Advertisement

Role of Ceramide in Cellular Senescence (∗)

  • Mark E. Venable
    Affiliations
    Department of Medicine and the Divisions of Geriatrics, Durham, North Carolina 27710
    Search for articles by this author
  • Joanna Y. Lee
    Affiliations
    Department of Medicine and the Divisions of Geriatrics, Durham, North Carolina 27710
    Search for articles by this author
  • Miriam J. Smyth
    Affiliations
    Department of Medicine and the Divisions of Geriatrics, Durham, North Carolina 27710
    Search for articles by this author
  • Alicja Bielawska
    Affiliations
    Hematology/Oncology, Duke University Medical Center, Durham, North Carolina 27710
    Search for articles by this author
  • Lina M. Obeid
    Correspondence
    Recipient of a Clinical Investigator Award from the NIA, National Institutes of Health. To whom correspondence should be addressed: Duke University Medical Center, Dept. of Medicine, Box 3345, Durham, NC 27710. Tel.: 919-684-2541; Fax: 919-681-8253
    Affiliations
    Department of Medicine and the Divisions of Geriatrics, Durham, North Carolina 27710

    GRECC VA, Durham, North Carolina 27710
    Search for articles by this author
  • Author Footnotes
    ∗ This work was supported in part by Grants AG00520 and GM43825 from the NIA, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
      Recently the sphingomyelin cycle, involving the hydrolysis of membrane sphingomyelin by an activated sphingomyelinase to generate ceramide, has emerged as a key pathway in cell differentiation and apoptosis in leukemic and other cell types. Here we investigate a role for this pathway in the senescence of WI-38 human diploid fibroblasts (HDF). We found that endogenous levels of ceramide increased considerably (4-fold) and specifically (compared with other lipids) as cells entered the senescent phase. Investigation of the mechanism of increased ceramide led to the discovery that neutral sphingomyelinase activity is elevated 8-10 fold in senescent cells. There were no changes in sphingomyelinase activity or ceramide levels as HDF entered quiescence following serum withdrawal or contact inhibition. Thus, the activation of the sphingomyelinase/ceramide pathway in HDF is due to senescence and supports the hypotheses that senescence represents a distinct program of cell development that can be differentiated from quiescence. Additional studies disclosed the ability of ceramide to induce a senescent phenotype. Thus, when exogenous ceramide (15 μM) was administered to young WI-38 HDF, it produced endogenous levels comparable to those observed in senescent cells (as determined by metabolic labeling studies). Ceramide concentrations of 10-15 μM inhibited the growth of young HDF and induced a senescent phenotype by its ability to inhibit DNA synthesis and mitogenesis. These concentrations of ceramide also induced retinoblastoma dephosphorylation and inhibited serum-induced AP-1 activation in young HDF, thus recapitulating basic biochemical and molecular changes of senescence. Sphingomyelinase and ceramide may thus be implicated as mediators of cellular senescence.

      INTRODUCTION

      Cellular senescence is defined as the limited capacity of cells to undergo population doublings(
      • Hayflick L.
      ); consequently, cells have a finite life span beyond which they can no longer proliferate. This finite life span correlates with the age of the organism and with the life expectancy of the species from which the cells were obtained; such that the older the age or the shorter the life span, the less the ability of the cells to undergo population doubling(
      • Hayflick L.
      ).
      Several important observations have been made in understanding the senescent phenotype. Senescence is a dominant process as demonstrated by cell fusion experiments demonstrating that the resultant heterokaryons have a finite life span (
      • Norwood T.H.
      • Pendergrass W.R.
      • Sprague C.A.
      • Martin G.M.
      ) and by the presence of factors from senescent cells that inhibit DNA synthesis in young cells (
      • Lumpkin C.K.
      • McClung J.K.
      • Pereira-Smith O.M.
      • Smith J.R.
      ). Senescence also appears to be an “irreversible” process, although it may be overridden by DNA tumor viruses leading to proliferation(
      • Rubelj I.
      • Pereira-Smith O.M.
      ). Several known biochemical parameters of senescence are beginning to shed light on the underlying mechanisms involved in this developmental program. These include lack of c-fos transcription (
      • Seshadri T.
      • Campisi J.
      ) and AP-1 activation(
      • Riabowol K.
      • Schiff J.
      • Gilman M.Z.
      ), presence of the Rb protein in a predominantly dephosphorylated form(
      • Stein G.H.
      • Beeson M.
      • Gordon L.
      ), and the occurrence of several alterations in cell cycle proteins(
      • Stein G.H.
      • Drullinger L.F.
      • Robetorye R.S.
      • Pereira-Smith O.M.
      • Smith J.R.
      ,
      • Dulic V.
      • Drullinger L.F.
      • Lees E.
      • Reed S.I.
      • Stein G.H.
      ,
      • Noda A.
      • Ning Y.
      • Venable S.F.
      • Pereira-Smith O.M.
      • Smith J.R.
      ), leading to inhibition of DNA synthesis and lack of cell cycle progression.
      Little is known about signal transduction pathways in cell senescence and less is known about lipid-mediated signaling pathways in senescence. We have recently demonstrated that senescent cells have a defect in the phospholipase D/diacylglycerol/protein kinase C pathway(
      • Venable M.E.
      • Blobe G.C.
      • Obeid L.M.
      ). Ceramide, a key molecule in sphingolipid metabolism and a candidate second messenger, has been shown to inhibit phospholipase D (
      • Venable M.E.
      • Blobe G.C.
      • Obeid L.M.
      ,
      • Gomez-Munoz A.
      • Martin A.
      • O'Brien L.
      • Brindley D.N.
      ,
      • Nakamura T.
      • Abe A.
      • Balazovich K.J.
      • Wu D.
      • Suchard S.J.
      • Boxer L.A.
      • Shayman J.A.
      ). This may implicate the recently described sphingomyelin cycle in cell senescence. This novel biochemical pathway has been shown to regulate cell differentiation and apoptosis(
      • Hannun Y.A.
      ). In this cycle, sphingomyelin is hydrolyzed by a neutral sphingomyelinase, and ceramide is generated in response to a number of extracellular inducers. Ceramide acts as a second messenger to mediate many of the effects of these inducers(
      • Hannun Y.A.
      ). Here we investigate the role of the sphingomyelinase/ceramide pathway in cellular senescence of fibroblasts, the prototypic model of senescence. We demonstrate that ceramide is a potent inhibitor of growth in young WI-38 HDF.
      The abbreviations used are: HDF
      human diploid fibroblasts
      C6-ceramide
      N-hexanoyl-sphingosine
      PBS
      phosphate-buffered saline
      PD
      population doublings
      Rb
      retinoblastoma
      FBS
      fetal bovine serum
      PDMP
      D-threo-1-phenyl-2-decanoylamine-3-morpholino-1-propanol
      YSS
      young serum-stimulated.
      We also demonstrate that endogenous ceramide levels and a neutral magnesium-dependent sphingomyelinase activity are markedly elevated in senescent cells. We show that exogenous ceramide is able to induce a senescent phenotype in young HDF at concentrations that mimic endogenous levels in senescent cells. This was demonstrated by the ability of ceramide to inhibit DNA synthesis, AP-1 activation, and Rb phosphorylation; all important biochemical markers of senescence. We propose that ceramide is a mediator of cellular senescence.

      EXPERIMENTAL PROCEDURES

       Materials

      D-Erythro-C6-ceramide and N-[14C]hexanoyl-D-erythro-sphingosine (C6-ceramide) were prepared as described previously(
      • Bielawska A.
      • Crane H.M.
      • Liotta D.
      • Obeid L.M.
      • Hannun Y.A.
      ). C6-ceramide was purchased from New England Nuclear as was the [γ32P]ATP. Silica gel 60 thin-layer chromatography plates were from Whatman, and solvents were supplied by Fisher Scientific. NTB-2 bulk emulsion was from Kodak. Anti-Rb antibodies were from Pharmingen, and the biotinylated rabbit anti-mouse IgG1 antibodies were from Amersham Corp., whereas the streptavidin-conjugated horseradish peroxidase and related reagents were from Bio-Rad. [3H]Thymidine was from ICN. D-threo-PDMP was from Matreya, Inc. Other reagents were purchased from Sigma.

       Cell Culture

      Young or senescent WI-38 HDF (NIA Aging Cell Repository catalog number AG06814E) were grown in Dulbecco's modified Eagle's medium with 4.5 g/liter glucose and 10% FBS. Cells were passaged to increasing population doublings until they reached senescence at approximately PD (55). Cells were maintained at 37°C in 5% CO2 and fed two times/week. Routine tests verified that the cells were free of mycoplasma.

       Thymidine Incorporation and Cell Growth

      Young and senescent WI-38 HDF were seeded in 24-well plates at 1.5 × 104 cells/well. Cells were fed after 2 days and then refed 5 days later with fresh 10% FBS-containing medium for 16 h. [3H]Thymidine (0.5 μCi/ml) was added for the last 4 h, the cells were harvested, and the trichloroacetic acid-insoluble pellet was counted. For studies of ceramide's effect on growth and thymidine incorporation, cells were seeded as above and treated on day 3 with the indicated concentrations of C6-ceramide in ethanol vehicle (0.1%). PDMP was added as indicated at the same time as C6-ceramides. Cells were either used for thymidine incorporation studies performed as above or harvested at the indicated times and counted after trypan blue exclusion. Ceramide wash out experiments were performed as above except that after 48 h of incubation with ceramide, the cells were washed with medium containing 10% FCS. This brought ceramide levels back to base line as determined by studies using 14C-labeled C6-ceramide (see below). Thymidine incorporation was determined 24 h later.

       Apoptosis

      Young WI-38 HDF were seeded in 12-well plates at a concentration of 9 × 104 cells/well. After 48 h the cells were incubated with 0.4 μCi/well of [3H]thymidine for 24 h. Cells were washed twice with Dulbecco's modified Eagle's medium containing 10% FCS and incubated with medium containing 2% FCS. Cells were treated with the indicated concentrations of C6-ceramide in ethanol vehicle for 24 h. The medium was aspirated and counted. The cells were lysed with PBS containing 1% Triton X-100 and 0.2 μM EDTA. The cells were centrifuged for 15 min at 14,000 rpm in a microcentrifuge. The supernatant was counted, and the remaining pellet containing larger DNA fragments was counted. The percentage of apoptosis was calculated by adding the counts in the medium and the supernatant and dividing by the total counts(
      • Obeid L.M.
      • Linardic C.M.
      • Karolak L.A.
      • Hannun Y.A.
      ). Cells were treated with C6-ceramide as above for trypan blue exclusion studies except that after 24 h cells were washed with PBS, treated with trypsin, and mixed with trypan blue (0.4%).

       Ceramide Measurements

      WI-38 cells at different PD were seeded at a concentration of 4 × 104 cells/well in 12-well plates. Cells were harvested before confluence and 5 days after the last feeding. The medium was aspirated, and the cells were scraped into 1 ml of methanol. Lipids in the suspension were extracted as described(
      • Bligh E.G.
      • Dyer W.J.
      ), assayed for ceramide (80% of sample) by the diacylglycerol kinase method(
      • Preiss J.
      • Loomis C.R.
      • Bishop W.R.
      • Stein R.
      • Niedel J.E.
      • Bell R.M.
      ), normalized to total cellular phospholipid (20% of sample)(
      • Rouser G.
      • Siakotos A.N.
      • Fleischer S.
      ), and represented as pmol of ceramide/nmol of lipid phosphorous. Exponentially growing cells were harvested 2 days after seeding. Quiescent cells were either serum-deprived for 48 h or allowed to become contact-inhibited in the presence of medium containing 10% FCS.

       Sphingomyelin Mass Measurements

      Cultures of young and senescent WI-38 HDF were washed with PBS scraped into methanol and extracted(
      • Bligh E.G.
      • Dyer W.J.
      ). The solvent was evaporated using a stream of N2. The lipids were resuspended in 1 ml of CHCl3, and a portion of the sample was removed and assayed for lipid phosphorous(
      • Rouser G.
      • Siakotos A.N.
      • Fleischer S.
      ). 2 N methanolic NaOH (100 μl) was added to the remainder of the sample as well as known standards of sphingomyelin treated in parallel. This was incubated at 37°C for 1 h and then neutralized while vortexing with 2 N HCl (100 μl). The reaction mixture was then extracted(
      • Bligh E.G.
      • Dyer W.J.
      ). The solvent was evaporated, and the lipids were resuspended in CHCl3 and chromatographed on silica gel 60 TLC plates (Whatman) in CHCl3:methanol:acetic acid:water (50:30:8:4). Sphingomyelin was visualized with I2 vapors then scraped and quantitated by lipid phosphorous analysis(
      • Rouser G.
      • Siakotos A.N.
      • Fleischer S.
      ).

       Sphingomyelinase Activity Measurement

      WI-38 HDF were prepared as described(
      • Suzuki K.
      ). Cells (approximately 5 × 106 cells) were washed with PBS, scraped, centrifuged (300 × g for 5 min), and resuspended in 20 μM Tris-HCl (pH 7.4), 1 μM ATP, 5 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 1 μg/ml pepstatin, 12 μg/ml leupeptin, 10 μg/ml soybean trypsin inhibitor, 21 μg/ml Nα-tosyl-L-phenylalanine chloromethyl ketone, 21 μg/ml N-tosyl-L-lysine chloromethyl ketone. Cells were broken (greater than 95%) by four cycles of freezing and thawing, and protein was quantitated by the Bio-Rad microassay. N-[14C]Methylsphingomyelin (130,000 dpm, 20 nmol in chloroform:methanol (9:1)) was prepared as described (
      • Stoffel W.
      • LeKim D.
      • Tschung T.S.
      ) and added with Triton X-100 (Sigma, 100 μg in chloroform:methanol (9:1)) to each polypropylene assay tube. The solvents were evaporated using N2, and the substrate was resuspended in 0.2 M sodium acetate (pH 5.0, “acid sphingomyelinase”) or 0.2 M Tris-HCl (pH 7.4, “neutral sphingomyelinase”) with or without 10 μM MgCl2 with vortexing (
      • Jayadev S.
      • Linardic C.M.
      • Hannun Y.A.
      ). Cell homogenates were added to the substrate, and the volume was adjusted to 200 μl with the respective buffer. Incubations proceeded at 37°C for 30 min and were stopped with 1.5 ml of chloroform:methanol (1:2). Water soluble [14C]phosphorylcholine product was separated from lipid substrate by extraction(
      • Folch J.
      • Lees M.
      • Stanley G.H.S.
      ), and 50% of each phase was dried in an 80°C oven and quantified by liquid scintillation spectrometry. Neutral magnesium-dependent activity was calculated by subtracting dpm from neutral activity in the presence of magnesium from those in the absence of magnesium.

       Metabolic Labeling Studies

      WI-38 cells were seeded at a density of 9.0 × 104 cells/well in 12-well plates. Cells were treated with 15 μM (120 μCi/mmol) [14C]hexanoyl]-C6-D-erythro-ceramide. After the indicated time the medium was removed, and the cells were washed two times with PBS and scraped into 1 ml of PBS. The cells were sonicated and then centrifuged in a TLA-100.3 rotor for 1.5 h at 40,000 rpm. Pellets were resuspended in 1 ml of PBS. Counts were obtained from 10% of each of medium, cytosol, and membrane. Lipids were extracted from the remaining 90% of the different fractions. 10% of the lipids were spotted on Silica Gel 60 (Whatman) thin-layer chromatography plates and resolved in chloroform:methanol (80:20). Metabolites were located by autoradiography, scraped, and quantitated by liquid scintillation counting. Samples (20%) were also analyzed for total phospholipid phosphate. For ceramide wash out experiments, [14C]C6-ceramide was washed out after 48 h with medium containing 10% FCS. The cells were then placed in this medium for 24 h and extracted as described above.

       Labeling Indices

      WI-38 HDF were grown to 50% confluence. Cells were then growth arrested by serum deprivation in the presence or the absence of the indicated concentrations of C6-ceramide. Cells were then stimulated with medium containing 10% FBS in the presence of 40 μCi/ml of [3H]thymidine and in the continued presence of C6-ceramide in ethanol vehicle where indicated. At 48 h, cells were fixed with acetic acid:methanol (1:3). Cells were coated with NTB-2 bulk emulsion (Kodak), exposed for 3 days at 4°C, developed, and stained with Wright-Giemsa exactly as described (
      • Afshari C.A.
      • Barrett J.C.
      ).

       AP-1 Gel Shift Analysis

      Young or senescent WI-38 HDF were grown to preconfluence. The cells were then placed in 0.1% FBS-containing medium for 48 h. Ethanol vehicle or ceramide (10 μM) was added to the young cells for that time. The cells were then either kept in the same medium or stimulated with 10% FBS-containing medium for 30 min. The cells were harvested and the nuclear extracts were obtained exactly as described(
      • Dbaibo G.
      • Obeid L.M.
      • Hannun Y.A.
      ). Gel shift assays were performed utilizing either the labeled AP-1 oligonucleotide 5′-CGCTTGATGAGTCAGCCGGAA-3′ (Promega) or the labeled oligonucleotide 5′-CGCTTGATGACTTGGCCGGAA-3′ with a mutated AP-1 site (Santa Cruz Biotechnology) as described by the manufacturer. Where indicated an antibody for c-Fos (Santa Cruz Biotechnology catalog number sc-52) was used in the gel shift reaction.

       Rb Western Blots

      WI-38 cells (PD 20) were grown to late log phase (except for exponentially growing young cells and senescent cells, which were harvested 2 days after refeeding) on 100 × 20-mm tissue culture dishes in Dulbecco's modified Eagle's medium supplemented with 10% FBS. For determination of dose response of ceramide inhibition of Rb phosphorylation, cells were serum-deprived by incubation for an additional 48 h in medium containing 0.1% FBS with or without the indicated ceramide concentration. Where indicated the cells were restimulated with medium containing 10% FBS and the appropriate ceramide concentration for 20 h. For determination of the time course of ceramide-induced Rb dephosphorylation, C6-ceramide (15 μM) was added to randomly proliferating cells. Cells were rinsed in PBS and lysed in hot solubilizing buffer. Proteins were fractionated on 6% gels by SDS-polyacrylamide gel electrophoresis, immunoblotted, and detected with anti-Rb antibodies, biotinylated rabbit anti-mouse IgG1, and streptavidin-conjugated horseradish peroxidase(
      • Chao R.
      • Khan W.
      • Hannun Y.A.
      ).

      RESULTS

      We utilized WI-38 HDF at various PD. Young HDF (PD 20-30) were able to grow rapidly in culture as indicated by their ability to incorporate thymidine in response to serum stimulation into newly synthesized DNA. Senescent cells (PD 55-60) were unable to proliferate as indicated by their inability to grow even at 3 weeks after seeding and by their inability to incorporate significant [3H]thymidine into DNA in response to serum (Fig. 1A).
      Figure thumbnail gr1
      Fig. 1Ceramide inhibits growth of young WI-38 HDF and induces cell death. A, young WI-38 HDF, unlike their senescent counterparts, are able to proliferate as indicated by their ability to undergo [3H]thymidine incorporation into newly synthesized DNA in response to serum stimulation. Experiments were performed in triplicate, and the data are representative of three experiments. B, growth of young WI-38 HDF is progressively inhibited by D-erythro-C6-ceramide (5 (○), 10 (•), 20 (Δ), or 30 μM (▴)). Growth is expressed as the percentage of control. The data are from single point determinations and are representative of three experiments. C, top panel, apoptosis of young HDF is induced by high concentrations of C6-ceramide. The data are from one experiment performed in duplicate and are representative of at least two experiments. ∗, p < 0.05; ∗∗, p < 0.001. Bottom panel, cell death was quantitated by trypan blue staining as described under “Experimental Procedures.” The data are from one experiment performed in triplicate and are representative of two experiments.
      Ceramide potently inhibited growth of WI-38 HDF. Young, rapidly proliferating cells grown in the presence of D-erythro-C6-ceramide (a cell-permeable ceramide), at concentrations of 5-15 μM, underwent complete growth arrest analogous to senescent cells (Fig. 1B). High concentrations of C6-ceramide appeared cytotoxic, whereby concentrations of greater than 20 μM induced apoptosis as measured by thymidine release (Fig. 1C, top) and as confirmed by trypan blue uptake (Fig. 1C, bottom).
      Because ceramide was able to cause growth arrest in WI-38 HDF and induce terminal differentiation in HL-60 cells (
      • Okazaki T.
      • Bielawska A.
      • Bell R.M.
      • Hannun Y.A.
      ) and because senescent cells appear to behave similar to growth-arrested and terminally differentiated cells(
      • Peacocke M.
      • Campisi J.
      ), we investigated the hypothesis that sphingomyelinase and endogenous ceramide may play a role in inducing cell senescence. Initially, we measured ceramide levels in young, old, and senescent WI-38 HDF and correlated ceramide levels to PD elapsed. Ceramide levels remained stable with increasing population doublings as long as the cells were able to proliferate. The capacity to proliferate may also be quantitated as the number of PD remaining. In the experiments shown in Fig. 2, HDF demonstrated a finite PD of 61 after which cells remained metabolically active but failed to grow or take up thymidine. Fig. 2A shows that ceramide levels in young cells (40 PD remaining) were 3.5 ± 0.05 pmol/nmol lipid phosphate. As the PD remaining decreased to 20, ceramide levels remained unchanged. However, when PD remaining decreased further, ceramide levels began to increase to 6.0 ± 0.5 pmol/nmol lipid phosphate. When all cells became senescent, ceramide increased to 14.7 ± 0.8 pmol/nmol lipid phosphate or 4.2-fold over young cells. On the other hand, diacylglycerol levels measured concurrently with ceramide levels increased only modestly with cellular senescence (Fig. 2A). These levels are corrected for total membrane phospholipid content and therefore reflect specific changes in ceramide. In addition, sphingomyelin mass measurements were performed on young and senescent Wi38 HDF and were 58.6 ± 2.9 and 42.5 ± 2.9 pmol/nmol lipid phosphate, respectively.
      Figure thumbnail gr2
      Fig. 2Endogenous ceramide levels and neutral sphingomyelinase activity increase in cell senescence but not in quiescence. Cells at increasing population doublings were assayed for levels of endogenous ceramide, diacylglycerol, neutral, and lysosomal acidic sphingomyelinase activity. A, ceramide levels increase slowly with increasing population doublings and peak at senescence (represented by the last point). This was attained when cells showed no growth for four weeks (i.e. no remaining PD) and did not incorporate [3H]thymidine (see A). In contrast, diacylglycerol levels increased only modestly with cellular senescence. The experiment shown was performed in duplicate and is representative of three separate experiments. B, neutral magnesium-dependent sphingomyelinase activity is significantly increased in senescent HDF whereas acidic sphingomyelinase activity appears relatively unchanged. The data represent the means ± standard error of two duplicate experiments. C, ceramide levels do not change significantly between exponentially growing (EG), quiescent serum-deprived (SD), or contact-inhibited (CI) HDF. The data for EG and SD represent the mean ± S.E. from three separate experiments performed in duplicate. CI represents the mean ± range from one experiment performed in duplicate. D, neutral sphingomyelinase activity remains unchanged in EG, SD, or CI HDF. Blank, no enzyme. The data represent the mean ± range from one experiment performed in duplicate.
      A neutral sphingomyelinase has been implicated in the generation of ceramide in response to inducers of differentiation and growth arrest (
      • Okazaki T.
      • Bell R.M.
      • Hannun Y.A.
      ,
      • Okazaki T.
      • Bielawska A.
      • Domae N.
      • Bell R.M.
      • Hannun Y.A.
      ,
      • Chatterjee S.
      ). Indeed, we demonstrate an 8-fold increase in neutral magnesium-dependent sphingomyelinase activity in protein extracts from senescent cells (Fig. 2B) as compared with young HDF. In contrast, acid sphingomyelinase, a well characterized lysosomal enzyme(
      • Spence M.W.
      ), had higher base-line activity in young HDF but increased only modestly in senescent extracts (Fig. 2B). These remarkable increases in ceramide levels, in ceramide:diacylglycerol ratio, and in neutral sphingomyelinase activity were even more marked than those observed in cell differentiation(
      • Okazaki T.
      • Bielawska A.
      • Bell R.M.
      • Hannun Y.A.
      ). They also represent permanent (stable and prolonged) increases in ceramide and sphingomyelinase activity as opposed to the transient signaling increases in response to inducers of apoptosis or differentiation such as tumor necrosis factor α (
      • Kim M.-Y.
      • Linardic C.
      • Obeid L.
      • Hannun Y.
      ) and dihydroxyvitamin D3(
      • Okazaki T.
      • Bielawska A.
      • Bell R.M.
      • Hannun Y.A.
      ).
      In order to evaluate if those increases in ceramide levels and sphingomyelinase activity were specific to senescence or were a consequence of growth arrest, we performed the same measurements in exponentially growing and in quiescent HDF. Exponentially growing cells were harvested 2 days after seeding, whereas quiescent cells were rendered so either by 48 h of incubation in medium containing 0.1% FBS or by contact inhibition in the presence of 10% FCS and then harvested. The conditions for quiescence were determined as follows: WI-38 cells were seeded at 1 × 106 cells in 10-cm plates. 2 days later the cells were placed in medium containing 0.1% FBS for 48 h. The cells were then harvested for determination of either their labeling index or for flow cytometry studies as described (
      • Walker P.R.
      • Kwast-Welfeld J.
      • Gourdeau H.
      • Leblanc J.
      • Neugebauer W.
      • Sikorska M.
      ) using a Becton Dickinson FACStarPlus flow cytometer (San Jose, CA). The cells showed a labeling index of <2%, and flow studies demonstrated that 97% of the cells were in the G1 phase of the cell cycle. Cells allowed to become quiescent by contact inhibition in the presence of 10% FCS demonstrated a similar cell cycle profile. Upon determination of ceramide levels (Fig. 2C), we demonstrated that there was no difference between quiescent and exponentially growing young HDF; also there was no difference in neutral sphingomyelinase activity between quiescent and exponentially growing young HDF (Fig. 2D). Therefore, the changes in sphingomyelinase and ceramide seen were specific to senescence and not a mere consequence of growth arrest.
      These increases in ceramide levels and sphingomyelinase activity coupled with the inability of senescent cells to respond to mitogenic stimuli led us to investigate whether ceramide imparts the senescent phenotype to young HDF. We examined several biological and biochemical parameters of senescence. Senescent cells, unlike young quiescent cells, are unable to undergo DNA synthesis and are unable to divide in response to serum or growth factor stimulation (Fig. 1)(
      • Peacocke M.
      • Campisi J.
      ). In order to mimic the persistent elevation of endogenous ceramide in senescent cells, young HDF were treated with concentrations of exogenous C6-ceramide for 1-3 days, which resulted in a persistent elevation in cellular ceramide levels comparable to those seen in senescent cells. This was demonstrated by uptake and metabolism studies utilizing [14C]hexanoyl-sphingosine (15 μM) (Table 1) or hexanoyl-[3H]sphingosine (not shown). The cells took up 16% of the label within 1 h. By 24 and 48 h, 44 and 63%, respectively, of the C6-ceramide radiolabel (Rf 0.73 in chloroform:methanol (80:20)) was metabolized to two predominant compounds, Rf = 0.37 and 0, which were observed first as cell-associated but then accumulated in the medium. These products have not been identified but do not appear to be the simple sphingosine, triglyceride, sphingomyelin, or glycosphingolipid metabolites. Membrane-associated radiolabel was 35% ceramide at 24 h and only 26% by 48 h. To compare effective added C6-ceramide levels, we calculated the ratio of C6-ceramide to total lipid. At 24 h, cell membranes contained 22 pmol of ceramide/nmol of lipid phosphate. This decreased to 18 pmol/nmol of lipid phosphate by 48 h. These levels closely resemble endogenous ceramide levels found in senescent fibroblasts (see Fig. 2A).
      We next elected to evaluate the ability of ceramide to impart senescence by evaluating inhibition of DNA synthesis, a known biological parameter of senescence. C6-ceramide (5-20 μM) treatment for 24 h progressively inhibited [3H]thymidine incorporation in young HDF in response to serum (Fig. 3A). There was no difference in the degree of inhibition of new DNA synthesis if the cells were kept in the presence of ceramide for up to 60 h (data not shown). These data demonstrate that ceramide is a potent inhibitor of DNA synthesis in young HDF. When C6-ceramide was washed out with 10% FCS-containing medium, intracellular ceramide levels returned to base line as documented by studies using 14C-labeled C6-ceramide, and the cells were able to resume growth and undergo DNA synthesis at the same rate as control cells (data not shown). This raised the intriguing possibility that senescence may be reversible if ceramide levels can be brought down to levels seen in young cells.
      Figure thumbnail gr3
      Fig. 3Ceramide inhibits DNA synthesis. Young WI-38 HDF were treated for 24 h with C6-ceramide (5-20 μM) in the absence or presence of PDMP (20 or 50 μM) (A) or 5 or 10 μM of C6-ceramide, dihydro-C6-ceramide, or dioctanoylglycerol (B). Thymidine incorporation in response to 24 h of serum stimulation was determined as described under “Experimental Procedures.” The data are from one experiment performed in triplicate and are representative of two separate experiments.
      Because a family of ceramide metabolites, glycosphingolipids, have been reported to be growth inhibitory(
      • Barbour S.
      • Edidin M.
      • Felding-Habermann B.
      • Taylor-Norton J.
      • Radin N.S.
      • Fenderson B.A.
      ), we used a ceramide glycosylation inhibitor, PDMP(
      • Radin N.S.
      • Shayman J.A.
      • Inokuchi J.-I.
      ), to rule out the possibility that C6-ceramide was acting through one of these products. Although PDMP alone had little effect up to 20 μM, it was growth inhibitory at 50 μM (Fig. 3A). PDMP at 20 μM had little effect on C6-ceramide-mediated DNA synthesis inhibition, but at 50 μM the PDMP effect was additive with C6-ceramide (Fig. 3A). This indicates that glycosylation of C6-ceramide is not required for growth inhibition.
      To test the specificity of this effect to C6-ceramide, cells were treated in parallel with C6-ceramide, 4,5-dihydro-C6-ceramide, and dioctanoylglycerol. Fig. 3B demonstrates that DNA synthesis is not inhibited by dihydro-C6-ceramide, which is desaturated at the 4,5 double bond and that the glycerol based analog is also inactive in this assay.
      In order to determine if the inhibition of DNA synthesis induced by ceramide imparted a senescent phenotype, we measured labeling indices in young HDF in the presence of ceramide. Young HDF were able to label 67.5 ± 9.5% of their nuclei by [3H]thymidine in response to serum stimulation (Fig. 4A). In contrast, senescent cells had a labeling index of <2% (Fig. 4B), cells grown in 10 μM C6-ceramide had a labeling index of 12 ± 5% (Fig. 4C), and cells grown in 15 μM C6-ceramide had a labeling index of 2.5 ± 0.5% (Fig. 4D) and morphologically resembled senescent cells.
      Figure thumbnail gr4
      Fig. 4Ceramide induces a senescent phenotype as determined by labeling indices. A, young WI-38 HDF were serum deprived for 48 h followed by stimulation with 10% FBS for 48 h. This resulted in a labeling index (% cells with stained nuclei) of 75%. (B) Senescent WI-38 HDF did not grow for at least four weeks in culture after seeding. They were stimulated with 10% FBS for 48 h and had a labeling index of <2%. C and D, young cells were treated with C6-ceramide (10 μM) (C) or 15 μM (D) for 48-72 h and then serum-stimulated as above, resembled senescent cells morphologically. Labeled nuclei were counted and expressed as the percentage of total nuclei. 500 nuclei were counted from random fields. The data are from single determinations and are representative of three experiments.
      Several biochemical parameters of cell aging and cellular senescence have been defined(
      • Peacocke M.
      • Campisi J.
      ). An important discovery came with the finding that senescent cells cannot undergo AP-1 activation in response to mitogenic stimuli(
      • Riabowol K.
      • Schiff J.
      • Gilman M.Z.
      ), presumably because of an inability to transcribe c-Fos(
      • Seshadri T.
      • Campisi J.
      ). This may be an important mechanism in preventing senescent cells from entering the cell cycle. Young HDF treated with C6-ceramide (10 μM) were unable to undergo AP-1 activation in response to serum stimulation as demonstrated by gel shift analysis of specific AP-1 complexes. Fig. 5A demonstrates that upon serum stimulation young HDF are able to respond by activation of AP-1 (lanes 1 and 2), whereas ceramide-treated cells (lanes 5 and 6) like senescent cells (lanes 7 and 8) were unable to respond by activation of AP-1. The specificity of AP-1 binding was demonstrated by the ability of cold wild type probe to compete off the complex (Fig. 5A, lane 3) as well as the lack of complex formation with a mutated AP-1 probe (Fig. 5A, lane 4) (see “Experimental Procedures” for exact sequences). Fig. 5B demonstrates the presence of Fos protein in the complex found in young but not senescent or ceramide-treated cells. In other studies from our laboratory we showed that ceramide inhibits phospholipase D activation leading to lack of diacylglycerol generation and protein kinase C activation in young HDF in response to serum stimulation, again mimicking alterations occurring in cellular senescence(
      • Venable M.E.
      • Blobe G.C.
      • Obeid L.M.
      ). These studies demonstrate that ceramide accumulation inhibits mitogenic stimulation, similar to what occurs in senescent cells.
      Figure thumbnail gr5
      Fig. 5Ceramide inhibits AP-1 activation. A, gel shift analysis utilizing a labeled AP-1 oligonucleotide and nuclear extracts obtained from young quiescent WI-38 HDF (lane 1), young serum-stimulated (YSS) WI-38 HDF demonstrating induction of AP-1 binding (lane 2), YSS extract preincubated with cold wild type AP-1 oligonucleotide demonstrating specificity of binding (lane 3), YSS extract incubated with labeled mutant AP-1 oligonucleotide demonstrating no specific binding (lane 4), young quiescent extract from cells pretreated with C6-ceramide (10 μM) (lane 5), YSS extract from cells pretreated with C6-ceramide (10 μM) demonstrating no induction of AP-1 binding (lane 6), senescent WI-38 HDF (lane 7), and senescent serum-stimulated WI-38 HDF demonstrating no induction of AP-1 binding (lane 8). B, the AP-1 complex contains c-Fos: YSS extract (lane 1) or YSS extract preincubated with c-Fos antibody at 4°C for 30 min (lane 2) partial competion AP-1 binding.
      Another important biochemical marker came with the discovery that senescent HDF are unable to phosphorylate the retinoblastoma protein in response to mitogenic stimuli(
      • Stein G.H.
      • Beeson M.
      • Gordon L.
      ). Remarkably, C6-ceramide in concentrations ranging from 3 to 20 μM was able to progressively inhibit Rb phosphorylation in young HDF in response to serum stimulation as indicated by Western blot analysis (Fig. 6, top). In addition, C6-ceramide (15 μM) was able to induce complete Rb dephosphorylation by 20-24 h, thus mimicking senescent HDF (Fig. 6, bottom).
      Figure thumbnail gr6
      Fig. 6Ceramide inhibits Rb phosphorylation and induces Rb dephosphorylation. Quiescent young WI-38 HDF treated with increasing concentrations of ceramide were progressively unable to undergo Rb phosphorylation in response to serum stimulation. Exponentially growing cells (EG), serum-deprived cells (SD), and serum-deprived and restimulated cells treated with ethanol vehicle or the indicated concentration of C6-ceramide were harvested. The proteins were analyzed by gel electrophoresis and immunoblotted with anti-Rb antisera. EG cells showed various forms of Rb phosphoprotein. Rb in SD cells became completely dephosphorylated. Upon serum stimulation, cells were able to phosphorylate Rb. However, cells treated with increasing concentrations of ceramide were unable to respond to serum stimulation, and Rb remained dephosphorylated (top panel). C6-ceramide (15 μM) induced complete Rb dephosphorylation by 20-24 h in young WI-38 HDF, thus mimicking senescent cells (SC) (lower panel).

      DISCUSSION

      In this study, evidence is provided for a role for ceramide in cellular senescence. First, ceramide levels and sphingomyelinase activity are significantly and specifically elevated in senescent cells and not in quiescent cells. Indeed, the changes in sphingomyelinase and ceramide appear to specifically distinguish senescence from quiescent growth arrest in WI-38 HDF. Second, the addition of ceramide to young HDF recapitulates many of the established parameters of cell senescence. These include the ability of ceramide to: 1) inhibit DNA synthesis; 2) inhibit growth; 3) inhibit AP-1 activation; and 4) activate Rb through dephosphorylation. Taken together the activation of sphingomyelinase and the ability of ceramide to induce parameters of senescence begin to point to a role for this pathway in the regulation of senescence.
      These remarkable increases in ceramide levels, in the ceramide:diacylglycerol ratio, and in neutral sphingomyelinase activity are even more marked than those observed in cell differentiation(
      • Okazaki T.
      • Bielawska A.
      • Bell R.M.
      • Hannun Y.A.
      ). They also represent permanent (stable and prolonged) increases in ceramide and sphingomyelinase activity as opposed to the transient signaling increases in response to inducers of apoptosis or differentiation, such as tumor necrosis factor α (
      • Kim M.-Y.
      • Linardic C.
      • Obeid L.
      • Hannun Y.
      ) and dihydroxyvitamin D3(
      • Okazaki T.
      • Bielawska A.
      • Bell R.M.
      • Hannun Y.A.
      ). Remarkably, there is no difference in ceramide levels or in sphingomyelinase activity between quiescent and exponentially growing young HDF. Therefore, the changes in sphingomyelinase and ceramide seen in the WI-38 cells are specific to senescence and not a mere consequence of growth arrest.
      The elevation of sphingomyelinase appeared to be specific to the neutral, Mg2+-dependent enzyme. This enzyme has been shown to be specifically (compared with the acidic enzyme) activated by various agonists(
      • Hannun Y.A.
      ). Sphingomyelin and particularly the hydrolyzable pool of sphingomyelin has been shown to be concentrated in the plasma membrane(
      • Linardic C.M.
      • Hannun Y.A.
      ), where it would be accessible to the neutral, Mg2+-dependent sphingomyelinase. Although the acid sphingomyelinase exhibits a higher specific activity in the WI-38 cell-free assay, it does not appear to be significantly increased in senescent cells. This enzyme is sequestered in the lysosomes that generally function in a degradative or recycling capacity. It is known that Nieman-Pick cells, which lack the acid sphingomyelinase, accumulate sphingomyelin in their lysosomal membranes (
      • Brady R.O.
      ). These cells also undergo senescence,
      C. J. Gamard and L. M. Obeid, unpublished observations.
      indicating that acid sphingomyelinase is not necessary for WI-38 senescence. Entry of cells into senescence is accompanied by induction of neutral sphingomyelinase, although we do not know the mechanism of that activation. However, many changes occur during cellular senescence (reviewed in (
      • Kirkland J.L.
      )), such as decreased membrane fluidity, increased protein oxidation, decreased DNA methylation, telomere shortening, and defects in mitogenic signaling; one or more of these events may act as a signal for activation of sphingomyelinase.
      The ability of exogenously administered ceramide to induce a senescent phenotype raises the possibility that one of its metabolites could be the biologically active species. We observe the formation of two minor metabolites, which are produced over the course of several hours. Hexanoyl-[3H]sphingosine (data not shown) and [14C]hexanoyl-sphingosine both showed the production of these same two spots on TLC. Our preliminary analysis demonstrates that these compounds do not represent predictable products such as glycosphingolipids, sphingosine, or sphingomyelin. In this context Rani et al.(
      • Rani C.S.S.
      • Abe A.
      • Chang Y.
      • Rosenzweig N.
      • Saltiel A.R.
      • Radin N.S.
      • Shayman J.A.
      ) have demonstrated that PDMP induces cell cycle arrest in NIH-3T3 fibroblasts; however, in that study this effect was not totally mimicked by exogenous ceramide, probably due to administration of C2-ceramide in serum-containing medium, which attenuates its effects(
      • Bielawska A.
      • Linardic C.M.
      • Hannun Y.A.
      ). However, because ceramide is elevated selectively in senescence, we feel that ceramide is either the active species itself or an immediate precursor of the active species.
      Several molecular changes that occur with cellular senescence have been characterized(
      • Peacocke M.
      • Campisi J.
      ). In this study we demonstrate that several of these biochemical events occur in response to C6-ceramide. The inability of senescent cells to activate the transcription factor AP-1 in response to mitogenic stimuli (
      • Riabowol K.
      • Schiff J.
      • Gilman M.Z.
      ) appears to be one of the critical events in preventing senescent cells from entering the cell cycle. Young HDF treated with C6-ceramide were unable to undergo AP-1 activation in response to serum stimulation. In a previous study we found that a mitogenic pathway mediated by PLD activation and DAG generation is inhibited in senescent HDF and that C6-ceramide inhibits this pathway(
      • Venable M.E.
      • Blobe G.C.
      • Obeid L.M.
      ), again mimicking alterations occurring in cellular senescence. Another important biochemical marker is that senescent HDF are unable to phosphorylate the retinoblastoma protein in response to mitogenic stimuli(
      • Stein G.H.
      • Beeson M.
      • Gordon L.
      ). This may be due to the overexpression of the recently described senescent cell-derived inhibitor of DNA synthesis (Sdi1 or p21), a protein involved in binding and inhibiting the kinase activity of cyclin-Cdk complexes(
      • Noda A.
      • Ning Y.
      • Venable S.F.
      • Pereira-Smith O.M.
      • Smith J.R.
      ). Remarkably, C6-ceramide was able to inhibit Rb phosphorylation in young HDF in response to serum stimulation and to induce complete Rb dephosphorylation; in addition preliminary data from our laboratory demonstrate that ceramide is also able to induce Sdi1 protein levels. Thus the addition of ceramide has a dual role in inducing a senescent phenotype in HDF by: 1) activation of a growth suppressor pathway through Rb dephosphorylation and 2) inhibition of the mitogenic pathway mediated by c-Fos and AP-1. Although sphingomyelinase activation and ceramide generation may not “cause” senescence, our data presented here support a role for ceramide in mediating or maintaining senescent cells in a nonreplicative state.
      Ceramide has been shown to activate a protein phosphatase 2A-like protein phosphatase termed ceramide-activated protein phosphatase(
      • Dobrowsky R.T.
      • Kamibayashi C.
      • Mumby M.C.
      • Hannun Y.A.
      ). Okadaic acid has been shown to inhibit ceramide-activated protein phosphatase, to be mitogenic, and to partially reverse cellular senescence(
      • Afshari C.A.
      • Barrett J.C.
      ). This allows us to speculate that inhibition of the sphingomyelinase/ceramide pathway might reverse cell senescence. Our studies showing that washing out ceramide allows cells to re-enter the cell cycle and escape senescence support this hypothesis. Thus, further understanding of sphingomyelinase and of the mechanism of action of ceramide should allow important insight into the molecular mechanisms involved in cellular senescence.
      In summary, cellular senescence is a state of growth arrest that partially resembles quiescent growth arrest. Very few biochemical events are able to distinguish senescence from quiescent growth arrest. Ceramide appears to be one of the first biologically active molecules that do so. We have previously shown that ceramide increases significantly after serum deprivation of Molt-4 cells, which induces a terminal event of growth arrest characterized by apoptosis and cell cycle arrest(
      • Jayadev S.
      • Liu B.
      • Bielawska A.E.
      • Lee J.Y.
      • Nazaire F.
      • Pushkareva M.Y.
      • Obeid L.M.
      • Hannun Y.A.
      ). The events occurring in senescence are analogous in the sense that senescence is a terminal event. Fibroblast quiescence, on the other hand, is a reversible event, and consequently it is expected that ceramide levels are not elevated. Taken together, significant new evidence now points to ceramide as an inducer of terminal events of cell biology, i.e.. apoptosis, differentiation, and now cellular senescence. These results provide an important first step into a new avenue of research on the biology of senescence and aging.

      Acknowledgments

      We thank Dr. Yusuf A. Hannun for expert advice, Dr. Cynthia Afshari for assistance with labeling indices studies, Linda Karolak for expert technical assistance, and Cynthia Jones and Andrea Oakley for expert secretarial assistance.

      REFERENCES

        • Hayflick L.
        Exp. Cell Res. 1965; 37: 614-636
        • Norwood T.H.
        • Pendergrass W.R.
        • Sprague C.A.
        • Martin G.M.
        Proc. Natl. Acad. Sci. U. S. A. 1974; 71: 2231-2235
        • Lumpkin C.K.
        • McClung J.K.
        • Pereira-Smith O.M.
        • Smith J.R.
        Science. 1986; 232: 393-398
        • Rubelj I.
        • Pereira-Smith O.M.
        Exp. Cell Res. 1994; 211: 82-89
        • Seshadri T.
        • Campisi J.
        Science. 1990; 247: 205-209
        • Riabowol K.
        • Schiff J.
        • Gilman M.Z.
        Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 157-161
        • Stein G.H.
        • Beeson M.
        • Gordon L.
        Science. 1990; 249: 666-669
        • Stein G.H.
        • Drullinger L.F.
        • Robetorye R.S.
        • Pereira-Smith O.M.
        • Smith J.R.
        Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 11012-11016
        • Dulic V.
        • Drullinger L.F.
        • Lees E.
        • Reed S.I.
        • Stein G.H.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 11034-11038
        • Noda A.
        • Ning Y.
        • Venable S.F.
        • Pereira-Smith O.M.
        • Smith J.R.
        Exp. Cell Res. 1994; 211: 90-98
        • Venable M.E.
        • Blobe G.C.
        • Obeid L.M.
        J. Biol. Chem. 1994; 269: 26040-26044
        • Gomez-Munoz A.
        • Martin A.
        • O'Brien L.
        • Brindley D.N.
        J. Biol. Chem. 1994; 269: 8937-8943
        • Nakamura T.
        • Abe A.
        • Balazovich K.J.
        • Wu D.
        • Suchard S.J.
        • Boxer L.A.
        • Shayman J.A.
        J. Biol. Chem. 1994; 269: 18384-18389
        • Hannun Y.A.
        J. Biol. Chem. 1994; 269: 3125-3128
        • Bielawska A.
        • Crane H.M.
        • Liotta D.
        • Obeid L.M.
        • Hannun Y.A.
        J. Biol. Chem. 1993; 268: 26226-26232
        • Obeid L.M.
        • Linardic C.M.
        • Karolak L.A.
        • Hannun Y.A.
        Science. 1993; 259: 1769-1771
        • Bligh E.G.
        • Dyer W.J.
        Can. J. Biochem. Physiol. 1959; 37: 911-917
        • Preiss J.
        • Loomis C.R.
        • Bishop W.R.
        • Stein R.
        • Niedel J.E.
        • Bell R.M.
        J. Biol. Chem. 1986; 261: 8597-8600
        • Rouser G.
        • Siakotos A.N.
        • Fleischer S.
        Lipids. 1966; 1: 85-86
        • Suzuki K.
        Methods Enzymol. 1987; 138: 727-739
        • Stoffel W.
        • LeKim D.
        • Tschung T.S.
        Hoppe-Seyler's Z. Physiol. Chem. 1971; 352: 1058-1064
        • Jayadev S.
        • Linardic C.M.
        • Hannun Y.A.
        J. Biol. Chem. 1994; 269: 5757-5763
        • Folch J.
        • Lees M.
        • Stanley G.H.S.
        J. Biol. Chem. 1957; 226: 497-509
        • Afshari C.A.
        • Barrett J.C.
        Cancer Res. 1994; 54: 2317-2321
        • Dbaibo G.
        • Obeid L.M.
        • Hannun Y.A.
        J. Biol. Chem. 1993; 268: 17762-17766
        • Chao R.
        • Khan W.
        • Hannun Y.A.
        J. Biol. Chem. 1992; 267: 23459-23462
        • Okazaki T.
        • Bielawska A.
        • Bell R.M.
        • Hannun Y.A.
        J. Biol. Chem. 1990; 265: 15823-15831
        • Peacocke M.
        • Campisi J.
        J. Cell. Biochem. 1991; 45: 147-155
        • Okazaki T.
        • Bell R.M.
        • Hannun Y.A.
        J. Biol. Chem. 1989; 264: 19076-19080
        • Okazaki T.
        • Bielawska A.
        • Domae N.
        • Bell R.M.
        • Hannun Y.A.
        J. Biol. Chem. 1994; 269: 4070-4077
        • Chatterjee S.
        Adv. Lipid Res. 1993; 26: 25-47
        • Spence M.W.
        Adv. Lipid Res. 1993; 26: 3-23
        • Kim M.-Y.
        • Linardic C.
        • Obeid L.
        • Hannun Y.
        J. Biol. Chem. 1991; 266: 484-489
        • Walker P.R.
        • Kwast-Welfeld J.
        • Gourdeau H.
        • Leblanc J.
        • Neugebauer W.
        • Sikorska M.
        Exp. Cell Res. 1993; 207: 142-151
        • Barbour S.
        • Edidin M.
        • Felding-Habermann B.
        • Taylor-Norton J.
        • Radin N.S.
        • Fenderson B.A.
        J. Cell. Physiol. 1992; 150: 610-619
        • Radin N.S.
        • Shayman J.A.
        • Inokuchi J.-I.
        Adv. Lipid Res. 1993; 26: 183-211
        • Linardic C.M.
        • Hannun Y.A.
        J. Biol. Chem. 1994; 269: 23530-23537
        • Brady R.O.
        Stanbury J.B. Wyngaarden J.B. Fredrickson D.S. Goldstein J.L. Brown M.S. The Metabolic Basis of Inherited Disease. McGraw-Hill Book Co., New York1983: 831-841
        • Kirkland J.L.
        Clin. Biochem. 1992; 25: 61-75
        • Rani C.S.S.
        • Abe A.
        • Chang Y.
        • Rosenzweig N.
        • Saltiel A.R.
        • Radin N.S.
        • Shayman J.A.
        J. Biol. Chem. 1995; 270: 2859-2867
        • Bielawska A.
        • Linardic C.M.
        • Hannun Y.A.
        FEBS Lett. 1992; 307: 211-214
        • Dobrowsky R.T.
        • Kamibayashi C.
        • Mumby M.C.
        • Hannun Y.A.
        J. Biol. Chem. 1993; 268: 15523-15530
        • Jayadev S.
        • Liu B.
        • Bielawska A.E.
        • Lee J.Y.
        • Nazaire F.
        • Pushkareva M.Y.
        • Obeid L.M.
        • Hannun Y.A.
        J. Biol. Chem. 1995; 270: 2047-2052