Elsevier

Molecular Aspects of Medicine

Volume 24, Issues 4–5, August–October 2003, Pages 263-272
Molecular Aspects of Medicine

4-Hydroxynonenal from pathology to physiology

https://doi.org/10.1016/S0098-2997(03)00021-9Get rights and content

Abstract

4-Hydroxynonenal is a major product of lipid peroxidation. It was firstly studied under the point of view of its toxicity, as it is an easily diffusable substance, thought to be able to explain the “far damages” seen in conditions of increased lipid peroxı́dation. Really, when used at concentration from 10 μM to 1 mM, usually referred to as high concentrations, the aldehyde is able to produce strong inhibitions of several enzymatic activities. When used, however, at concentration of 1 μM or lower, it displays a lot of activities regarding especially cell multiplication and differentiation. As the concentrations indicated above are usually found in normal tissues, these effects may be considered as physiological. As a low level of lipid peroxidation exists in normal tissues, the aldehyde displays signalling activities in normal cells. Among them, it is to consider the stimulation of neutrophil chemotaxis, the strong activation of plasmamembrane adenylate kinase, the strong activation of membrane phospholipase C, both in hepatocytes and neutrophils, the block in the expression of the oncogene c-myc in human leukemic cells, accompanied by differentiation of the same cells, the effects on the cyclins and the activity of E2F transcription factor, the strong increase of the expression of the gene for procollagen alfa1(I), occurring due to the activation of the c-jun/junkinases/AP-1 pathway. Moreover, it is able to block the activity of the PDGF-beta receptor. The last facts allow to think that a hydroxynonenal pathway works in the production of fibrosis.

Introduction

4-hydroxynonenal (HNE) is a member of the 4-hydroxy-2,3-trans-alkenal series, found among the final aldehydic products of unsaturated fatty acids. It was firstly identified, together with the other members of the series (4-hydroxyhexenal, 4-hydroxyoctenal, 4-hydroxyundecenal, 4,5-dihydroxydecenal) by Schauenstein and Esterbauer group in Graz (Schauenstein et al., 1977). The authors studied the toxicity for different aldehydes and considered HNE as one of the interesting products, able even to produce partial tumor regression when deeply injected into the tumour mass.

My personal approach to this aldehyde arose after the discovery, done practically at the same time in 1965 in my laboratory in Siena (Comporti et al., 1965) and in Recknagel laboratory in Cleveland (Ghoshal and Recknagel, 1965), that CCl4 is a strong stimulator of lipid peroxidation in the liver. The idea to study lipid peroxidation in CCl4 poisoning came to me after having observed a pro-oxidant effect of CCl4 on mitochondrial proteins (Dianzani, 1961). Substances containing thiol groups gave a partial protection against the aggregation of mitochondrial proteins seen in mitochondria isolated from CCl4-treated rats, as well as the same type of aggregation seen when mitochondria from normal liver were incubated at room temperature for 30 min.

In my previous experiments, I had shown that mitochondria isolated from the liver of CCl4-treated rats were swollen and unable to display oxidative phosphorylation (Dianzani, 1954). Moreover, they release cofactors, like diphosphopyridine nucleotides (Dianzani, 1955) and cytochrome c (Dianzani and Viti, 1955) in the incubation medium. In the same 1954 paper, it was also described a damage to the so called “light mitochondria” described by De Duve, now referred to as lysosomes. In fact, I found that after CCl4 treatment, they lost acid phosphatase into the surrounding medium. This result was confirmed by De Duve himself, so he invited me to participate to the meeting on lysosomes he organized in the Ciba Foundation of London in February 1963. I described there the changes I had seen, that were confirmed in the same meeting also by a young english scientist, whose name was Trevor Slater. This started a long term collaboration and friendship.

Another invited speaker was Albert Tappel from California, who reported lysosomal damages in the muscles of vitamin E-defı́cient rabbits. When discussing my results, Tappel observed that the changes I had described were similar to those seen by him. In vitamin E-deficiency, however, there is a strong increase of lipid peroxidation. So, he asked me if I had checked lipid peroxidation in my rats. I was only able to reply about the pro-oxidant effect of CCl4 of mitochondrial proteins.

Shortly after my return in Italy, I moved from Cagliari, where I had done the observations on the pro-oxidant effect of CCl4, to Siena.

I met there Mario Comporti, who had experience about vitamin E-deficiency. So I asked him to check the behaviour of lipid peroxidation after CCl4. We found that it was strongly increased both in vivo and in vitro. A few months later, Trevor Slater did in London experiments upon the action of CCl4 on isolated liver microsomes, and postulated that the strong increase in lipid peroxidation he was also able to seen was due to CCl4 metabolism, generating possibly radical dotCCl3 free radical (Slater, 1966). The definite proof of radical dotCCl3 formation came however several years after, when Trevor used phenylbutylnitrone as a spin trap (Albano et al., 1982). Afterwards, Trevor was also able to show that another CCl4-derived free radical, CCl3Oradical dot2, was formed. Reaction with lipids is prevented in the presence of promethazine, preventing also lipid peroxidation. CCl3Oradical dot2 is therefore the real effector of the stimulation of lipid peroxidation, whereas radical dotCCl3 is mostly concerned in covalent binding with lipids, proteins and nucleic acids, such effects being totally unaffected by promethazine.

In Turin, we were able to separate, by using promethazine or vitamin E-pretreatments, the damages produced in the liver by covalent binding from those produced by lipid peroxidation. Among the last damages, we have to consider the acute cell death as well as the inhibition microsomal glucose-6-phosphatase and galactosyl-transferase, that are totally prevented by antioxidant pretreatments.

Block in protein synthesis, block in the secretion of lipoproteins producing fatty liver, and inhibition of the oxidative chains was, however, not influenced by antioxidant pretreatments, then these effects may be considered as related to covalent binding. Free radicals are supposed to react immediately after their production. So, their short life span cannot explain the “far damage”, i.e. that occurring in sites distant from endoplasmic reticulum. The aldehydes could be responsible for such “far damage”. Our interest was especially centered on the 4-hydroxy-2,3-trans-alkenals. In our experiment we started from 4-hydroxypentenal, a non-natural product that was however rather easy to synthesize. We found that it displays inhibitory activity on several SH-containing enzymes and also on cancer cell proliferation. No anticancer activity was however found in vivo, due to the fact that the aldehyde, when injected in animals, is quickly destroyed by its interaction with SH- or NH2-containing shields.

Trevor and I decided therefore to attract Schauenstein and his group in the collaboration. They accepted, and it was so possible to start a long-term successful work. The first result was the identification of several different carbonyl compounds originating from the lipoperoxidative degradation of microsomal lipids (Esterbauer et al., 1982). The aldehydes belonged to three main classes: (1) saturated compounds; (2) 2,3-trans-unsaturated; (3) 4-hydroxy-2,3-trans-unsaturated. The most represented in the last group, that contained most of toxic compounds, was HNE. Benedetti et al. (1980) were able to identify this substance as the most toxic among the aldehydes coming from the peroxidizing microsomal lipids. They used for separation a paper chromatographic method. Our attention was so especially centered to this aldehyde.

In Turin, we found that it is able to inhibit in vitro a lot of different enzymes and functions (Table 1) when used at concentrations ranging from 10 μM to 1 mM. 10–15 μM were really found in the liver after CCl4 poisoning, so the idea that it could act as a messenger of “far damage” was confirmed (Poli et al., 1985). With concentrations of 1 μM or less, however, we got results that might be defined as physiological. Such concentration are really present in the so called “normal” cell.

Historically, the first one of such “physiological” effects was the demonstration that the aldehyde displays chemotactic activity on neutrophils (Curzio et al., 1982, Curzio et al., 1986). All the members of the 4-hydroxy-2,3-trans-unsaturated series were active. The mechanism of the chemotactic effect was different from that displayed by the bacterial tripeptide, that produces “phagocytic burst” and release of the superoxide anion. No such effects were detectable with HNE. Moreover, cells treated with bacterial tripeptide do not respond to a second stimulation by the same substance, but they still respond to HNE and viceversa. By using tritiated HNE, we were able to show that it becomes fixed to an intracellular target (receptor) with a relatively low Kd. The labelled HNE is displaced by further treatment with excess unlabeled HNE, but is not displaced by the corresponding saturated or not 4-hydroxylated unsaturated aldehydes.

Schaur et al. (1994) have shown that HNE accumulates in the phlogystic sites even in vivo; so there is a substantial proof that its activity is not restricted to in vitro experiments.

A second demonstration that HNE may act as a physiological messenger was given by Paradisi et al., 1985), who found that addition of 1 μM aldehyde to isolated liver plasmamembranes results in 100–200% stimulation of adenylate cyclase activity. We explored deeply the mechanism of this effect. It is well known that adenylate cyclase complex is formed by two types of receptors (Rs stimulatory, and Ri inhibitory), by two types of G proteins (Gs stimulatory and Gi inhibitory) and by the catalytic subunit. By using different inhibitors and stimulators of the single proteins forming adenylate cyclase, we were able to show that the target of HNE is Gi that is strongly inhibited, with resulting predominance of Gs.

Rossi et al. (1988) have found that another key enzyme of cell functions, i.e. inositol phosphatides phospholipase C is activated strongly in the presence of 0, 1–2 μM HNE. Activation of protein kinase C was therefore expected, and this was really found in Genoa by the group of Marinari. The activation was however restricted to one of the isoforms of this enzyme.

Lipid peroxidation is very low in tumours, whose cells contain very small amounts of HNE. We had therefore the idea to study the effects on such cells of HNE, that is not produced by the cells themselves. For this purpose, we used three tumour lines, i.e. the human erythroleukemic K562 strain, the human myeloblastic-promyelocytic HL-60 strain and the murine leukemic MEL strain.

All these cell lines cultivated in vitro have a rather high proliferation rate, are very undifferentiated and have a high expression of the oncogene c-myc. A single treatment with 1 μM HNE depresses cell proliferation, induces differentiation and blocks c-myc expression (Barrera et al., 2000, Barrera et al., 2002).

Further investigations were done on the expression of cyclins regulating cell cycle in HL-60 cells. The expression of cyclins D1, D2 and, to a minor extent, A, i.e. the cyclins promoting the passage from G1 to S phase of the cycle, have been found consistently decreased, whereas the cyclin-dependent kinases (cdk) were unaffected; other cyclins, working in different stages of cell cycle (B, C, E) are also unaffected. Morphologically, the inhibition of cyclins D1, D2 and A expression corresponds to a big increase of the cells in the stages G0–G1. The complex cyclins-cdk produces normally the hyperphosphorylation of the pRb (retinoblastoma) protein, normally binding the transcription factor E2F in the hypophosphorylated form. So, hyperphosphorylation provokes the release from the complex of E2F (in our cells especially E2F1 and E2F4, the other forms of E2F having not being found) (Fig. 1, Fig. 2).

Free E2F binds to members of the DP family and this complex is able to activate the c-myc promoter. In our cells displaying, after HNE treatment, a decrease in cyclin expression, we found also a block in pRb phosphorylation. So, E2F transcription factors remained bound to this protein, their content in the cytosol being strongly decreased. This results in a block in c-myc activation.

Since it is known that several other proteins may regulate cyclins expressions, usually referred to as p15, p16, p18, p19, p21 and p27, we decided to study even their expression after HNE treatments. The first four proteins act by inhibiting the cyclin D/cdk-4/-6 kinases, whereas the inhibitory effect of p21 and p27 is displayed towards several different cdks. The expression of all these proteins was found unaffected in HL-60 cells after HNE treatment, with the only exception of p21, that became activated in late times of the treatment, when the effects on cyclins, pRb and E2F described above were already consistent.

These results strengthen the hypothesis that HNE is involved as a signal at the levels of cell growth and differentiation.

HNE is able even to influence the production of fibrosis in the liver, as well as in other tissues. Parola et al., 1992a, Parola et al., 1992b discovered that liver cirrhosis occurring in the rat liver after long-term CCl4 treatment is partially, but consistently prevented by vitamin E overloading of the animals. In the livers of such rats, a significant reduction in lipid peroxidation was detected. Even the expression of the chemokine TGF β1, that is high in CCl4-treated animals, is strongly depressed in rats protected by vitamin E loading (Parola et al., 1992a, Parola et al., 1992b). Hepatic stellate cells of the liver, that are strongly involved in fı̀brosis, are susceptible to lipid peroxidation stimulated by the addition to their cultures of ascorbate/iron. After this treatment, they show a strongly increased expression of the gene of procollagen alfa 1 (I) that is prevented by preventing lipid peroxidation by addition of antioxidants (Parola et al., 1993). The addition of HNE to the stellate cells (1 μM) also produces a big increase in the procollagen gene expression, as well as of its protein product. In that case, addition of antioxidants remains without any effect.

HNE treated cells display also an increased expression of the AP-1 transcription factor; this is related to an increase in the expression of the c-jun oncogene. It has been shown that HNE becomes bound in the cytosol to c-jun amino-terminal kinases isoforms (JNK-1 and -2), and that the two adducts are quickly transferred inside the nucleus (Parola et al., 1998). The formed AP-1 is essentially a homodimer of c-jun product, at least in the early phases of the treatment, whereas the involvement of c-fos occurs in later stages.

By these experiments, we think to have shown the mechanisms by which lipid peroxidation is involved in procollagen production, that is one of the first steps in fibrosis. Possibly, this is not the only mechanism by which lipid peroxidation influences the development of fibrosis. In fact, a further effect of HNE is the block in the expression of the PDGF-beta receptor. This may produce a decline in hepatic stellate cell proliferation. Moreover, we are studying now the effect of HNE treatment on the formation of connective tissue matrix. The two major metalloproteinases regulating the digestion of the matrix (MMP-1 and MMP-2) seem to be not influenced, whereas a significant increase in the expression of TIMP-1 (tissue inhibitor of metalloproteinases-1) was detected (Zamara et al., 2002).

It is noteworthy that all the effects of HNE so far described occur at least 1 h after HNE addition, with the only exception of the effects on adenylate cyclase and of phospholipase C, that are produced within a few minutes. They are transient, as they disappear after 4–8 h. They can be prolonged for more time if the treatments are repeated several times. Added HNE, however, disappears within 15–30 min, due to the activity of enzymes and the formation of adducts. So, HNE, behaves like a signal for the described effects that do not require further HNE stimulation.

Other authors have described different effects of HNE on several other cell functions (Table 1, Table 2).

What is the final conclusion of these researches? Is HNE a real biological signal working in normal cells? Most of the experiments showing these properties were done in vitro. In my knowledge, the only demonstration of an activity in vivo is that of chemotaxis, reported by Schaur et al. (1994). I consider therefore this demonstration as a very important one.

In order to accept the idea that HNE is a biological signaling substance, we have to accept firstly that lipid peroxidation occurs in normal tissues. In my opinion, this is not difficult to accept. Malonyl-dialdehyde and HNE are both present in normal cells, as well as in plasma, of normal subjects. Moreover, subcellular organelles have life spans shorter than the cells bearing them. Segregation of the old structures within cytolysosomes, where myelin figures, as well as HNE itself, can be easily demonstrated, is a physiological method to remove these remnants. Substances deriving from lipid peroxidation are normally detected in the expired air, and lipohydroxyperoxides can be identified in normal cells and plasma. HNE itself in now easily demonstrable by immunological techniques in normal structures.

It is also well known that oxygen free radicals are normally produced in the mitochondrı́al respiratory chain, as well as in microsomal mono-oxygenase chain. The production site is at the level of FAD, a cofactor that is also present in flavoenzymes. The oxygen free radicals produced during these normal functions can therefore stimulate “physiological” lipid peroxidation.

Of course, all this is speculation, but I start to be more and more convinced that this speculation is solid. So, lipid peroxidation, that we considered at its very beginning as an important mechanism of cell damage, may be also a physiological regulating mechanism in normal cells. The multiplicity of the functions of HNE is not, in my opinion, an intriguing problem. cAMP, the most studied messenger, does the same. The effect depends upon the target, and not by the signal.

References (41)

  • M. Rinaldi et al.

    4-Hydroxynonenal-induced MEL cell differentiation involves PKC activity translocation

    Biochem. Biophys. Res. Commun.

    (2000)
  • S. Spycher et al.

    4-Hydroxy-2,3-trans-nonenal induces transcription and expression of aldose reductase

    Biochem. Biophys. Res. Commun.

    (1996)
  • E. Tamagno et al.

    Oxidative stress increases expression and activity of BACE in NT(2) neurons

    Neurobiol. Dis.

    (2002)
  • E. Zamara et al.

    4-Hydroxynonenal selectively up-regulates TIMP-1 gene expression in stellate cells

    J. Hepatol.

    (2002)
  • E. Albano et al.

    Spin trapping studies on the free-radical products formed by metabolic activation of carbon tetrachloride in rat liver microsomal fractions, isolated hepatocytes and in vivo

    Biochem. J.

    (1982)
  • G. Barrera et al.

    Effects of 4-hydroxynonenal, a product of lipid peroxidation, on cell proliferation and ornithine decarboxylase activity

    Free Radical Res. Commun.

    (1991)
  • G. Barrera et al.

    Cancer Detect. Prevent.

    (2000)
  • F. Cajone et al.

    The action of 4-hydroxynonenal on heat shock gene expression in cultured hepatoma cells

    Free Radical Res. Commun.

    (1989)
  • S. Camandola et al.

    Biogenic 4-hydroxynonenal activates transcription factor AP-1 but not NF-kB in cells of the macrophage lineage

    BioFactors

    (1997)
  • S. Camandola et al.

    The lipid peroxidation product 4-hydroxy-2,3-nonenal increases AP-1-binding activity through caspase activation in neurons

    J. Neurochem.

    (2000)
  • Cited by (122)

    • Mitochondrially-targeted treatment strategies

      2020, Molecular Aspects of Medicine
    View all citing articles on Scopus
    View full text