Biochemical Journal

Research article

Aldehyde stress and up-regulation of Nrf2-mediated antioxidant systems accompany functional adaptations in cardiac mitochondria from mice fed n−3 polyunsaturated fatty acids

Ethan J. Anderson , Kathleen Thayne , Mitchel Harris , Kristen Carraway , Saame Raza Shaikh


Diets replete with n−3 PUFAs (polyunsaturated fatty acids) are known to have therapeutic potential for the heart, although a specifically defined duration of the n−3 PUFA diet required to achieve these effects remains unknown, as does their mechanism of action. The present study was undertaken to establish whether adaptations in mitochondrial function and stress tolerance in the heart is evident following short- (3 weeks) and long- (14 weeks) term dietary intervention of n−3 PUFAs, and to identify novel mechanisms by which these adaptations occur. Mitochondrial respiration [mO2 (mitochondrial O2)], H2O2 emission [mH2O2 (mitochondrial H2O2)] and Ca2+-retention capacity [mCa2+ (mitochondrial Ca2+)] were assessed in mouse hearts following dietary intervention. Mice fed n−3 PUFAs for 14 weeks showed significantly lower mH2O2 and greater mCa2+ compared with all other groups. However, no significant differences were observed after 3 weeks of the n−3 PUFA diet, or in mice fed on an HFC (high-fat control) diet enriched with vegetable shortening, containing almost no n−3 PUFAs, for 14 weeks. Interestingly, expression and activity of key enzymes involved in antioxidant and phase II detoxification pathways, all mediated by Nrf2 (nuclear factor E2-related factor 2), were elevated in hearts from mice fed the n−3 PUFA diet, but not hearts from mice fed the HFC diet, even at 3 weeks. This increase in antioxidant systems in hearts from mice fed the n−3 PUFA diet was paralleled by increased levels of 4-hydroxyhexenal protein adducts, an aldehyde formed from peroxidation of n−3 PUFAs. The findings of the present study demonstrate distinct time-dependent effects of n−3 PUFAs on mitochondrial function and antioxidant response systems in the heart. In addition, they are the first to provide direct evidence that non-enzymatic oxidation products of n−3 PUFAs may be driving mitochondrial and redox-mediated adaptations, thereby revealing a novel mechanism for n−3 PUFA action in the heart.

  • aldehyde
  • glutathione
  • lipid peroxidation
  • mitochondrion
  • n−3 polyunsaturated fatty acid
  • permeability transition pore


Considerable interest in the relationship between dietary fatty acids and cardiac function has arisen from a number of reports showing that low-glycaemic/high-fat diets preserve contractile force in experimental models of heart failure [13]. Notably, some of the most dramatic effects have been demonstrated by inclusion of n−3 PUFAs (polyunsaturated fatty acids) to the diet [4,5], findings that have been mirrored and compounded in importance by clinical studies showing that dietary supplementation with low-to-moderate doses of the fish oil n−3 PUFAs DHA (docosahexaenoic acid) and EPA (eicosapentaenoic acid) improve left ventricular function [6] and decrease hospitalization rates and overall mortality [7] in patients with heart failure. However, despite the potential benefits and improved cardiac performance brought about by n−3 PUFAs in these studies, the proper dosage and duration of intake required to achieve these effects remain controversial, and their mechanism of action remain unknown.

Mitochondria are of paramount importance in the heart for a multiplicity of reasons in addition to ATP production. Owing to their role as critical loci of necrotic and apoptotic cell death programmes within the cardiomyocyte, mitochondria have repeatedly been implicated to play significant roles in cardiomyocyte cell death, a singular feature of heart failure [8,9]. Much focus has also been directed at cardiolipin, the phospholipid that is exclusive and crucial to mitochondrial inner-membrane functionality and integrity, as a decrease in its concentration and peroxidation have been associated with increased cell death and mitochondrial dysfunction [1012]. It has also been shown that dietary supplementation with fish oil, particularly DHA, increases mitochondrial tolerance to Ca2+ overload (i.e. decreased necrotic susceptibility), and alters the cardiolipin composition of the mitochondrial inner membrane [13,14].

Aside from the direct effects of the n−3 PUFAs on mitochondria and other targets in the heart, indirect mechanisms may also be at work in these systems and must be considered. The high degree of unsaturation in all PUFAs makes them particularly vulnerable to lipid peroxidation, and cardiomyocytes have a number of adaptive processes with which to prevent lipid peroxidation and counteract the products formed from these reactions [1517]. Indeed, the α,β-unsaturated hydroxyalkenal HNE (4-hydroxynonenal) is a well-known and widely studied product of n−6 PUFA peroxidation. As a result of its chemical stability, lipophilic structure and electrophilic reactivity, HNE readily reacts with proteins and phospholipids to trigger alterations in cellular processes and enzymatic activities [18]. In contrast with HNE, far less is known about HHE (4-hydroxyhexenal), the aldehyde formed from non-enzymatic peroxidation of n−3 PUFAs [19].

Since an n−3 PUFA diet has been shown to up-regulate antioxidant systems in various tissues, including the heart [20,21], we speculated that a cause of this may be a low-level sub-toxic stress imposed by PUFA peroxidation which may induce protective adaptations over time (i.e. hormesis) in the heart, possibly via redox-mediated signalling pathways. The present study was undertaken to establish whether adaptations in mitochondrial function in the heart is evident following a short- (3 weeks) and long- (14 weeks) term dietary intervention of n−3 PUFAs, and to determine the mechanism by which these adaptations occur. We provide evidence that long-term dietary intervention with n−3 PUFAs causes a substantial decrease in mitochondrial ROS (reactive oxygen species) emission, combined with improved tolerance to Ca2+ overload. Intriguingly, these alterations are accompanied by increased expression and activity of a number of key enzymes in antioxidant and phase 2 detoxification systems known to be regulated by Nrf2 (nuclear factor E2-related factor 2). We also provide completely novel data regarding the formation of HHE through dietary n−3 PUFAs, and how HHE protein adduct formation in the heart precedes, in chronological order, these mitochondria and redox-mediated adaptations during the course of the diet.


Animals, reagents and diet

Chemicals used in the present study were all obtained from Sigma–Aldrich, with the exception of Amplex Red and Calcium Green 5-N, which were purchased from Invitrogen. All experiments and interventions in mice performed in the present study were approved by the Animal Care and Use Committee at East Carolina University. Male C57BL/6 mice, aged 4–6 weeks (~18 g), were given custom purified diets (Table 1) from Harlan–Teklad. The standard chow control diet (5% fat by weight) provides ~13% of the total energy from fat. The n−3 PUFA and HFC (high-fat control) diets (20% fat by weight) provide ~41% of the total energy from fat. For the n−3 PUFA diet, 10.5% of the total energy was from α-linolenic acid, 3% of the total energy was from EPA and 2% of the total energy was from DHA, with the remaining energy from n−6 PUFAs [22]. The fatty acid composition of the control and n−3 PUFA diets was analysed by GC using commercial standards from NuCheck Prep. These results are shown in Supplementary Table S1 (at and also described in [23]. Analysis revealed that short-chain fatty acids (shorter than myristic acid) were not detectable. Mice were fed ad libitum for 3 weeks or 14 weeks as indicated, at which time they were killed by CO2 inhalation (in a fed state) and the hearts were removed. The hearts were rinsed in saline and a portion of the left ventricle was used for mitochondrial experiments. The remaining left ventricle and right ventricle were snap-frozen in liquid nitrogen for biochemical analysis.

View this table:
Table 1 Composition of the diets

Numbers in the columns represent the weight (in g) of each component of the diet, assuming a final weight of 1 kg. TBHQ, t-butylhydroquinone.

Preparation of permeabilized cardiac myofibres

The technique used for preparation of permeabilized cardiac myofibres has been described previously by our group [24] and others [25]. For the present study, small portions (<30 mg) of left ventricle tissue was dissected and placed in ice-cold buffer X [50 mM Mes, 7.23 mM K2EGTA, 2.77 mM CaK2EGTA, 20 mM imidazole, 0.5 mM DTT (dithiothreitol), 20 mM taurine, 5.7 mM ATP, 14.3 mM PCr (phosphocreatine) and 6.56 mM MgCl2·6H2O (pH 7.1, 290 mOsm)]. The muscle was trimmed of connective and vascular tissue. Five or six small muscle bundles (~4–6 mm in length, 1.0–2.5 mg wet weight) were prepared from each animal. These bundles were gently separated along their longitudinal axis with a pair of needle-tipped forceps under magnification (MX6 Stereoscope, Leica Microsystems). Cardiac fibre bundles were then treated with 50 μg/ml saponin in ice-cold buffer X, and incubated on a rotor for 30 min at 4°C. Following permeabilization, the fibres used for mO2 (mitochondrial O2) consumption experiments were placed in buffer Z [105 mM K-Mes, 30 mM KCl, 1 mM EGTA, 10 mM K2HPO4, 5 mM MgCl2·6H2O, 0.005 mM glutamate and 0.002 mM malate with 5.0 mg/ml BSA (pH 7.4, 290 mOsm)]. The fibres used for mH2O2 (mitochondrial H2O2) and mCa2+ (mitochondrial Ca2+) uptake were placed in ice-cold buffer Y (250 mM sucrose, 10 mM Tris/HCl, 20 mM Tris base, 10 mM KH2PO4, 2 mM MgCl2·6H2O and 0.5 mg/ml BSA). Permeabilized fibres remained in buffer Z on a rotator at 4°C until analysis (<100 min).

Measurement of mO2, mH2O2 and mCa2+ in permeabilized cardiac myofibres

All mitochondrial measurements were performed at 30°C. The O2K Oxygraph system (Oroboros Instruments) was used for all mO2 consumption measurements. The mH2O2 and mCa2+ measurements were obtained using a spectrofluorometer (Photon Technology Instruments), equipped with a thermo-jacketed cuvette chamber. The mO2 experiments were performed in buffer Z. During the course of the experiments, substrates, nucleotides and respiratory inhibitors were provided as indicated in the Figure legends. Both mCa2+ and mH2O2 experiments in the present study were all performed in buffer Y with 100 μM ADP, 5 mM glucose and 1 unit/ml hexokinase present to keep the mitochondria in a permanent submaximal phosphorylating state. For mH2O2 measurements, buffer Y contained 10 μM Amplex Red (Invitrogen), 1 unit/ml horseradish peroxidase, 5 mM pyruvate, 2 mM malate and 5 mM succinate, and mH2O2 emission was calculated as outlined previously [24]. For mCa2+ measurements, buffer Y contained 1 μM Calcium Green 5-N (Invitrogen), 5 mM pyruvate and 2 mM malate. At the start of mCa2+ experiments, 1 μM thapsigargin was added to inhibit SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase), and 40 μM EGTA was added to chelate residual Ca2+ and to establish Fmin. Pulses of 4 nmol of Ca2+ (CaCl2) were added sequentially, and Ca2+ uptake was followed until mPTP (mitochondrial permeability transition pore) opening, as described previously [26]. At the end of the experiment, 1 mM CaCl2 was added to saturate the probe and establish Fmax. Changes in free Ca2+ in the cuvette during mCa2+ uptake were then calculated using the known Kd for Calcium Green 5-N and the equations established by Tsien [27] for calculating free ion concentrations using ion-sensitive fluorophores. At the conclusion of all experiments, fibres were rinsed in double-distilled water, freeze-dried for >2 h and weighed on a micro-scale. Data are expressed as pmol·min−1·(mg of dry weight)−1 (for mH2O2) or nmol·(mg of dry weight)−1 (for mCa2+).

Catalase, glutathione (GSH/GSSG) and related enzyme activity measurements

Myocardial samples frozen in liquid nitrogen were homogenized in 10× (w/v) TEE buffer (10 mM Tris base, 1 mM EDTA and 1 mM EGTA) and 0.5% Tween 20, using a glass grinder (Kimble Chase), and protein samples were prepared for glutathione measurements as described previously [28] using a modification of a previously described method [29]. For measurements of GSSG, samples were prepared in the presence of 1-methyl-2-vinylpyridium triflate. GR (glutathione reductase) activity in myocardial tissue was determined in TEE buffer containing 1 mM GSSG and 0.5 mM NADPH, where activity was calculated from the linear decrease in NADPH absorbance with time [30]. GPx (glutathione peroxidase) activity was determined in TEE buffer containing 1 mM GSH, 100 m-units/ml GR enzyme and 0.5 mM NADPH. The reaction was initiated with a nominal amount of t-butyl-hydroperoxide and the activity of GPx was calculated from the linear decrease in NADPH absorbance with time [31]. Catalase activity was determined using a commercially available kit (Percipio Biosciences) according to the manufacturer's protocol.

Immunoblot analysis of myocardial protein

Samples of myocardium were homogenized in 10× (w/v) TEE buffer to which Complete™ protease inhibitor cocktail (Roche) was added. Protein was then separated using SDS/PAGE, transferred on to PVDF membrane and subjected to immunoblot analysis using a MitoProfile® OxPhos antibody cocktail (Mitosciences), which recognizes discrete subunits of mitochondrial oxidative phosphorylation complexes I (~20 kDa), II (~30 kDa), III (~39 kDa), IV (~45 kDa) and V (~50 kDa) in the electron transport system, and an anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) antibody(Advanced Immunochemical). For immunoblot analysis of aldehyde protein adducts, monoclonal antibodies against HHE adducts (Genox) and HNE adducts (Percipio Biosciences) were used. The densitometry of immunoblots was performed using ImageJ software (National Institutes of Health). Quantification of total HHE and HNE adduct levels were normalized to GAPDH for each group.

Real-time quantitative PCR of Nrf2-mediated gene expression

Total RNA from myocardial samples was extracted using the RNeasy Micro Kit (Qiagen). Reverse transcription and the mRNA content of a cassette of Nrf2-regulated genes was determined by fluorescence-based real-time PCR using SsoAdvanced™ SYBR® Green Supermix (Bio-Rad Laboratories). Forward and reverse primer pairs for GSTA1 (glutathione transferase A1), GCLC (γ-glutamylcysteine ligase catalytic subunit), NQO1 [NAD(P)H:quinone oxidoreductase 1] and HO-1 (haem oxygenase-1; HUGO Gene Nomenclature Committee symbol HMOX1) were identical with those published previously [33]. β-Actin was used as a control mRNA in these experiments to normalize data obtained with Nrf2 target genes.


Results are presented as means±S.E.M. Data were normally distributed which allowed for the use of parametric statistics. Therefore interval variables between groups were compared using a two-tailed Student's t test (a 3 week intervention) or one-way ANOVA with post-hoc Dunnett's t test for significance (GraphPad Prism), as appropriate. Differences between groups was considered statistically significant for P<0.05.


The body weights were identical between mice fed the normal control diet and mice fed the n−3 PUFA or HFC diets at 3 weeks, but, by 14 weeks, animals had gained more weight on the n−3 PUFA diet than the control diet (Table 2). Cardiac mass was not significantly affected by diet (results not shown). In a previous study using these same cohorts of mice, we demonstrated that levels of n−3 PUFAs in tissues were substantially increased as a result of this diet [22], so we have not reported the levels in the present paper.

View this table:
Table 2 Body weight values of mice

Values are means±S.E.M. (n=6); *P<0.05.

Time-dependent effects of the n−3 PUFA diet on mO2, mH2O2 and mCa2+ in mouse heart

To keep cardiac mitochondria permanently in a submaximal phosphorylating state (a most physiological state), we used 100 μM ADP+glucose/hexokinase in our mH2O2 and mCa2+ experiments [26]. Mitochondria in permeabilized cardiac myofibres from mice fed on an n−3 PUFA diet for 3 weeks showed no significant differences in mH2O2 when supported by pyruvate, malate and succinate, or mCa2+ when supported by pyruvate and malate (Figures 1A and 1B, left-hand panels). However, under identical experimental conditions, cardiac mitochondria from mice fed on an n−3 PUFA diet for 14 weeks showed significant changes, reflected by decreased rates of mH2O2 (Figure 1A, right-hand panel) and increased mCa2+-retention capacity (Figure 1B, right-hand panel). Both basal and maximal ADP-stimulated rates of mO2 supported by pyruvate, malate, glutamate and succinate in permeabilized cardiac myofibres from mice fed on an n−3 PUFA diet for both 3 and 14 weeks were not significantly increased, although a directional trend was observed (Table 3). Basal and maximal ADP-stimulated mO2 supported by palmitoyl-L-carnitine and malate were also unchanged. Importantly, no significant differences were observed in the RCR (respiratory control ratio), an indicator of mitochondrial coupling, between mice fed on the control and n−3 PUFA diets at either 3 or 14 weeks (Table 3), regardless of whether pyruvate/malate or palmitoyl-L-carnitine were used. Furthermore, immunoblot analysis of complexes I–V in the mitochondrial electron transport system from mice fed the control and n−3 PUFA diets for 3 and 14 weeks showed no differences in protein content (results not shown). To establish that the differences observed in cardiac mH2O2 and mCa2+ after 14 weeks of the n−3 PUFA diet were due to n−3 PUFAs and not simply a reflection of a high-fat diet, an additional set of mice fed on an HFC diet comprising a mixture of fatty acids was studied under identical experimental conditions. Mice fed the HFC diet for 14 weeks were found to have no significant differences compared with the normal control diet in any parameter of cardiac mitochondrial function measured (Figure 1 and Table 3).

Figure 1 Dietary n−3 PUFA intake causes time-dependent adaptations in mitochondrial function in mouse hearts, resulting in decreased mH2O2 emission (A) and increased tolerance to mCa2+ overload (B) by 14 weeks

Values are means±S.E.M. with n=7 for the 3 week diet and n=5 for the 14 week diet. *P<0.05 compared with the control diet.

View this table:
Table 3 Diet effects on mitochondrial O2 consumption in permeablized left ventricle myofibres supported by a wide array of substrates

Values are means±S.E.M. with n=7 for the 3 week n−3 PUFA diet and n=5 for the 14 week n−3 PUFA diet. Rates of respiration are in the presence of 5 mM pyruvate/2 mM malate (P/M); 5 mM pyruvate/2 mM malate+5 mM ADP (P/MADP); 5 mM pyruvate/2 mM malate+5 mM ADP+5 mM glutamate and 5 mM succinate (P/M/G/SADP); 5 mM pyruvate/2 mM malate+5 mM ADP+5 mM glutamate and 5 mM succinate+1 μM rotenone (P/M/G/SADP+Rot); 100 μM palmitoyl-L-carnitine/2 mM malate (PC/M); and 100 μM palmitoyl-L-carnitine/2 mM malate+5 mM ADP (PC/MADP).

Effects of the n−3 PUFA diet on cardiac glutathione and its related enzymes

We speculated that the decreased mH2O2 and increased mCa2+ tolerance observed in the hearts of mice fed on the n−3 PUFA diet might be due to increased antioxidant capacity (e.g. glutathione) in the tissue, as other groups have shown that the rates of mH2O2 efflux and opening of the mPTP can be greatly affected by glutathione status [3436]. No significant differences were observed in levels of total (GSHt) or oxidized (GSSG) glutathione following 3 or 14 weeks of the n−3 PUFA diet, or 14 weeks of the HFC diet (Figures 2A and 2B). However, the cardiac GSH/GSSG ratio was significantly higher in the mice fed on the n−3 PUFA diet for 14 weeks compared with the control mice (Figure 2C). To determine whether the differences in GSH/GSSG ratio in the n−3 PUFA group at 14 weeks were due to altered activity of the enzymes responsible for maintaining this ratio, we measured activities of GPx and GR in cardiac tissue from these mice. As shown in Figure 3(A), no significant differences were observed in cardiac GPx activity at either 3 or 14 weeks, but cardiac GR activity was nearly 2-fold greater in mice fed on the n−3 PUFA diet for 14 weeks, although this was unchanged at 3 weeks (Figure 3B). Both GPx and GR activity remained similar in cardiac tissue from mice fed on the HFC diet for 14 weeks, as compared with the control diet (Figure 3).

Figure 2 Time-dependent effects of dietary n−3 PUFAs on levels of total glutathione (GSHt) (A), oxidized glutathione (GSSG) (B) and the GSH/GSSG ratio (C) in mouse hearts

Values are means±S.E.M. with n=7 for the 3 week diet and n=6 for the 14 week diet. *P<0.05 compared with the control diet.

Figure 3 Rates of GPx (A), GR (B) and catalase (C) activity in mouse hearts after dietary intervention

Values are means±S.E.M. with n=7 for the 3 week diet and n=4–5 for the 14 week diet. *P<0.05 compared with the control diet and **P<0.05 compared with the HFC and control diet for that respective group.

Nrf2-mediated gene expression and enzyme activity following dietary n−3 PUFA intervention

Since a number of glutathione-related enzymes and other antioxidant systems are mediated by Nrf2 [33], we wanted to determine whether the increase in GR following the n−3 PUFA diet may reflect a broad increase in all enzymes in this redox signalling pathway. Catalase, an enzyme whose expression is regulated by Nrf2, had substantially greater activity in hearts of mice fed on the n−3 PUFA diet for 14 weeks (Figure 3C). Furthermore, gene expression analysis of a small cassette of Nrf2-mediated genes revealed that the mice fed on the n−3 PUFA diet had augmented expression of GSTA1 and GCLC, even after 3 weeks of diet, although expression of NQO1 and HO-1 were not significantly changed with the diet in this cohort of animals (Figure 4). Notably, this effect was specific to the n−3 PUFA diet, as the HFC diet did not up-regulate any of these genes or enzymes, compared with the control diet.

Figure 4 Effects of dietary intervention on the Nrf2-mediated genes GSTA1 (A), NQO1 (B), GCLC (C) and HO-1 (D) in mouse hearts

Values are means±S.E.M. of the fold-change in mRNA compared with the control diet for each respective time point (3 week and 14 week). n=7 for the 3 week diet and n=4–5 for the 14 week diet. *P<0.05 compared with the HFC and control diet for that respective group.

Aldehyde stress in myocardium of mice fed on the n−3 PUFA diet

Other groups have shown that electrophilic lipids, such as the α,β-unsaturated aldehydes HNE and HHE, specifically up-regulate glutathione-synthesizing enzymes and other antioxidant systems in cardiovascular tissues via Nrf2 and other transcription factors [17,37,38]. Thus we speculated that the increases in glutathione and other antioxidant enzymes following the n−3 PUFA diet may be due to increased aldehyde stress precipitated by the diet. Interestingly, the levels of HHE adduct formation were substantially higher in hearts from mice fed on the n−3 PUFA diet at 3 weeks compared with the control. At 14 weeks, the levels of HHE adducts in mice fed on the n−3 PUFA diet were still significantly greater than in the mice fed on the control and HFC diets, and no significant differences existed between mice fed on the control and HFC diets (Figures 5A and 5B). Unlike HHE, however, levels of HNE adducts did not significantly change under any of the dietary treatments examined (Figure 5C).

Figure 5 Analysis of protein adducts in mice fed control, HFC and n−3 PUFA diets

Immunoblots of HHE protein adducts in hearts from mice fed on an n−3 PUFA diet are shown in representative blots (A) and quantified by densitometry (B). (C) Quantified levels of HNE protein adducts. Values are means±S.E.M. with n=6 for the 3 week diet and n=4 for the 14 week diet. *P<0.05 compared with the control diet. Ctl, control diet; Hi Fat Ctl, HFC diet.


The therapeutic application of dietary n−3 PUFAs in cardiac disease remains controversial despite years of experimental and clinical work in this area. The origin of much of this controversy can be attributed to the wide disparity in approaches used by investigators with respect to the source of n−3 PUFA (fish compared with flaxseed oil) and the dose required to achieve the observed effects. Even less is known about the time-dependent kinetics of these effects. In this context, the novelty of the present study is that (i) hearts from mice fed on an n−3 PUFA diet have decreased mH2O2 emission and increased Ca2+-retention capacity; (ii) these mitochondrial adaptations are dependent upon time; (iii) these mitochondrial adaptations are accompanied by increased expression and activity of a number of antioxidant and phase 2 detoxification enzymes known to be regulated by Nrf2; and (iv) accumulation of HHE precedes these mitochondrial and redox enzyme adaptations during the course of the diet. These results show a unique and important chronological sequence of biochemical events which serve to establish a novel mechanism underlying n−3 PUFA-mediated cardioprotection, and suggest that products formed from peroxidation of n−3 PUFAs have an important role in this process.

Previous studies directed at identifying mechanisms behind the cardioprotective effect of n−3 PUFAs have often focused on how these long unsaturated fatty acids may be altering plasma membrane composition and electrophysiology [39,40]. Specific focus has also been directed at mitochondrial phospholipid fatty acid composition, particularly cardiolipin, and the impact that this remodelling may have on the biophysical/biochemical interactions between cardiolipin and key proteins involved in mCa2+ uptake and retention capacity. Indeed, an enhanced capacity of cardiac mitochondria to buffer Ca2+ could be expected to lead to cardioprotection in a number of ways, such as prevention of arrhythmia [41], enhanced oxidative phosphorylation and mitochondrial release of ATP [42], and decreased susceptibility to Ca2+-induced opening of the mPTP and cell death [43,44]. Previously it was shown that isolated cardiac mitochondria from fish-oil-treated rats had increased Ca2+-retention capacity before mPTP opening [13,14], findings which support other studies involving global ischaemia/reperfusion injury in ex vivo rat hearts where fish-oil-fed animals had increased recovery of contractile force and decreased infarct sizes compared with chow-fed or n−6 PUFA-fed animals [45,46]. Measurements of cardiolipin composition were not performed in the present study and cannot be excluded as a potential factor that may have contributed towards the effects observed with the n−3 PUFA diet. However, clear mechanistic evidence linking altered cardiolipin structure (e.g. more fatty-acyl unsaturation) to altered mitochondrial protein function and enzyme activity has yet to be demonstrated. Thus we directed the present study towards other potential novel mediators of Ca2+-induced mPTP.

It has been shown that oxidative stress and lipid peroxidation specifically alter key components of the mPTP such as adenine nucleotide translocase [47,48], cyclophilin D [49] and cardiolipin [50,51]. Moreover, glutathione and cell redox balance have been demonstrated, in elegant studies, to be critical for maintaining proper mitochondrial Δψ (membrane potential) [34,52] and Ca2+ dynamics [53] in cardiomyocytes. Importantly, it was previously shown that cardiac glutathione depletion and redox imbalance caused fatal arrhythmias in isolated rat hearts [54], underscoring the importance of redox balance in maintaining cardiac function. In the present study we observed a substantial decrease in mH2O2 emission in hearts from mice fed on an n−3 PUFA diet that occurred simultaneously with increased Ca2+ tolerance, so focus was directed at investigating glutathione and other redox-related enzymes in these hearts. We found that hearts from mice fed on an n−3 PUFA diet at 14 weeks had a significantly higher GSH/GSSG ratio, due in part to the significantly increased activity of GR. In addition to GR, a number of other enzymes whose expression is under the control of Nrf2 [55] were found to be increased in the hearts of mice fed n−3 PUFAs, including catalase, GST and GCLC. The latter two enzymes are both critical for detoxification of aldehydes [18] and synthesis of glutathione [56] respectively. These increased enzyme acitivties may explain, in part, the changes in mH2O2 emission and Ca2+ retention seen in hearts from mice fed n−3 PUFAs [57].

Because of the well-known links between electrophilic lipids and Nrf2, we decided to investigate whether evidence of lipid peroxidation was increased in the mice fed on the n−3 PUFA diet. Owing to their unique and highly unsaturated chemical structure, PUFAs are highly prone to peroxidation, which proceeds through both enzymatic and non-enzymatic pathways. Enzymatic peroxidation of PUFAs, controlled largely by COXs (cyclo-oxygenases) and LOXs (lipoxygenases), are formed during the production of eicosanoids. These pathways have been a well-known and intensely studied component of the systemic immune and inflammatory response for decades [58]. Non-enzymatic peroxidation of PUFAs occurs as well, and increases when levels of ROS/RNS (reactive nitrogen species) increase [59]. The peroxidation of PUFAs results in formation of a number of bioactive intermediates, and these intermediates, if left unchecked, ultimately form chemically stable and highly electrophilic aldehydes called 4-hydroxyalkenals [60]. An aldehyde formed in this manner that has received the most scrutiny from investigators is HNE, a nine-carbon aldehyde formed from peroxidation and degradation of n−6 PUFAs, such as arachidonic acid and linoleic acid [18]. Like many electrophiles, HNE exerts bi-directionality in its effect on cells and tissues. At low subtoxic levels, HNE is cardioprotective and induces up-regulation of many enzymes responsible for ROS de-toxification and cell survival, but it is cytotoxic at high levels [61,62]. Non-enzymatic degradation of lipid peroxidation products derived from n−3 PUFAs (α-linolenic acid, DHA and EPA) results in the formation of HHE, a six-carbon aldehyde with many similarities to HNE [19]. In contrast with HNE, almost nothing is known about HHE and its role in cell signalling and cytotoxicity.

Both HNE and HHE covalently react with proteins to form adducts. In the present study, we have demonstrated for the first time that levels of HHE adducts are increased in hearts of mice fed on an n−3 PUFA diet. Conversely, no significant changes in HNE adducts were observed. These findings are significant because they reveal a potential clue as to the underlying signals driving these adaptations in hearts from mice fed on an n−3 PUFA diet. Specifically, both HNE [15] and HHE [38] have been shown to activate Nrf2 in cardiovascular cell types, leading to increased GR activity and other antioxidant enzymes. It is not surprising that levels of HNE adducts were not different between mice fed on control, n−3 PUFA or HFC diets in the present study because the relative levels of n−6 PUFAs in proportion to the total fatty acid content in each diet was identical. The finding that HHE levels, which were already present at 3 weeks, preceded the up-regulation of antioxidant enzyme activity and mitochondrial adaptations supports a causal role for HHE in this process. In addition to Nrf2, other signalling pathways responsible for these changes in glutathione and antioxidant systems may be a factor here. A comprehensive study by Fukuda and co-workers [17] outlined a novel signalling pathway whereby aldehyde stress in the heart increases glutathione synthesis and enzyme activity through the eIF (eukaryotic initiation factor) 2α/Atf4 (activating transcription factor 4) pathway [17]. It is also important to emphasize that the present study is not the first to associate n−3 PUFA therapy with increased antioxidant enzymes. This is an effect that has been observed in a number of other studies, as well as in various tissue types, including the heart [20,21,63].

To conclude, the findings of the present study provide evidence for a completely novel paradigm regarding the action of n−3 PUFAs in the heart, in that it outlines a time-dependent framework of adaptive processes that occur in myocardium during the course of n−3 PUFA treatment, leading to up-regulation of antioxidant systems, decreased mitochondrial ROS and improved mitochondrial Ca2+ tolerance. Furthermore, it illustrates a role for n−3 PUFA lipid peroxidation and subsequent activation of Nrf2 and other pathways involved in redox homoeostasis as potential mediators of n−3 PUFA cardioprotection.


Ethan Anderson conceptualized the study, performed the mitochondrial function and redox enzyme assays, and wrote the paper. Kathleen Thayne performed the real-time PCR, and the immunoblot and enzyme assays. Mitchel Harris performed the GC analysis of the diet and maintained the diet. Kristen Carraway maintained the dietary intervention and weighed the animals. Saame Raza Shaikh conceptualized the dietary interventions and assisted in the preparation of the paper.


This study was supported, in part, by the National Institutes of Health [grant numbers R21HL098780 (to E.J.A.), R15AT006122 (to S.R.S.).


We thank Benjamin Drew Rockett for his assistance in feeding the mice.

Abbreviations: DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GCLC, γ-glutamylcysteine ligase catalytic subunit; GPx, glutathione peroxidase; GR, glutathione reductase; GSTA1, glutathione transferase A1; HFC, high-fat control; HHE, 4-hydroxyhexenal; HNE, 4-hydroxynonenal; HO-1, (HMOX1), haem oxygenase-1; mCa2+, mitochondrial Ca2+, mH2O2, mitochondrial H2O2; mO2, mitochondrial O2, mPTP, mitochondrial permeability transition pore; NQO1, NAD(P)H:quinone oxidoreductase 1; Nrf2, nuclear factor E2-related factor 2; PUFA, polyunsaturated fatty acid; RCR, respiratory control ratio; ROS, reactive oxygen species


View Abstract