Preliminary experiments showed that hypoxia for 1 and 2 h did not induce a marked oxygen stress response. Accordingly, hypoxia duration was set at 0, 3, 6, 12, and 24 h. Our data showed that HIF-1α suppression exacerbated the hypoxia-induced decline in cell proliferation and promoted caspase-3 activity; however, contrary to our expectation, NRF-1 expression increased along with BCL-2 and BCL-xL protein levels. These results indicate that similar to NIF-1, HIF-1α acts as an anti-apoptotic protein; furthermore, a mutually inhibitory relationship might exist between NRF-1 and HIF-1α, and the increased BCL-2 and BCL-xL levels could result from the increase in NRF-1 expression caused by HIF-1α inhibition. To further evaluate these possibilities, HIF-1α and NRF-1 expression was inhibited. Compared with HIF-1α inhibition (sh-HIF1α) under hypoxia, the dual inhibition of NRF-1 and HIF-1α relieved the decline in cell proliferation and lowered caspase-3 activity. Unexpectedly, with the reduction in NRF-1 expression, the levels of BCL-2 and BCL-xL decreased, whereas the levels of HIF-1α increased (Fig. 4A, B, C). These results indicate that although NRF-1 primarily regulates the expression of BCL-2 and BCL-xL, this is not the only mechanism that modulates apoptosis. Therefore, HIF-1α may regulate apoptosis under hypoxia in a BCL-2/BCL-xL-independent manner.

NRF-1 competes with HIF-1α to bind HIF-1β
Previous results suggested that HIF-1α might exert the same anti-apoptotic effect as NRF-1. Concurrently, the results also showed that NRF-1 and HIF-1α exhibited mutual expression inhibition. Next, we analyzed the effect exerted by this mutual inhibition. HIF-1α forms a dimer with HIF-1β, and this complex has several biological functions. This led us to investigate NRF-1 and HIF-1β interaction and its effect on other molecular events. Moreover, the effect of this interaction on HIF-1α and HIF-1β binding warranted investigation. First, the binding among HIF-1α, NRF-1, and HIF-1β following NRF-1 interference (sh-NRF1) was evaluated. The results showed that similar to HIF-1α, NRF-1 could bind to HIF-1β, and the level of this binding gradually decreased with the prolongation of hypoxia. Compared with the empty vector group (pGreenPuro), the NRF-1 inhibition group (sh-NRF1) showed a decrease in NRF-1-HIF-1β binding, and HIF-1α-HIF-1β binding increased simultaneously (Fig. 5A). These results indicated that a competitive binding interaction might exist between NRF-1 and HIF-1α for HIF-1β under hypoxia.
These ndings further con rmed the competitive binding between NRF-1 and HIF-1α.

PGC-1α participates in the competitive binding of NRF-1 and HIF-1α with HIF-1β
The regulatory effect of NRF-1 on cells is inseparable from the synergistic effect of PGC-1α. The role of PGC-1α was thus investigated. PGC-1α expression decreased gradually with an increase in hypoxia duration, whereas NRF-1 inhibition did not alter PGC-1α expression. Immunoprecipitation con rmed a stable binding relationship between PGC-1α and NRF-1 that did not change with hypoxia duration. Then, the effect of this binding relationship on HIF-1α was evaluated. DMOG inhibited PGC-1α, and PGC-1α could bind to HIF-1β; however, PGC-1α binding reduced, similar to NRF-1, with an increase in DMOG concentration and HIF-1α accumulation. Subsequently, the cells were treated with ZLN005 (HY-17538; MedChemExpress, NY, USA), a drug that activates PGC-1α transcription and expression. This demonstrated that for different durations of hypoxia (0, 6, 12, and 24 h), PGC-1α intervention increased NRF-1 expression and decreased HIF-1α expression, without affecting HIF-1β expression. However, the increase in PGC-1α expression induced by ZLN005 enhanced the binding between NRF-1 and HIF-1β and suppressed the binding between HIF-1α and HIF-1β (Fig. 6A, B, C). These results indicate a novel binding mechanism among NRF-1/PGC-1α, HIF-1α, and HIF-1β.

Discussion
Apoptosis accompanies ischemia-and hypoxia-induced myocardiocyte injury [38]. Cardiomyocyte apoptosis is common to almost all types of heart diseases [38][39][40][41][42]. To alleviate the effect or avoid the occurrence of cardiomyocyte apoptosis, especially the extensive damage it causes under hypoxia injury, extensive studies were conducted. NRF-1 overexpression alleviated cardiomyocyte apoptosis induced by chemical hypoxia. To further investigate effects of NRF-1 on cardiomyocyte apoptosis, especially under actual hypoxic conditions, we used three-gas incubators to reduce the oxygen concentration to 1%. We found that NRF-1 overexpression and inhibition affected cell growth and proliferation under normoxic conditions. While NRF-1 overexpression promoted cell growth, its inhibition suppressed cell proliferation. Thus, NRF-1 can promote cell proliferation, and NRF-1 inhibition affects cell division and growth, consistent with the nding that NRF-1 knockout is lethal in mouse embryos [43]. This established the important role of NRF-1 in the cell viability and growth of individual organisms. Subsequently, we studied the effect of NRF-1 on cardiomyocyte proliferation under different durations of hypoxia. In the rst few hours (6 h) under hypoxia, the cells maintained a certain proliferation level that may be attributed to the residual oxygen in the solution. However, progressive depletion of oxygen with increase in hypoxia duration decreased the cardiomyocyte proliferation in each group. Notably, NRF-1 overexpression signi cantly slowed the decline in cell proliferation, whereas its inhibition signi cantly decreased the cell proliferation. This indicates that NRF-1 can protect cardiomyocytes from hypoxia and highlights the anti-apoptotic mechanism adopted by NRF-1.
The decrease in cell proliferation under hypoxia may be caused by cell necrosis, apoptosis, and autophagy. To further identify the role of apoptosis in hypoxia, we observed the levels of apoptosis-related molecules, including caspase-3-the primary executor of apoptosis. In transgenic mice, caspase-3 overexpression increased the infarct size of cardiomyocytes caused by oxygen stress and increased the likelihood of death [44]. Conversely, caspase-3 downregulation reduced the apoptotic index and improved cardiac function after myocardial infarction [45,46]. We rst assessed caspase-3 activity in different groups of cardiomyocytes under hypoxia to determine whether the decrease in cardiomyocyte proliferation was caused by apoptosis. Caspase-3 activity gradually increased with increase in the duration of hypoxia. Additionally, NRF-1 overexpression alleviated the high caspase-3 activity; after NRF-1 inhibition, caspase-3 activation became more pronounced. Subsequently, real-time dynamic observation of cell growth showed that the nuclear red uorescence became more evident with increase in caspase-3 activity. As the severity of nuclear membrane damage increased, nuclear condensation increased as well. However, NRF-1 overexpression signi cantly inhibited this process. These results further indicate that hypoxia-induced damage caused to cardiomyocytes is partly triggered by apoptosis; since NRF-1 can alleviate apoptosis, it can prevent the damage caused to cardiomyocytes under hypoxia.
Previous studies on NRF-1 have primarily focused on its effect on mitochondrial function; NRF-1 regulates the expression of mitochondrial respiratory chain complex gene family members, affecting mitochondrial biogenesis, and increases mitochondrial ATP production [47,48]. Reportedly, NRF-1 signi cantly improves the mitochondrial membrane potential in cardiomyocytes under hypoxia and enhances the mitochondrial respiratory capacity to increase cardiomyocyte viability [5]. Therefore, the anti-apoptotic effect of NRF-1 may be achieved by regulating apoptosis-related proteins associated with the mitochondria. Most of these molecules belong to the BCL-2 family, including pro-apoptotic molecules, such as Bax, Bak, and Bid, and anti-apoptotic molecules, such as BCL-2 and BCL-xL. At high levels, Bax can form dimers with other pro-apoptotic molecules, such as Bak and Bad, thus forming pore channels in the mitochondrial membrane, resulting in cytochrome c release and altered mitochondrial membrane potential, eventually triggering apoptosis.
BCL-2 and BCL-xL competitively bind to Bax to reverse the effects of Bax binding to other apoptotic triggers, thus achieving an anti-apoptotic effect. The interaction between anti-apoptotic and pro-apoptotic BCL-2 family proteins can directly determine the fate of different cardiac pathological processes, including myocardial infarction, dilated cardiomyopathy, and ischemic heart disease [49]. For example, BCL-2 can signi cantly reduce the infarct size caused by apoptosis [16,50], while BCL-xL can inhibit the expression of the pro-apoptotic molecules Bax and Bid through different mechanisms [51]. Our results showed that decreased NRF-1 expression caused the expression of the anti-apoptotic molecules BCL-2 and BCL-xL to decrease, while that of the pro-apoptotic molecule Bax was not affected. Since NRF-1 expression can be upregulated via human intervention, this increase can induce subsequent upregulation of BCL-2 and BCL-xL expression, thereby preventing or delaying their hypoxia-induced downregulation. Although Bax level did not increase signi cantly with hypoxia, NRF-1 could inhibit its expression. Thus, the anti-apoptotic effect of NRF-1 on cardiomyocytes under hypoxia is achieved by promoting the expression of BCL-2/BCL-xL and inhibiting Bax expression.
Our results indicate that NRF-1 exerts an anti-apoptotic effect in hypoxia-induced cardiomyocyte injury. The speci c mechanism or the molecular control method adopted by NRF-1 for regulating the process remains to be studied. The central element in hypoxia is the reduction in oxygen concentration. Studying the speci c genes related to oxidative stress and the regulation of cardiomyocyte apoptosis by proteins encoded by these genes may help understand the regulatory mechanism. HIF family proteins are extremely sensitive to oxygen; they are degraded by proteases under normoxia, but remain stable under hypoxia. HIF-1 affects apoptosis under hypoxia and in cardiac diseases in various ways [52,53]. Therefore, we rst assessed whether NRF-1 targets HIF-1α. Transcription factors generally modulate molecular regulation by controlling the transcription levels of genes. Consistently, NRF-1 inhibited Hif-1α mRNA expression and led us to determine whether the effect is directly inhibited by binding or mediated by other regulatory processes. Although NRF-1 exhibits a competitive relationship with methylation, it could promote DNMT-1 expression [35] to maintain the methylation level and further regulate spermatogenesis [13]. In addition to maintaining methylation levels, DNMT-1 initiates methylation [52,54]. Since NRF-1 inhibits the expression of HIF-1α, we investigated the effect of increased methylation levels caused by methyltransferase-1 on the above relationship and found that NRF-1 could bind to Dnmt1 promoter region and regulate its expression, consistent with previous reports [13]. Subsequent studies showed that DNMT-1 expression decreases under prolonged hypoxia, whereas NRF-1 overexpression delayed this decline. Notably, according to an existing prediction website (http://alggen.lsi.upc.es/cgibin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3) and the contributions of previous reports [13], an NRF-1-binding sequence is found in human and mouse Dnmt1. Furthermore, we found a corresponding sequence in rats, suggesting the conservation of this binding sequence and highlighting the important regulatory role of NRF-1 in DNMT-1 expression. However, even though Hif-1α promoter region predictably possessed a methylation site, there was no signi cant change in the methylation status before and after hypoxia, as shown by bisul te sequencing PCR. Possibly, methylation is not the primary regulatory event affecting HIF-1α expression. Based on these results, we conclude that NRF-1 negatively regulates HIF-1α expression by directly binding to Hif-1α promoter region.
Previous studies have shown that HIF-1α acts as a protective molecule in cardiomyocytes under stress [55,56]. It enhances cardiac tolerance to hypoxia in various ways, for instance, by enhancing anaerobic respiration and nucleotide metabolism and reducing cellular oxidative stress [57][58][59][60]. However, HIF-1α can also induce apoptosis, increase the myocardial infarct area, and promote damage [61,62]. Therefore, the effect of HIF-1α on the heart remains to be completely deciphered. We con rmed the effect of HIF-1α on apoptosis of rat cardiomyocytes under hypoxia through experimental studies. HIF-1α inhibition reduced cardiomyocyte proliferation under normoxia; however, following hypoxia, the proliferation of cardiomyocytes in the HIF-1α inhibition group decreased more signi cantly, and caspase-3 activity increased. These results indicate that HIF-1α exerts a protective effect on cardiomyocytes under hypoxia. To elucidate the negative regulatory effect of NRF-1 on HIF-1α and the effect of NRF-1 on cardiomyocyte apoptosis, we speci cally inhibited NRF-1 expression in coordination with HIF-1α inhibition. Under normoxia, the inhibition of both molecules reduced cardiomyocyte proliferation; however, compared with that in the HIF-1α inhibition group, the reduction in cardiomyocyte proliferation level was alleviated in the NRF-1 inhibition group, and caspase-3 activity was suppressed. Further analysis showed that HIF-1α inhibition increased BCL-2 and BCL-xL expression, and BCL-2/BCL-xL levels were relatively lower in the NRF-1 and HIF-1α coinhibition group than that in the HIF-1α inhibition group. Interestingly, HIF-1α inhibition led to an increase in NRF-1 expression, which has not been previously reported. This novel nding could also explain the simultaneous increase in BCL-2/BCL-xL expression with HIF-1α and NRF-1 inhibition. Super cially, HIF-1α inhibition leads to cardiomyocyte injury and apoptosis under hypoxia, indicating that HIF-1α serves as an anti-apoptotic molecule under hypoxia. However, under the simultaneous inhibition of NRF-1, the cell morphology appeared qualitatively better than that in the HIF-1α inhibition group; this may be related to the fact that NRF-1 inhibition can partially restore HIF-1α levels and alleviate, to a certain extent, cardiomyocyte hypoxia-induced injury. The results are concordant with the protective effect of HIF-1α on cardiomyocytes previously reported and suggest that HIF-1α plays a more signi cant anti-apoptotic role than NRF-1. The speci c inhibition of HIF-1α expression was accompanied by an increase in BCL-2/BCL-xL expression, whereas after the simultaneous inhibition of NRF-1, BCL-2/BCL-xL expression decreased with a relative increase in HIF-1α expression; these ndings were consistent with those reported by Choi et al. [63], Menrad et al. [64], and Zhao et al. [62], and indicated that BCL-2/BCL-xL is primarily affected by NRF-1 rather than HIF-1α. These results suggest that apoptosis caused by NRF-1-mediated inhibition of HIF-1α may not involve BCL-2 family proteins but occurs via hitherto unknown mechanisms.
Since NRF-1 and HIF-1α exert the same anti-apoptotic effect, it is worth investigating why they exhibit mutual inhibition. Whether the mutual inhibition is associated with the similar anti-apoptotic effects and the antagonism between the two molecules remains unclear. HIF-1α regulates downstream genes by forming a dimer with HIF-1β. The involvement of NRF-1 in this dimer formation remains to be determined. However, our results demonstrated that NRF-1 could also bind to HIF-1β, and the binding levels gradually decreased under prolonged hypoxia, which reciprocally affected the binding with HIF-1α. Additionally, the inhibition of NRF-1 expression led to an increase in HIF-1α and HIF-1β binding, suggesting a possibility of competitive binding between HIF-1α and HIF-1β. We used DMOG and BAY 87-2243 to verify our hypothesis and found that these drugs did not affect HIF-1β expression while regulating HIF-1α expression. DMOG promoted HIF-1α expression, inhibited NRF-1 expression, and suppressed the binding between NRF-1 and HIF-1β. However, when BAY 87-2243 was used to inhibit HIF-1α, a contrasting yet consistent phenomenon was observed-HIF-1α inhibition increased NRF-1 expression, as well as the binding levels of NRF-1 and HIF-1β. This phenomenon may be related to the adaptation of cells to hypoxic stress. Thus, under normal conditions, NRF-1 acts as a key transcription factor that exerts multiple effects on cardiomyocyte molecular regulation. In the absence of external stimulation, NRF-1 and other nuclear molecules, including HIF-1β, form a key complex and participate in the regulation of mitochondrial function, cell growth, and metabolism. However, under prolonged hypoxia, the expression of NRF-1 decreases gradually, leading to the loss of cell function. As an adaptation to hypoxia and to promote survival, HIF-1α accumulates, gradually replacing NRF-1, thereby promoting the binding between HIF-1α and HIF-1β. Certain key molecules, such as CD39, CD73, p53, and LDHA, are expressed, facilitating cell tolerance to hypoxia. This could explain why HIF-1α can promote hypoxia tolerance in cardiomyocytes and aggravate cardiomyocyte injury under different hypoxic conditions [65,66]. However, the exact reason for aggravation of cardiomyocyte injury needs to be investigated.
Lastly, we examined the role of PGC-1α that is reportedly involved in processes related to apoptosis regulation, such as p53 gene-mediated apoptosis, enhancement of mitochondrial recovery to reduce apoptosis, and regulation of the expression of apoptosis-related molecules [67][68][69]. We found that PGC-1α levels gradually decreased with the increase of hypoxia duration, and this change was not affected by NRF-1 inhibition, indicating that PGC-1α is a transcription factor operating upstream of NRF-1. This nding was consistent with previous reports [70,71]. Our results also showed that NRF-1 and PGC-1α formed stable dimers, and the binding levels of NRF-1 and PGC-1α did not change as the duration of hypoxia was prolonged. Upon DMOG stimulation, the same effect observed for NRF-1 was observed. Promoting PGC-1α expression with ZLN005 treatment, resulted in increased levels of NRF-1 as well, and the levels of NRF-1 and HIF-1β binding increased during the initial stages of hypoxia (0 and 6 h), whereas the binding between HIF-1α and HIF-1β was suppressed.
However, the enhanced binding effect was gradually reversed by HIF-1α because the levels of the two molecules decreased with the increase in the duration of hypoxia. These results suggest that PGC-1α, as an upstream transcription regulator of NRF-1, is also involved in regulating the competitive binding between NRF-1 and HIF-1α. However, whether PGC-1α directly binds to HIF-1α or indirectly binds to HIF-1β by binding with NRF-1 needs to be studied further.

Conclusion
Our results indicate that NRF-1 can alleviate cardiomyocyte apoptosis and improve cardiomyocyte viability under hypoxia. Moreover, HIF-1α serves as an important anti-apoptotic molecule under hypoxia. To the best of our knowledge, this is the rst study to report that NRF-1 cooperatively acts with PGC-1α and competes with HIF-1α to bind HIF-1β. This molecular regulatory process may be related to cardiomyocyte adaptation to hypoxia and may promote cell survival under low-oxygen stress. This study provides a novel theoretical framework for improving the protective mechanisms for hypoxia-induced myocardial injury; however, the speci c molecular process warrants further investigation.  B. Followed by the NRF-1 overexpression (pCDH-NRF1), compared to those in pCDH-CMV, the levels of NRF-1, BCL-2, BCL-xL, and Bax changed after the cells were subjected to hypoxia (at 1% O2) for 0, 1, 2, 3, 6, 12, and 24 h. C. Following NRF-1 inhibition (sh-NRF1), the levels of NRF-1, BCL-2, BCL-xL, and Bax changed after the cells were subjected to hypoxia (at 1% O2) for 0, 1, 2, 3, 6, 12, and 24 h. Bad: BCL-associated agonist of cell death; Bak: BCL-2 homologous antagonist/killer; Bax: BCL-2-associated X protein; BCL-2: B-cell lymphoma 2; BCL-xl: BCL-extra-large; NRF-1: nuclear respiratory factor-1.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.