Keloid Fibroblasts Resist Ceramide-Induced Apoptosis byOverexpression of Insulin-Like Growth Factor I Receptor

Hiroshi Ishihara, Hiroshi Yoshimoto, Masaki Fujioka, Ryuuichi Murakami, Akiyoshi Hirano, Tohru Fujii,Akira Ohtsuru,* Hiroyuki Namba,* and Shunichi Yamashita*Department of Plastic and Reconstructive Surgery, *Department of Nature Medicine, Atomic Bomb Disease Institute, Nagasaki University School of

Medicine, Japan

Keloids are benign dermal tumors, characterized byovergrowth of lesions, invasiveness beyond the origi-nal boundary of the insult, and recurrence of lesions.The exact etiology is unknown, however. Ourhypothesis is that keloids are acquired as a result ofan abnormal or prolonged wound healing process,with persistent proliferation and extracellular matrixproduction of ®broblasts that should otherwisediscontinue in normal wound healing. In this study,we examined the response of keloid ®broblasts toproapoptotic signaling. Cell-permeable ceramide,N-acetyl-D-sphingosine, induced apoptosis of dermal®broblasts in a dose- and time-dependent manner,which was detected by phase contrast microscopy,¯uorescent microscopy, the TUNEL method, ¯owcytometric analysis, and WST-1 assay. In contrast,keloid ®broblasts resisted apoptosis induced by

N-acetyl-D-sphingosine (percent survival with 40 mM

ceramide treatment for 12 h, normal versus keloid:9.6% 6 6.6% vs 66.8% 6 5.5%). Western blotting ana-lysis showed insulin-like growth factor I receptoroverexpression in keloid ®broblasts, but not innormal ®broblasts. Exogenously added insulin-likegrowth factor I enhanced the resistance of keloid®broblasts to ceramide-induced apoptosis. Wort-mannin, a phosphatidylinositol 3 kinase inhibitor,suppressed the antiapoptotic action of insulin-likegrowth factor I in keloid ®broblasts. Our resultssuggest that keloid ®broblasts overexpressinginsulin-like growth factor I receptor are resistant toapoptosis, thus allowing persistent proliferation andproduction of excessive extracellular matrix. Keywords: programmed cell death/wound healing. J InvestDermatol 115:1065±1071, 2000

Keloids are characterized by proliferation of dermal®broblasts, overproduction of extracellular matrix bydermal ®broblasts, invasiveness beyond the originalboundary of the insult, and recurrence, and areclassi®ed as benign dermal tumors (Tredget et al,

1997). Although a benign disorder, surgical resection of keloids iscontraindicated because of the likelihood of local invasion andrecurrence. Accordingly various conservative therapies have beenused such as steroid, tranilast, and pressure treatment, but de®niteand effective therapy has not yet been established.

The normal wound healing process consists of three phases:in¯ammation, proliferation, and remodeling. From the in¯amma-tion phase to the proliferation phase, in¯ammatory cells and smallvessels as well as ®broblasts accumulate and proliferate. Fibroblastsproduce extracellular matrix such as collagen and ®bronectin, andthey form granulation tissue. In the late phase of wound healing,both cellularity and extracellular matrix decrease and the granula-tion tissue evolves into a mature scar. We speculated thatin¯ammation and proliferation phases are sustained in keloid

formation and keloid ®broblasts become more resistant to cellelimination by some unknown mechanism.

In a previous study, we reported that keloids overexpress insulin-like growth factor I receptor (IGF-IR), and the IGF-I-IGF-IRsignal mediates invasiveness of keloid ®broblasts (Yoshimoto et al,1999). IGF-I is also a well-known major survival factor in theserum (O'Connor, 1998) and protects ®broblasts from apoptosisunder a wide variety of conditions, including serum deprivation,Fas ligand, tumor necrosis factor a (Ortiz et al, 1997),chemotherapeutic drugs (Sell et al, 1995), and ceramide(Gerritsen et al, 1998).

Ceramide is a lipid second messenger that is well known as amediator of extracellular signals. It is formed by cell membranesphingomyelin and in¯uences proapoptotic signal pathwaysinduced by various stimuli such as Fas antigen (Tepper et al,1995), tumor necrosis factor a (Kaipia et al, 1996), chemother-apeutic drugs (Jaffrezou et al, 1996), and radiation (Haimovitz,1998), and can also induce apoptosis by itself. So we consideredthat ceramide might be one of the key mediators for apoptosis andchose it as the inducer in our experiments. In this study, wedetermined the expression of IGF-IR in keloid ®broblasts andinvestigated whether it in¯uences the response to proapoptoticsignaling by N-acetyl-D-sphingosine (C2 ceramide).

MATERIALS AND METHODS

Materials Keloid samples were obtained from three different Japanesepatients after surgical excision. Three normal skin samples were obtained fromthree Japanese volunteers. Informed consent was obtained from each subject.

Manuscript received September 15, 1999; revised May 22, 2000;accepted for publication August 22, 2000.

Reprint requests to: Dr. Hiroshi Ishihara, Department of NatureMedicine, Atomic Bomb Disease Institute, Nagasaki University School ofMedicine, 1-12-4 Sakamoto, Nagasaki 852-8523, Japan.

Abbreviations: C2 ceramide, N-acetyl-D-sphingosine; PI, propidiumiodide; PI3 kinase, phosphatidylinositol 3 kinase; TUNEL, terminal-deoxynucleotidyl-transferase-mediated dUTP-biotin nick end labeling.

0022-202X/00/$15.00 ´ Copyright # 2000 by The Society for Investigative Dermatology, Inc.

1065

Reagents C2 ceramide and dihydro-C2-ceramide were purchased fromCalbiochem (La Jolla, CA). They were dissolved in ethanol and used as10 mM stock solution. Hoechst 33342 and propidium iodide (PI) werepurchased from Sigma Chemical (St. Louis, MO). Human recombinantIGF-I was kindly provided by Fujisawa Pharmaceutical (Tokyo, Japan).Wortmannin was purchased from Wako Pure Chemical (Osaka, Japan) anddissolved in phosphate-buffered saline (PBS) as 100 mM stock solution. Allother chemicals and reagents were of the highest purity available fromcommercial sources.

Cell cultures Primary cultures of dermal ®broblasts were established aspreviously described (Arakawa et al, 1990). Explants were maintained inDulbecco's modi®ed Eagle's medium (DMEM) with 10% fetal bovineserum (Gibco BRL, Life Technologies), 100 U per ml penicillin, and100 mg per ml streptomycin at 37°C in a humidi®ed incubator with 5%CO2. Fibroblasts at passages 3±7 were used in this study.

Western blot analysis for IGF-IR Expression of IGF-IR was alsodetected by western blotting. Cells were washed twice in ice-cold PBSbefore lysis in sodium dodecyl sulfate (SDS) sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 50 mM dithiothreitol, and 0.1%bromophenol blue). The lysates were sonicated for 2 s, boiled for 5 min,and then microcentrifuged for 5 min. The protein concentration of thelysates was determined by the Bio-Rad assay. Proteins were separated on7.5% SDS polyacrylamide gels and transferred to nitrocellulose membranes(Hybond-p; Amersham, Arlington Heights, IL). The membranes werepreincubated with a blocking buffer [1 3 Tris-buffered saline (TBS), 0.1%Tween-20, and 5% nonfat dry milk] overnight at 4°C. They were thenincubated with a primary antibody to IGF-IR (mouse monoclonalantibody for a subunits of human IGF-IR; Santa Cruz Biotechnology)with gentle agitation overnight at 4°C. They were washed three times withTBS-T, and incubated with the secondary antibody (horseradish peroxidaseconjugated goat antimouse IgG; ABC kit, Vector Laboratories). Proteinswere detected by enhanced chemiluminescence (Pierce) using the protocolprovided by the supplier.

Induction and detection of cell death Dermal ®broblasts werepreincubated for 48 h with serum-free medium. In the next step variousconcentrations of C2 ceramide were added and the ®broblasts were furtherincubated for an additional 12 h. The cells were viewed with a Nikon phasecontrast microscope and then trypsinized and washed with PBS, ®xed with2% glutaraldehyde, and stained with Hoechst 33342 (1 mM) and PI (1 mgper ml). Specimens were analyzed with an Olympus BH-2 ¯uorescentmicroscope.

Terminal-deoxynucleotidyl-transferase -mediated dUTP-biotinnick end labeling (TUNEL) method For the detection of apoptosis,the procedure is based upon the method described by Gavrieli et al (1992).To investigate the response of dermal ®broblasts to C2 ceramide, culturesreaching 80%±90% con¯uence in the complete medium were starved for

Figure 1. Upregulation of IGF-I receptor expression in keloid®broblasts compared with normal ®broblasts. Immunoblotting ofIGF-IR a subunits. Total cell lysates were prepared from the threeindependent strains of normal and keloid ®broblasts, and equal amounts ofproteins were subjected to 7.5% SDS-polyacrylamide gel electrophoresisfollowed by western blotting against IGF-IR a subunits. Lanes 1±3, normal®broblasts; lanes 4±6, keloid ®broblasts.

Figure 2. Dose-dependent induction of celldeath by C2 ceramide in normal and keloid®broblasts. (a-f) Phase contrast microscopy ofdermal ®broblasts cultured with DMEM contain-ing various concentrations of C2 ceramide for12 h. (a-c) Normal ®broblasts treated with 0±50 mM C2 ceramide; (d-f) keloid ®broblasts treatedwith 0±50 mM C2 ceramide; g, normal ®broblaststreated with ethanol only; (h) normal ®broblaststreated with 50 mM dihydro-C2-ceramide; (i)¯uorescent microscopy (Hoechst 33342) of normal®broblasts cultured with DMEM containing30 mM C2 ceramide for 12 h. Typical morpholo-gic changes of apoptotic cells are observed amongnormal ®broblasts (arrow). Magni®cations: (a±h)3 40; (i) 3 200.

1066 ISHIHARA ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

48 h in serum-free DMEM. Then various concentrations of C2 ceramidewere added and the cultures were further incubated for an additional 6 h.Apoptosis induction by C2 ceramide was detected in situ by the TUNELmethod, using the Apoptosis in situ Detection Kit (Wako, Osaka, Japan)according to the manufacturer's protocol.

Flow cytometry For ¯ow cytometric analysis, cells (5 3 105) were ®xedwith 70% ethanol and washed with PBS. After preincubation with RNaseA (0.1 mg per ml) at 37°C, cells were stained with PI (25 mg per ml), and0.1% bovine serum albumin was added. Fluorescence was measured byusing a FACScan ¯ow cytometer (Becton Dickinson, Mountain View,CA).

Cell survival assay We also assessed cell survival by using a modi®edMTT assay, WST-1 assay, a colorimetric assay based on the cleavage of thetetrazolium salt WST-1 to a formazan dye by the mitochondrialdehydrogenase of viable cells. Fibroblasts suspended in DMEMcontaining 10% fetal bovine serum were seeded at 5 3 103 cells per wellinto 96-well plates. The cells were preincubated for 24 h with serum-freemedium at 37°C. The medium was replaced with DMEM containing C2ceramide (0±50 mM) with or without IGF-I (100 ng per ml) for 0±48 h at

37°C. The ready-to-use WST-1 reagent was added to the cells during thelast 2 h and incubated at 37°C.

Statistical analysis All values were expressed as mean 6 SD of triplicatesamples and were representative of at least three experiments. Differencesbetween groups were examined for statistical signi®cance using analysis ofvariance. A p-value less than 0.01 denoted the presence of a statisticallysigni®cant difference.

RESULTS

Expression of IGF-IR on dermal ®broblasts In our previousstudy, we reported that keloids overexpress IGF-IR (Yoshimotoet al, 1999). In this study we used different strains of keloid andnormal dermal ®broblasts, however, and therefore we ®rst checkedthe expression of IGF-IR in these cells. In western blotting, highexpression of IGF-IR was detected in three keloid ®broblast strains(Fig 1, lanes 4±6) whereas only a relatively weak expression of IGF-IR was present in three normal dermal ®broblast strains (Fig 1, lanes1±3). These results con®rmed our earlier ®ndings (Yoshimoto et al,1999).

Figure 3. Determination of apoptotic ®bro-blasts by the TUNEL method in vitro. (a, b)Normal ®broblasts treated with 30 and 50 mM C2ceramide; (c, d) keloid ®broblasts treated with 30and 50 mM C2 ceramide; (e) normal ®broblastswith no treatment; (f) normal ®broblasts treatedwith ethanol only; (g) normal ®broblasts treatedwith 50 mM dihydro-C2-ceramide. Typical apop-totic cells were observed in a concentration-de-pendent manner by C2 ceramide in normal®broblasts, but only few TUNEL-positive cellswere detected in keloid ®broblasts under the sameconditions. Magni®cations: 3100.

VOL. 115, NO. 6 DECEMBER 2000 KELOID FIBROBLASTS AND APOPTOSIS 1067

C2-ceramide-induced apoptosis of dermal ®broblasts After12 h of treatment, C2 ceramide induced a dose-dependent celldeath of primary cultured human dermal ®broblasts (Fig 2a-c). Todiscriminate between apoptosis and necrosis, we examined theDNA fragmentation of C2-ceramide-treated dermal ®broblasts by¯uorescent microscopy analysis and the TUNEL method. In¯uorescent microscopy, typical apoptotic cells were easily observedin normal ®broblasts (Fig 2i), but necrotic cells were very rare.From this result, we considered that C2 ceramide mainly inducedapoptosis but not necrosis in dermal ®broblasts within the range 10-50 mM. In contrast, keloid ®broblasts were more resistant to C2-ceramide-induced cell death compared with normal dermal®broblasts and they maintained their original structure andviability (Fig 2d-f). In the TUNEL method, normal ®broblaststreated with C2 ceramide underwent apoptosis in a concentration-dependent manner, but only few TUNEL-positive cells weredetected in keloid ®broblasts (Fig 3). FACScan analysis showed ahigh proportion of apoptotic cells among normal ®broblasts(48.61% 6 8.82%) (Fig 4a, b). In contrast, keloid ®broblastsresisted apoptosis (7.76% 6 3.48%) and underwent G2 arrest(Fig 4c, d). We also checked cell viability of dermal ®broblasts byWST-1 assay. In this assay, C2 ceramide induced loss of viability innormal dermal ®broblasts in a concentration- and time-dependentmanner, but keloid ®broblasts resisted loss of viability comparedwith normal ®broblasts (Figs 5, 6).

IGF-I inhibits ceramide-induced apoptosis IGF-I is wellknown as a major survival factor in serum and can protect cellsagainst apoptosis under a wide variety of conditions. In the nextstep, we examined the effects of IGF-I on normal and keloid®broblasts treated with C2 ceramide. IGF-I (100 ng per ml)inhibited apoptosis of normal ®broblasts slightly, but theprotective effects were more marked in keloid ®broblasts (Fig 7).Furthermore, the protective effect correlated with the level of IGF-IR expression.

Phosphatidylinositol 3 kinase (PI3-kinase) mediates theantiapoptotic signal of IGF-IR in keloid ®broblasts Finally,we investigated the mechanisms of antiapoptotic effects of IGF-IRin keloid ®broblasts. PI3-kinase and Akt are known to mediate theantiapoptotic signal of IGF-IR in ®broblasts (Kulik et al, 1997). Inthe next step, we used wortmannin as a PI3-kinase inhibitor andexamined its effect on the antiapoptotic inhibitory action of IGF-Iin keloid ®broblasts. Preincubation with 100 nM wortmanninblocked the antiapoptotic signal of IGF-IR (Fig 8).

DISCUSSION

Apoptosis is de®ned as programmed cell death. It is a geneticallycontrolled process to eliminate cells in response to a variety ofstimuli and provides protection against cancer, viral infections,cytotoxic drugs, and radiation as well as maintaining homeostasis

Figure 4. DNA contents of normal and ke-loid ®broblasts after treatment with C2 cera-mide. Cells were cultured with serum-freeDMEM for 24 h and incubated with DMEM con-taining 25 mM C2 ceramide for 12 h. Followingnuclear staining with PI, ¯ow cytometric analysiswas performed as described in Materials and Meth-ods. (a, b) Normal ®broblasts; (c, d) keloid ®bro-blasts; (e, f) negative controls of normal ®broblasts.Note the increased population of apoptotic cells(48.61% 6 8.82%) in normal ®broblasts after 12 h.In contrast, keloid ®broblasts underwent a G2 ar-rest and resisted apoptosis (7.76% 6 3.48%).

1068 ISHIHARA ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

under physiologic conditions (Cuff et al, 1996; Blank et al, 1997;Dixon et al, 1997; Kinloch et al, 1999). Apoptosis also plays animportant role in the wound healing process (Desmouliere et al,1995), and this study provided clear evidence for the resistance ofkeloid ®broblasts to apoptosis, as evidenced by increased cellproliferation and decreased cell apoptosis. We also demonstratedthat such properties correlated with overexpression of IGF-IR onkeloid ®broblasts. To our knowledge, there are no previous reportson the role of IGF-IR in the apoptosis of keloid ®broblasts, exceptfor one study on apoptosis in keloid tissues (Appleton et al, 1996).In contrast to our ®ndings, the above study showed increasedapoptosis in keloid tissues by using in situ end labeling (the TUNELmethod). The exact explanation of the different ®ndings is not clearat present. The number of TUNEL-positive cells relative to totalcell number might be low, however, because of the markedly highcell density in keloid tissues. Also Appleton et al (1996) did notprovide any functional studies using a cell culture system.

In this study, western blotting analysis showed that theexpression of IGF-IR on keloid ®broblasts was 1.4±2.9 timeshigher than that on normal ®broblasts. Furthermore, we also

demonstrated resistance of keloid ®broblasts to C2-ceramide-induced apoptosis compared with normal dermal ®broblasts. Theaddition of exogenous IGF-I enhanced the survival of keloid®broblasts in a manner proportionate to IGF-IR expression.Comparison of cells derived from null mice of the IGF-IR genewith stable transfectants overexpressing the human IGF-IRdemonstrated the importance of IGF-IR expression in resistingapoptosis (Nakamura et al, 1997). Although we did not examine theacquisition of resistance in normal skin ®broblasts by IGF-IRtransfection, our results also suggest that the number of functionalIGF-IR is important in the regulation of antiapoptotic signals.

The antiapoptotic signal pathway of IGF-IR in ®broblasts andother cells has already been reported (Datta et al, 1997; Kulik et al,1997). PI3-kinase is a major effector for signaling by IGF-IR (DeMeyts et al, 1994; Carpenter and Cantley, 1996; Mendez et al,1996; Myers and White, 1996). Furthermore, Akt is identi®ed asthe downstream component of survival signaling through PI3-kinase (Kulik et al, 1997; Dudek et al, 1997; Kauffmann-Zeh et al,1997; Kennedy et al, 1997; Khwaja et al, 1997), which is agrowth factor that regulates serine/threonine kinase. Furthermore,

Figure 5. Cell viability of normal and keloid®broblasts following C2 ceramide treatment± dose-response. Each of three primary culturecell lines of normal and keloid ®broblasts seededat 5 3 105 cells per well into 96-well plates werecultured with the indicated concentrations of C2ceramide, and the cell viability was monitored byWST-1 assay as described in Materials and Methods.The dose-dependent effect of C2 ceramide on thecell viability of normal and keloid ®broblasts 12 hafter incubation. Note that keloid ®broblasts werestill resistant with the 40 mM concentration of C2ceramide even though the same dose of C2 cera-mide strongly abolished the cell viability of normal®broblasts.

VOL. 115, NO. 6 DECEMBER 2000 KELOID FIBROBLASTS AND APOPTOSIS 1069

activated Akt provides a survival signal that protects cells fromapoptosis induced by various stresses. Bad, a proapoptotic memberof the bcl-2 family, was identi®ed as a substrate of Akt when theintersection point of proapoptotic and antiapoptotic regulatory

cascades was investigated (Ladin et al, 1998; del Peso et al, 1997).This signal pathway has been recognized as a major antiapoptoticsignal pathway for IGF-IR. Kulik et al reported the existence of asurvival signaling pathway downstream of PI3-kinase, Akt/PKB, in®broblasts overexpressing the IGF-IR (Kulik and Weber, 1998). Inthis study, wortmannin, a PI3-kinase inhibitor, completely blockedthe antiapoptotic effect of IGF-I and IGF-IR in keloid ®broblasts.Thus, PI3-kinase may also play a major role in the antiapoptoticsignal pathway in keloid ®broblasts.

The underlying mechanisms for the increased number of IGF-IR in keloid ®broblasts remain unclear at this stage. The expressionof IGF-IR gene is modulated by a number of physiologic andpathologic factors, such as developmental stage, nutritional status,hormones, growth disorders, and malignancy (Werner et al, 1995).The tumor suppressor gene p53 is a regulatory factor oftranscription of IGF-IR gene. Wild-type p53 suppresses IGF-IRpromoter activity whereas mutant p53 stimulates IGF-IR genetranscription (Werner et al, 1996). Sead et al (1998) analyzed p53gene mutation in keloids using polymerase-chain-reaction-basedsingle-strand conformational polymorphism and DNA sequencing.They found mutations in exon 4 of p53 gene in all samples (codon72, CGC®CCC, arginine®proline). We also checked thismutation of p53 gene; however, we could not ®nd such mutationsin any of our keloid samples (data not shown). Viral proteininteracts with p53 and changes the protein conformation to besimilar to that of mutant p53. It is possible that changes in keloidp53 might occur without any genetic mutation.

In conclusion, we demonstrated in this study that keloid®broblasts overexpress IGF-IR, resist proapoptotic signals, anddisturb the balance between cell death and proliferation in thelesions, eventually producing excessive extracellular matrix, result-ing in the unique pathologic features of keloid tissues. Identi®cationof the factors responsible for controlling the expression of IGF-IRand apoptotic signal in keloids may therefore allow the design ofnovel therapies for the treatment of keloids.

Human recombinant IGF-I was a kind gift from Fujisawa Pharmaceutical Co.,

Tokyo. This work was supported by Grants-in-Aid (10671682) from the Ministry

of Education, Science, Sports and Culture, Japan.

REFERENCES

Appleton I, Brown NJ, Willoughby DA: Apoptosis, necrosis, and proliferation:possible implications in the etiology of keloids. Am J Pathol 149:1441±1447,1996

Arakawa M, Hamamochi A, Takeda K, Ueki H: Increased collagen synthesisaccompanying elevated m-RNA levels in cultured Werner's syndrome®broblasts. J Invest Dermatol 94:187±190, 1990

Blank KR, Rudoltz MS, Kao GD, Muschel RJ, McKenna WG: The molecularregulation of apoptosis and implications for radiation oncology. Int J Radiat Biol71:455±466, 1997

Carpenter CL, Cantley LC: Phosphoinositide kinases. Curr Opin Cell Biol 8:153±158,1996

Cuff S, Ruby J: Evasion of apoptosis by DNA viruses. Immunol Cell Biol 74:527±537,1996

Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, Greenburg ME: Aktphosphorylation of BAD couples survival signals to the cell-intrinsic deathmachinery. Cell 91:231±241, 1997

De Meyts P, Wallach B, Christoffersen CT, et al: Insulin-like growth factor-Ireceptor. Structure, ligand binding mechanism and signal transduction. HormRes 42:152±159, 1994

Desmouliere A, Redard M, Darby I, Gabbiani G: Apoptosis mediates the decrease incellularity during the transition between granulation tissue and scar. Am J Pathol146:56±66, 1995

Dixon SC, Soriano BJ, Lush RM, Borner MM, Figg WD: Apoptosis: its role in thedevelopment of malignancies and its potential as a novel therapeutic target. AnnPharmacother 31:76±82, 1997

Dudek H, Datta SR, Franke TF, et al: Regulation of neuronal survival by the serine-threonine protein kinase Akt. Science 275:661±665, 1997

Gavrieli Y, Sherman Y, Ben-Sasson SA: Identi®cation of programmed cell death insitu via speci®c labeling of nuclear DNA fragmentation. J Cell Biol 119:493±501, 1992

Gerritsen ME, Shen CP, Perry CA: Synovial ®broblasts and the sphingomyelinasepathway: sphingomyelin turnover and ceramide generation are not signaling

Figure 6. Cell viability of normal and keloid ®broblasts followingC2 ceramide treatment ± time course. Normal and keloid ®broblastsseeded at 5 3 105 cells per well into 96-well plates were cultured with theindicated concentrations of C2 ceramide, and the cell viability wasmonitored by WST-1 assay in a similar method to Fig 5. The time courseof cell viability in cells treated with 0 mM, 30 mM, and 40 mM C2 ceramide.The viability of normal ®broblasts decreased in a time-dependent manner,whereas the reduction rate of viability was small in keloid ®broblasts.

Figure 7. IGF-I rescues C2-ceramide-induced apoptosis of keloid®broblasts but not of normal ®broblasts. Normal and keloid ®broblastswere cultured with DMEM containing various concentrations (0±50 mM)of C2 ceramide with or without 100 ng per ml IGF-I for 12 h, and the cellviability was studied by WST-1 assay. Incubation of keloid ®broblasts withIGF-I enhanced the resistance to C2-ceramide-induced apoptosis in keloid®broblasts but not in normal ®broblasts.

Figure 8. PI3-kinase is involved in IGF-I-mediated resistance toapoptosis in keloid ®broblasts. Keloid ®broblasts were cultured withDMEM containing 40 mM C2 ceramide with or without 100 ng per mlIGF-I, and the cell viability at the indicated time was monitored by WST-1assay. Preincubation with 100 nM wortmannin for 30 min blocked IGF-I-mediated resistance of keloid ®broblasts against C2-ceramide-inducedapoptosis.

1070 ISHIHARA ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

mechanisms for the actions of tumor necrosis factor-alpha. Am J Pathol152:505±512, 1998

Haimovitz-Friedman A: Radiation-induced signal transduction and stress response.Radiat Res 150:s102±108, 1998

Jaffrezou JP, Levade T, Bettaieb A, et al: Daunorubicin-induced apoptosis: triggeringof ceramide generation through sphingomyelin hydrolysis. EMBO J 15:2417±2424, 1996

Kaipia A, Chun SY, Eisenhauer K, Hsueh AJ: Tumor necrosis factor-alpha and itssecond messenger, ceramide, stimulate apoptosis in cultured ovarian follicles.Endocrinology 137:4864±4870, 1996

Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C, Coffer P, DownwardJ, Evan G: Suppression of c-Myc-induced apoptosis by Ras signaling throughPI (3) K and PKB. Nature 385:544±548, 1997

Kennedy SG, Wagner AJ, Conzen SD, Jordan J, Bellacosa A, Tsichlis PN, Hay N:The PI3-kinase/Akt signaling pathway delivers an anti-apoptotic signal. GenesDev 11:701±713, 1997

Khwaja A, Rodriguez-Viciana P, Wennstrom S, Warne PH, Downward J: Matrixadhesion and Ras transformation both activate a phosphoinositide 3-OH kinaseand protein kinase B/Akt cellular survival pathway. EMBO J 16:2783±2793,1997

Kinloch RA, Treherne JM, Furness LM, Hajimohamadreza I: The pharmacology ofapoptosis. Trends Pharmacol Sci 20:35±42, 1999

Kulik G, Weber MJ: Akt-dependent and -independent survival signaling pathwaysutilized by insulin-like growth factor I. Mol Cell Biol 18:6711±6718, 1998

Kulik G, Klippel A, Weber MJ: Antiapoptotic signalling by the insulin-like growthfactor I receptor, phosphatidylinositol 3-kinase, and Akt. Mol Cell Biol17:1595±1606, 1997

Ladin DA, Hou Z, Patel D, McPhail M, Olson JC, Saed GM, Fivenson DP: p53 andapoptosis alterations in keloids and keloid ®broblasts. Wound Repair Regen 6:28±37, 1998

Mendez R, Myers MG Jr, White MF, Rhoads RE: Stimulation of protein synthesis,eukaryotic translation initiation factor 4E phosphorylation, and PHAS-Iphosphorylation by insulin requires insulin receptor substrate I andphosphatidylinositol 3-kinase. Mol Cell Biol 16:2857±2864, 1996

Myers MG Jr, White MF: Insulin signal transduction and the IRS proteins. Annu RevPharmacol Toxicol 36:615±658, 1996

Nakamura S, Watanabe H, Miura M, Sasaki T: Effect of the insulin-like growthfactor I receptor on ionizing radiation-induced cell death in mouse embryo®broblasts. Exp Cell Res 235:287±294, 1997

O'Connor R: Survival factors and apoptosis. Adv Biochem Eng Biotechnol 62:137±166,1998

Ortiz A, Lorz C, Gonzalez-Cuadrado S, Garcia del Moral R, O'Valle F, Egido J:Cytokines and Fas regulate apoptosis in murine renal interstitial ®broblasts. JAm Soc Nephrol 8:1845±1854, 1997

del Peso L, Gonzalez-Garcia M, Page C, Herrera R, Nunez G: Interleukin-3-induced phosphorylation of BAD through the protein kinase Akt. Science2278:687±689, 1997

Sead GM, Ladin D, Olson J, Han X, Hou Z, Fivenson D: Analysis of p53 genemutations in keloids using polymerase chain reaction-based single-strandconformational polymorphism and DNA sequencing. Arch Dermatol 134:963±967, 1998

Sell C, Baserga R, Rubin R: Insulin-like growth factor I (IGF-I) and the IGF-Ireceptor prevent etoposide-induced apoptosis. Cancer Res 55:303±306, 1995

Tepper CG, Jayadev S, Liu B, et al: Role for ceramide as an endogenous mediator ofFas-induced cytotoxicity. Proc Natl Acad Sci USA 92:8443±8447, 1995

Tredget EE, Nedelec B, Scott PG, Ghahary A: Hypertrophic scars, keloids, andcontractures. The cellular and molecular basis for therapy. Surg Clin North Am77:701±730, 1997

Werner H, Hernandez-Sanchez C, Karnieli E, Leroith D: The regulation of IGF-Ireceptor gene expression. Int J Biochem Cell Biol 27:987±994, 1995

Werner H, Karnieli E, Rauscher FJ III, Leroith D: Wild-type and mutant p53differentially regulate transcription of the insulin-like growth factor I receptorgene. Proc Natl Acad Sci USA 93:8318±8323, 1996

Yoshimoto H, Ishihara H, Ohtsuru A, et al: Overexpression of insulin-like growthfactor-I (IGF-I) receptor and the invasiveness of cultured keloid ®broblasts. AmJ Pathol 154:883±889, 1999

VOL. 115, NO. 6 DECEMBER 2000 KELOID FIBROBLASTS AND APOPTOSIS 1071