Factors affecting expression and transcription of uncoupling protein 2 gene.

Previous studies suggest a negative relationship between hepatic oxidative stress and productivity in beef cattle. Uncoupling protein 2 (UCP2) is involved in the disappearance of reactive oxygen species, suggesting the defensive role of UCP2 against oxidative stress. The present study examined the relationship between oxidative stress and expression levels of UCP2/Ucp2 in cultured human and mouse liver-derived cells. We also explored factors regulating bovine Ucp2 transcription. As oxidative stress inducers, hydrogen peroxide, ethanol, and cumene hydroperoxide (CmHP) were used. Expression levels of hemoxygenase 1 (HMOX1), a representative gene induced by oxidative stress, were not affected by any oxidative stress inducers in HepG2 human liver-derived cells. The levels of UCP2 mRNA were also unaffected by the oxidative stress inducers. Treatment with CmHP increased expression of Hmox1 in Hepa1-6 mouse liver-derived cells, but Ucp2 expression was not changed. Stimulus screening for regulator of transcription (SSRT) revealed that expression of p50 or p65, transcription factors conferring response to oxidative stress, did not stimulate bovine Ucp2 transcrition in HepG2 cells. SSRT also showed 11 molecules that induced Ucp2 transcription more than 4-fold; among them, endoplasmic reticulum (ER) stress-related transcription factors such as XBP1, c-JUN, JUNB, and C/EBPβ were identified. However, treatment with ER stress inducers did not increase Ucp2 expression in HepG2 and Hepa1-6 cells. The present results suggest that 1) neither oxidative stress nor ER stress induces Ucp2 expression in liver-derived cells, and 2) Ucp2 transcription is stimulated by several transcription factors.

Oxidative stress triggers the onset of various diseases in humans, including cancer, diabetes, cardiovascular diseases, and neurodegenerative diseases [20]. Upon induction of oxidative stress, various antioxidant responses are stimulated [8]; the defensive responses to oxidative stress include the disappearance of reactive oxygen species (ROS) via uncoupling protein 2 (UCP2) [5, 6]. Originally, UCP2 was suggested to be involved in nonshivering thermogenesis, regulatory secretion of insulin, and the onset of diabetes [35]. But the primary function of UCP2 is now thought to control the amount of mitochondria-derived ROS [5, 6]. It was revealed in a study in response to infection with Toxoplasma gondii where macrophages from Ucp2-knock out mice generated more ROS than those from wild-type mice [1]. UCP2-mediated removal of ROS is partly regulated by the transcript level of UCP2 [18, 27]. Thus, oxidative stress may directly induce Ucp2 expression, which leads to defensive responses to oxidative stress. However, previous results on regulatory expression of UCP2 related to oxidative stress are not always consistent. Song et al. [27] reported that treatment with H2O2 increased UCP2 expression in A549 human alveolar basal epithelial cells. In contrast, Liu et al. [16] showed down-regulation of UCP2 expression in response to H2O2 treatment in ARPE-19 human retinal pigment epithelium cells.

Increased oxidative stress has been suggested in the liver of cattle utilizing feed inefficiently for its production (cattle with high residual feed intake), as compared with that of cattle utilizing feed more efficiently (cattle with low residual feed intake) [30]. Gene expression profile revealed that hepatic expression of molecules involved in detoxification of products of oxidative stress, including glutathione S-transferase and Ucp2 was higher in cattle with low residual feed intake than in those with high residual feed intake [10, 30]. These results can be understood that higher expression of hepatic UCP2 is involved in reduced oxidative stress, leading to high feed efficiency (a ratio of body weight gain to feed intake). The oxidative stress is also suggested to be increased during fattening and by feeding dietary restriction of antioxidant reagents such as β-carotene and vitamin E in Japanese Black beef cattle [34]. The regulation of bovine Ucp2 transcription is not known so far.

The final objective of the study was to clarify whether oxidative stress directly induces Ucp2 expression in bovine hepatocytes. At present useful cell line originated from bovine hepatocyte is not available. Thus, as the first step of the study, we explored 1) whether oxidative stress induces UCP2/Ucp2 expression in human/murine liver-derived cells, and 2) factors affecting bovine Ucp2 transcription and expression.

MATERIALS AND METHODS

Cell culture

HepG2 human hepatoma cells (TKG0205) were provided by Cell Resource Center for Biomedical Research, Institute of Development, Aging, and Cancer, Tohoku University. Hepa1-6 mouse hepatoma cells (RCB1638) was obtained from RIKEN BioResource Research Center. 3T3-L1 mouse preadipocytes (JCRB9014) were were obtained from JCRB Cell Bank.

HepG2, Hepa1-6, and 3T3-L1 cells were cultured in Dulbecco’s modified Eagle medium supplemented with 10% heat‐inactivated fetal bovine serum (Sigma, St. Louis, MO, USA, Lot: BCBV9382) and antibiotics; HepG2 and Hepa1-6 cells were cultured in the presence of high glucose (4,500 mg/l), whereas 3T3-L1 cells were in low glucose (1,000 mg/l). The cells were treated with H2O2 (0.5, 1, or 2 mM), ethanol (EtOH: 1, 2, or 3%), or cumene hydroperoxide (CmHP: 25 or 50 µM) for 1 hr. The mRNA levels of hemoxygenase 1 (HMOX1/Hmox1), UCP2/Ucp2, and forkhead box protein A1 (FOXA1/Foxa1) were examined. In addition, HepG2 and Hepa1-6 cells were treated with thapsigargin (100 nM), A23187 (1 µM), or tunicamycin (2 µg/ml) for 6 hr. The mRNA levels of the spliced form of X-box binding protein 1 (XBP1/Xbp1), C/EBP homologous protein (CHOP/Chop), and UCP2/Ucp2 were examined. Further, transcriptional regulation of bovine Ucp2 was evaluated by luciferase-based reporter assays, as described below.

RNA isolation and reverse transcription-quantitative PCR (RT-qPCR)

Total RNA isolation and real-time RT-qPCR were performed as previously described [23]. Following qPCR primers were used; 5′ggcagagggtgatagaagagg-3′ and 5′-agctcctgcaactcctcaaa-3′ to detect human HMOX1, 5′-ttccgcatacaaccagtgagtgg-3′ and 5′-aaatcctggggcatgctgtcgg-3′ to detect mouse Hmox1, 5′-tgaaggaggccaccaaggaggt-3′ and 5′- ggtacaaggaggccatcaccagctt-3′ to detect both human HMOX1 and mouse Hmox1, 5′-tgaaagccaacctcatgaca-3′ and 5′-gatgacagtggtgcagaagc-3′ to detect human UCP2, 5′-accggcagctttgaagaac-3′ and 5′-agaaacggggaccttcaatc-3′ to detect mouse Ucp2, 5′-tctgtcaaacagttctacaccaagg-3′ and 5′-gccagggcacctgtggt-3′ to detect both human UCP2 and mouse Ucp2, 5′-cgactggaacagctactacg-3′ and 5′-tggtgttcatggtcatgtaggtgtt-3′ to detect both human FOXA1 and mouse Foxa1, 5′-tctgctgagtccgctacaggtg-3′ and 5′-gctggcaggctctggggaag-3′ to detect spliced form of human XBP1, 5′-tctgctgagtccgctacaggtg-3′ and 5′-ccatgggaagatgttctggg-3′ to detect spliced form of mouse Xbp1, 5′-gggagctggaagcctggtatgag-3′ and 5′-gagtrgtgaagrtttttgattcttcctct-3′ to detect both human CHOP and mouse Chop, 5′-tgaccttgatttattttgcatacc-3′ and 5′-cgagcaagacgttcagtcct-3′ to detect human hypoxanthine phosphoribosyltransferase 1 (HPRT1), 5′-tcctcctcagaccgctttt-3′ and 5′-cctggttcatcatcgctaatc-3′ to detect mouse Hprt1, 5′- attcctatgactgtagattttatcagctgaagagcta-3′ and 5′-ccagttaaagttgagagatcatctccaccaa-3′ to detect both human HPRT1 and mouse Hprt1. The ΔΔCt method was used to normalize the levels of target transcripts to HPRT1/Hprt1 levels. Expression levels in control cells were set at 1.

Stimulus screening for regulator of transcription (SSRT)

We recently developed a screening to explore factors affecting the transcription of the target gene named as SSRT [23]. In this screening, the luciferase-based reporter gene is transfected with expression vectors to stimulate various signaling pathways: the expression vectors include plasmids encoding transcription (co-) factors, cytosolic kinases, and extracellular ligands [23]. Factors involved in the regulatory transcription of the target gene are explored by measuring luciferase activity. Evaluated expression vectors do not necessarily contain coding region originated from the bovine genes but human and mouse genes.

The reporter with a bovine Ucp2 promoter fragment (nt −3,000 to +25, Ucp2 (−3,000)-luc) was prepared using pGL4, a luciferase-based reporter plasmid; the translation initiation site was numbered as +1. The nucleotide sequence of the constructs was verified by DNA sequencing. SSRT were performed in HepG2 cells, because plasmids can be easily transfected in HepG2; HepG2 cells were transiently transfected with the individual expression vector and Ucp2 (−3,000)-luc. At 4 hr post-transfection, cells were cultured with a new medium for further 24 hr. Luciferase activity in cells transfected with pcDNA3 (empty vector) was set at 1.

Statistical analysis

Data are expressed as mean ± S.E. Data on relative gene expression were log-transformed to provide an approximation of a normal distribution before analysis. Differences between groups in each tissue were analyzed by one-way analysis of variance (ANOVA). When the effect of group in one-way ANOVA was significant, differences among groups were evaluated by Dunnett’s test or Tukey test. P<0.05 was considered significant.

RESULTS

Oxidative stress does not induce the expression of human and mouse UCP2/Ucp2

To clarify the relationship between oxidative stress and expression levels of UCP2/Ucp2, we first treated HepG2 cells with several oxidative stress inducers, including H2O2, EtOH, and CmHP for 1 hr [11, 14, 25, 33]. Expression levels of HMOX1, a representative gene induced by oxidative stress [17], were not affected by any oxidative inducers (Fig. 1A). We also treated with H2O2 and EtOH for various duration up to 24 hr, but expression levels of HMOX1 were not affected (data not shown). The long-term treatment with CmHP induced cell death in HepG2 cells (data not shown). The oxidative inducers did not affect UCP2 expression (Fig. 1B). Treatment with CmHP increased expression of Hmox1 in Hepa1-6 cells, another liver-derived cell line, whereas H2O2 and EtOH did not affect Hmox1 expression (Fig. 1C). These results suggest that unlike HepG2, CmHP induced oxidative stress in Hepa1-6 cells. However, Ucp2 expression was unaffected by CmHP treatment (Fig. 1D). The other oxidative stress inducers did not modify the expression level of Ucp2 in Hepa1-6 cells, either.

Fig. 1.

Oxidative stress inducers do not induce uncoupling protein 2 (UCP2/Ucp2) in HepG2 cells and Hepa1-6 cells. HepG2 cells (A and B) and Hepa1-6 cells (C and D) were treated with or without H2O2 (0.5, 1, or 2 mM), ethanol (EtOH: 1, 2, or 3%), or cumene hydroperoxide (CmHP: 25 or 50 µM) for 1 hr. Expression levels of hemoxygenase 1 (HMOX1/Hmox1) (A and C) and UCP2/Ucp2 (B and D) were examined by reverse transcription-quantitative PCR (RT-qPCR) analysis, and the expression levels in the control cells were set at 1. The data are presented as the mean ± S.E. (n=3). **: P<0.01 vs. the control group.

A previous study revealed that treatment with H2O2 induced FOXA1 expression in A549 human alveolar basal epithelial cells, which negatively regulates the excess expression of Ucp2 [27]. Thus, it is possible that the inability of H2O2 to induce Ucp2 expression results from increased expression of FOXA1/Foxa1. However, treatment with H2O2 had no effect on FOXA1/Foxa1 expression in HepG2 and Hepa1-6 cells (Fig. 2).

Fig. 2.

Oxidative stress inducers do not apparently induce forkhead box protein A1 (FOXA1/Foxa1) in HepG2 cells and Hepa1-6 cells. HepG2 cells (A) and Hepa1-6 cells (B) were treated with or without H2O2 (0.5, 1, or 2 mM), ethanol (EtOH: 1, 2, or 3%), or cumene hydroperoxide (CmHP: 25 or 50 µM) for 1 hr. Expression levels of FOXA1/Foxa1 were examined by reverse transcription-quantitative PCR (RT-qPCR) analysis, and the expression levels in the control cells were set at 1. The data are presented as the mean ± S.E. (n=3). *: P<0.05 vs. the control group.

Next, we compared expression levels of HMOX1/Hmox1 and UCP2/Ucp2 between cells (Fig. 3); to enable the comparison, we designed primers for qPCR to detect both human and mouse HMOX1/Hmox1 and UCP2/Ucp2 mRNAs. Expression levels of HMOX1 in HepG2 cells were significantly higher than those of Hmox1 in Hepa1-6 cells. HMOX1 expression in HepG2 cells was also higher than that in another linage cells, 3T3-L1 preadipocytes (Fig. 3A). The higher expression levels of UCP2 were also detected in HepG2 cells (Fig. 3B). In view of lower expression of Hmox1 in 3T3-L1 cells than in HepG2 cells, we hypothesized that the oxidative stress inducers could modulate expression levels of Hmox1 in 3T3-L1 cells; CmHP treatment increased Hmox1 expression (Fig. 4A). Thus, we consider that HepG2 cells are constitutively exposed to oxidative stress at a significant level under the basal condition of culture, even if cells were cultured by the recommended method. As a result, additional oxidative stress induction led to an unaffected expression of HMOX1. In 3T3-L1 cells, Ucp2 expression was not also affected by any oxidative stress inducers (Fig. 4B).

Fig. 3.

Expression levels of hemoxygenase 1 (HMOX1/Hmox1) and uncoupling protein 2 (UCP2/Ucp2) gene are higher in HepG2 cells than in Hepa1-6 cells and 3T3-L1 cells. Expression levels of HMOX1/Hmox1 (A) and UCP2/Ucp2 (B) in HepG2, Hepa1-6, and 3T3-L1 cells were examined by reverse transcription-quantitative PCR (RT-qPCR) analysis, and the expression levels in HepG2 cells were set at 1. The data are presented as the mean ± S.E. (n=3). a, b, c: P<0.05.

Fig. 4.

Cumene hydroperoxide (CmHP) induces oxidative stress but does not affect expression level of uncoupling protein 2 (Ucp2) in 3T3-L1 cells. 3T3-L1 cells were treated with or without H2O2 (0.5, 1, or 2 mM), ethanol (EtOH: 1, 2, or 3%), or cumene hydroperoxide (CmHP: 25 or 50 µM) for 1 hr. Expression levels of Hmox1 (A) and Ucp2 (B) were examined by reverse transcription-quantitative PCR (RT-qPCR) analysis, and the expression levels in the control cells were set at 1. The data are presented as the mean ± S.E. (n=3). **: P<0.01 vs. the control group.

SSRT reveals novel regulators for Ucp2 transcription

Recently, we developed a screening system for transcriptional regulators named as SSRT [23]. This screening enables us to identify molecule(s) to regulate transcription of the target gene; effect of 71 expression vectors including oxidative stress-related molecules was evaluated. Forced expression of p50 or p65, members of nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) transcription factor, did not stimulate bovine Ucp2 transcription (<4-fold)- the NF-κB pathway centrally regulates responses to oxidative stress [3] (Fig. 5). We also examined bovine Ucp2 transcription using oxidative stress inducer (hydrogen peroxide) and antioxidant agent (ascorbic acid phosphate), and observed no changes in luciferase activity in response to treatment with these reagents (data not shown).

Fig. 5.

Stimulus screening for regulator of transcription (SSRT) identifies novel regulators of uncoupling protein 2 (Ucp2) transcription. HepG2 cells were transfected with various expression vector encoding transcription (co-) factors, cytosolic kinases, and extracellular ligands with Ucp2 (−3,000)-luc (n=2). Luciferase activity was measured and the expression level in cells transfected with pcDNA3 (empty vector) was set at 1. The data are presented as the mean ± S.E. (n=2).

SSRT also revealed 11 molecules that stimulated bovine Ucp2 transcription more than 4-fold; among them, spliced form of XBP1, c-JUN, JUNB, and C/EBPβ have been shown to be involved in the endoplasmic reticulum (ER) stress induction [13, 21]. Thus, we speculated that ER stress leads to the up-regulation of UCP2/Ucp2 expression in HepG2 and Hepa1-6 cells. As expected, treatment with ER stress inducers, thapsigargin, A23187, and tunicamycin, increased expression levels of spliced XBP1/Xbp1 and CHOP/Chop (Fig. 6). In contrast, expression levels of UCP2/Ucp2 were unaffected by ER stress induction in HepG2 cells (Fig. 6A) and Hepa1–6 cells (Fig. 6B).

Fig. 6.

Uncoupling protein 2 (Ucp2) expression is not affected by induction of endoplasmic reticulum (ER) stress. HepG2 (A) and Hepa1-6 cells (B) were treated with thapsigargin (100 nM), A23187 (1 µM), or tunicamycin (2 µg/ml) for 6 hr. Expression levels of spliced form of X-box binding protein 1 (XBP1/Xbp1), C/EBP homologous protein (CHOP/Chop), and UCP2/Ucp2 were examined by reverse transcription-quantitative PCR (RT-qPCR) analysis, and the expression levels in the control cells were set at 1. The data are presented as the mean ± S.E. (n=4). * and **: P<0.05 and P<0.01, respectively, vs. the control group.

DISCUSSION

The present study explored 1) whether oxidative stress induces Ucp2 expression in human/murine liver-derived cells, and 2) factors affecting bovine Ucp2 transcription and expression. We show that oxidative stress does not directly increase Ucp2 expression in HepG2, Hepa1-6, and 3T3-L1 cells. Results of SSRT also suggest that direct transactivation of Ucp2 gene is not induced by oxidative stress, because molecules for the NF-κB pathway did not stimulate bovine Ucp2 transcrition. We also reveal several novel regulators to stimulate bovine Ucp2 transcription; especially, molecules stimulated in response to ER stress induced the Ucp2 transcription and expression. However, ER stress induction did not affect UCP2/Ucp2 expression. The present results provide basic information on regulatory Ucp2 expression. UCP2 serves the removal of ROS [5, 6]. More factors than considered ever are likely to be involved in the control of ROS amount by UCP2.

Ucp2 expression was not induced by CmHP in Hepa1-6 cells and 3T3-L1 cells (Figs. 1D and 4B), irrespective of the induction of Hmox1 gene. The results suggest that the defensive response to oxidative stress by lipid peroxidation is not achieved through the regulatory expression of Ucp2. Previous studies have shown that oxidative stress modulated UCP2 expression. However, the responses were not consistent. The oxidative stress increased UCP2 expression in A549 human alveolar basal epithelial cells [27], whereas oxidative stress-induced down-regulation of UCP2 expression in ARPE-19 human retinal pigment epithelium cells line [16]. Expression levels of Ucp2 in response to oxidative stress may be regulated in a cell-context dependent manner. Previous studies also showed post-translational regulation of uncoupling activity via UCP2 [9, 36]; the fatty acid and superoxide have been known to increase UCP2 activity, whereas GDP and genipin inhibit UCP2 activity. The cell defense to oxidative stress by CmHP may be achieved through modulation of UCP2 activity in Hepa1-6 cells and 3T3-L1 cells.

Gene induction of HMOX1/Hmox1 in response to oxidative stress inducers depended on cultured cells; HepG2 cells were resistant to HMOX1 expression to the oxidative stress inducers, whereas expression of Hmox1 was increased by CmHP treatment in Hepa1-6 cells, despite both being liver-derived cells. As the reason why expression levels of HMOX1 were unaffected by any oxidative stress inducers in HepG2 cells but not in Hepa1-6 cells, we speculate that the degree of oxidative stress is higher in HepG2 cells than in Hepa1-6 cells under the basal condition of culture; consequently, expression levels of HMOX1 may not be changed to the addition of the oxidative stress inducers to the culture medium in HepG2 cells. In fact, expression levels of HMOX1/Hmox1 were significantly higher in HepG2 cells than Hepa1-6 cells (Fig. 3). Previous studies showed that H2O2 (200 µM) or EtOH (2.5%) did not increase HMOX1 protein level in HepG2 cells [15, 19]. These results are conceptually consistent with the present study.

Up-regulation of Hmox1 expression depended on oxidative stress inducer in Hepa1-6 cells and 3T3-L1 cells; CmHP but not H2O2 and EtOH effectively increased expression of Hmox1. Previous studies also revealed that the extent of induced oxidative stress was different among oxidative stress inducers [2]. EtOH inhibited a mitochondrial electron transfer system, leading to increased production of superoxide production in gastric epithelial cells [29]. H2O2 is directly linked to superoxide production and damages the whole cell, including mitochondria. On the contrary, the primary action of CmHP occurred at the cell plasma membrane [32]. The mode to induce oxidative stress is likely to lead to the oxidative stress inducer-dependent Hmox1 expression. CmHP was a more potent oxidant than H2O2 in neurons [12], and this can be explained by the difference of hydrophobicity; CmHP with more hydrophobicity could initiate lipid peroxidation predominantly, leading to ROS-induced cell membrane injury [12]. In addition, considering that Hmox1 expression is transcriptionally regulated by transcription factor NF-E2-related factor 2 (NRF2) [17], the amount of nuclear NRF2 may vary depending on oxidative stress inducers.

Previous studies have shown that UCP2/Ucp2 expression is regulated by various factors. Ucp2 mRNA was increased by PPARα in rat primary hepatocytes [22]; our results showed that forced expression of PPARα did not increase bovine Ucp2 transcription- the reason is currently unclear. Ucp2 transcription was also stimulated by PPARγ [18], and SIRT1 inhibits PPARγ-induced mouse Ucp2 transcription in 293T cells [4]. Transcription of the human UCP2 gene was increased by co-treatment with triiodothyronine and forced expression of PGC-1α, although the effect of PGC-1α expression alone was minimal (<2-fold) in INS-1E pancreatic β cells [24]. FOXA1 negatively regulated the mRNA level of mouse Ucp2 probably through binding to the Ucp2 promoter [27, 31]. Treatment with TGF-β also decreased expression of human UCP2, which may relate to SMAD4 binding to UCP2 promoter [26]. Further, microRNAs have been reported to be involved in mouse Ucp2 mRNA levels [7, 28]. These results are partly consistent with the present results on SSRT that showed stimulatory effects of PPARγ2 (4.1-fold) and PGC-1α (2.1-fold) on bovine Ucp2 transcription. The present study revealed additional factors to stimulate bovine Ucp2 transcription, i.e., C/EBPβ, C/EBPδ, ERα, ERRβ, SMAD3, FOXA3, TAZ, XBP1, c-JUN, and JUNB. The present results showed that induction of ER stress did not affect UCP2/Ucp2 expression, irrespective of stimulation of bovine Ucp2 gene transcription by ER stress-related molecules. It is possible that the molecules induced by ER stress other than XBP1, C-JUN, JUNB, and C/EBPβ have an inhibitory effect on UCP2/Ucp2 expression.

In the present study, the effects of oxidative stress and ER stress were examined in human- and mouse-originated cells. Also, molecules to evaluate bovine Ucp2 transcription were not always originated from the bovine gene transcript, and SSRT was performed in HepG2 human liver-derived cells; use of different cells may lead to different responses. Future studies are needed to clarify how these factors regulate Ucp2 transcription using bovine gene transcript and bovine cells.

POTENTIAL CONFLICTS OF INTEREST

The authors have nothing to disclose.