ERK1/2 Regulates Hepatocyte Trib1 in Response to Mitochondrial Dysfunction

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1833 (2013) 3405–3414

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

ERK1/2 regulates hepatocyte Trib1 in response to mitochondrial dysfunction

Sébastien Soubeyrand a,ThetNainga,AmyMartinuka, Ruth McPherson a,b,⁎

a Atherogenomics Laboratory, University of Ottawa Heart Institute, Ottawa, Canada b Division of Cardiology, University of Ottawa Heart Institute, Ottawa, Canada

article info abstract

Article history: The TRIB1 locus (8q24.13) is a novel locus identiﬁed and replicated by several genome-wide association studies Received 5 June 2013 for associations with plasma triglycerides, apolipoprotein B and coronary artery disease. The TRIB1 protein prod- Received in revised form 13 September 2013 uct, tribbles-like protein 1 (Trib1), regulates MAPK activity. MAP kinases transduce a large variety of external sig- Accepted 2 October 2013 nals, leading to a wide range of cellular responses, including growth, differentiation, inﬂammation and apoptosis. Available online 24 October 2013 Importantly, Trib1 has been shown to regulate hepatic lipogenesis and very low density lipoprotein production. Despite the relevance of hepatocyte Trib1 to lipid metabolism and atherosclerosis, little is known about the Keywords: TRIB1 mechanisms regulating Trib1 itself. Here, we identify the mitochondria axis as a regulator of Trib1. Treatment Mitochondria of HepG2 cells with a short pulse of a low oligomycin concentration led to a potent and prolonged increase in Hepatocytes the Trib1 mRNA, an effect that was shared with other mitochondria stressors. HuH7 cells as well murine hepato- Oligomycin cytes were also responsive albeit to a weaker extent. The upregulation appeared largely independent of reactive MAPK oxygen species generation or metabolic stress and was mainly under transcriptional control, with ERK1/2 playing HepG2 an important regulating role in the process. While the presence of the Trib1 protein could be inferred, attempts to correlate the increased mRNA to changes in protein level were unsuccessful due to the lack of recognizable Trib1 signal. Our data enrich the current paradigm of Trib1 as an activator of the MAPK pathway by uncovering a role for MAPK in regulating Trib1. © 2013 Elsevier B.V. All rights reserved.

1. Introduction At the protein level, TRIB1 (also called C8FW or SKIP1) is a member of a conserved family of 3 proteins in mammals, with pleiotropic effects Coronary disease (CAD) is a complex disorder with multiple etiolo- ranging from homeostasis control to cellular proliferation and neurode- gies and a strong heritability component. In the last few years genome generation [26,36,57]. All 3 members harbor kinase-like domains but wide association studies (GWAS) from multiple groups have identified appear inactive due to differences at a key residue [26]. Rather than common alleles robustly linked to CAD [13,22,24,47]. Among the more functioning as bona fide kinase they appear to interface with, and regu- than 40 loci recently associated with CAD, is a region located 30 kb late the activity of, effector kinases. More specifically TRIB1 has been downstream of the Tribbles 1 gene (TRIB1) [47].TheTRIB1 locus has associated with the anti-apoptotic mitogen-activated protein kinase also been associated with increased circulating triglycerides and choles- (MAPK) pathway [57]. TRIB1 overexpression leads to enhanced ERK1/ terol, decreased high density lipoprotein cholesterol concentration as 2 (MAPK1/3) phosphorylation in various cell types likely via its interac- well as with hepatic steatosis [9,24,45,48,52]. Consistent with a role in tion with MEK1 (MAP2K1) and MKK4 (MAP2K4), two regulators of the CAD, TRIB1 protein was reported to be elevated in atherogenic lesions ERK1/2 and JNK pathway, respectively [21,27,46].Inadditiontoassoci- [46]. Additional efforts using murine models have provided strong ating with MEKs, the three Tribbles family members can recruit the support for a role for TRIB1 in lipid homeostasis and demonstrated ubiquitin ligase COP1 (constitutive photomorphogenic 1 protein) via a increased lipolysis in a TRIB1 deficient model system [7,40]. conserved C-terminal [12]. Recruitment of COP1 by TRIB3 has been shown to play a pivotal role in lipid mobilization in adipose tissue [37]. While one can only speculate as to the implications of this interaction for normal TRIB1 function, it is entirely consistent with a role as an Abbreviations: GWAS, genome-wide association study; CAD, coronary artery disease; adaptor platform promoting ubiquitination and degradation of cognate TRIB1, tribbles homolog 1; VLDL, very low density lipoprotein; ERK1/2, extracellular targets, as demonstrated for CEBPα in leukemic cells [12,56]. signal-regulated kinase 1/2; OLGM, oligomycin A; MAPK, mitogen-activated protein TRIB1 mRNA is widely expressed in a variety of tissues albeit to vary- kinase ing degrees and is relatively unstable, with a reported half-life of approx- ⁎ Corresponding author at: 40 Ruskin St-H4203, Ottawa K1Y 4W7, Canada. Tel.: +1 613 761 5256; fax: +1 613 761 4676. imately 1 h [27,54]. This instability may help explain why the protein is E-mail address: [email protected] (R. McPherson). expressed at a very low level. While studies so far indicate that all 3

0167-4889/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbamcr.2013.10.001 3406 S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 proteins have important functions in several biological systems, relative- 2.3. Western blot analysis ly little is known about what regulates their gene expression [57]. The mitochondrial network is involved in a plethora of essential cel- Whole cell extracts were lysed in RIPA buffer (50 mM Tris-HCl, lular processes ranging from energy generation to apoptosis [3,31].In 0.15 M NaCl, 1% Triton X-100, 0.5% NaDeoxycholate, 0.1% sodium dode- addition mitochondria are both physically and functionally intertwined cyl sulfate, pH7.4) supplemented with 1 X Complete Protease Inhibitor with the endoplasmic reticulum [14,15,32]. Although the full extent of Tablets and 1 X phoSTOP cocktail (Roche). Lysates (30 μg) were cleared this integration is still incompletely understood, the juxtaposition of by centrifugation (13,000 ×g) for 5 min at 4 °C and resolved by electro- these two organelles appears of paramount importance in multiple reg- phoresis on 4–15% gradient gels. Protein standards were from Bio-Rad ulatory systems including calcium signaling and lipid biosynthesis [14]. (Precision plus All blue). Proteins were then transferred to nitrocellu- As the site of ongoing oxidative phosphorylation, the mitochondrial lose membranes for 90 min at 90 V. Membranes were blocked in 5% compartment is continuously generating oxidative radicals, which can milk and incubated with the various primary antibodies (1:1000 dilu- potentially damage multiple cellular targets [58]. Cells have evolved tion) for 16h at 4°C while incubations with the matching secondary an- adaptive systems to deal with reactive oxygen species (ROS) by neutral- tibodies (1:10,000 dilution, LI-COR) were performed for 1 h. Following izing enzymatically or chemically the reactive species themselves or each incubation, blots were washed 4 × 5 min in phosphate buffer saline their oxidized products [3]. The mitochondria is itself at the forefront (PBS). Acquisition was performed on an Odyssey imaging station. Quan- of this line of defense and interfering with normal mitochondrial func- tification was performed using their proprietary software (version tion can generate ROS by uncoupling the respiration complexes present 3.0.21). For multiple isoforms (ERK and JNK), the entire population was in the inner mitochondrial membrane or tempering with the ability of quantified. Primary antibodies used are described in the supplemental the mitochondria to appropriately address escaped ROS [3]. section. In view of the importance of mitochondria in the orchestration of cellular metabolism as well as the implication of TRIB1 in lipid homeos- 2.4. MitoSOX, Dichlorodihydrofluorescein diacetate, tetramethylrhodamine tasis, this work investigates the involvement of mitochondria in TRIB1 ethyl (TMRE) and ATP quantification assays regulation. Here, we demonstrate that mitochondria dysfunction leads to increased TRIB1 expression in human hepatoma cells and murine hepa- HepG2 cells were seeded 24 h prior to assay and fluorescence was tocyte cultures and strive to clarify its mechanistic underpinnings. In the measured in 24 well plates on a SynergyMX microplate reader (Biotek). process, we have identifiedacontributionofERK1/2inregulatingboth Total ROS were measured using the Image-iT LIVE green Reactive Oxy- basal and stimulated TRIB1levels. Thus our work uncovers an unexpected gen Species Detection Kit (which employs 5-(and-6)-carboxy-2′,7′- roleforERK1/2asaregulatorofTRIB1expression, broadening the func- dichlorodihydrofluorescein diacetate) while mitochondrial oxidation tional link between the two entities by demonstrating the presence of a was measured with MitoSOX Red. Methodology was essentially as sug- reciprocal regulatory relationship. gested by the manufacturer (Molecular Probes). Total fluorescence was measured in triplicates and corrected for Hoechst 33342 fluorescence, immediately following the ROS indicators measurements. For mito- 2. Materials and methods chondria membrane potential measurement, TMRE (0.1 μM) was added for an additional 15 min following the incubation of HepG2 2.1. Cell culture cells with FCCP and/or OLGM. Wells were briefly rinsed with PBS and fluorescence was immediately measured using 549 and 575 nm for ex- HepG2 and HuH7 cells (A.T.C.C.) were maintained in low glucose citation and emission respectively. Correction for cell count was DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% performed as above, using Hoechst fluorescence. ATP levels were mea- FBS (fetal bovine serum) and Penicillin/Streptomycin. A mixed mouse he- sured with an ATP determination kit from Molecular Probes. Briefly patocyte population was obtained by dissection of murine livers followed HepG2 were lysed in RIPA buffer supplemented with 1 X Complete Pro- by plating on fibronectin-coated plates (BD Falcon). Cells were allowed to tease Inhibitor Tablets, 1 X phoSTOP cocktail (Roche) and 5mM NaF The recover for 24 h in the above media prior to treatment. TRIB1 silencing cleared supernatants were diluted 100-fold in H O and immediately was performed for 48 h using either with a 2′-O-(2-methoxy) ethyl anti- 2 assayed as described by the kit manufacturer (Molecular Probes). sense oligonucleotide (ISIS Pharmaceuticals) targeting TRIB1 mRNA or a non-target control (NT) using Lipofectamine RNAiMAX (Invitrogen) essentially as per manufacturer's instructions for RNAi except that a final ol- 3. Results igonucleotide concentration of 20 nM was used. Antimycin A, Oligomycin A, rotenone and the antioxidants were from Sigma. All other inhibitors 3.1. Exposure of HepG2 cells to oligomycin lead to increased were from Cell Signaling. HepG2 stable cell lines were generated by TRIB1mRNA levels lentiviral infection with viruses carrying the untagged human TRIB1 ORF or control vector (PLVX). In an initial attempt to identify regulators of TRIB1, we screened HepG2 cells, a widely used model of hepatic liver cells with modulators of various cellular pathways; Since mitochondria are parts and parcels 2.2. RNA isolation and qRT-PCR of multiple cellular pathways, well characterized inhibitors of the organelle were included in the analysis. Oligomycin (OLGM), that inhibits RNA was isolated and converted to cDNA using the High Pure Iso- the mitochondrial ATP synthase, was tested first. Low doses of lation Kit (Roche) and the Transcriptor First Strand cDNA Synthesis oligomycin led to a rapid and potent increase in TRIB1 mRNA levels as Kit (Roche), respectively, following the supplier's guidelines. Ampli- assessed by qRT-PCR (Fig. 1, left). Levels were noticeably increased by fication was performed using the SYBR Green I Master reaction mix 2 h and were maximal by 5 h, remaining high until at least 20 h post (Roche) on a Roche LightCycler 480. Relative amounts of target treatment (Fig. 1, right). Note that the extent of the maximal stimula- cDNAs were normalized to peptidylprolyl isomerase A (PPIA). tion observed appeared to vary somewhat, from 3 to 6 times basal, Sequences of the oligonucleotides used are described in the supple- which may reflect slight differences in OLGM batch and/or cell density. mental Materials and Methods. Values presented are the mean of 3 Given the level of redundancy among the Tribbles family members, we biological replicates performed in duplicates (±S.D.). Data are nor- next compared the effect of TRIB1 to the other members TRIB2 and malized to PPIA and are expressed relative to the control sample TRIB3. While all 3 Tribbles mRNA were upregulated, TRIB1 was signifi- values, set as 100%. cantly more affected by OLGM (Supplemental Fig. 1). S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 3407

600 700

500 600

500 400 400 300 300 200 200

100 100 Relative Trib1 expression (% of Ctl) 0 0 Ctl 0 h 2 h 5 h 20 h

Fig. 1. TRIB1 mRNA is upregulated in response to oligomycin treatment in HegG2 cells. Left, dose response. HepG2 cells were treated with oligomycin (OLGM) at the indicated doses (μM) for 1 h and trib1 mRNA levels quantiﬁed 5 h post treatment. Right, time course. TRIB1 mRNA levels were measured at various time points following a 1 h treatment with 0.3 μMOLGM.

3.2. TRIB1 mRNA upregulation occurs in response to multiple OXPHOS Both compounds interfere with respiration by interfering with complex inhibitors and is conserved I (Rot) or III functions (Antim) [29,51]. Treating HePG2 cells with Antim or Rot stimulated TRIB1 levels as well, to similar extent similar to what In addition to targeting mitochondrial ATP synthase, OLGM has the was observed with OLGM (Fig. 2A). Hence interfering with mitochon- potential to affect the Na/K-ATPase, albeit at higher doses (IC50 drial respiration lead to an increase in TRIB1 expression in HepG2 cells. 4.5 μM) or perhaps other uncharacterized OLGM targets as previously The prevalence of this response was then examined in another hep- suggested [38,43]. In order to ascertain the putative functional link atoma cell line (HuH7) as well as in primary (murine) hepatocytes between mitochondria and TRIB1, HepG2 were treated with two other (Fig. 2B, C). HuH7 displayed a TRIB1 response comparable, albeit of mitochondria speciﬁc drugs Rotenone (Rot) and Antimycin A (Antim). lower magnitude, to that of HepG2 with regard to all three drugs. By

A B 450 200 400 180 350 160 300 140 120 250 100 200 80 (% of Ctl) 150 (% of Ctl) 60 100 40 50 20 Relative Trib1 expression 0 Relative Trib1 expression 0 Ctl Antim Rot Ctl OLGM Antim Rot C 250 2 h 5h 200 ** *

150

100

50 Relative Trib1 expression (% of Ctl) 0 Ctl Antim 0.6 Antim 2 OLGM 0.3 OLGM 2 Rot 2

Fig. 2. TRIB1 mRNA upregulation occurs in response to other mitochondria-specifc drugs and is conserved. A, Hepg2 cells were treated with Antimycin A (Antim; 0.6 μM) or Rotenone (Rot; 2 μM) for 1 h and allowed to recover for 5 h prior to RNA isolation. B, HuH7 were treated as in A as well as with OLGM (0.3 μM). C, Cells isolated from mouse liver were treated with the indicated drugs (in μM) for 1 h and chased for either 2 or 5 h, as indicated. RNA was isolated and TRIB1 abundance measured by qRT-PCR. Asterisks indicate statistically signiﬁcant differences between samples and control (p b 0.05). 3408 S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 comparison, primary murine hepatocytes displayed a more nuanced A and transient response to the drugs. 700 600 3.3. TRIB1 upregegulation is independent of reactive oxygen species 500

Next, we turned our attention to the identiﬁcation of proximal 400 events mediating the observed upregulation. Reactive oxygen species 300 (ROS) appeared to be a promising initial avenue of investigation since 200 all three drugs are associated with increased ROS to some extent. Relative Trib1

Antimycin A and rotenone treatments of isolated mitochondria lead to expression (% of Ctl) 100 increased ROS generation in the form of superoxide radicals that are 0 rapidly converted to membrane-permeant hydrogen peroxide by superoxide dismutase in the mitochondrial matrix [29]. Reports on the impact of Oligomycin treatment on ROS levels differ somewhat, likely as a function of the system studied [23,29,33]. Contribution of ROS to the TRIB1 increase was tested by the inclusion of well characterized antioxidants: Trolox, a peroxyl radical scavenger B and Tiron, a superoxide radical anion scavenger [30] and the broad spec- 500 800 ificity N-acetylcysteine [2]. As shown in Fig. 3A, all three compounds 700 400 failed to attenuate the response, suggesting that the TRIB1 increase Ctl) 600 following OLGM treatment cannot be ascribed to oxidative damage. 500 To evaluate more directly whether oxidative stress alone 300 can lead to TRIB1 upregulation, HepG2 cells were treated with tert- 400 butylHydroperoxide (tBHP), a non-specific ROS generator. An effect on 200 300 Normalized oxidative stress was ascertained by measuring 2,7-dichlorofluorescein Relative Trib1 200

fl 100 fluorescence (AU) diacetate (DCFDA) and mitoSOX uorescence, indicators of global cellular expression (% of 100 H O level and mitochondria-specific superoxide anions respectively. Im- 2 2 0 0 portantly, incubation with tBHP did not affect TRIB1 mRNA level despite leading to increased total H202 and a milder albeit statistically significant Ctl b fi fl (p 0.05) increase in mitochondria-speci c uorescence (Fig. 3B). Taken tBHP together these data indicate that increased ROS levels are neither neces- OLGM sary nor sufficient to induce TRIB1. C D

3.4. Oligomycin induction occurs independently of mitochondria inner 600 100 membrane potential ) 80 Oligomycin may also exert its effect on TRIB1 via a direct impact on 400 mitochondrial respiration and/or mitochondrial function by causing hy- 60 perpolarization across the inner mitochondrial membrane [10]. Thus 40 TRIB1 upregulation could be a direct result of hyperpolarization or 200 (Ctl = 100%) arise from general pertubations in inner mitochondrial polarity. To in- Relative Trib1 20

vestigate the contribution of polarity changes to TRIB1 mRNA, mito- expression (% of Ctl Relative TMRE signal chondria were depolarized with p-trifluoromethoxy carbonylcyanide 0 0 phenyl hydrazone (FCCP), in the presence or absence of OLGM. FCCP OLGM -+--+ also leads to a transient and potent stimulation of respiration as mito- FCCP (μM) 55050 FCCP FCCP chondria scramble to regenerate the dissipated proton gradient across + OLGM the mitochondrial inner membrane [6]. Interestingly FCCP treatment Fig. 3. TRIB1 mRNA levels are insensitive to antioxidants and exhibit a nuanced response to alone led to a marginal increase in TRIB1 mRNA albeit at relatively mitochondrial membrane potential perturbations. A, HepG2 cells were treated with the high doses (Fig. 3C). The increase was associated with a 60% decrease indicated amounts of antioxidants in conjunction with OLGM or vehicle (DMSO) for 1 h. in membrane potential as assessed by TMRE fluorescence (Fig. 4D). Im- RNAs were harvested 5 h later and TRIB1 mRNA levels were measured by qRT-PCR. B, portantly, FCCP did not affect the extent of the OLGM-induced increase HepG2 were treated OLGM or tert-Butyl Hydroperoxide (tBHP) for 60 min and either fi in TRIB1 message indicating that these compounds operate via function- TRIB1 (left) or total cellular ROS levels or mitochondria speci c ROS (right) were measured by DCFDA or mitoSOX, respectively. C, HepG2 cells were pre-treated for 30 min with FCCP ally distinct mechanisms. followed, where appropriate, by the inclusion of a 60 min OLGM treatment. RNA was isolated 5 h later post-treatment and assessed for TRIB1 and PPIA. D, FCCP controls showing 3.5. TRIB1 increase in response to OLGM is not mediated by established that a 50 μM FCCP pulse results in a significant decrease in membrane potential as assessed energy sensors by TMRE fluorescence. Fluorescence was measured immediately after the 1 h drug incubation. Data are the average of n = 3, ±S.D. While short, the OLGM challenge could conceivably impact the cellular energy stores by blocking the mitochondrial ATP synthase sufficient- ly to induce a starvation response. Indeed we observe that ATP levels (Fig. 4B). AICAR and compound C, which respectively stimulate and in- decreased rapidly, approaching roughly half of its control levels within hibit AMPK, had minimal effects on basal TRIB1 levels: AICAR had no sig- 1 h and remained suppressed for at least another 5 h (Fig. 4A). This sug- nificant impact on TRIB1 while compound C marginally increased TRIB1 gested that TRIB1 upregulation could conceivably arise in response to mRNA. Similarly there was a significant but weak stimulation of basal declining energy stores. In order to assess the relative contribution of TRIB1 levels associated with the mTOR inhibitor Rapamycin. However, energy depletion to the TRIB1 increase, we evaluated the contributions none of these modulators affected the OLGM response noticeably, argu- of two pivotal regulators of energy homeostasis AMPK and mTOR ing that the TRIB1 upregulation either does not result from the energetic S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 3409 A B 120 600

100 500

80 400

60 300 40 200 **

Relative ATP level (% of Ctl) 20 100 Relative Trib1 expression (% of Ctl)

0 0 Ctl 0 h 5 h

Fig. 4. Interfering with key cell energy sensors fails to curb TRIB1 upregulation. A, Transient OLGM treatment decreases ATP levels in HepG2 cell. ATP levels from HepG2 cells treated with OLGM or vehicle were assessed using luciferin conversion into oxyluciferin; ATP concentrations were normalized to the vehicle treated control. Data represent the mean of 5 independent experiments ±S.D. B, Relative amount of TRIB1 mRNA in HepG2 pre-treated with rapamycin (RAP), AICAR or compound C (CMPNDC) prior to the addition of OLGM as indicated for 1 h. RNA was harvested 5 h post treatment. Among the pre-treated samples statistically signiﬁcant differences with Ctl are indicated by an asterisk (P b 0.05, 2-tailed Student t-test).

imbalance, is upregulated by a distinct pathway or occurs prior to mTOR or AMPK activation.

3.6. Implication of the MAPKs in TRIB1 regulation by OLGM Ctl 0 h 2 h 5 h

In light of the tight association of the MAPK pathway with TRIB1 JNK (and the Trib proteins in general) and the speciﬁc implication of JNK in the cellular response to oxidative stress [42], we then examined the pJNK involvement of the various MAPKs in the response of TRIB1 to OLGM. MAPK activation was ﬁrst assessed by western blot analysis of JNK, Ponceau p38 and ERK, using the phosphorylated form as a proxy for the pathway activation status (Fig. 5). Phosphorylation of all 3 MAPKs increased rapidly in response to OLGM, although JNK exhibited a relative delay in this p38 regard. Kinetically therefore, phosphorylation preceded the maximal TRIB1 increase (5 h post-treatment), consistent with a role for these pp38 MAPKs in the control of TRIB1 mRNA level. In order to identify which of these pathways, if any, regulates TRIB1 regulation, cells were treated with inhibitors targeting the individual MAPK pathways: SB203580 for p38, SP600125 for JNK and U0126 for Ponceau MEK1/2 (ERK pathway). Among the 3 inhibitors used, only U0126 was able to abrogate TRIB1 upregulation (Fig. 6A). While SP600125 treatment had little effect, arguing against a role of JNK in the process, SB203580 increased both basal and OLGM-induced TRIB1 levels to a ERK1/2 small extent, suggesting that the p38 axis may function as a negative regulator of TRIB1 transcription and/or mRNA stability. Pathway inhibi- pERK1/2 tion was then evaluated by western blot by examining the phosphorylation status of the MAPKs (Fig. 6B). The analysis revealed evidence of Ponceau interplay between the inhibitors. U0126, which blocks the ERK pathway upstream of ERK, interfered with both p38 and ERK phosphorylation, suggesting that ERK may play a dominant role over p38 in the response Fig. 5. Activation of MAPKs by OLGM. Western blot analysis of whole cell lysates obtained from HepG2 treated with OLGM for 1 h and chased for 0 to 5 h. Ponceau of each blot is to OLGM. As expected, phosphorylation of p38 and JNK which depends shown as a protein load indicator. Only the p54 form of JNK was detectable whereas on upstream kinases, remained unaffected by their respective inhibitors both forms of ERK (p42 and p44) were detected. Data is representative of 3 independent SB203580 and SP600125. experiments. 3410 S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 AB

1400 *

1200 U0126 SP600125 1000 -+-+-+-+SB203580 OLGM 800 JNK pJNK 600 p38

pp38 400 *

ERK1/2

200 * * pERK1/2 Relative Trib1 expression (% of Ctl) 0 Ponceau OLGM -+ -+ -+ -+ DMSO SB SP U0126

Fig. 6. ERK1/2 activation is required for both basal and OLGM-induced TRIB1 mRNA levels. A, HepG2 cells were concomitantly treated with the indicated drugs (SB, 20 μM SB203580; SP, 40μM SP600125; U0126: 10 μM) for 1 h. RNA was harvested 5 h post treatment and subjected to qRT-PCR to assess TRIB1 levels; Pre-treating cells for 1 h with the MAPK inhibitors yielded highly similar results (n = 2, data not shown). Asterisks indicate statistically signiﬁcant differences between samples treated with inhibitors or matching (±OLGM) DMSO controls. B, Samples treated as in A were analyzed by western blot. Data is representative of 3 independent experiments.

3.7. Increased TRIB1 is the result of transcriptional activation and does not or a combination thereof. Of interest, TRIB1 mRNA has been reported to require de novo protein synthesis have a short half-life [54] and ERK has been associated with translation regulation via the ribosomal S6 kinase family [1]. In order to address the Having identiﬁed the plausible basis of the TRIB1 mRNA increase, we role of transcription in this process, cells previously treated with OLGM next turned our attention to more proximal inquiries, namely the na- were allowed to recover in the presence of the transcription inhibitor ture of the TRIB1 increase. Increased TRIB1 levels could presumably Actinomycin D. As expected in view of the short half-life of TRIB1 ensue from a net stabilization of the message or enhanced transcription, mRNA, Actinomycin D led to a pronounced decrease in TRIB1 message

A B 500 450 450 400 400 350 350

300 300

250 250

200 200

150 150

100 Relative Trib1 expression (% of Ctl) 100 Relative Trib1 expression (% of Ctl) 50 50

0 0

Fig. 7. TRIB1 mRNA stability is regulated at the transcription level and is independent of protein translation. A, HepG2 cells treated or not with OLGM were allowed to recover for 5 h in the presence of Actinomycin D (ActinoD; 5μg/ml) as indicated. B, Cells were treated with cycloheximide (5μg/ml) and OLGM, either singly or concurrently, as indicated. TRIB1 mRNA level was assessed by qRT-PCR and normalized to PPIA. S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 3411 abundance over the course of the 5 h treatment (down to 3% of control immunoprecipitation with either TRIB1 antibody prior to SDS-PAGE value) (Fig. 7A). Actinomycin D had a similar effect on the OLGM- were also unsuccessful (data not shown). We did however observe treated samples, indicating that the magnitude of the TRIB1 upregula- that TRIB1 down-regulation was always correlated with a decrease in tion is largely, if not entirely, under transcriptional control. recovered RNA (47% ± 15%, n = 4), consistent with the presence of a The role of de novo protein synthesis and/or translational elongation functional TRIB1 protein capable of promoting MAPK and proliferation. in the stabilization of TRIB1 was then investigated by exposing HepG2 Indeed examining the status of ERK following TRIB1 downregulation re- cells to the translation inhibitor cycloheximide. Cycloheximide treat- vealed a blunted basal ERK phosphorylation (Fig. 8D), confirming the ment led to a slight increase in basal TRIB1 level but had no significant role of TRIB1 as an important modulator of the MAPK axis and its func- impact on the OLGM response excluding a requirement for de novo pro- tional status in HepG2 cells. tein synthesis in the process (Fig. 7B). As part of this investigation, we systematically noted the presence of a cross-reacting signal that closely matched the migration position of re- combinant TRIB1 by western blot. In order to determine its nature, the 3.8. Evidence for a functional TRIB1 protein in HepG2 cells corresponding epitope was partially purified by FPLC and analyzed by mass spectrometry (Supplemental Fig. 2). The analysis revealed no Although TRIB1 mRNA was increased in response to OLGM, effects at trace of TRIB1-derived peptides. Rather, the best match (52% coverage, the protein level remained to be investigated. Attempts were then made 11% of total spectra, 12 more abundant than its nearest neighboring can- to directly measure TRIB1 protein levels by western blot. Importantly didate) corresponded to elongation factor 1-alpha-like 3 (EF1A3). De- TRIB1 has been reported to be expressed at a very low level, from an un- spite sharing no obvious identity, a low stringency alignment with stable RNA [54]. Treating HepG2 cells with OLGM did not reveal any NCBI Basic Local Alignment Search Tool (BLAST) revealed a weak iden- change in protein intensity using two antibodies with distinct specific- tity corresponding to amino acids 3 to 22 and 302 to 333 of TRIB1 which ities (Fig. 8A). Morover, a profound downregulation of TRIB1 mRNA may explain the cross-reactivity observed. levels achieved via the inclusion of U0126 either as a pulse or as a con- tinuous treatment for 48 h, failed to impact the western blot signal noticeably. Integrity and specificities of the antibodies was confirmed by 4. Discussion detection of TRIB1 in 293 T extracts transiently overexpressing TRIB1 (Fig. 8C). In addition, depletion of TRIB1 mRNA to 40% of its normal TRIB1 plays an important role in hepatocyte MAPK signaling and level in HepG2 using TRIB1 specific antisense oligonucleotide (60% lipid metabolism, and genetic variation at the TRIB1 locus has been mRNA knock-down, 48 h) did not alter any detectable signal, even linked to plasma triglycerides, hepatic steatosis and CAD [9,24,45, upon prolonged exposure (Fig. 8C). Attempts to enrich TRIB1 by 48,52]. Despite its importance, little is known about the regulation of

A B Ctl 0 h 2 h 5 h 24 h 1

Ctl 48 h P 48 h C 0.9 50- αTrib1 (R) 0.8 50- αTrib1 (R) 37- 0.7 37- 0.6 50- αTrib1 (G) 0.5 50- αTrib1 (G) 0.4 37- 37- 0.3 50- 0.2 50- Overlay (G + R) Overlay (G + R) 0.1 37- 0 37- 50- Relative Trib1expression (Ctl =1) Ponceau Ctl 50- Ponceau 37- U0126 P U0126 C 37- C D 100 80 NT

Trib1 60 Trib1 NT H293T+ 100 ERK 40 50 αTrib1 (R) 80 (% of NT) 20 p-ERK

60 Relative pERK 0 50 αTrib1 (G) 40 β-Tubulin 20

Relative Trib1 0 50 Ponceau expression (% of NT)

Fig. 8. TRIB1 epitope in HepG2 cells is insensitive to drug treatments. A, Western blot analysis of OLGM-treated HepG2 cells. HepG2 cells were treated with a pulse of OLGM (0.3 μM) for 1 h and allowed to recover for the indicated time (in hours). Presence of the TRIB1 protein was evaluated by western blot analysis of whole cell lysates. Detection was performed with antibodies against 2 distinct regions of TRIB1; red and green signals represent the goat and rabbit antibodies, respectively. Yellow is indicative of co-migration of proteins recognized by both antibodies. B as in A, except that cells were treated with either a 1 h pulse (P) or continuously (C) with U0126 (10 μM). Samples were harvested 48 h post-onset of treatment. TRIB1 expression (mRNA) was measured by qRT-PCR. The blot was stained with Ponceau to visualize protein load. C, Cells were transfected for 48 h either with an antisense oligonucleotide targeting TRIB1 or a non-target control (NT). Lysates we analyzed by western blot using antibodies targeting the core region (rabbit, R) or the C terminal region of TRIB1 (Goat, G). The matching Ponceau stain is shown to visualize protein loading. Messenger RNA knock-down as assessed by qRT-PCR is shown on the right. For comparison, a lysate (4 μg) of 293 T cells overexpressing TRIB1 was included; No corresponding TRIB1 signal could be obtained from 4 μg of untransfected H293T extracts (data not shown). D, Cells treated as in C but were analyzed for pERK level. Quantiﬁcation represents the ratio of pERK to ERK, normalized to the NT, from 2 representative experiments (±S.D.). β Tubulin signal was used as loading control. 3412 S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414

TRIB1 expression. Our results start to address this point by describing levels in HepG2 cells have yielded no evidence for a protection from the regulation of TRIB1 expression by the MAPK pathway, more specifi- OLGM however (Supplementary Fig. 3). While this does not fully ex- cally the ERK1/2 branch. This observation enriches the current model of clude a pro-survival role for TRIB1, it indicates at the very least that it TRIB1 as a MAPK regulator by establishing a positive feedback loop: is not be sufficient in this regard, at least in HepG2 cells. TRIB1 activates ERK1/2 which in turn increases TRIB1 levels. Evidently We demonstrate that the TRIB1 mRNA increase in response to OLGM the feed forward cycle needs to be kept in check and the p38 axis may is largely dependent on steady state transcription. One cannot exclude be parts and parcel of this regulatory network. This is suggested by at present the possibility that some of the OLGM effect can be ascribed the observation that p38 inhibition with SB203580 led to increased to mRNA stabilization. Consistent with this idea, mRNA upregulation, TRIB1 levels both basally and following OLGM exposure. Yet its role expressed as fold increase due to the presence of OLGM is virtually iden- may be somewhat subservient to that of ERK1/2 as blocking ERK tical with or without Actinomycin D. Indeed the 3′ UTR of TRIB1 has been seemed to block p38 activation, as judged by decreased p38 phosphor- shown to stabilize TRIB1 when expressed in trans [28]. Alternatively, the ylation. Clearly, unraveling the full extent of its contribution to TRIB1 residual transcription enhancement could be entirely attributable to a regulation will require further investigation. low level of residual, Actinomycin D-resistant transcription. In any event The implication of ERK1/2 also appears to place TRIB1 upregulation OLGM treatment cannot reverse the inherent instability of the message, in a pathway distinct from the ROS-sensitive JNK/RARα axis [17].Acti- the presence of which remains heavily dependent on transcription. vation of this recently described pathway, observed in several model In silico analysis using publicly available data (UCSC gene browser/ systems [17,44] results in the inactivation of RARα via a proteasome- ENCODE data) indicates the presence of multiple conserved transcrip- mediated degradation and an ensuing inhibition of its cognate tran- tion factor binding sites within or in the vicinity of the TRIB1 promoter scription targets. In one study performed in a hepatocyte model region. Of particular interest in view of the involvement of ERK1/2, the (HuH7), oxidative stress was shown to activate JNK, which in turn led promoter contains a putative binding site for the cAMP responsive ele- to RARα degradation [17]. A similar axis, this time involving both RAR ment binding protein 1 (CREB1), a known target of the ERK pathway via and RXR was demonstrated in cardiomyocites subjected to hyperglyce- the RSK kinase [11]. Activation of CREB1 is a key regulatory step distal to mic conditions and associated radical stress [44]. By contrast, intrinsical- a diverse range of triggers and ultimately results in the transcriptional ly compromised oxidative phosphorylation was shown to trigger a activation of multiple, pathway specific, genes (the CREB regulon) retrograde signaling from the mitochondria to the nucleus involving [19]. Recent work using an epilepsy model has demonstrated that acti- RXRα pathway activation [8]. The characteristics of the TRIB1 response vation of ERK1/2 and CREB1 do correlate with TRIB1 upregulation fur- differ from the JNK/RARα response in 3 major ways. First it is a relatively ther lending credence to the role of CREB1 as a distal effector of the rapid, whereas the RAR degradation is a longer term process (N6h)and OLGM response [4]. The CEBP familly might also play a role in TRIB1 ex- second, whereas TRIB1 upregulation was clearly ERK1/2 dependent, the pression: the promoter region contains a cognate binding site that can RARα degradation is not. Lastly, the absence of effect of the JNK inhibitor be enriched by CHIP using antibodies specific for CEBPα and δ (ChIP- appears to exclude JNK activity from the process. Seq Uniform Peaks, ENCODE). Furthermore, CEBPα has been reported Another question that remains open for future investigations is the na- to promote the expression of a Trib3 promoter construct and can be ac- ture of the initial event that ultimately results in MAPK activation and tivated in a ERK1-dependent manner [18,39,41]. TRIB1 upregulation following the OLGM insult. The low dose of OLGM One limitation of these experiments is that the endogenous protein used, together with the phenocopying with two other mitochondria- could not be specifically detected by western blot or immunocytochem- specific drugs as well as the impact of membrane potential dissipation istry (data not shown). This may be due in part to an unstable mRNA by FCCP, argues in favor of a primary role for the organelle in this respect. and/or protein or perhaps an inef ficient translation. In addition the 5′ The rapid decline in ATP concentration in response to OLGM might consti- UTR of TRIB1 is pyrimidine-rich, consistent with a poor translatability tute a plausible trigger. Surprisingly however we found that AMPK activ- of the RNA [54]. Current efforts are ongoing in our laboratory to isolate ity was not required for the OLGM-mediated increase. AMPK plays a and characterize endogenously expressed TRIB1. central role in regulating cellular energy expenditure, is exquisitely sensi- Both ERK1/2 and TRIB1 play important roles in lipid metabolism: in- tive to changes in ATP levels and has been shown to be rapidly activated terference with either ERK1/2 or TRIB1 leads to profound changes in by OLGM [16,34,49]. Although our data are inconsistent with a role for lipid profiles [5,7]. Recently, both loci have been linked to adipocyte li- AMPK (or mTOR) in TRIB1 regulation following OLGM exposure, they do polysis, albeit in very different ways [20,40]. One experiment uncovered not exclude that ATP depletion might elicit the response albeit via distinct a novel contribution for macrophage-specific TRIB1 as its deletion was mediators which remain to be defined. sufficient to promote adipose tissue lipolysis and predispose mice to Alternatively, or in addition, mitochondrial stress may be directly at high-fat induced hypertriglyceridemia [40]. In a second study, Jiao play. While targeting a common pathway, namely the OXPHOS chain, et al have demonstrated that adipocyte lipolysis and liver beta oxidation the 3 mitochondrial inhibitors used in this study bind to different com- were stimulated by liver-specific ERK activation [20]. In both studies ad- plexes in the chain (complex I, III and V) and as a result somewhat di- ipose tissue lipolysis occurred as a secondary response to distinct organs vergent outcomes are to be expected [29,51]. In light of this, our data and processes, namely aberrant glucose metabolism in the liver and an seem most consistent with a model where distinct mitochondrial altered macrophage population. Hence the underlying regulative path- stresses are recognized and integrated by signaling machineries into a ways likely differ significantly from each other as well as from our common response ultimately implicating TRIB1. Of interest, activated liver model system. Examination of ERK (in macrophages) and TRIB1 ERK1/2 (and JNK) as well as the MAPK upstream regulatory kinases (in liver) as a result of TRIB1 deficiency or ERK activation respectively Raf and MEK1/2 have been shown to relocate to stressed mitochondria will help clarify the relationship. and may act as early mediators [25,53,55,59]. Our study does not rule With respect to HepG2 cells, ERK1/2 pathway inhibition lead to a out the contribution of other mediators: we have focused on the re- 60% decrease in basal TRIB1 mRNA levels, suggesting that some of the sponse to OLGM and whether ERK1/2 mediates TRIB1 upregulation in ERK effects could be mediated by TRIB1. In view of the fact that response to other mitochondrial stressors (including Rotenone and inhibiting ERK1/2 can restore an inherent defect in VLDL assembly par- Antimycin A) remains to be tested. It remains unclear at present what ticular to HepG2 cells [50], TRIB1 may act in normal hepatocytes as a the functional implications of the TRIB1 increase are, but the ability of negative regulator of VLDL generation via its involvement in the ERK TRIB1 to activate the MAPK pathway suggests it may be part of a pro- pathway. According to this scenario MAPK activation would lead to in- survival response. Indeed we demonstrate that TRIB1 downregulation creased TRIB1 levels as well as a repression of VLDL secretion which impedes cell proliferation and TRIB1 overexpression protects leukemic could account for the lower circulating VLDL associated with the over- cells from apoptosis [21]. Our attempts to similarly overexpress TRIB1 expression of TRIB1 in a mouse model system [7]. According to this S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414 3413 model, the contribution of TRIB1would be most likely at the level of [10] J.B. Chappell, G.D. Greville, Effects of oligomycin on respiration and swelling of iso- fi lated liver mitochondria, Nature 190 (502-4) (1961) 502. VLDL assembly and/or secretion since liver-speci c MAPK activation [11] R.H. Chen, C. Sarnecki, J. Blenis, Nuclear localization and regulation of erk- and has been shown to decrease expression of genes favoring beta oxidation rsk-encoded protein kinases, Mol. Cell. Biol. 12 (1992) 915. [20] which should in turn promote increased VLDL assembly as a result [12] P.H. Dedhia, K. Keeshan, S. Uljon, L. Xu, M.E. Vega, O. Shestova, M. Zaks-Zilberman, C. of fatty acid accumulation. Romany, S.C. Blacklow, W.S. Pear, Differential ability of Tribbles family members to promote degradation of C/EBPalpha and induce acute myelogenous leukemia, Blood This appears consistent with the HuH7 data as well. While the re- 116 (2010) 1321. sponsiveness to OLGM is shared to some extent with HuH7 (and prima- [13] A.C. Edmondson, P.S. Braund, I.M. Stylianou, A.V. Khera, C.P. Nelson, M.L. Wolfe, S.L. ry murine hepatocytes), HepG2 cells are particularly responsive in this Derohannessian, B.J. Keating, L. Qu, J. He, M.D. Tobin, M. Tomaszewski, J. Baumert, N. Klopp, A. Doring, B. Thorand, M. Li, M.P. Reilly, W. Koenig, N.J. Samani, D.J. Rader, regard. In addition, whereas VLDL lipidation can be corrected in Dense genotyping of candidate gene loci identifies variants associated with HepG2 by inhibiting ERK1/2, the picture is different in HuH7 where high-density lipoprotein cholesterol, Circ. Cardiovasc. Genet. 4 (2011) 145. lipidation remains deficient [35]. Conceivably, these differences could [14] A.R. English, G.K. Voeltz, Endoplasmic reticulum structure and interconnections with other organelles, Cold Spring Harb. Perspect. Biol. 5 (2013) a013227. be due to differences in the TRIB1/ERK interplay between the two cell [15] J.R. Friedman, L.L. Lackner, M. West, J.R. DiBenedetto, J. Nunnari, G.K. Voeltz, ER tu- types. Indeed, we did observe that TRIB1 mRNA (relative to both PPIA bules mark sites of mitochondrial division, Science 334 (2011) 358. and total RNA input) was more abundant in HepG2 than HuH7 (5 to [16] D.G. Hardie, The AMP-activated protein kinase pathway—new players upstream and downstream, J. Cell Sci. 117 (2004) 5479. 8-fold by qRT-PCR, data not shown) implying that HepG2 cells may [17] Y. Hoshikawa, K. Kanki, A.A. Ashla, Y. Arakaki, J. Azumi, T. Yasui, Y. Tezuka, Y. rely more on TRIB1 for VLDL maturation. Matsumi, H. Tsuchiya, A. Kurimasa, I. Hisatome, T. Hirano, J. Fujimoto, H. Kagechika, K. Shomori, H. Ito, G. Shiota, c-Jun N-terminal kinase activation by oxidative stress suppresses retinoid signaling through proteasomal degradation of Acknowledgements retinoic acid receptor alpha protein in hepatic cells, Cancer Sci. 102 (2011) 934. [18] J. Hu, S.K. Roy, P.S. Shapiro, S.R. Rodig, S.P. Reddy, L.C. Platanias, R.D. Schreiber, D.V. Kalvakolanu, ERK1 and ERK2 activate CCAAAT/enhancer-binding protein-beta- We are grateful to Richard Lee (ISIS Pharmaceuticals, Inc) for dependent gene transcription in response to interferon-gamma, J. Biol. Chem. 276 providing us with the antisense oligonucleotides. This work was sup- (2001) 287. ported by grant #T-7268 from the Heart & Stroke Foundation of Canada. [19] S. Impey, S.R. McCorkle, H. Cha-Molstad, J.M. Dwyer, G.S. Yochum, J.M. Boss, S. McWeeney, J.J. Dunn, G. Mandel, R.H. Goodman, Defining the CREB regulon: a genome-wide analysis of transcription factor regulatory regions, Cell 119 (2004) 1041. Appendix A. Supplementary data [20] P. Jiao, B. Feng, Y. Li, Q. He, H. Xu, Hepatic ERK activity plays a role in energy metabolism, Mol. Cell. Endocrinol. 375 (2013) 157. Supplementary data to this article can be found online at http://dx. [21] G. Jin, Y. Yamazaki, M. Takuwa, T. Takahara, K. Kaneko, T. Kuwata, S. Miyata, T. Nakamura, Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogen- doi.org/10.1016/j.bbamcr.2013.10.001. esis, Blood 109 (2007) 3998. [22] C.T. Johansen, S. Kathiresan, R.A. Hegele, Genetic determinants of plasma triglycerides, J. Lipid Res. 52 (2011) 189. References [23] L. Karawajew, P. Rhein, G. Czerwony, W.D. Ludwig, Stress-induced activation of the p53 tumor suppressor in leukemia cells and normal lymphocytes requires mito- [1] R. Anjum, J. Blenis, The RSK family of kinases: emerging roles in cellular signalling, chondrial activity and reactive oxygen species, Blood 105 (2005) 4767. Nat. Rev. Mol. Cell Biol. 9 (2008) 747. [24] S.Kathiresan,C.J.Willer,G.M.Peloso,S.Demissie,K.Musunuru,E.E.Schadt,L.Kaplan,D. [2] O.I. Aruoma, B. Halliwell, B.M. Hoey, J. Butler, The antioxidant action of Bennett,Y.Li,T.Tanaka,B.F.Voight,L.L.Bonnycastle,A.U.Jackson,G.Crawford,A.Surti, N-acetylcysteine: its reaction with hydrogen peroxide, hydroxyl radical, superoxide, C.Guiducci,N.P.Burtt,S.Parish,R.Clarke,D.Zelenika,K.A.Kubalanza,M.A.Morken,L.J. and hypochlorous acid, Free Radic. Biol. Med. 6 (1989) 593. Scott, H.M. Stringham, P. Galan, A.J. Swift, J.Kuusisto,R.N.Bergman,J.Sundvall,M. [3] R.S. Balaban, S. Nemoto, T. Finkel, Mitochondria, oxidants, and aging, Cell 120 (2005) Laakso,L.Ferrucci,P.Scheet,S.Sanna,M.Uda,Q.Yang,K.L.Lunetta,J.Dupuis,P.I.de 483. Bakker,C.J.O'Donnell,J.C.Chambers,J.S.Kooner,S.Hercberg,P.Meneton,E.G.Lakatta, [4] T.L. Beaumont, B. Yao, A. Shah, G. Kapatos, J.A. Loeb, Layer-specific CREB target gene A. Scuteri, D. Schlessinger, J. Tuomilehto, F.S. Collins, L. Groop, D. Altshuler, R. Collins, induction in human neocortical epilepsy, J. Neurosci. 32 (2012) 14389. G.M.Lathrop,O.Melander,V.Salomaa,L.Peltonen,M.Orho-Melander,J.M.Ordovas, [5] F. Bost, M. Aouadi, L. Caron, B. Binetruy, The role of MAPKs in adipocyte differentia- M. Boehnke, G.R. Abecasis, K.L. Mohlke, L.A. Cupples, Common variants at 30 loci con- tion and obesity, Biochimie 87 (2005) 51. tribute to polygenic dyslipidemia, Nat. Genet. 41 (2009) 56. [6] M.D. Brand, C. Affourtit, T.C. Esteves, K. Green, A.J. Lambert, S. Miwa, J.L. Pakay, N. [25] S. Kharbanda, S. Saxena, K. Yoshida, P. Pandey, M. Kaneki, Q. Wang, K. Cheng, Y.N. Parker, Mitochondrial superoxide: production, biological effects, and activation of Chen, A. Campbell, T. Sudha, Z.M. Yuan, J. Narula, R. Weichselbaum, C. Nalin, D. uncoupling proteins, Free Radic. Biol. Med. 37 (2004) 755. Kufe, Translocation of SAPK/JNK to mitochondria and interaction with Bcl-x(L) in re- [7] R. Burkhardt, S.A. Toh, W.R. Lagor, A. Birkeland, M. Levin, X. Li, M. Robblee, V.D. sponse to DNA damage, J. Biol. Chem. 275 (2000) 322. Fedorov, M. Yamamoto, T. Satoh, S. Akira, S. Kathiresan, J.L. Breslow, D.J. Rader, [26] E. Kiss-Toth, Tribbles: ‘puzzling’ regulators of cell signalling, Biochem. Soc. Trans. 39 Trib1 is a lipid- and myocardial infarction-associated gene that regulates hepatic li- (2011) 684. pogenesis and VLDL production in mice, J. Clin. Invest. 120 (2010) 4410. [27] E. Kiss-Toth, S.M. Bagstaff, H.Y. Sung, V. Jozsa, C. Dempsey, J.C. Caunt, K.M. Oxley, [8] S. Chae, B.Y. Ahn, K. Byun, Y.M. Cho, M.H. Yu, B. Lee, D. Hwang, K.S. Park, A systems D.H. Wyllie, T. Polgar, M. Harte, L.A. O'neill, E.E. Qwarnstrom, S.K. Dower, Human approach for decoding mitochondrial retrograde signaling pathways, Sci. Signal. 6 tribbles, a protein family controlling mitogen-activated protein kinase cascades, J. (2013) rs4. Biol. Chem. 279 (2004) 42703. [9] J.C. Chambers, W. Zhang, J. Sehmi, X. Li, M.N. Wass, H.P. Van der, H. Holm, S. Sanna, [28] E. Kiss-Toth, D.H. Wyllie, K. Holland, L. Marsden, V. Jozsa, K.M. Oxley, T. Polgar, E.E. M. Kavousi, S.E. Baumeister, L.J. Coin, G. Deng, C. Gieger, N.L. Heard-Costa, J.J. Qwarnstrom, S.K. Dower, Functional mapping and identification of novel regulators Hottenga, B. Kuhnel, V. Kumar, V. Lagou, L. Liang, J. Luan, P.M. Vidal, L. Mateo, I.P.F. for the Toll/Interleukin-1 signalling network by transcription expression cloning, O'Reilly, J.F. Peden, N. Rahmioglu, P. Soininen, E.K. Speliotes, X. Yuan, G. Cell. Signal. 18 (2006) 202. Thorleifsson, B.Z. Alizadeh, L.D. Atwood, I.B. Borecki, M.J. Brown, P. Charoen, F. [29] W.J. Koopman, L.G. Nijtmans, C.E. Dieteren, P. Roestenberg, F. Valsecchi, J.A. Smeitink, Cucca, D. Das, E.J. de Geus, A.L. Dixon, A. Doring, G. Ehret, G.I. Eyjolfsson, M. P.H. Willems, Mammalian mitochondrial complex I: biogenesis, regulation, and reac- Farrall, N.G. Forouhi, N. Friedrich, W. Goessling, D.F. Gudbjartsson, T.B. Harris, A.L. tive oxygen species generation, Antioxid. Redox Signal. 12 (2010) 1431. Hartikainen, S. Heath, G.M. Hirschfield, A. Hofman, G. Homuth, E. Hypponen, H.L. [30] C.M. Krishna, J.E. Liebmann, D. Kaufman, W. DeGraff, S.M. Hahn, T. McMurry, J.B. Janssen, T. Johnson, A.J. Kangas, I.P. Kema, J.P. Kuhn, S. Lai, M. Lathrop, M.M. Lerch, Mitchell,A.Russo,Thecatecholicmetalsequestering agent 1,2-dihydroxybenzene- Y. Li, T.J. Liang, J.P. Lin, R.J. Loos, N.G. Martin, M.F. Moffatt, G.W. Montgomery, P.B. 3,5-disulfonate confers protection against oxidative cell damage, Arch. Biochem. Munroe, K. Musunuru, Y. Nakamura, C.J. O'Donnell, I. Olafsson, B.W. Penninx, A. Biophys. 294 (1992) 98. Pouta, B.P. Prins, I. Prokopenko, R. Puls, A. Ruokonen, M.J. Savolainen, D. [31] J. Lee, S. Giordano, J. Zhang, Autophagy, mitochondria and oxidative stress: Schlessinger, J.N. Schouten, U. Seedorf, S. Sen-Chowdhry, K.A. Siminovitch, J.H. cross-talk and redox signalling, Biochem. J. 441 (2012) 523. Smit, T.D. Spector, W. Tan, T.M. Teslovich, T. Tukiainen, A.G. Uitterlinden, M.M. [32] J.W. Lee, W.H. Kim, J. Yeo, M.H. Jung, ER stress is implicated in mitochondrial Van der Klauw, R.S. Vasan, C. Wallace, H. Wallaschofski, H.E. Wichmann, G. dysfunction-induced apoptosis of pancreatic beta cells, Mol. Cells 30 (2010) 545. Willemsen, P. Wurtz, C. Xu, L.M. Yerges-Armstrong, G.R. Abecasis, K.R. Ahmadi, D.I. [33] Y. Liu, D.R. Schubert, The specificity of neuroprotection by antioxidants, J. Biomed. Boomsma, M. Caulfield, W.O. Cookson, C.M. van Duijn, P. Froguel, K. Matsuda, M.I. Sci. 16 (2009) 98. McCarthy, C. Meisinger, V. Mooser, K.H. Pietilainen, G. Schumann, H. Snieder, M.J. [34] A.S. Marsin, L. Bertrand, M.H. Rider, J. Deprez, C. Beauloye, M.F. Vincent, B.G. Van Sternberg, R.P. Stolk, H.C. Thomas, U. Thorsteinsdottir, M. Uda, G. Waeber, N.J. den, D. Carling, L. Hue, Phosphorylation and activation of heart PFK-2 by AMPK Wareham, D.M. Waterworth, H. Watkins, J.B. Whitfield, J.C. Witteman, B.H. has a role in the stimulation of glycolysis during ischaemia, Curr. Biol. 19 (10) Wolffenbuttel, C.S. Fox, M. la-Korpela, K. Stefansson, P. Vollenweider, H. Volzke, (2000) 1247. E.E. Schadt, J. Scott, M.R. Jarvelin, P. Elliott, J.S. Kooner, Genome-wide association [35] S.J. Meex, U. Andreo, J.D. Sparks, E.A. Fisher, Huh-7 or HepG2 cells: which is the study identifies loci influencing concentrations of liver enzymes in plasma, Nat. better model for studying human apolipoprotein-B100 assembly and secretion? Genet. 43 (2011) 1131. J. Lipid Res. 52 (2011) 152. 3414 S. Soubeyrand et al. / Biochimica et Biophysica Acta 1833 (2013) 3405–3414

[36] A. Ostertag, A. Jones, A.J. Rose, M. Liebert, S. Kleinsorg, A. Reimann, A. Vegiopoulos, Witteman, J.F. Wilson, G. Willemsen, H.E. Wichmann, J.B. Whitfield, D.M. D.M. Berriel, D. Strzoda, M. Yamamoto, T. Satoh, S. Akira, S. Herzig, Control of adi- Waterworth, N.J. Wareham, G. Waeber, P. Vollenweider, B.F. Voight, V. Vitart, A.G. pose tissue inflammation through TRB1, Diabetes 59 (2010) 1991. Uitterlinden, M. Uda, J. Tuomilehto, J.R. Thompson, T. Tanaka, I. Surakka, H.M. [37] L. Qi, J.E. Heredia, J.Y. Altarejos, R. Screaton, N. Goebel, S. Niessen, I.X. Macleod, C.W. Stringham, T.D. Spector, N. Soranzo, J.H. Smit, J. Sinisalo, K. Silander, E.J. Sijbrands, Liew, R.N. Kulkarni, J. Bain, C. Newgard, M. Nelson, R.M. Evans, J. Yates, M. A. Scuteri, J. Scott, D. Schlessinger, S. Sanna, V. Salomaa, J. Saharinen, C. Sabatti, A. Montminy, TRB3 links the E3 ubiquitin ligase COP1 to lipid metabolism, Science Ruokonen, I. Rudan, L.M. Rose, R. Roberts, M. Rieder, B.M. Psaty, P.P. Pramstaller, I. 312 (2006) 1763. Pichler, M. Perola, B.W. Penninx, N.L. Pedersen, C. Pattaro, A.N. Parker, G. Pare, B.A. [38] T. Arato-Oshima, H. Matsui, A. Wakizaka, H. Homareda, Mechanism responsible for Oostra, C.J. O'Donnell, M.S. Nieminen, D.A. Nickerson, G.W. Montgomery, T. oligomycin-induced occlusion of Na + within Na/K-ATPase, J. Biol. Chem. 271 Meitinger, R. McPherson, M.I. McCarthy, W. McArdle, D. Masson, N.G. Martin, F. (1996) 25604. Marroni, M. Mangino, P.K. Magnusson, G. Lucas, R. Luben, R.J. Loos, M.L. Lokki, G. [39] S.K. Roy, J. Hu, Q. Meng, Y. Xia, P.S. Shapiro, S.P. Reddy, L.C. Platanias, D.J. Lindner, Lettre, C. Langenberg, L.J. Launer, E.G. Lakatta, R. Laaksonen, K.O. Kyvik, F. P.F. Johnson, C. Pritchard, G. Pages, J. Pouyssegur, D.V. Kalvakolanu, MEKK1 plays Kronenberg, I.R. Konig, K.T. Khaw, J. Kaprio, L.M. Kaplan, A. Johansson, M.R. a critical role in activating the transcription factor C/EBP-beta-dependent gene Jarvelin, A.C. Janssens, E. Ingelsson, W. Igl, H.G. Kees, J.J. Hottenga, A. Hofman, A.A. expression in response to IFN-gamma, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) Hicks, C. Hengstenberg, I.M. Heid, C. Hayward, A.S. Havulinna, N.D. Hastie, T.B. 7945. Harris, T. Haritunians, A.S. Hall, U. Gyllensten, C. Guiducci, L.C. Groop, E. Gonzalez, [40] T. Satoh, H. Kidoya, H. Naito, M. Yamamoto, N. Takemura, K. Nakagawa, Y. Yoshioka, C. Gieger, N.B. Freimer, L. Ferrucci, J. Erdmann, P. Elliott, K.G. Ejebe, A. Doring, A.F. E. Morii, N. Takakura, O. Takeuchi, S. Akira, Critical role of Trib1 in differentiation of Dominiczak, S. Demissie, P. Deloukas, E.J. de Geus, F.U. de, G. Crawford, F.S. Collins, tissue-resident M2-like macrophages, Nature 495 (2013) 524. Y.D. Chen, M.J. Caulfield, H. Campbell, N.P. Burtt, L.L. Bonnycastle, D.I. Boomsma, [41] E. Selim, J.T. Frkanec, R. Cunard, Fibrates upregulate TRB3 in lymphocytes indepen- S.M. Boekholdt, R.N. Bergman, I. Barroso, S. Bandinelli, C.M. Ballantyne, T.L. dent of PPAR alpha by augmenting CCAAT/enhancer-binding protein beta (C/EBP Assimes, T. Quertermous, D. Altshuler, M. Seielstad, T.Y. Wong, E.S. Tai, A.B. beta) expression, Mol. Immunol. 44 (2007) 1218. Feranil, C.W. Kuzawa, L.S. Adair, H.A. Taylor Jr., I.B. Borecki, S.B. Gabriel, J.G. [42] L.A. Sena, N.S. Chandel, Physiological roles of mitochondrial reactive oxygen species, Wilson, H. Holm, U. Thorsteinsdottir, V. Gudnason, R.M. Krauss, K.L. Mohlke, J.M. Mol. Cell 48 (2012) 158. Ordovas, P.B. Munroe, J.S. Kooner, A.R. Tall, R.A. Hegele, J.J. Kastelein, E.E. Schadt, [43] L.A. Shchepina, O.Y. Pletjushkina, A.V. Avetisyan, L.E. Bakeeva, E.K. Fetisova, D.S. J.I. Rotter, E. Boerwinkle, D.P. Strachan, V. Mooser, K. Stefansson, M.P. Reilly, N.J. Izyumov, V.B. Saprunova, M.Y. Vyssokikh, B.V. Chernyak, V.P. Skulachev, Oligomycin, Samani, H. Schunkert, L.A. Cupples, M.S. Sandhu, P.M. Ridker, D.J. Rader, C.M. van inhibitor of the F0 part of H + -ATP-synthase, suppresses the TNF-induced apopto- Duijn, L. Peltonen, G.R. Abecasis, M. Boehnke, S. Kathiresan, Biological, clinical and sis, Oncogene 21 (2002) 8149. population relevance of 95 loci for blood lipids, Nature 466 (2010) 707. [44] A.B. Singh, R.S. Guleria, I.T. Nizamutdinova, K.M. Baker, J. Pan, High glucose-induced [49] M.C. Towler, D.G. Hardie, AMP-activated protein kinase in metabolic control and in- repression of RAR/RXR in cardiomyocytes is mediated through oxidative stress/JNK sulin signaling, Circ. Res. 100 (2007) 328. signaling, J. Cell. Physiol. 227 (2012) 2632. [50] J. Tsai, W. Qiu, R. Kohen-Avramoglu, K. Adeli, MEK-ERK inhibition corrects the defect [45] E.K. Speliotes, L.M. Yerges-Armstrong, J. Wu, R. Hernaez, L.J. Kim, C.D. Palmer, V. in VLDL assembly in HepG2 cells: potential role of ERK in VLDL-ApoB100 particle as- Gudnason, G. Eiriksdottir, M.E. Garcia, L.J. Launer, M.A. Nalls, J.M. Clark, B.D. sembly, Arterioscler. Thromb. Vasc. Biol. 27 (2007) 211. Mitchell, A.R. Shuldiner, J.L. Butler, M. Tomas, U. Hoffmann, S.J. Hwang, J.M. [51] J.F. Turrens, Superoxide production by the mitochondrial respiratory chain, Biosci. Massaro, C.J. O'Donnell, D.V. Sahani, V. Salomaa, E.E. Schadt, S.M. Schwartz, D.S. Rep. 17 (1997) 3. Siscovick, B.F. Voight, J.J. Carr, M.F. Feitosa, T.B. Harris, C.S. Fox, A.V. Smith, W.H. [52] A. Varbo, M. Benn, A. Tybjaerg-Hansen, P. Grande, B.G. Nordestgaard, TRIB1 and Kao, J.N. Hirschhorn, I.B. Borecki, Genome-wide association analysis identifies vari- GCKR polymorphisms, lipid levels, and risk of ischemic heart disease in the general ants associated with nonalcoholic fatty liver disease that have distinct effects on population, Arterioscler. Thromb. Vasc. Biol. 31 (2011) 451. metabolic traits, PLoS Genet. 7 (2011) e1001324. [53] H.G. Wang, U.R. Rapp, J.C. Reed, Bcl-2 targets the protein kinase Raf-1 to mitochon- [46] H.Y. Sung, H. Guan, A. Czibula, A.R. King, K. Eder, E. Heath, S.K. Suvarna, S.K. Dower, dria, Cell 87 (1996) 629. A.G. Wilson, S.E. Francis, D.C. Crossman, E. Kiss-Toth, Human tribbles-1 controls pro- [54] F. Wilkin, N. Suarez-Huerta, B. Robaye, J. Peetermans, F. Libert, J.E. Dumont, C. liferation and chemotaxis of smooth muscle cells via MAPK signaling pathways, J. Maenhaut, Characterization of a phosphoprotein whose mRNA is regulated by the Biol. Chem. 282 (2007) 18379. mitogenic pathways in dog thyroid cells, Eur. J. Biochem. 248 (1997) 660. [47] E.S. Tai, X.L. Sim, T.H. Ong, T.Y. Wong, S.M. Saw, T. Aung, S. Kathiresan, M. [55] Z. Yao, R. Seger, The ERK signaling cascade—views from different subcellular com- Orho-Melander, J.M. Ordovas, J.T. Tan, M. Seielstad, Polymorphisms at newly identi- partments, Biofactors 35 (2009) 407. fied lipid-associated loci are associated with blood lipids and cardiovascular disease [56] T. Yokoyama, Y. Kanno, Y. Yamazaki, T. Takahara, S. Miyata, T. Nakamura, Trib1 links in an Asian Malay population, J. Lipid Res. 50 (2009) 514. the MEK1/ERK pathway in myeloid leukemogenesis, Blood 116 (2010) 2768. [48] T.M. Teslovich, K. Musunuru, A.V. Smith, A.C. Edmondson, I.M. Stylianou, M. Koseki, [57] T. Yokoyama, T. Nakamura, Tribbles in disease: Signaling pathways important for J.P. Pirruccello, S. Ripatti, D.I. Chasman, C.J. Willer, C.T. Johansen, S.W. Fouchier, A. cellular function and neoplastic transformation, Cancer Sci. 102 (2011) 1115. Isaacs, G.M. Peloso, M. Barbalic, S.L. Ricketts, J.C. Bis, Y.S. Aulchenko, G. [58] R.J. Youle, A.M. van der Bliek, Mitochondrial fission, fusion, and stress, Science 337 Thorleifsson, M.F. Feitosa, J. Chambers, M. Orho-Melander, O. Melander, T. (2012) 1062. Johnson, X. Li, X. Guo, M. Li, C.Y. Shin, G.M. Jin, K.Y. Jin, J.Y. Lee, T. Park, K. Kim, X. [59] Q. Zhou, P.Y. Lam, D. Han, E. Cadenas, c-Jun N-terminal kinase regulates mitochon- Sim, O.R. Twee-Hee, D.C. Croteau-Chonka, L.A. Lange, J.D. Smith, K. Song, Z.J. Hua, drial bioenergetics by modulating pyruvate dehydrogenase activity in primary cor- X. Yuan, J. Luan, C. Lamina, A. Ziegler, W. Zhang, R.Y. Zee, A.F. Wright, J.C. tical neurons, J. Neurochem. 104 (2008) 325.