Search

Premature senescence of IMR-90 human diploid fibroblasts expressing telomerase (hTERT) establishes after exposure to an acute sublethal concentration of H2O2. We showed herein that p38(MAPK) was phosphorylated after exposure of IMR-90 hTERT cells to H2O2. Selective inhibition of p38(MAPK) activity attenuated the increase in the proportion of cells positive for senescence associated beta-galactosidase activity. We generated a low density DNA array to study gene expression profiles of 240 senescence-related genes. Using this array, p38(MAPK) inhibitor and p38(MAPK) small interferent RNA, we identified several p38(MAPK)-target genes differentially expressed in H2O2-stressed IMR-90 hTERT fibroblasts.

Premature senescence of human fibroblasts is established after exposure to an acute sublethal concentration of H(2)O(2). Overexpression of transforming growth factor-beta1 (TGF-beta1) was shown to be responsible for the appearance of the biomarkers of senescence in these conditions. Other studies have shown that incubation of human fibroblasts with TGF-beta1 leads to overexpression of H(2)O(2). In this work, we show an increased production of H(2)O(2) by human fibroblasts as premature senescence is established after an initial exposure to H(2)O(2).

Premature senescence of human diploid fibroblasts (HDFs) can be induced by exposures to a variety of oxidative stress and DNA damaging agents. In this study we developed a robust model of UVB-induced premature senescence of skin HDFs. After a series of 10 subcytotoxic (non-proapoptotic) exposures to UVB at 250 mJ/cm2, the so-called biomarkers of senescence were markedly expressed: growth arrest, senescence-associated beta-galactosidase activity, senescence-associated gene overexpression, deletion in mitochondrial DNA. A set of 44 stress- and senescence-associated genes were found to be differentially expressed in this model, among which clusterin/apolipoprotein J (apo J) and transforming growth factor-beta1 (TGF-beta1). Transfection of apo J cDNA provided protection against premature senescence-inducing doses of UVB and other stressful agents. Neutralizing antibodies against TGF-beta1 or its receptor II (TbetaRII) sharply attenuated the senescence-associated features, suggesting a role for TGF-beta1 in UVB-induced premature senescence. Both the latent and active forms of TGF-beta1 were increased with time after the last UVB stress. Proteasome inhibition was ruled out as a potential mechanism of UVB-induced stress-induced premature senescence (SIPS). This model represents an alternative in vitro model in photoaging research for screening potential anti-photoaging compounds.

We compared the DNA-binding activity of transcription factors and gene expression patterns in BJ human diploid fibroblasts (HDFs) expressing or not telomerase (hTERT) in stress-induced premature senescence (SIPS). Senescent BJ cells were also studied. Hydrogen peroxide (H2O2)-induced SIPS modulated gene expression in both BJ and hTERT-BJ1 cells. Increased p21(WAF-1) mRNA level was amongst the common gene expression changes in BJ and hTERT-BJ1 cells induced by SIPS. Telomerase expression markedly changed gene expression in non-stressful conditions. Expression patterns of senescent BJ cells partially overlapped those of BJ and hTERT-BJ1 cells in SIPS. The basal levels of DNA-binding activity of NF-kappaB and phosphorylated ATF-2 were different in BJ and hTERT-BJ1 cells. Both cell lines displayed a higher DNA-binding activity of p53 and HIF-1 72 h after H2O2 exposure. Our results indicate that similar mechanisms involving p21(WAF-1) and probably p53 are at work in BJ and hTERT-BJ1 HDFs under H2O2-induced SIPS, suggesting that generalized DNA damage rather than telomere length/telomerase plays a crucial role in H2O2induced SIPS. We propose that H2O2-induced SIPS involves a rearrangement of proliferative and apoptotic pathways. The marked changes in gene expression induced by telomerase suggest that apart from immortalization of HDFs, telomerase also alters the normal cellular functions but does not protect against SIPS., (Copyright 2004 Elsevier Inc.)

Normal human diploid fibroblasts (HDFs) exposed to a single H(2)O(2) subcytotoxic stress display features of premature senescence, termed stress-induced premature senescence (SIPS). In this work, our aim was to study SIPS in Werner syndrome (WS) fibroblasts, derived from a patient with WS, a disease resembling accelerated aging. The subcytotoxic dose for WS fibroblasts was found to be inferior to that of normal HDFs, indicating WS fibroblasts are more sensitive to hydrogen peroxide than normal HDFs. SA beta-gal activity has been shown to occur both in vitro and in vivo, and we studied the proportion of WS cells positive for SA beta-gal. Intriguingly, the percentage of positive cells did not increase with the dose of H(2)O(2) used. Contrary to other HDFs, the DNA-binding activity of p53 in WS fibroblasts did not increase in SIPS. We found, based on our results, that WS fibroblasts feature an altered stress response and do not reach SIPS from H(2)O(2). We suggest that the proportion of cells that in normal HDFs would enter SIPS instead die in WS fibroblasts. Last, we propose that aging derives from a loss of integrity of the chromatin structure, which occurs faster in WS patients.

Peroxiredoxin VI (PrxVI) is a bifunctional enzyme with non-selenium glutathione peroxidase and Ca2+-independent acidic phospholipase A2 activities. We demonstrate that transfection-mediated PrxVI overexpression protects immortalized human WI-38 and murine NIH3T3 fibroblasts against cytotoxic doses of tert-butylhydroperoxide and H2O2. Mutants for either glutathione peroxidase or phospholipase A2 activity show that glutathione peroxidase but not phospholipase A2 activity is required to promote cell survival after stress. Also, ectopic PrxVI overexpression does not protect telomerase-stabilized WI-38 fibroblasts against stress-induced premature senescence.

In this work, stable overexpression of peroxiredoxin VI was generated in WI-38 human diploid fibroblasts using a retrovirus-mediated transfection system. Estimation of cell survival showed that peroxiredoxin VI provides a significant protection against tert-butylhydroperoxide- or UVB-caused cytotoxicity. No protection was found against ethanol- or H(2)O(2)-caused cytotoxicity. These effects are correlated with the known functions of Prx VI.

A stress-induced senescence-like phenotype is induced by exposure of human diploid fibroblasts to subcytotoxic H2O2 stress. Previous studies showed that TGF-beta1 is responsible for the induction of several biomarkers of replicative senescence within 72 h after stress: senescence-like morphology, senescence-associated beta-galactosidase activity, and an increase in the mRNA steady state level of four senescence-associated genes. Other studies showed that the retinoblastoma protein is responsible for the appearance of these biomarkers in the same conditions. Here we show that sustained p38(MAPK) phosphorylation is responsible for both H2O2-induced overexpression of TGF-beta 1 and subsequent TGF-beta 1-induced appearance of these biomarkers. p38(MAPK) phosphorylation is shown to be necessary for a self-sustained TGF-beta 1 overexpression after H2O2 stress through the activation of ATF-2 transcription factor, thereby creating a regulatory loop between sustained p38(MAPK) activation and sustained TGF-beta 1 overexpression after stress. p38(MAPK) activation is also shown to be responsible in part for the growth arrest observed in stress-induced senescence-like phenotype. At 48 h after stress, ATF-2 starts to interact with hypophosphorylated Rb, which allows the biomarkers of stress-induced senescence-like phenotype to appear. This report gives an overall explanation of how a senescence-like phenotype is established after subcytotoxic H2O2 stress.

In this work we show that the steady state mRNA level of TGF-beta1 is significantly increased in UVB-induced premature senescence.

In this work, we show that repeated stresses with UVB (290-320 nm) induce stress-induced premature senescence (SIPS) of skin human diploid fibroblasts (HDFs). HDFs at early cumulative population doublings were exposed three or five times to increasing subcytotoxic doses of UVB with one stress per day. After 2 days of recovery, several biomarkers of replicative senescence were established. First, there was an increase in the proportion of cells positive for senescence-associated beta-galactosidase activity. Second, there was a loss of replicative potential as assessed by a very low level of [3H]-thymidine incorporation. Third, the steady-state level of the mRNA of three senescence-associated genes, i.e. fibronectin, osteonectin and SM22, was increased in HDFs at 72 h after three and five exposures to UVB. In conclusion, these results suggest that it is possible to induce SIPS in HDFs after repeated exposures to subcytotoxic doses of UVB. This model could be used to test whether HDFs in UVB-induced premature senescence are able to promote epithelial cell growth and tumorigenesis in skin, as shown recently with HDFs in H(2)O(2)-induced premature senescence.

To test the involvement of the telomeres in the senescent phenotype, we used telomerase-immortalized human foreskin fibroblasts (hTERT-BJ1). We exposed hTERT-BJ1 and parental BJ cells to either UVB or H(2)O(2) subcytotoxic stress(es). Both cell lines developed biomarkers of replicative senescence: loss of replicative potential, increase in senescence-associated beta-galactosidase activity, typical senescence-like morphology, overexpression of p21(WAF-1) and p16(INK-4a), and decreased level of the hyperphosphorylated form of pRb. Telomere shortening was slightly higher under stress for both BJ and hTERT-BJ1 but still much lower than that reported for other cell lines. We conclude that pathways alternative to telomere shortening must cause the appearance of the senescence phenotype.

Human diploid fibroblasts (HDFs) exposed to subcytotoxic stresses under H2O2, tert-butylhydroperoxide (t-BHP), and ethanol (EtOH) undergo stress-induced premature senescence (SIPS) characterized by many biomarkers of HDFs replicative senescence. Among these biomarkers are a growth arrest, an increase in the senescence-associated beta-galactosidase activity, a senescent morphology, an overexpression of p21waf-1 and the subsequent inability to phosphorylate pRb, the presence of the common 4977-bp mitochondrial deletion, and an increase in the steady-state level of several senescence-associated genes such as apolipoprotein J (apo J). Apo J has been described as a survival gene against cytotoxic stress. In order to study whether apo J would be protective against cytotoxicity SIPS and replicative senescence in human fibroblasts, a full-length complementary deoxyribonucleic acid of apo J was transfected into WI-38 HDFs and SV40-transformed WI-38 HDFs. The overexpression of apo J resulted in an increased cell survival after t-BHP and EtOH stresses at cytotoxic concentrations. In addition, when WI-38 HDFs were exposed to 5 subcytotoxic stresses with EtOH or t-BHP, in conditions that were previously shown to induce SIPS, a lower induction of 2 biomarkers of SIPS was observed in HDFs overexpressing apo J. No effect of apo J overexpression was observed on the proliferative life span of HDFs, even if apo J overexpression triggered osteonectin (SPARC) overexpression, which was shown to decrease the mitogenic potential of platelet-derived growth factor but not of other common growth-inducing conditions. Apo J senescence-related overexpression is proposed to have antiapoptotic rather than antiproliferative effects.

WI-38 human diploid fibroblasts underwent accelerated telomere shortening (490 bp/stress) and growth arrest after exposure to four subcytotoxic 100 microM tert-butylhydroperoxide (t-BHP) stresses, with a stress at every two population doublings (PD). After subcytotoxic 160 microM H2O2 stress or five repeated 30 microM t-BHP stresses along the same PD, respectively a 322 +/- 55 and 380 +/- 129 bp telomere shortening was observed only during the first PD after stress. The percentage of cells resuming proliferation after stress suggests this telomere shortening is due to the number of cell divisions accomplished to reach confluence during the first PD after stress.

Our understanding of the interleukin-1 (IL-1) signaling molecular mechanisms has recently made considerable progress, with the discovery of the IL-1 receptor-associated kinase and the downstream enzymatic cascade that leads to the activation of nuclear factor-kappaB (NF-kappaB). IL-1 signaling and especially NF-kappaB activation are thought to be redox-sensitive, even though the precise nature and the molecular targets of the oxidants/antioxidants involved remain largely unknown. Here, we investigated the possible role of cellular oxidized/reduced glutathione (GSSG/GSH) balance in IL-1 signaling. We describe a quantitative method based on capillary electrophoresis designed to assay both intracellular GSH and GSSG in adhering fibroblasts. This method allows the GSSG/GSH balance to be followed during IL-1 stimulation. Our data show that IL-1 induces rapid and transient oxidation of intracellular glutathione in human fibroblasts. Using various antioxidants, including pyrrolidine dithiocarbamate and curcumin, we were unable to show a direct relationship between this IL-1-induced glutathione oxidation and NF-kappaB activation. Of the five antioxidants tested, only curcumin was able to inhibit IkappaBalpha degradation upstream and, hence, NF-kappaB DNA-binding activity and NF-kappaB-dependent expression of IL-6 downstream.

Sublethal oxidative stresses increase the proportions of human fibroblasts positive for senescence associated beta-galactosidase activity and accelerate the transition in the fibroblast morphotypes characterising fibroblast ageing. Stimulation of fibroblasts with TNF-alpha or IL-1alpha transiently increases the production of reactive oxygen species (ROS) in human fibroblasts. Here we propose that repeated stimulation of WI-38 fibroblasts with TNF-alpha or IL-1alpha can generate enough ROS to accelerate the transition in the fibroblast morphotypes and increase the proportion of cells positive for senescence associated beta-galactosidase activity. The involvement of ROS is suggested by experiments where the stimulation of fibroblasts with TNF-alpha or IL-1alpha are performed in the presence of N-acetylcysteine which increases the intracellular antioxidant potential. It is proposed that the decrease in the proportions of morphotypes I and II, and the increase in the proportions of morphotypes III to VI observed after successive stimulation with TNF-alpha or IL1-alpha is attributed to an increased ROS production occurring during the stimulation.

We tested the long-term effects of sublethal oxidative stresses on replicative senescence. WI-38 human diploid fibroblasts (HDFs) at early cumulative population doublings (CPDs) were exposed to five stresses with 30 microM tert-butylhydroperoxide (t-BHP). After at least 2 d of recovery, the cells developed biomarkers of replicative senescence: loss of replicative potential, increase in senescence-associated beta-galactosidase activity, overexpression of p21(Waf-1/SDI-1/Cip1), and inability to hyperphosphorylate pRb. The level of mRNAs overexpressed in senescent WI-38 or IMR-90 HDFs increased after five stresses with 30 microM t-BHP or a single stress under 450 microM H(2)O(2). These corresponding genes include fibronectin, osteonectin, alpha1(I)-procollagen, apolipoprotein J, SM22, SS9, and GTP-alpha binding protein. The common 4977 bp mitochondrial DNA deletion was detected in WI-38 HDFs at late CPDs and at early CPDs after t-BHP stresses. In conclusion, sublethal oxidative stresses lead HDFs to a state close to replicative senescence.

Ageing of WI-38 fibroblasts in culture was used as a model in order to investigate the evolution and the alteration of the key antioxidant enzyme glutathione peroxidase. The activity of glutathione peroxidase is influenced by the presence of selenium in the culture medium and we have also shown that the specific activity of this enzyme does not decrease during ageing, but rather slightly increases. No alteration could be detected by immunotitration. Also the kinetic parameter Km for tert-butyl hydroperoxide has not changed. However, the heat resistance of the enzyme dramatically decreases with ageing. Dilutions of the enzyme preparations had the same influence on the thermosensitivity of the enzyme. This dilution effect is most probably linked to the dissociation of the enzyme subunits into dimers and monomers. Moreover, the kinetic of thermoinactivation curves are best explained by consecutive reactions of inactivation with an intermediary enzyme form. These observations strongly support the hypothesis that ageing is associated with an increased dissociation constant of the tetrameric glutathione peroxidase leading to an easier dissociation of the enzyme in old cells.

Lipid peroxidation occurs during oxidative stress and leads to the formation of various active compounds. However, controversy remains about its importance in the events leading to cell death. One approach to estimate their role in cell death would be to test the toxicity of oxidative products generated during the stress. In this work, three of these products were incubated with human fibroblasts and their toxicities were compared. The three compounds tested are: linoleic acid peroxide (LOOH), malondialdehyde (MDA) and 4-hydroxynonenal (HNE). Three cellular parameters were assayed: viability, DNA synthesis estimated by thymidine incorporation and protein synthesis measured by leucine incorporation. Protection against cellular damages was also tested adding alpha-tocopherol in the culture medium. The results showed that the peroxide was more toxic than HNE and much more than MDA. The possibility of initiation and propagation of the free radical chain reaction could explain this highest toxicity. The fibroblasts seem to be protected by alpha-tocopherol against LOOH. These effects emphasize the crucial role of this lipophilic antioxidant to protect cells against peroxidation damages.

The protection of human diploid fibroblasts against high oxygen tension was investigated using various combinations of the three major antioxidant enzymes: superoxide dismutase, catalase and glutathione peroxidase. alpha-Tocopherol, a well-known hydrophobic antioxidant, was also tested in combination with the different enzymes. Microinjection of solutions containing different combinations of the three enzymes was compared with the injection of each single enzyme. We observed that the protections given by catalase or superoxide dismutase on the one hand, and by glutathione peroxidase on the other hand, were additive. Surprisingly, the combinations of catalase and superoxide dismutase were less effective than catalase alone and was even toxic at low SOD concentrations. Addition of alpha-tocopherol following the injection of any of the three enzymes was highly beneficial, but the strongest synergistic effect was obtained with glutathione peroxidase. These results stress the importance of membrane protection by alpha-tocopherol and indirectly by glutathione peroxidase. They also showed that any injection leading to the decrease in the O2.- or H2O2 concentration combined with one of these two protectors is very beneficial for the cells probably by decreasing the OH concentration. This is also proven by the very good protective effect obtained with desferrioxamine.

Antioxidant enzymes (catalase, superoxide dismutase and glutathione peroxidase) have been injected into human fibroblasts exposed to 2 atm O2 in order to test if the threshold of oxidative damage versus antioxidant defenses could be modulated and if the damage remains reversible beyond the threshold. Cell damage was estimated by thymidine incorporation and cell survival curves. The proportion of dividing cells, measured by thymidine incorporation, rapidly decreased after O2 incubation: no cells could divide after 15 h of hyperoxia. However, cells incubated for a short time and injected with a high concentration of any of the three enzymes divided like non-oxygen-incubated cells: the enzymes could protect the cells against their loss of division potential. However, when cells were incubated for a longer period and/or when the injected enzyme concentration was lower, cells were either less or not protected and could no longer divide. These results suggest the presence of a threshold for the oxidative damage which cannot be totally repaired and which impairs the cell division; this threshold can, however, be modulated by supplementation of antioxidant enzymes, glutathione peroxidase being the most efficient.

A three-step method for the purification of plasma membranes from WI-38 fibroblasts was developed thus allowing the recovery of 36--44% of the plasma membrane. Except in the case of galactosyltransferase, the activity of the contaminating enzymes was very low. Morphological observations confirm the presence of a homogeneous population of vesicles. Preparations obtained from young and old cell cultures were compared for their enzymatic and protein contents. With ageing the activity of 5-nucleotidase significantly increases whereas that of alkaline phosphodiesterase I decreases. Out of the 26 components detected after sodium dodecyl sulphate polyacrylamide gel electrophoresis, four decreased but only one increased. Cellular ageing seems to fulfil a specific and localized effect on the plasma membrane.

An experimental model was developed in order to study the protective effect of antioxidant molecules. Human diploid WI-38 fibroblasts were cultivated under 2 atm of 95% O2. Antioxidants like alpha-tocopherol or superoxide dismutase (SOD) were added respectively in the culture medium or directly inside the cell through a microinjection technique. With both antioxidant molecules a protection was observed. In the control experiment, cells died within 6 or 8 days depending on the confluency and the malonaldehyde content increased sharply. This model represents a new tool in order to test other antioxidant systems towards an oxidative stress.

WI-38 fibroblasts cultivated in vitro were homogenized and their subcellular organelles analysed by the techniques of differential centrifugation and isopycnic equilibration in density gradient. In these experiments, the assayed enzymes were known to be specifically associated with subcellular components in other cells types. In most cases, their behaviour and properties corresponded with observations made in earlier studies and we could consider them as being representative of the specific subcellular organelles. Some significant differences were observed between young and old fibroblasts. The specific activity of alkaline phosphodiesterase was lower in the old cells whereas for the other enzymes it was identical or higher, especially for the 5'-nucleotidase; also the particulate fractions obtained by differential centrifugation contained more material. After equilibration in density gradient, the average density of the 5'-nucleotidase, alkaline phosphodiesterase and N-acetyl-beta-D-glucosaminidase was less in the old than in the young cells, whereas that of the galactosyltransferase of Golgi apparatus was greater. For mitochondria, endoplasmic reticulum and peroxisomes, the differences observed were small.

Procaine is a local anesthetic, also used in experimental gerontology and has been tested in cultivated human WI-38 fibroblasts. This molecule was found to enhance growth rate and cell densities in actively dividing cultures. As the cells aged, however, this stimulatory effect diminished and finally vanished. In a long term experiment the enhancement of growth of procaine treated cultures was finally replaced by a toxic effect even at low concentration. The amount of the thermolabile enzyme found in phase III cells did not change when procaine was added to the culture medium. In this cellular aging model, procaine behaved like a metabolic stimulator of actively dividing cells but not as an "antiaging" molecule as it is sometimes assumed.

Chemical inhibitors of the different antioxidant enzymes were systematically testet either on purified enzymes of after incubation with human fibroblasts in culture. Inhibition values were obtained for catalase with aminotriazole, for superoxide dismutase with diethyldithiocarbamate, for glutathione peroxidase with mercaptosuccinate, for glutathione reductase with bischloroethylnitrosourea and for glutathione synthesis with buthionine sulfoximine. Viability of cells incubated with these inhibitors was then tested under normal conditions and under high oxygen pressure; the data were correlated with the above-mentioned inhibitory values. Cell viability was particularly affected when the glutathione-related enzymes, especially glutathione peroxidase, were inhibited.

The alteration of superoxide dismutase (SOD) defined as the change occurring in its thermostability and observed in ageing cells, concerned the cytoplasmic but not the mitochondrial enzymes. Altered SOD disappeared if it was incubated in a supernatant from the young cells whereas supernatants from the old cells induced the alteration. The alteration induced on purified SOD was found to be associated with the appearance of thermolabile tetramers. Cytoplasmic SOD tetramers were also observed in the supernatants of the old cells. The addition of NADPH into the incubation medium could reverse this alteration; also when cultivated in the presence of vincamine the alteration, normally present in the old cells, disappears. The alteration of SOD is therefore associated with the formation of tetramers; it is a reversible process influenced by the cytoplasmic composition of the old cells.

Glucose 6-phosphate dehydrogenase has been found to be altered in ageing human fibroblasts in culture. In this article, we have studied the effects of incubation conditions on glucose 6-phosphate dehydrogenase present in isolated cells or homogenates from young and old cells. We show that incubation at 4 degrees C and pH 7.4 induced the appearance of a heat-labile enzyme; this was not the case at pH 6.5. Also NADP+, when present, protects the enzyme from being modified. The kinetics of the modification indicates that the altered form of the enzyme is an intermediary stage between the active and inactive forms. We now have a model system in which the appearance of an altered enzyme can be induced in only a few hours; this model will be used in our next paper to study the mechanism of this alteration.

Differential centrifugation and isopycnic equilibration in density gradients were used to localize glutathione (GSH), glutathione peroxidase and glutathione reductase in the subcellular organelles of WI-38 fibroblasts. GSH was present in all the subcellular fractions, whereas the glutathione peroxidase and reductase activities were restrained to the cytoplasm and the mitochondrial fractions. After equilibration in density gradients, the results showed the presence of GSH, glutathione peroxidase and glutathione reductase in both the cytoplasm and mitochondria. GSH was also located in plasma membranes and probably in peroxisomes, endoplasmic reticulum and lysosomal membranes. Evolution of GSH in ageing fibroblasts showed a sudden increase of its concentration just before cell death. The glutathione peroxidase activity already decreases in the early passages, while the decrease of the glutathione reductase activity was constant and reached a drastic low level at the end of the culture. In conclusion, GSH is probably involved in the cell degeneration associated with ageing but because of its multiple functions and its ubiquitous localization, it is difficult to assert to which extent this metabolite is implicated in the ageing process.

Fibroblasts from hamster embryos cultivated in vitro present the typical ageing process of other fibroblastic lines, but they also suddenly give rise to giant non dividing cells which could be considered to represent terminally differentiated cells [36]. We investigated the latter mechanism, first by showing that microtubules in these cells depolymerized from the centrosome and not from the cell periphery as in other cells; secondly we analysed the structure of the centrosome on serial sections and found a diminished pericentriolar material; finally time lapse sequence studies of cell division confirmed that this process sometimes aborts giving rise to these giant polyploid cells. As a consequence, what first appeared as a differentiation process is in fact the result of an environmental deterioration which probably reaches a critical level thus creating a catastrophic consequence for the cell.

The appearance of a heat-labile glucose 6-phosphate dehydrogenase fraction can be induced by incubation of human fibroblasts at 4 degrees C and pH 7.4. Using this system we first studied the effects of pH and NADP+ on the reversibility of the alteration. In young cells, the induced alteration of glucose 6-phosphate dehydrogenase disappeared if the pH of the medium was lowered to 6.5 or if NADP+ was added. In old cells, the disappearance of the natural heat-labile glucose 6-phosphate dehydrogenase was possible only if both pH 6.5 and NADP+ were present. Secondly, we studied the effects of the alteration and its reversibility on the equilibrium between the enzyme subunits; we showed that alteration and inactivation are linked to a dissociation of the enzyme subunits into inactive monomers. We also proposed a model where the heat-labile enzyme is a transient form between the active dimer and inactive monomer. Our results are in perfect agreement with the theories of post-translational modifications now proposed to explain the presence of the altered enzymes in old cells.