Proteins modified by aldehydes generated from oxidized lipids accumulate in cells during oxidative stress and are commonly detected in diseased or aged tissue. The mechanisms by which cells remove aldehyde-adducted proteins, however, remain unclear. Here, we report that products of lipid peroxidation such as 4-HNE (4-hydroxynonenal) and acrolein activate autophagy in rat aortic smooth-muscle cells in culture. Exposure to 4-HNE led to the modification of several proteins, as detected by anti-protein–4-HNE antibodies or protein-bound radioactivity in [3H]4-HNE-treated cells. The 4-HNE-modified proteins were gradually removed from cells. The removal of 4-HNE-modified proteins was not affected by the oxidized protein hydrolase inhibitor, acetyl leucine chloromethyl ketone, or lactacystin, although it was significantly decreased by PSI (proteasome inhibitor I), the lysosome/proteasome inhibitor MG-132 (carbobenzoxy-L-leucyl-L-leucyl-leucinal), insulin or the autophagy inhibitor 3-MA (3-methyladenine). Pre-incubation of cells with rapamycin accelerated the removal of 4-HNE-modified proteins. Treatment with 4-HNE, nonenal and acrolein, but not nonanal or POVPC (1-palmitoyl-2-oxovaleroyl phosphatidyl choline), caused a robust increase in LC3-II (microtubule-associated protein 1 light chain 3-II) formation, which was increased also by rapamycin, but prevented by insulin. Electron micrographs of 4-HNE-treated cells showed extensive vacuolization, pinocytic body formation, crescent-shaped phagophores, and multilamellar vesicles. Treatment with 3-MA and MG-132, but not proteasome-specific inhibitors, induced cell death in 4-HNE-treated cells. Collectively, these results show that lipid peroxidation-derived aldehydes stimulate autophagy, which removes aldehyde-modified proteins, and that inhibition of autophagy precipitates cell death in aldehyde-treated cells. Autophagy may be an important mechanism for the survival of arterial smooth-muscle cells under conditions associated with excessive lipid peroxidation.
Oxidative degradation of polyunsaturated lipids generates a variety of bioactive intermediates and end-products. These include lipid hydroperoxides and carbonyl end-products that arise from fragmentation of peroxyl linkages and radical-elimination reactions . In the presence of oxygen, these reactions are autocatalytic and self-sustaining and are terminated only by radical–radical annihilation. Non-radical metastable products such as alkanes, hydroxides or alkyl carbonyls are by-products of such lipid oxidation reactions. The appearance of lipid peroxidation products in live tissue is indicative of unquenched radical reactions and their presence has been documented under a variety of toxicological and pathological states associated with radical injury such as carbon tetrachloride poisoning , Alzheimer's disease  and atherosclerosis .
Reactive products of oxidized lipids could amplify and prolong oxidative stress induced by free radicals. In addition to being footprints of radical presence, lipid peroxidation products may be in themselves mediators of oxidative damage, propagators of tissue injury or triggers of stress signalling . Nonetheless, the specific role of lipid peroxidation products is poorly understood, and it remains unclear whether the association of lipid peroxidation products with diseased tissue is merely indicative of cumulative injury or whether it is in fact reflective or causative of ongoing pathology. Products of lipid peroxidation interfere with signal transduction , energy production [7,8], ion channel function  and cell death pathways . In addition, aldehydes generated from lipid oxidation could interfere with cell function by forming covalent adducts with proteins and thereby disrupt their function.
Proteins modified by lipid peroxidation-derived aldehydes such as malondialdehyde, 4-HNE (4-hydroxynonenal) or POVPC (1-palmitoyl-2-oxovaleroyl phosphatidylcholine) have been detected in animal and human tissues under several pathological conditions including myocardial ischaemia , atherosclerosis [12,13], restenosis , Parkinson's disease  and Alzheimer's disease . Sustained presence of protein–aldehyde adducts in diseased lesions suggests that these products are generated continuously or are associated with cells that have either died, remain uncleared or are otherwise unable to cope with damaged proteins.
Mechanisms responsible for removing proteins modified by lipid peroxidation products are not well understood. As a result, it is difficult to delineate the contribution of such a protein modification to disease severity or progression. Previous studies report conflicting results. In lens epithelial cells, 4-HNE-modified proteins have been shown to be preferentially ubiquitinated and degraded by the lysosome with little or no contribution by the proteasomal pathway of protein degradation . In the kidney of ferric nitrilotriacetate-treated animals, proteasomal degradation has been suggested to be a crucial pathway for the removal of protein–4-HNE adducts . A predominant role for the proteasome in the removal of protein–4-HNE adducts is, however, inconsistent with several studies showing that oxidative stress decreases proteasome activity and that 4-HNE inhibits the proteasome by direct covalent modification [18,19]. In addition, in vitro studies show that 4-HNE-cross-linked proteins inhibit proteasomal activity , suggesting that protein degradation pathways other than the proteasome may be important for the removal of protein–4-HNE adducts. It is likely that different pathways of protein removal are engaged by different cells and that their contribution varies with the extent of lipid peroxidation. The present study was designed to examine the major pathways for the degradation of protein–4-HNE adducts in VSMCs (vascular smooth-muscle cells). These cells make up the medial layer of adult arteries and, under pathological conditions such as arterial restenosis , vasculitis  and atherosclerosis [4,13], accumulate high levels of proteins modified by products of lipid peroxidation.
Reagent 4-HNE, [3H]4-HNE and POVPC were synthesized as described in [22,23]. All other chemicals were obtained from Sigma (St. Louis, MO, U.S.A.) unless otherwise stated. Electrophoresis supplies were purchased from Bio-Rad. Primary antibodies against LC3 (microtubule-associated protein 1 light chain 3) were obtained from MBL International (Woburn, MA, U.S.A.). Polyclonal antibodies against KLH (keyhole-limpet haemocyanin)–HNE (protein–4-HNE) were raised and tested as previously described . ECL® reagents were purchased from GE Healthcare (Amersham Biosciences, Pittsburgh, PA, U.S.A.). The Oxyblot kit was obtained from Chemicon International (Temecula, CA, U.S.A.). ALCK (acetyl leucine chloromethyl ketone) was synthesized as described in .
Cell culture experiments
Primary VSMCs obtained from Sprague–Dawley rat aortic explants were grown in DMEM (Dulbecco's modified Eagle's medium; Life Technologies–Invitrogen) supplemented with 10% (v/v) FBS (foetal bovine serum) and 0.1% streptomycin/penicillin. Only passages 4–12 were used for study. At 95–100% confluency, the culture medium (DMEM containing 10% FBS) was removed and the cells were washed twice with HBSS (Hanks balanced salt solution; pH 7.4) containing 20 mM Hepes, 135 mM NaCl, 5.4 mM KCl, 1.0 mM MgCl2, 2.0 mM CaCl2, 2.0 mM NaH2PO4 and 5.5 mM glucose. To avoid reactions between 4-HNE and nucleophilic constituents of the culture medium (lysine, albumin etc.), cells were treated with 4-HNE in HBSS. After indicated treatments, the cells were washed twice with fresh HBSS and scraped with a rubber policeman in lysis buffer containing 25 mM Hepes (pH 7.0), 1 mM EDTA, 1 mM EGTA, 1% Nonidet P40, 0.1% SDS, 1:100 protease inhibitor cocktail (Sigma) and 1:100 phosphatase inhibitor cocktail (Pierce, Rockford, IL, U.S.A.). The cell suspension was further lysed by sonication, and lysates were centrifuged at 14000 g for 15 min at 4 °C. Total protein concentration was measured using a commercially available kit (Bio-Rad).
Measurement of protein-bound 4-HNE
[3H]4-HNE [50 μM; 2.7×106 cpm (counts per minute) per well] was added to cultured VSMCs. After 30 min, the 4-HNE-containing medium was removed, and the cells were washed extensively with HBSS and were either collected or incubated further in 4-HNE-free culture medium. At the indicated times, cells were collected in 5% (v/v) HClO4 and sonicated, and the protein precipitates were sedimented by centrifugation. Radioactivity in the supernatant was used for HPLC analysis of acid-soluble metabolites (as described in the Supplementary material at http://www.BiochemJ.org/bj/410/bj4100525add.htm). The protein pellet was washed extensively with acetone and solubilized in 0.5 M Tris (pH 8.8) containing 1 mM EDTA and 1% SDS. Protein-bound 4-HNE was measured by scintillation counting and normalized to total protein.
Immunological detection and quantification of protein–4-HNE adducts
SDS/PAGE, Western blotting and slot blotting were performed as previously described . For quantification of protein–4-HNE adducts, protein from tissue homogenate (2.0 μg) was loaded on to the wells of the Bio-Dot SF apparatus (Bio-Rad, Hercules, CA, U.S.A.) and microfiltered through nitrocellulose membranes under vacuum. The nitrocellulose membranes were subjected to standard immunodetection techniques using anti-protein–4-HNE antibodies. Western and slot blots were developed using ECL® plus reagents and a Typhoon 9400 variable mode imager (Amersham Biosciences). Intensity of the immunoreactive bands was quantified using ImageQuant TL software (Amersham Biosciences).
VSMCs were grown on Arclar coverslips (Ladd Research, Williston, VT, U.S.A.) and treated in HBSS with 50 μM 4-HNE for 30 min. The medium was removed and the cells were incubated in culture medium for an additional 2.5 h. The cells were washed with PBS and fixed with 3% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 2 h at room temperature (25 °C). The cells were then post-fixed with 1% osmium tetroxide, sectioned and embedded in LX112 plastic. Ultrathin sections were stained with uranyl acetate and lead citrate, and electron micrographs were taken using a Philips CM10 transmission electron microscope operating at 60 kV (×1000–72000 magnification).
Photomicrography and estimation of cell survival
Cells were treated as indicated, and a Nikon 990 digital camera was used to capture ×10 inverted microscope images. Adherent cells were counted in each photomicrograph and used for cell survival analyses. All treatments were performed at least in triplicate.
Results are reported as means±S.E.M. The unpaired Student's t test was used to compare two treatment groups. A P value of <0.05 was considered statistically significant.
4-HNE metabolism and protein–4-HNE modification in VSMCs
To examine the formation and the removal of protein–4-HNE adducts, VSMCs were exposed to 50 nmol of reagent 4-HNE in 1 ml of HBSS for 30 min. This treatment corresponds to ∼50 nmol of 4-HNE per 500000 cells. After 30 min of incubation, the medium was removed and replaced with 4-HNE-free culture medium. At the indicated times, cells were lysed and protein–4-HNE adducts were measured by immunoblot analysis using anti-protein–4-HNE antibodies. As shown in Figures 1(A) and 1(B), the protein–4-HNE adducts formed 30 min after 4-HNE exposure were almost completely removed within 8 h (P<0.0001). Notably, 4-HNE did not form adducts with proteins during shorter time exposures, indicating that at least 30 min are needed for immunologically detectable increases in protein modification (results not shown). No protein adducts were detected at 4-HNE concentrations <50 μM (Figure 3Bii).
Formation and removal of protein–4-HNE adducts in VSMCs
Antibodies raised against protein–4-HNE adducts recognize both Michael adducts and Schiff bases formed between 4-HNE and protein cysteine, lysine and arginine residues. Therefore, to identify whether Michael adducts of 4-HNE (in which the carbonyl group remains free) are also removed with a similar time course, the proteins recovered from 4-HNE-treated cells were derivatized with DNPH (2,4-dinitrophenylhydrazine), which derivatizes the aldehyde moiety of 4-HNE–Michael adducts to a hydrazone. Modified proteins were detected by immunoblotting with anti-DNP (2,4-dinitrophenol) antibodies. Treatment with 4-HNE led to an increase in protein-bound DNP. Upon removal of 4-HNE, the abundance of DNP adducts decreased with a time course similar to that observed with anti-protein–4-HNE antibodies (Figure 1C).
To confirm further that immunopositive bands were indeed due to 4-HNE-bound proteins, the VSMCs were treated with [3H]4-HNE (2.7×106 cpm per well). The cells were then either collected after 30 min of exposure to 4-HNE or incubated further for the indicated times in 4-HNE-free culture medium. As shown in Figure 1(D), the abundance of 4-HNE-bound proteins, as determined by scintillation counting, decreased with time after [3H]4-HNE exposure (P<0.0001). Some detectable modification persisted after 8 h as indicated by protein-bound radioactivity. The close agreement with tracer studies suggests that immunological measurements reflect valid and quantifiable rates of protein–4-HNE removal. Most of the 4-HNE-modified proteins within the SDS/PAGE resolving range were removed at similar rates (see Supplementary Figure 1 at http://www.BiochemJ.org/bj/410/bj4100525add.htm). Proteins migrating to 250, 150, 80 and 50 kDa that were immunoreactive with anti-protein–4-HNE antibodies had a calculated half-life of 169±16 min. Because most 4-HNE-modified proteins were lost with similar rates, immunological assessments of the rate of removal of protein–HNE adducts appear to reflect the loss of most 4-HNE-modified proteins and are not dominated by one heavily modified protein.
Our previous studies show that VSMCs rapidly metabolize 4-HNE . Hence, the rate of loss of protein adducts after bolus exposure to 4-HNE could be confounded by HNE metabolites that could persistently and recurrently induce new modifications. To determine whether metabolites contribute to adduct formation, VSMCs were treated with [3H]4-HNE for 30 min and the medium was collected for analysis of 4-HNE metabolites by HPLC (see Supplementary Figure 2A at http://www.BiochemJ.org/bj/410/bj4100525add.htm). Approx. 60% of the radioactivity in the medium was assigned to unmetabolized reagent 4-HNE. Other metabolites found in the HBSS medium were: GS-X (glutathione conjugates of 4-HNE; ∼8%), DHN (dihydroxynonene; ∼5%), 4-HNA (4-hydroxynonenoic acid; ∼12%) and an unidentified metabolite (∼13%) eluting with a τR (retention time) of 48–49 min (see Supplementary Table 1 at http://www.BiochemJ.org/bj/410/bj4100525add.htm). To examine metabolites that remained inside cells, cells were collected in HClO4 and sedimented by centrifugation. Acid-soluble metabolites in the supernatant were separated by HPLC. The major intracellular metabolite after 0.5 h of exposure to 4-HNE eluted with the τR of a glutathione conjugate (Supplementary Figure 2B). Significant radioactivity recovered from the extracellular medium 4 and 8 h after exposure to 4-HNE was also due to the glutathione conjugates of HNE (Supplementary Table 1). Collectively, these results show that 4-HNE metabolites and unmetabolized 4-HNE are rapidly extruded from VSMCs. Glutathione conjugates are the major intracellular metabolites and, although extruded more slowly, are quantitatively recovered in the extracellular medium. These observations suggest that protein adducts are formed directly from 4-HNE and are the longest-lived products of 4-HNE.
Removal of protein–4-HNE adducts
In most cells, major pathways for the removal of modified proteins are proteasomal degradation and lysosomal autophagy . In addition, oxidized proteins have also been shown to be degraded by oxidized protein hydrolase . To examine how VSMCs remove protein–4-HNE adducts, we examined each of these pathways. First, cells were treated with MG-132 (the proteasome inhibitor carbobenzoxy-L-leucyl-L-leucyl-leucinal), which inhibits both the lysosomal/autophagic and proteasomal pathways , and the formation and loss of protein–4-HNE adducts was followed by slot-immunoblot analysis. As shown in Figure 2(A), treatment with MG-132 decreased adduct removal, such that 4 h after 4-HNE treatment, significantly higher levels (38.0±6.1%; P<0.02) of protein–4-HNE adducts were observed in treated compared with untreated cells. 4-HNE co-treatment with MG-132 also increased adduct formation by 29.2±4.5% after 30 min of 4-HNE exposure (P<0.05), suggesting that adduct formation and removal may be concurrent processes, initiated simultaneously upon 4-HNE exposure. Because interpretation of the results shown in Figure 2(A) could be confounded due to the presence of MG-132 during the adduct formation phase, MG-132 was added after [3H]4-HNE exposure and protein-bound radioactivity was measured. As shown in Figure 2(A) (inset), MG-132 added to the culture medium after 4-HNE exposure modestly but significantly increased the amount of protein-bound [3H]4-HNE at 4 h, suggesting that MG-132 inhibits removal of protein–4-HNE adducts.
Removal of protein–4-HNE adducts by proteasomal and autophagic pathways
The specific contribution of the proteasome in mediating protein–4-HNE adduct removal in VSMCs was assessed by treatment with the proteasome-selective inhibitors, lactacystin  and Z (benzyloxycarbonyl)-Ile-Glu(OtBu)-Ala-Leu-CHO [PSI (proteasome inhibitor I)] , immediately after exposure of cells to [3H]4-HNE. Treatment with lactacystin did not affect protein–4-HNE removal (Figure 2B), whereas treatment with PSI led to a small, statistically significant, increase in protein-bound radioactivity at 4 h (Figure 2B, inset), suggesting that the proteasome plays a minor role in the degradation of protein–4-HNE adducts. To determine the role of the lysosomal–autophagy pathway, VSMCs were treated with 4-HNE for 30 min and the medium was replaced with culture medium containing the autophagy inhibitor 3-MA (3-methyladenine) . As shown in Figure 2(C), cells treated with 3-MA contained significantly more protein–4-HNE adducts than non-3-MA-treated cells. The relative contribution of autophagy was also examined by measurement of protein-bound [3H]4-HNE. Cells were exposed to [3H]4-HNE followed by treatment with 3-MA. Comparable with immunological measurements, inhibition of autophagy significantly decreased removal of protein–[3H]4-HNE adducts (Figure 2C, inset). In contrast, treatment with the oxidized protein hydrolase inhibitor, ALCK (10 μM), did not affect the abundance of adducts 4 h after 4-HNE treatment (results not shown). Taken together, these results suggest that autophagy plays a significant role in the removal of protein–4-HNE adducts in VSMCs.
Unsaturated aldehydes activate autophagy
We next hypothesized that the electrophilic stress induced by aldehyde exposure stimulates autophagy. To test this, we measured the processing of LC3 in aldehyde-treated VSMCs. Formation of LC3-II (phosphatidylethanolamine-conjugated form) is an essential step in autophagosome formation; the abundance of LC3-II correlates with the number of autophagosomes and is therefore a practical index of autophagic activity in mammalian cells [32,33]. Untreated cells maintained in culture showed detectable levels of LC3-II, indicating baseline autophagy (Figure 3A). 4-HNE treatment caused a robust and significant increase in the abundance of the autophagy-indicative form of LC3 (16 kDa; LC3-II) (Figure 3A). Similar increases in LC3-II were also observed with other aldehydes. Stimulation with the C3 unsaturated aldehyde, acrolein, led to a more profound increase in LC3-II formation than did 4-HNE; however, the saturated phospholipid aldehyde, POVPC, only marginally increased LC3-II formation.
Unsaturated electrophilic aldehydes promote LC3-II formation
Exposure to 4-HNE led to a concentration-dependent increase in LC3-II formation in VSMCs (Figure 3Bi). Appreciable increases in LC3-II were observed at concentrations of 4-HNE ≥50 μM. The concentration dependence of LC3-II formation was similar to that of protein–4-HNE adduct formation (Figure 3Bii). While some protein–4-HNE adducts were also detected in untreated cells, treatment with low concentrations of 4-HNE (<50 μM) led to a decrease in protein–4-HNE adducts. This may be a reflection of increased proteolysis stimulated by low concentrations of 4-HNE. Indeed, it has been shown that, at low concentrations, carbonyls such as 4-HNE activate the proteasome by as much as 400% [34–36].
Because acrolein is more electrophilic than 4-HNE or POVPC, it appears that the ability of the aldehydes to induce autophagy may be dependent on their electrophilicity. A proportional relationship between electrophilicity and autophagy stimulation was supported by a more systematic evaluation of the role of aldehyde electrophilicity in LC3-II activation. As shown in Figure 3(C), the electrophilic C9 aldehyde 4-HNE was more potent in stimulating LC3-II formation than the less electrophilic C9 aldehyde nonenal, which in turn was more active than the least reactive C9 saturated aldehyde nonanal. On the basis of these observations, we conclude that highly electrophilic aldehydes such as 4-HNE and acrolein are likely to be more potent in stimulating autophagosome formation.
Previous studies show that the mTOR (mammalian target of rapamycin) is a negative regulator of autophagy and that inhibition of mTOR by rapamycin activates autophagy , whereas activation of mTOR by insulin decreases autophagy . As shown in Figure 4(A), cells pretreated with rapamycin accumulated less protein–4-HNE adducts after 30 min, and rapamycin pretreatment decreased the abundance of protein–4-HNE adducts after 4 h by 50% when compared with cells treated with 4-HNE alone (P<0.05). Both rapamycin pretreatment and 4-HNE exposure increased LC3-II formation; the combination of the two treatments appeared to have a synergistic effect and increased LC3-II formation by 2.5±0.3-fold (P<0.05; Figure 4B). In contrast, treatment with insulin resulted in a small but statistically significant increase in the amount of adducts retained in the cells at 4 h (P<0.05; Figure 4C), and insulin prevented the formation of LC3-II by 4-HNE (P<0.0005; Figure 4D). Interestingly, when rapamycin was added to the cells at the same time as 4-HNE, no change in adduct removal was observed (results not shown), as opposed to an increase in adduct removal in cells that were pretreated with rapamycin 30 min before 4-HNE treatment (Figure 4A). These results suggest that exposure to 4-HNE results in prompt and early activation of autophagic responses.
Regulation of protein–HNE adduct removal by rapamycin and insulin
Ultrastructural changes induced by 4-HNE
The autophagic programme involves multiple signalling pathways that converge on the formation of autophagosomes. These collect targeted cell constituents and subsequently fuse with lysosomes . While many biochemical markers could be used to follow autophagosome formation, electron microscopy is considered the gold standard for documenting autophagy. Hence, we examined ultrastructural changes in VSMCs upon 4-HNE treatment. As shown in Figures 5(A) and 5(B), 4-HNE-exposed cells showed mild to extensive autophagic vacuolization, phagocytotic/pinocytic body formation and loss or thinning of the glycocalyx. Such changes were not apparent in untreated cells (see Supplementary Figure 4 at http://www.BiochemJ.org/bj/410/bj4100525add.htm). Several 4-HNE-treated cells showed isolation membranes, representing immature phagophores, that were observed near mitochondria (Figure 5C), and mature phagophores and autophagosomes were observed surrounding bulk cytoplasm or organelles such as the mitochondria and rough ER (endoplasmic reticulum) (Figures 5D and 5E). All stages of autophagosome formation were observed. These ranged from early double-membraned vacuoles (Figure 5D), to crescent-shaped phagophores (Figures 5D and 5E) and single-membrane vacuoles containing electron-dense ingested material and late multilamellar membranes (Figure 5D). Cells exposed to 4-HNE also demonstrated mild to extensive membrane ruffling (Figure 5F), indicative of processes leading to the internalization of extracellular cargoes via vesicular intermediates.
Ultrastructure of HNE-treated VSMCs
Inhibition of autophagy causes 4-HNE-induced cell death
The results obtained so far indicated to us that exposure to 4-HNE results in the accumulation of several proteins modified by 4-HNE and that these proteins are gradually and systemically removed by proteasomal proteolysis and autophagy. To determine whether removal of modified proteins is essential for protection against 4-HNE toxicity, we investigated whether inhibiting adduct removal would affect cell viability. As shown in Figure 6(A), cells treated with MG-132 (Figure 6Aii) or 4-HNE (Figure 6Aiii) alone showed no signs of cell death 4 h after treatment; however, treatment with MG-132 triggered significant cell death in 4-HNE-exposed cells (Figure 6Aiv), with only 57.2±7.1% (Figure 6B) of the cells remaining viable after the treatment (P<0.005). In contrast, treatment of the cells with PSI did not induce cell death in 4-HNE-treated cells (see Supplementary Figure 5 at http://www.BiochemJ.org/bj/410/bj4100525add.htm). Treatment of 4-HNE-exposed cells with 3-MA also induced cell death. As shown in Figure 6(C), cells treated with 3-MA (Figure 6Cii) or 4-HNE (Figure 6Ciii) alone showed little to no cell death 4 h after treatment; however, significant cell death occurred in 3-MA-treated cells that were exposed to 4-HNE, with only 63.6±2.5% (Figure 6D) of the cells remaining viable (P<0.005) after treatment.
Inhibition of proteolysis triggers cell death in HNE-treated VSMCs
The present study demonstrates that proteins modified by electrophilic products of lipid peroxidation such as 4-HNE are efficiently removed from VSMCs. We find that unsaturated aldehydes stimulate autophagic responses and that inhibition of autophagy inhibits the removal of protein–4-HNE adducts and induces cell death. These observations indicate that proteins modified by aldehydes, if not removed, are likely to be cytotoxic. To the best of our knowledge, this is the first demonstration of autophagy stimulated by lipid peroxidation products.
Aldehydes generated from oxidized lipids are highly reactive . They induce a variety of stress responses and interfere with cell signalling , metabolism  and function . To protect against their potential cytotoxicity, these aldehydes are rapidly metabolized by several biochemical pathways that include reduction, oxidation and the formation of glutathione conjugates . Nonetheless, when metabolic detoxification is overwhelmed, these aldehydes form covalent adducts with nucleophilic side chains of proteins, particularly cysteine, lysine and histidine residues . As shown by the present study, covalent protein adducts are the most sustained outcome of 4-HNE exposure. Greater than 80% of the radioactivity remaining in [3H]4-HNE-treated cells after 4 h was protein-bound (Supplementary Figure 2). Abundant protein adducts of 4-HNE and related aldehydes have been detected in several tissues under conditions of high oxidative stress associated with cardiovascular [4,11–14,21] and neurodegenerative [15,16] diseases. Several toxicological states are also associated with the accumulation of protein–aldehyde adducts [2,40]. Moreover, as shown in Figure 1, detectable levels of protein–4-HNE adducts are also present in untreated cells, indicating that basal levels of lipid peroxidation in otherwise healthy cells are sufficient to induce protein modifications. Nevertheless, the metabolic fate of aldehyde-modified proteins remains obscure, and the cytotoxic potential of such protein modification reactions has not been assessed. Results of the present study suggest that healthy cells progressively remove 4-HNE-modified proteins and that their accumulation in diseased tissue may be a reflection of metabolic failure resulting from either uncontrolled lipid peroxidation or sustained inhibition of cell processes that remove modified proteins.
We investigated several mechanisms by which protein–4-HNE adducts are removed in VSMCs. Results of these investigations show that ALCK does not prevent protein–4-HNE adduct removal, indicating that degradation via oxidized protein hydrolase is not a significant fate of protein–4-HNE adducts in these cells. Similarly, no inhibition of protein–4-HNE removal was observed with lactacystin, although slight inhibition was observed with PSI, suggesting that proteasome-mediated degradation may be a minor pathway for the removal of proteins modified by 4-HNE. This observation is consistent with previous work showing that 4-HNE inhibits the proteasome and that the proteasome is unable to degrade proteins heavily modified by 4-HNE. Nevertheless, some of the proteins modified by 4-HNE may be removed initially by the proteasome before it is inhibited. Further experiments are required to delineate fully the role of the proteasome in the removal of proteins modified by 4-HNE and related aldehydes. In contrast, inhibition of the lysosomal–autophagy pathway either by MG-132 or by 3-MA led to significantly greater accumulation of protein–4-HNE adducts (Figure 2), indicating that protein–aldehyde adducts may be degraded as part of the autophagic response. In addition to inhibitor data, several other lines of evidence suggest that autophagy is an important mechanism for the removal of protein–4-HNE adducts. These include the observations that several unsaturated aldehydes (4-HNE, nonenal and acrolein) led to robust stimulation of LC3-II formation, which is the first committed step in autophagosome formation and autophagy (Figure 3). Moreover, the removal of protein–4-HNE adducts was accelerated by the autophagy stimulator rapamycin  and decreased by the inhibitor of autophagy insulin  (Figure 4). In 4-HNE-exposed cells, treatment with rapamycin also led to an increase in LC3-II levels and insulin prevented LC3-II formation. Thus inhibition or activation of autophagy was found to exert corresponding effects on protein–4-HNE removal and LC3-II formation. Significantly, 4-HNE-treated cells displayed extensive vacuole formation, double-membrane vacuoles, multilamellar vesicles, crescent-shaped phagophores, membrane blebs, invaginations and ruffling, features that were not observed in untreated cells and are ultrastructural signatures of autophagy (Figure 5). Thus, taken together, this evidence supports the notion that degradation by autophagy is a significant fate of 4-HNE-modified proteins in VSMCs. Removal of 4-HNE-modified proteins may also be due to a general increase in protein turnover. While this possibility could not be excluded by the current data, basal mechanisms of proteolysis are unlikely to be sensitive to 3-MA or insulin or associated with increased LC3-II formation.
Mechanisms by which 4-HNE or protein–4-HNE adducts trigger autophagy remain unclear. It has been suggested that one mechanism by which nutrient starvation induces autophagy is by increasing ROS (reactive oxygen species) generation . In support of this idea, it was demonstrated that lipid-soluble antioxidants such as resveratrol and vitamin E are more potent inhibitors of autophagy than soluble ROS antioxidants such as Tiron and NAC (N-acetylcysteine) . However, neither NAC nor Tiron was able to prevent LC3-II formation in 4-HNE-treated cells (results not shown). Hence, stimulation of LC3-II formation by proteins modified by lipid peroxidation products may be a downstream event in ROS-mediated autophagy. Our data show that several products of lipid peroxidation stimulate LC3-II formation. On a mole-per-mole basis, acrolein, the most reactive member of the α,β-unsaturated aldehyde series, was the most effective, followed by 4-HNE and nonenal. In contrast, the saturated aldehyde nonanal was inactive, whereas the phospholipid aldehyde POVPC was only marginally effective. These observations suggest that the processes that activate the autophagic programme may be dependent on Michael addition-type reactions and that strongly electron-deficient aldehydes that modify and cross-link proteins are more likely to trigger autophagy. Alternatively, products of lipid peroxidation could directly trigger autophagic signalling. However, in our studies, we found little or no free 4-HNE (or its metabolites DHN or 4-HNA) in cells after HNE exposure (Supplementary Figure 2). The glutathione conjugates of 4-HNE were found to persist and could trigger LC3-II formation; however, the concentration or the activity of 4-HNE, acrolein or nonenal conjugates is unlikely to be different enough to cause different levels of LC3-II formation. Finally, 3-MA prevented LC3-II formation after HNE exposure (see Supplementary Figure 3 at http://www.BiochemJ.org/bj/410/bj4100525add.htm), indicating that significant accumulation of protein adducts was required to trigger autophagy. Hence, it appears that electrophile-modified or cross-linked products rather than oxidative damage caused by cytosolic ROS promote autophagy. Our observations are consistent with the idea that protein cross-linking (rather than simple protein modification or free aldehydes) may be one signal common to the activation of autophagy by different products of lipid peroxidation. Nevertheless, further experiments are required to determine whether the accumulation of aldehyde-modified proteins also triggers chaperone-mediated autophagy or whether processes related to macrophagy of entire organelles are activated specifically.
Autophagy is a carefully orchestrated pathway for bulk degradation of protein aggregates or damaged organelles . During starvation, autophagy is employed to recycle proteins to generate energy and cell constituents. Such a response may also be an attempt to establish homoeostasis after potentially lethal cytotoxic insults. In addition, autophagy may be a cellular defence mechanism to remove damaged organelles or to protect against the toxic effects of protein aggregates. That autophagy is a pro-survival mechanism is consistent with several studies suggesting a protective effect of autophagy during atherosclerosis, myocardial ischaemia and diabetes mellitus [27,43]. Furthermore, loss of autophagy has been shown to lead to the accumulation of ubiquitin-positive inclusions in the nervous system, which are pathological indicators of neurodegenerative disease . Interestingly, age-related increases in protein carbonyls in rats correlate with the age-related decline in lysosomal proteolysis , indicating that the process of autophagy may be a primary route for the removal of protein carbonyls that accumulate in aged tissues. The results of the present study are consistent with the idea that autophagy is an attempt to restore homoeostasis and to support survival. Our data show that inhibition of the autophagic pathway induces cell death in 4-HNE-treated cells at concentrations of 4-HNE that were by themselves not lethal (Figure 6). These observations suggest that 4-HNE cytotoxicity can be attributed in part to protein modifications.
Modification of proteins by 4-HNE may be harmful not only because it disrupts the function of the protein but also because it leads to the accumulation of inactive or cross-linked proteins, which must be removed to prevent further toxicity. Hence, the removal of protein or organelle detritus to prevent further spread of injury may be a protective function of autophagy. Yet it appears that autophagic signalling may have additional survival benefits. In our experiments, inhibition of the proteasome by PSI did not induce cell death, even though it prevented the removal of 4-HNE-modified proteins, whereas inhibition of the lysosome–autophagy pathway by 3-MA or MG-132 precipitated cell death. While this may be due to the variable efficacy of the inhibitors used to prevent proteolysis, our data show that only those compounds that could inhibit autophagy proved lethal in cells exposed to 4-HNE. In addition to preventing the removal of damaged proteins, inhibition of autophagy may be attenuating the stimulation of additional pathways required for cell survival. What these pathways may be remains unclear, but possibilities include the modulation of downstream effectors of the PI3K (phosphoinositide 3-kinase)–PKB (protein kinase B; also called Akt) pathway and the dissociation of Beclin-1 from Bcl-2 . Additionally, the accumulation of modified proteins could activate the UPR (unfolded protein response) and trigger cell death pathways. In this regard, it has been recently demonstrated that ER stress and UPR are linked to autophagy [47,48]; hence, the UPR may be triggered in an attempt to inhibit protein synthesis and to activate adaptive stress signalling to remove modified proteins by autophagy.
Stimulation of autophagic responses may be a significant component of the stress response in VSMCs and may be a particularly important determinant of intimal proliferation or the stabilization of atherosclerotic plaques. The formation of myelin figures and severe vacuolization typical of autophagy have been detected by electron microscopy in smooth-muscle cells in the fibrous cap of atherosclerotic plaques . Autophagy has also been shown to be triggered by oxidized LDL (low-density lipoprotein) : 7-ketocholesterol, an oxysterol found in oxidized LDL, promotes protein–4-HNE modification, vacuolization, protein ubiquitination and LC3 conversion in human VSMCs , indicating that the local environment surrounding smooth-muscle cells in atherosclerotic plaques could promote the intracellular generation of protein–aldehyde adducts that may be cleared by autophagy.
In summary, the results of the present study provide evidence that electrophilic products generated by the oxidation of lipids activate autophagic responses in VSMCs and that proteins modified by these aldehydes are removed by autophagy. The removal of protein adducts by autophagy appears to be a protective mechanism, because inhibition of autophagy triggered cell death. On the basis of these observations, we speculate that autophagic responses stimulated by lipid peroxidation products may be a common feature of several conditions where such adducts accumulate, e.g. aging, diabetes and several etiologically unrelated cardiovascular and neurological diseases.
We acknowledge Daniel Riggs for technical assistance with HPLC, and Cathie Caple for assistance with EM. This work was supported in part by the NIH (National Institutes of Health) grants HL55477, HL59378, HL65618, ES11594 and ES11860, Philip Morris USA and an American Heart Association predoctoral fellowship to B. G. H.
acetyl leucine chloromethyl ketone
counts per minute
Dulbecco's modified Eagle's medium
foetal bovine serum
Hanks balanced salt solution
microtubule-associated protein 1 light chain 3
mammalian target of rapamycin
proteasome inhibitor I
reactive oxygen species
unfolded protein response
vascular smooth-muscle cell