Background: Hereditary gingival fibromatosis (HGF) is a fibroticgingival enlargement. In previous work, HGF fibroblasts grew fasterand produced more collagen and fibronectin (FN) than normalgingival (GN) fibroblasts. HGF FN and collagen production,but not proliferation, were under autocrine transforming growthfactor (TGF)-β control, suggesting other means of activation ofHGF proliferation. Elevated/prolonged expression of the proto-oncogene c-myc is implicated in disregulation of cell growth. Theobjectives of this study were to: 1) determine if c-myc expres-sion is abnormal in quiescent and serum-stimulated HGF andGN fibroblasts and 2) determine the relationship between c-mycexpression and fibroblast proliferation using a c-myc antisenseoligonucleotide (ODN).

Methods: Proliferation was determined by enzyme-linkedimmunosorbent assay (ELISA), measuring incorporation of bromo-deoxyuridine into DNA. Expression of c-myc was determined byquantitative polymerase chain reaction (PCR), using incorporationof fluorescent dCTP and detection via electrophoresis.

Results: Proliferation was minimal until 24 hours or more afterserum stimulation, when HGF proliferation was greater than GN(P ≤0.02). All cells expressed c-myc mRNA at quiescenceand ≥1 hour after serum stimulation. Expression of c-myc in quies-cent HGF fibroblasts was elevated, and it peaked and remainedhigher after serum stimulation than in GN cells. Proliferation ofan HGF cell line was inhibited by 4 µM c-myc antisense ODN(14% decrease; P ≤0.006) and 8 µM c-myc antisense ODN (~80%decrease; P ≤0.0001), but generally not by c-myc sense ODN. Thiseffect was reversed by hybridizing the c-myc antisense and senseODNs (P = 0.007).

Conclusion: Data suggest that elevated proliferation of anHGF fibroblast cell line is related to elevated c-myc expression.J Periodontol 2004;75:360-369.

KEY WORDSFibroblasts, gingival; fibromatosis, gingival; genes, myc.

Volume 75 • Number 3

360

Role of the c-myc Proto-Oncogene in the Proliferation of Hereditary Gingival Fibromatosis FibroblastsDavid A. Tipton,*† Ernest S. Woodard III,‡ Mark A. Baber,* and Mustafa Kh. Dabbous*‡§

* Dental Research Center, University of Tennessee Health Science Center, Memphis, TN.† Department of Periodontology, University of Tennessee Health Science Center.‡ Department of Oral and Maxillofacial Surgery, University of Tennessee Health Science

Center.§ Department of Molecular Sciences, University of Tennessee Health Science Center.

Hereditary gingival fibromatosis (HGF)is an idiopathic, progressive, fibroticenlargement of the maxillary and

mandibular gingivae.1 HGF gingiva istypical of fibrotic connective tissue withheavy accumulation of collagen and fewfibroblasts.2 It appears that HGF fibro-blasts are phenotypically distinct fromnormal human gingival fibroblasts andthat cell selection or activation of particu-lar fibroblast types may participate in HGFpathogenesis.2-5

Our previous investigation of HGF fibro-blasts showed that they had significantlyhigher rates of proliferation and elevatedproduction of the extracellular matrix mol-ecules collagen, fibronectin (FN), and gly-cosaminoglycans.5 These and other cellularfunctions are under the control of severalregulatory cytokines such as transform-ing growth factor-β (TGF-β), which is animportant regulator of normal and patho-genic fibrogenesis, especially fibroses.TGF-β stimulates fibroblast proliferationand induces the synthesis of FN and col-lagen.6-8 Our studies have also shown thatboth normal and HGF fibroblasts constitu-tively produce TGF-β, but the HGF fibro-blasts generally make greater amounts ofTGF-β than the normal fibroblasts.9

The elevated secretion of TGF-β by theHGF fibroblasts suggests that it may playa role in the autocrine stimulation oractivation of HGF fibroblasts. We foundthat treatment of HGF fibroblasts withneutralizing anti-TGF-β decreased theirproduction of FN and collagen, but had noeffect on their proliferation9 (and unpub-lished observations). This suggested that

3062.qxd 3/10/04 9:23 AM Page 360

361

J Periodontol • March 2004 Tipton, Woodard, Baber, Dabbous

elevated HGF fibroblast proliferation may be due toother mechanisms.

The proto-oncogene c-myc plays a central rolein the proliferation and differentiation of many celltypes.10,11 It has been implicated in the disregulationof cell growth in neoplastic cells and as an essentialelement of response to growth factors and the tran-scription of other genes.12 c-myc encodes a short-livednuclear phosphoprotein required for proliferation.13

Such nuclear proto-oncogene proteins are inducedwhen normal cells are stimulated to grow, indicatingtheir direct role in growth control.14 For example, treat-ment of quiescent fibroblasts with platelet-derivedgrowth factor causes a transient increase in c-mycexpression, and unregulated expression of c-myc cancause uncontrolled growth.14 Aberrant c-myc expres-sion has been implicated in the hyper-proliferation ofvascular smooth muscle cells cultured from athero-matous plaques.15 Moreover, in fibroblasts derived fromfibrotic scleroderma skin, there is elevated expressionof c-myc as well as abnormal response of the c-mycgene after stimulation of proliferation.12 Feghali et al.have shown that c-myc expression is elevated in fibro-blasts derived from the skin of patients with systemicsclerosis.16

Gene expression can be inhibited by using specificantisense oligonucleotides (ODNs), short syntheticDNA molecules with nucleotide sequences designed tobe complementary to a specific intracellular targetmRNA. This allows the functional analysis of individ-ual genes. The hybridization of the ODN and the targetmRNA inhibits the expression of the target gene byblocking mRNA processing, transport, or transla-tion.17-19 The purpose of the present study was toinvestigate possible links between increased HGFfibroblast proliferation and c-myc production using ac-myc antisense ODN.

MATERIALS AND METHODSHuman Gingival FibroblastsHuman gingival fibroblast cell lines were derived fromthe gingiva of three patients with HGF,9 using standardtechniques,20 and were designated GH 8, GH 10, andGH 11. Two of the HGF strains were derived frombrothers, ages 6 and 9; the third HGF strain was froman unrelated male patient, age 11. Control fibroblaststrains, designated GN 43, GN 45, and GN 46, werederived from the non-inflamed gingiva of gender- andrace-matched healthy individuals; because of difficultiesin obtaining normal tissue from age-matched healthypatients, it was necessary to derive these strains fromthe gingiva of older subjects (ages 30 to 58). Exceptfor one HGF strain, which was derived from the maxil-lary anterior area, all the fibroblasts were derived fromgingival tissue in the maxillary or mandibular molarareas. Family histories were too incomplete to determine

whether HGF segregated as an autosomal dominanttrait. The cells were maintained in Dulbecco’s modifiedEagle medium (DMEM)� supplemented with 10% new-born calf serum� and 100 µg/ml gentamicin¶ (completemedium), and were grown at 37°C in a humidifiedatmosphere of 5% CO2 in air. Fibroblasts used in thisstudy were between passages 2 and 10.

Oligodeoxynucleotides (ODNs)c-myc antisense and sense ODNs with phosphorothioatebackbones# were used in this study.

The ODNs were supplied in lyophilized form and werereconstituted with Tris-EDTA (TE) buffer (pH 7.2) sup-plied with the ODNs. The phosphorothioate ODNderivatives (purified by high pressure liquid chro-matography [HPLC], cross-flow dialysis, and ultrafiltra-tion) are more stable than underivatized ODNs todegradation by nucleases that can be present in culturemedium and serum used for in vitro experimentation.21

These oligonucleotides have a half-life in serum-con-taining medium of >48 hours.

Cell Proliferation ELISAProliferation was determined by measuring the incor-poration of 5-bromo-2′-deoxyuridine (BrdU) into cellu-lar DNA. This technique is based on the incorporationof the pyrimidine analog BrdU instead of thymidine intothe DNA of proliferating cells. After it is incorporated intoDNA, BrdU is detected by immunoassay using a colori-metric BrdU cell proliferation ELISA.** All the neededreagents for this experiment were supplied in the kit, andthe experiment was performed as instructed by the kitmanual. Fibroblasts (1 × 104) were seeded into wells of96-well plates, with each plate corresponding to a dif-ferent incubation period, and cultured in completemedium for 24 hours at 37°C. This medium wasremoved and serum-free medium was added and incu-bated for 48 hours to render the cells synchronous andquiescent (G0-arrested).22 This medium was thenremoved and complete medium containing BrdU wasadded to the cells and incubated for 0, 1, 6, 12, 24, 48,and 96 hours. The cells were then fixed and their DNAdenatured to improve the accessibility of the incorpo-rated BrdU for detection by an anti-BrdU peroxidase-conjugated monoclonal antibody (which has nocross-reactivity with any endogenous cellular compo-nents such as thymidine, uridine, or DNA). Peroxidasesubstrate was added, and the colored reaction productwas measured by determining absorbance at 450 nmusing an ELISA reader. The absorbance values directlycorrelate to the amount of DNA synthesis and therebyto the number of proliferating cells.

� Gibco, Grand Island, NY.¶ Sigma Chemical Co., St. Louis, MO.# Biognostik GmbH, Göttingen, Germany.** Boehringer Mannheim Corp., Indianapolis, IN.

3062.qxd 3/10/04 9:23 AM Page 361

362

c-myc Proto-Oncogene in Hereditary Gingival Fibromatosis Volume 75 • Number 3

Effect of c-myc ODNs on Cell ProliferationFor determination of the effect of c-myc ODNs on pro-liferation, the experiments were performed as describedabove, except that TE buffer, or c-myc sense or anti-sense ODNs at 2 µM, 4 µM, or 8 µM were added at thetime of addition of the serum-containing medium afterthe 48-hour serum starvation period. The cells wereexposed to the ODNs for 48, 96, 120, or 144 hours,and BrdU was added for the final 24 hours of incuba-tion. To confirm that c-myc antisense ODN effects weredue to target gene inhibition, reversal experiments wereperformed using the c-myc sense ODN. Sense ODNscan reverse the effects of antisense ODNs under differ-ent experimental conditions. For example, the senseand antisense ODNs can be preannealed prior to add-ition to media; the cells can be preloaded with senseODN prior to addition of the antisense; or the senseand antisense ODNs can be added simultaneously, withthe sense ODN at various levels of excess over theantisense.21,23 In this study we preincubated the c-mycantisense ODN with the c-myc sense ODN [equimolaramounts; or 2:1 or 3:1 (molar) sense:antisense] for30 minutes at 37°C before addition to the cultures.

Isolation of mRNA and Production of cDNARepresentative normal strains GN 43 and GN 46, andthe HGF strain GH 8 were chosen for further analysis.PolyA+ mRNA was isolated from the fibroblasts using anmRNA isolation kit.†† mRNA was isolated from syn-chronous, quiescent cells and from cells at specific timesafter they were stimulated to proliferate. The fibroblastswere seeded at 5 × 105 cells in 75 cm2 flasks and cul-tured in 10 ml complete medium for 24 hours to sub-confluent levels. They were then made synchronousand quiescent (G0-arrested) as described above. Toreverse the growth arrest, the serum-free DMEM wasremoved and DMEM containing gentamicin and 10%newborn calf serum was added to the cells. mRNA wasthen isolated from quiescent cells and from cells at 1,6, 24, 48, and 96 hours after reentry into the cell cycle.In other studies, normal mouse fibroblasts showed max-imum c-myc product approximately 1 hour after serumstimulation, and this appeared to fall to a constant levelby 6 hours after stimulation.24 By testing for c-myc attime periods beyond this, we determined if the c-myclevels in the HGF fibroblasts remained at an elevatedlevel for longer time periods compared to the GN fibro-blasts. To isolate mRNA, the cells were lysed in SDSbuffer, incubated at 45°C in a buffer containing pro-teases, and then applied directly to oligo (dT) cellulose.The non-polyadenylated RNAs, DNA, dissolved mem-branes, proteins, and cell debris were washed off the resinwith salt-containing buffers, and then the polyadeny-lated mRNA was eluted with salt-free buffer. The mRNAwas precipitated by incubating at −70°C in 100% ethanolcontaining glycogen and 0.1 M sodium acetate pH 4.5,

followed by centrifugation at 16,000 × g for 30 minutesat 4°C. The pellets were each resuspended in 10 µl ofthe salt-free elution buffer described above.

cDNA was generated by reverse transcription of thetotal mRNA isolated from the cells using random primers.Briefly, the mRNA was reverse transcribed at 42°C for1 hour using M-MLV reverse transcriptase. After treat-ment with RNase, the reaction was terminated by heat-ing at 100°C for 10 minutes. The samples were placedon ice, centrifuged briefly, and stored at −80°C until use.

Polymerase Chain Reaction (PCR)For PCR amplifications, an air thermo-cycler‡‡ wasused. Denaturation at 94°C for 1 second, annealing at55°C for 1 second, and extension at 72°C for 15 sec-onds was routinely performed for 50 cycles. Amplifica-tion of the cDNA with primers designed to detect theubiquitously expressed gene for hypoxanthine phos-phoribosyl transferase (HPRT) was first routinely per-formed to determine the success of the generation ofcDNA from mRNA by reverse transcription. These ampli-fications were performed in a stock buffered solutionincluding magnesium (1.5 mM) dNTPs, 5′ and 3′ HPRTprimers, cDNA, and Taq DNA polymerase.§§ Amplifi-cation of the cDNA was performed similarly to detect c-myc expression, using specific primers for c-myc � � (479bp product). In all PCR experiments negative controlscontaining HPLC water instead of cDNA were includedto rule out DNA contamination of the stock solutions. Apositive control fragment for amplifications was includedwith the c-myc primer set. The PCR products were elec-trophoresed on ethidium bromide-stained agarose gelsand photographed under ultraviolet light.

For quantitative PCR, stock reaction mixtures wereprepared as above, but containing both the HPRT primerset (used as an internal control) and the c-myc primerset, and dCTP coupled to the rhodamine dye.¶¶ Thec-myc PCR products were normalized to HPRT, used asa reference gene. These PCR amplifications were per-formed for 35 cycles, which is in the exponential rangeof amplification. After the amplification, unincorporatedlabeled dCTPs were removed by extraction with phenol/water/chloroform (68:18:14). The PCR products wereprecipitated by incubation for 30 minutes at −80°C in100% ethanol containing 0.1M sodium acetate pH 4.5.Following centrifugation at 16,000 × g for 15 minutes,the pellets were washed with 70% ethanol, dried by cen-trifugation under vacuum,## resuspended in 5 µl HPLCwater, and stored at 4°C until analysis. The sampleswere electrophoresed using a DNA sequencing sys-

†† Micro-FastTrack, Invitrogen Corp., San Diego, CA.‡‡ Idaho Technology, Inc., Boise, ID.§§ Fisher Scientific, Pittsburgh, PA.� � Clontech Laboratories, Inc., Palo Alto, CA.¶¶ TAMRA, Perkin-Elmer Corp., Foster City, CA.## Savant Speed VAC Concentration System, Global Medical

Instrumentation, Inc., Albertville, MN.

3062.qxd 3/10/04 9:23 AM Page 362

Table 1.

BrdU Incorporation by Normal and HGF Fibroblasts After Serum Stimulation

Time (hours)

Cell Line 1 6 12 24 48 72 96 120

GN 43 ND 0.01 (0.01) 0.08 (0.03) 0.7 (0.03) 1.5 (0.04) 1.61 (0.04) 1.5 (0) 1.49 (0.05)

GN 45 ND ND 0.06 (0) 0.14 (0.02) 0.7 (0.07) 0.8 (0.04) 0.73 (0.01) 0.9 (0.015)

GN 46 0.01 (0.01) ND 0.07 (0.001) 0.12 (0.03) 0.19 (0.02) 0.14 (0.04) 0.1 (0.02) 0.17 (0.025)

GH 8 0.02 (0.015) ND 0.03 (0.01) 0.9 (0.02) 1.6 (0) 1.64 (0.04) 1.6 (0.015) 1.51 (0.110)

GH 10 ND ND 0.07 (0.01) 1.3 (0.1) 1.6 (0.01) 1.65 (0.04) 1.6 (0) 1.6 (0.01)

GH 11 ND ND 0.06 (0.02) 0.76 (0.07) 1.6 (0.02) 1.64 (0.04) 1.54 (0.1) 1.6 (0.03)

Fibroblasts (1 × 104) were seeded into wells of 96-well plates, allowed to attach overnight, and rendered synchronous and quiescent in serum-free medium.After 48 hours, complete medium containing BrdU was added and incubated for 1, 6, 12, 24, 48, 72, 96, and 120 hours. BrdU incorporation was measured byELISA. The data are expressed as Abs.450 nm. The figures in parentheses represent the standard deviation of triplicate samples in multiple experiments.ND = not detected.

363

J Periodontol • March 2004 Tipton, Woodard, Baber, Dabbous

tem.*** The use of special software allowed detectionof the size of the PCR products and quantitation bycalculating peak height and area of the fluorescentsignals.††† The electrophoresis was performed in theMolecular Resource Center, University of TennesseeHealth Science Center.

Statistical AnalysisStatistical analysis was performed by one-way analysisof variance (ANOVA) using a software program.‡‡‡

RESULTSProliferationThe proliferation of three GN and three HGF cell linesat quiescence or at 1 to 120 hours after serum stimu-lation is shown in Table 1. None of the cell lines hadmore than minimal proliferation at periods ≤12 hoursafter stimulation. The normal cell lines GN 45 and GN 46grew more slowly than all of the HGF cell lines at alltime periods ≥24 hours (P ≤0.002). The HGF cell lineshad proliferation patterns very similar to one another.GN 43 had a proliferation pattern similar to the HGFcells; however, it proliferated significantly more slowlythan GH 8 at 24, 48, and 96 hours after stimulation(P ≤0.02); GH 10 at 24, 48, and 120 hours (P ≤0.03);and GH 11 at 48, 96, and 120 hours (P ≤0.03). GN 43,GN 46, and the HGF cell line GH 8 were selected for fur-ther analysis of c-myc expression at quiescence and 1,6, 24, 48, and 96 hours after stimulation of proliferation,and the proliferation of these cell lines at these specifictimes is shown in Table 2.

Expression of c-myc mRNAmRNA was isolated from GN 43, GN 46, and GH 8 atquiescence (time 0), and at 1, 6, 24, 48, and 96 hoursafter exposure to serum, and used to synthesize cDNA.

The cDNA samples were first amplified with HPRTprimers to determine the success of cDNA generationfrom mRNA by reverse transcriptase, and all sampleswere positive for HPRT expression (data not shown).Before quantitative PCR was performed, the cDNAsamples were evaluated qualitatively for c-myc expres-sion. Figure 1 shows that at quiescence and at all timeperiods after serum stimulation the normal and HGFcell lines all expressed c-myc.

Figure 2 shows the results of the quantitative PCRanalysis of the normal and HGF cell lines. While all threecell lines expressed c-myc at quiescence, GH 8 did soat almost twice the level of the two normal cell lines.Maximum expression of c-myc in each cell line occurred1 hour after stimulation of proliferation, and in GH 8,reached twice the peak height of expression in the nor-mal cell lines. c-myc expression in all cell lines decreasedto quiescent levels at ≥6 hours after stimulation, withGH 8 maintaining approximately the same 2:1 elevatedexpression ratio compared to the normal cells.

Effect of c-myc ODNs on HGF FibroblastProliferationThe addition of a single dose of c-myc antisense ODN(4 or 8 µM) caused dose- and time-dependent inhi-bition of GH 8 proliferation. Figure 3 shows that 2 µMc-myc antisense ODN did not significantly reduce pro-liferation at any time period tested, up to 144 hoursafter stimulation and exposure to the ODN, exceptat 48 hours (P = 0.03). The addition of 4 µM c-mycantisense ODN inhibited proliferation at periods of

*** 373 DNA Sequencer, Applied Biosystems, Foster City, CA.††† GENESCAN 672 Software, Applied Biosystems.‡‡‡ Statview SE + Graphics, Abacus Concepts, Inc., Berkeley, CA.

3062.qxd 3/10/04 9:23 AM Page 363

364

c-myc Proto-Oncogene in Hereditary Gingival Fibromatosis Volume 75 • Number 3

≥96 hours (P ≤0.007) (Fig. 4), and 8 µM c-myc anti-sense ODN inhibited proliferation at all time periodstested (P ≤0.003) (Fig. 5).

Specific decreases in proliferation (antisense versussense at a given time period) caused by 2, 4, or 8 µMODN are shown in Figure 6. In general there weregreater specific decreases in proliferation with longerexposure to 4 µM ODN (7% to 14%) and 8 µM ODN(57% to 82%). At 2 µM ODN, specific decreases in pro-liferation ranged from 0% to 8% and did not correlatewith time of exposure to the ODN. Exposure to 4 µM

ODN caused greater specific decreases inproliferation compared to 2 µM at 96 and 144hours (P ≤0.001). Exposure to 8 µM ODNcaused greater specific decreases in prolifer-ation compared to 2 µM and 4 µM at all timestested (P ≤0.0003).

Reversal of c-myc Antisense Effecton ProliferationThe sense and antisense c-myc ODNs werepreincubated for 30 minutes at 37°C beforeaddition to the cell cultures. The ODNs wereused at equimolar concentrations or at 2 or3 M sense excess. Figure 7 shows that 8 µMc-myc antisense ODN caused approximately80% inhibition of proliferation. Preincubat-ing the antisense ODN with an equimolaramount of sense ODN partially reversed the

antisense effects, resulting in only 55% decrease inproliferation. A 2- or 3-fold molar excess of sense ODNdid not reverse the antisense effects further (data notshown).

DISCUSSIONIn this study, the relationship between c-myc expres-sion and elevated proliferation in HGF fibroblasts wasinvestigated to determine whether the expression of thisproto-oncogene was also elevated in HGF fibroblastsduring quiescence and after stimulation of proliferation.

Figure 1.c-myc expression by normal and HGF fibroblasts at quiescence (0 hr) or at specific times after serum stimulation. cDNA was amplified using specificprimers for c-myc (479 bp product).The PCR products were electrophoresed on ethidium bromide-stained agarose gels and photographed underultraviolet light. A) Lane 1: bp standards; Lane 8: positive control; Lane 2: GN 43 0 hr; Lane 3: GN 46 0 hr; Lane 4: GH 8 0 hr; Lane 5: GN 43 1 hr; Lane6: GN 46 1 hr; Lane 7: GH 8 1 hr; B) Lane 1: bp standards; Lane 8: positive control; Lane 2: GN 43 6 hr; Lane 3: GN 46 6 hr; Lane 4: GH 8 6 hr; Lane 5:GN 43 24 hr; Lane 6: GN 46 24 hr; Lane 7: GH 8 24 hr; C) Lane 1: bp standards; Lane 8: positive control; Lane 2: GN 43 48 hr; Lane 3: GN 46 48 hr;Lane 4: GH 8 48 hr; Lane 5: GN 43 96 hr; Lane 6: GN 46 96 hr; Lane 7: GH 8 96 hr.

Table 2.

BrdU Incorporation by GN 43, GN 46, and GH 8Fibroblasts After Serum Stimulation

Time (hours)

Cell Line 1 6 24 48 96

GN 43 ND 0.01 (0.01) 0.7 (0.03) 1.5 (0.04) 1.5 (0)

GN 46 0.01 (0.01) ND 0.12 (0.03) 0.19 (0.02) 0.1 (0.02)

GH 8 0.02 (0.015) ND 0.9 (0.02) 1.6 (0) 1.6 (0.015)

Fibroblasts (1 × 104) were seeded into wells of 96-well plates, allowed to attach overnight,and rendered synchronous and quiescent in serum-free medium. After 48 hours, completemedium containing BrdU was added and incubated for 1, 6, 12, 24, 48, and 96 hours. BrdUincorporation was measured by ELISA. The data are expressed as Abs.450 nm. The figures inparentheses represent the standard deviation of triplicate samples in multiple experiments.ND = not detected.

3062.qxd 3/10/04 9:23 AM Page 364

365

J Periodontol • March 2004 Tipton, Woodard, Baber, Dabbous

Figure 2.Quantitation of c-myc expression in normal and HGF fibroblasts atquiescence and after serum stimulation. For quantitative PCR, cDNAwas amplified with HPRT and c-myc primers, and dCTP coupled to therhodamine dye. Unincorporated labeled dCTPs were removed byextraction with phenol/water/chloroform and PCR products wereprecipitated with 100% ethanol containing 0.1M sodium acetate pH4.5.The products were centrifuged, washed, dried, and finallyresuspended in HPLC water.The samples were electrophoresed on aDNA sequencer, and a software program allowed detection of the sizeof the PCR products and quantitation by calculating peak height andarea of the fluorescent signals.The c-myc PCR products werenormalized to HPRT, used as a reference gene.

Figure 3.Effect of c-myc sense or antisense ODN on GH 8 proliferation(2 µM). For determination of the effect of ODNs on proliferation, theexperiments were performed as described in Figure 2, except that TEbuffer, or 2 µM sense or antisense ODNs were added at the timeof addition of the serum-containing medium after the 48-hour serumstarvation period.The cells were exposed to the ODNs for 48, 96,120, or 144 hours, and BrdU was added for the final 24 hours ofincubation. BrdU incorporation was measured by ELISA.

Figure 4.Effect of c-myc sense or antisense ODN on GH 8 proliferation (4 µM).For determination of the effect of ODNs on proliferation, theexperiments were performed as described in Figure 2, except that TEbuffer, or 4 µM sense or antisense ODNs were added at the time ofaddition of the serum-containing medium after the 48-hour serumstarvation period.The cells were exposed to the ODNs for 48, 96, 120,or 144 hours, and BrdU was added for the final 24 hours ofincubation. BrdU incorporation was measured by ELISA.

Other studies have shown that elevated and abnormalc-myc expression is associated with increased fibroblastproliferation in both fibrotic scleroderma skin and skinfrom patients with systemic sclerosis.12,15

The HGF fibroblasts proliferated at higher rates thanGN fibroblasts in vitro, and in all cell lines, there wasa lag period of 12 to 24 hours after the addition ofserum before appreciable proliferation could be mea-sured. While the HGF cell lines had somewhat similarproliferation profiles, there was variability among thenormal cell lines. Fibroblast cell lines are heterogeneouswith regard to a number of functional activities.25-28

GN 46 and GN 45 proliferated more slowly than all ofthe HGF cell lines ≥24 hours after serum stimulation.While GN 43 behaved similarly to the HGF cell lines,it still proliferated significantly more slowly than thethree HGF cell lines at most times tested. There areconflicting reports in the literature of the proliferationof HGF fibroblasts in vitro. Consistent with the resultsof our work,5 Coletta et al.29 found that, compared tonormal gingival fibroblasts, proliferation rates were sig-nificantly higher in fibroblasts derived from HGF gin-giva. Similarly, other investigators have found enhancedgrowth rates and in vitro longevity in fibroblasts from

3062.qxd 3/10/04 9:23 AM Page 365

366

c-myc Proto-Oncogene in Hereditary Gingival Fibromatosis Volume 75 • Number 3

fibrotic lung tissue.25,30 Other studies report that HGFfibroblasts have proliferation rates that are the sameor slower than those of normal gingival fibroblasts.2,3

Our previous studies showed that TGF-β appeared tobe an HGF autocrine stimulatory factor for fibronectinand collagen production, but not proliferation, sug-

gesting other means of activation of HGF pro-liferation9 (and unpublished observations).This is in contrast to a study by de Andradeet al., which suggested a role for TGF-β1autocrine stimulation of HGF fibroblast pro-liferation.31

Inconsistencies in the description of HGFfibroblasts and their functions, such as prolif-eration, may, in part, be due to differences inculturing conditions, interindividual variationsin the patients from whom the biopsy mate-rials were obtained, or the genetic hetero-geneity of HGF and phenotypic differences infibroblasts from related yet etiologically dif-fering gingival fibroses. Gingival fibromatosisoccurs in several genetically related dis-eases,32-34 as do other genetic disorders thataffect the orofacial structures.35 Studies haveidentified loci for autosomal dominant, non-syndromic forms of HGF to chromosomes2p21-p2236,37 and 5q13-q22.38 Duplications,deletions, or other chromosomal abnormali-ties have been reported for syndromic formsof HGF.39-41 Hart et al.42 have also identifieda mutation in the Son of Sevenless-1 (SOS1)gene in members of a family with HGF.

Two normal (GN 43 and GN 46) cell linesand one HGF cell line (GH 8) were chosenfor more in-depth study, and all expressedc-myc at quiescence and at 1, 6, 24, 48, and96 hours after serum stimulation. c-myc is anuclear proto-oncogene and is induced whennormal cells are stimulated to grow, indicatingits direct role in growth control.14 It appearsto play a central role in response of cells togrowth factors and the transcription of othergenes.16 After stimulation with growth factors,in many types of cells there is an initial rise inc-myc expression, which later decreases, butremains somewhat elevated.14 Thus, c-mycappears to be expressed throughout the cellcycle in proliferating cells, suggesting that thec-myc product is important in initiating pro-liferation as well as in maintaining it. Quanti-tative PCR was used to deduce the startinglevels of c-myc among the different cDNAsamples. There can be obstacles in obtainingquantitative information using PCR, primarilydue to the fact there are two sequential enzy-matic steps involved: the synthesis of DNA

from the RNA template, and the polymerase chain reac-tion itself. In this study, the entire population of mRNAin each sample was converted to cDNA by the use ofrandom primers, which tend to normalize the efficiencyof cDNA synthesis. In addition, an endogenous sequencewas used as an internal standard in quantitative exper-

Figure 5.Effect of c-myc sense or antisense ODN on GH 8 proliferation (8 µM). Fordetermination of the effect of ODNs on proliferation, the experiments were performedas described in Figure 2, except that TE buffer, or 8 µM sense or antisense ODNswere added at the time of addition of the serum-containing medium after the48-hour serum starvation period.The cells were exposed to the ODNs for 48, 96,120, or 144 hours, and BrdU was added for the final 24 hours of incubation. BrdUincorporation was measured by ELISA.

Figure 6.Specific decreases in GH 8 proliferation caused by c-myc antisense ODN. Specificdecreases caused by c-myc antisense ODN were calculated by subtracting thepercentage decrease caused by antisense ODN from the specific decrease causedby sense ODN.

3062.qxd 3/10/04 9:23 AM Page 366

367

J Periodontol • March 2004 Tipton, Woodard, Baber, Dabbous

iments, and therefore the reference mRNA and the tar-get mRNA were processed together for the duration ofthe experiment, from RNA extraction through PCRamplification. This tends to minimize differences in RNAyield between samples, and this is important for smallsamples, such as those used in this study, in which RNAquantities are too small to measure by ultraviolet spec-trophotometry. The samples were amplified with the c-myc and reference (HPRT) primers simultaneously,which also tends to normalize results since the ampli-fications for the two sequences were carried out underidentical conditions. The yield of PCR product is pro-portional to the starting amount of template under con-ditions in which PCR amplification is proceedingexponentially, and this is typically in the range of 25 to35 cycles of amplification, which was used in theseexperiments.43

The quiescent HGF fibroblasts had greater basalexpression of c-myc than quiescent normal fibroblasts.Higher c-myc mRNA levels in the resting HGF fibro-blasts suggests that despite the lack of serum (andthus exogenous growth factors) some cells were stillproliferating (more so than the normal fibroblasts).Trojanowska et al.12 also found that scleroderma fibro-blasts expressed greater levels of c-myc at quiescence,and that 5% of those cells were proliferating under thoseconditions compared to 1.5% of the cells in the normalcultures. c-myc expression in all the fibroblast cell linespeaked approximately 1 hour after serum stimulation,but at 2-fold higher levels in the HGF cells, comparedto GN fibroblasts, and remained elevated compared tothe normal cells, as long as 96 hours after stimulation.

GN 43 had a growth profile more similar tothe HGF cell lines than the other normal ones,although it still proliferated significantly moreslowly than the HGF cell lines at most timestested. However, its c-myc expression wassimilar to the other normal cell line tested,suggesting that in this cell line, there may begrowth stimulatory factors other than c-mycexpression which result in its higher prolifer-ative rates in relation to the other normal celllines. The mechanism of the stimulation ofc-myc expression in HGF fibroblasts ispresently unknown, but could possibly includegene amplification, retrovirus insertion, orchromosomal translocation.44 In view of therole of the c-myc product in regulating geneexpression and cell growth, the elevatedexpression of c-myc may contribute to theincreased proliferation as well as to the over-all activated phenotype of HGF fibroblasts.

The HGF cell line GH 8 was further inves-tigated for the effects of c-myc antisense ODNon proliferation. A single exposure of thesecells to sufficient concentrations of antisense

ODN caused dose- and time-dependent inhibition of pro-liferation, consistent with another study which used HL-60 promyelocytic cells.45 The ODNs used in the presentstudy were phosphorothioate derivatives, which are effi-cacious and also have improved resistance to nucle-ases. However, such derivatives can have non-specificeffects on cells, due to the binding of phosphorothioatesto proteins and by the release of thioated nucleotidesafter degradation, which likely are toxic to cells. Thereasons why some control ODNs have strong effects onsome cell types is not well understood. In our study,non-specific effects of the sense ODNs were small. Fur-thermore, we calculated specific reductions in prolifer-ation, taking into account any effects of the control(sense) ODNs, and found that there were still specificinhibitory effects of the antisense ODN. The antisenseODN concentration required to have an effect on theHGF fibroblasts in our study was 4 to 8 µM. Huefeleret al.45 and Holt et al.24 also saw effects of c-myc anti-sense ODN over the range of 1 to 8 µM. Consistent withthe results of our study, Capeans et al.46 found that 2 µMc-myc antisense ODN had no effect on BrdU incorpo-ration by retinal pigment epithelial cells, but that 4 µMsignificantly inhibited it. Exposure periods of up to sev-eral days, depending on the ODN concentration, werenecessary to have an effect on both HGF and normalgingival fibroblast proliferation. Antisense ODNs inhibitthe de novo synthesis of proteins, and biological effectsare seen only after enough pre-existing protein has beendegraded while the antisense ODN is inhibiting the syn-thesis of new protein. In addition, 4 to 16 hours are usu-ally required for ODN uptake. ODNs are taken up into

Figure 7.Effect of c-myc sense, antisense, or sense + antisense ODN on GH 8 proliferation.Theexperiments were performed as described in Figure 5, except that some cells receivedantisense ODN with the sense ODN [equimolar amounts; or 2:1 or 3:1 (molar)sense:antisense] preincubated for 30 minutes at 37°C before addition to the cultures.The cells were exposed to the ODNs for 120 hours, and BrdU was added for the final24 hours of incubation. BrdU incorporation was measured by ELISA.

3062.qxd 3/10/04 9:23 AM Page 367

368

c-myc Proto-Oncogene in Hereditary Gingival Fibromatosis Volume 75 • Number 3

cells by active transport mechanisms and a putativereceptor has been identified. This type of uptake requiresmore time than passive diffusion, so it is not unusual forbiological effects such as inhibition of proliferation to beseen only after 2 to 7 days of exposure to the ODN. Ourresults [i.e., inhibition, which was incomplete (~80%maximum), after at least 2 days of exposure to the anti-sense ODN] are consistent with those of Bondurant et al.,who studied antisense c-myc inhibition of proliferation oferythroid progenitor cells.47 Cerutti et al.48 also observed80% inhibition of human thyroid carcinoma cells byc-myc ODN after 72 hours of exposure, with the senseODN causing <10% inhibition of proliferation. Incom-plete inhibition may be due to inefficient delivery of theODN into some cells, or division of some cells with littleor no c-myc protein.48 The degree of inhibition causedby most antisense ODNs is in the range of 50% to 90%(unpublished communication with Chemicon Interna-tional, Temecula, California).

Annealing the sense ODN to the antisense ODN onlypartially reversed the effects of the antisense ODN alone.This was true even in the presence of 2- or 3-fold molarexcess of sense ODN. A greater excess of sense ODN(up to 10-fold molar excess) may be required to fullyreverse the inhibition caused by antisense ODN alone.24

In addition, incomplete reversal may indicate that thereare some toxic effects due to impurities in the ODNpreparation, or toxicity may be caused by its binding toother proteins. The cause for the incomplete reversal inour system remains to be determined.

In summary, the elevated expression of c-myc mRNAin the GH 8 HGF fibroblast cell line, and the reversibleinhibition of its proliferation by c-myc antisense ODN,suggest a role for c-myc overexpression in the elevatedproliferation of HGF fibroblasts and in the clinical gingi-val enlargement characteristic of HGF.

ACKNOWLEDGMENTSThe authors thank Dr. William S. Walker, Departmentof Immunology, St. Jude Children’s Research Hospital,Memphis, Tennessee, and his laboratory personnel formaking his laboratory available for PCR experiments.The authors also thank Dr. Michael Dockter and KimPrince of the Molecular Resource Center, University ofTennessee Health Science Center, for helpful discus-sion and assistance in quantitative PCR. This work wassupported in part by USPHS grant DE07258 fromNIDCR.

REFERENCES1. Jorgenson R, Cocker M. Variation in inheritance and

expression of gingival fibromatosis. J Periodontol 1974;45:472-477.

2. Shirasuna K, Okura M, Watatami I, Hayashido Y, Saka M,Matsuya T. Abnormal cellular property of fibroblasts fromcongenital gingival fibromatosis. J Oral Pathol 1989;7:381-385.

3. Johnson B, el-Guindy M, Ammons W, Narayanan A,Page R. A defect in fibroblasts from an unidentified syn-drome with gingival hyperplasia as the predominantfeature. J Periodont Res 1986;21:403-413.

4. Collan Y, Ranta H, Vartio T, Perheentupa J, Raeste A.Histochemical and biochemical study of hereditaryfibrous hyperplasia of the gingiva. Scand J Dent Res1982;90:20-28.

5. Tipton DA, Howell KJ, Dabbous MKh. Increased prolifer-ation, collagen, and fibronectin production by hereditarygingival fibromatosis fibroblasts. J Periodontol 1997;68:524-530.

6. Varga J, Rosenbloom J, Jimenez S. Transforming growthfactor beta (TGF beta) causes a persistent increase insteady-state amounts of type I and type III collagen andfibronectin mRNAs in normal human dermal fibroblasts.Biochem J 1987;247:597-604.

7. Sporn M, Roberts A, Wakefield L, de Crombrugghe B.Some recent advances in the chemistry and biology oftransforming growth factor-beta. J Cell Biol 1987;105:1039-1045.

8. Ignotz R, Massague J. Transforming growth factor-betastimulates the expression of fibronectin and collagenand their incorporation into the extracellular matrix. J BiolChem 1986;261:4337-4345.

9. Tipton DA, Dabbous MKh. Autocrine transforminggrowth factor β stimulation of extracellular matrix pro-duction by fibroblasts from fibrotic human gingiva. JPeriodontol 1998;69:609-619.

10. Freytag SO. Enforced expression of the c-myc onco-gene inhibits cell differentiation by precluding entry intoa distinct predifferentiation state in G0/G1. Mol Cell Biol1988;8:1614-1624.

11. Heikkila R, Schwab G, Wickstrom E, et al. A c-myc anti-sense oligodeoxynucleotide inhibits entry into S phase butnot progress from G0 to G1. Nature 1987;328:445-449.

12. Trojanowska M, Wu L, LeRoy E. Elevated expression ofc-myc proto-oncogene in scleroderma fibroblasts. Onco-gene 1988;3:477-481.

13. Evan G, Littlewood T. The role of c-myc in cell growth.Curr Opin Genet Dev 1993;3:44-49.

14. Darnell J, Lodish H, Baltimore D. Cancer. In: MolecularCell Biology, 2nd ed. New York: W.H. Freeman and Co.;1990:955-1002.

15. Parkes JL, Cardell RR, Hubbard FC, Meltzer A, Penn A.Cultured human atherosclerotic plaque smooth musclecells retain transforming potential and display enhancedexpression of the myc proto-oncogene. Am J Pathol1991;138:765-775.

16. Feghali CA, Boulware DW, Ferriss JA, Levy LS. Expres-sion of c-myc, c-myb, and c-sis, in fibroblasts fromaffected and unaffected skin of patients with systemicsclerosis. Autoimmunity 1993;16:167-171.

17. Walder RY, Walder JA. Role of RNase H in hybrid-arrestedtranslation by antisense oligonucleotide. Proc Natl AcadSci (USA) 1988;85:5011-5015.

18. Kim SK, Wold BJ. Stable reduction of thymidine kinaseactivity in cells expressing high levels of antisense RNA.Cell 1985;42:129-138.

19. Scherczinger CA, Knecht DA. Systematic analysis ofantisense RNA inhibition of myosin II heavy chain geneexpression in Dictyostelium discoideum. Antisense ResDev 1993;3:191-205.

20. Tipton DA, Pabst MJ, Dabbous MKh. Interleukin-1β- andtumor necrosis factor α-independent monocyte stimu-lation of fibroblast collagenase activity. J Cell Biochem1990;44:253-264.

3062.qxd 3/10/04 9:23 AM Page 368

369

J Periodontol • March 2004 Tipton, Woodard, Baber, Dabbous

21. Robinson-Benion C, Holt JT. Antisense techniques. Meth-ods Enzymol 1995;254:363-375.

22. Tobey R, Valdez J, Crissman H. Synchronization ofhuman diploid fibroblasts at multiple stages of the cellcycle. Exp Cell Res 1988;179:400-416.

23. Holt JT, Redner RL, Nienhuis AW. An oligomer compli-mentary to c-myc mRNA inhibits proliferation of HL-60promyelocytic cells and induces differentiation. Mol CellBiol 1988;8:963-973.

24. Greenberg ME, Ziff EB. Stimulation of 3T3 cells inducestranscription of the c-fos proto-oncogene. Nature 1984;311:433-438.

25. Jordana M, Schulman J, McSharry C, et al. Heterogene-ous proliferative characteristics of human adult lung fibro-blast lines and clonally derived fibroblasts from controland fibrotic tissue. Am Rev Resp Dis 1988;137:579-584.

26. Bordin S, Page RC, Narayanan AS. Heterogeneity ofnormal human diploid fibroblasts: Isolation and charac-terization of one phenotype. Science 1984;223:171-173.

27. Hassell TM, Stanek EJ. Evidence that healthy humangingiva contains functionally heterogeneous fibroblastsubpopulations. Arch Oral Biol 1983;28:617-625.

28. Ko SD, Page RC, Narayanan AS. Fibroblast hetero-geneity and prostaglandin regulation of subpopulations.Proc Natl Acad Sci (USA) 1977;74:3429-3432.

29. Coletta RD, Almeida OP, Graner E, Page RC, Bozzo L.Differential proliferation of fibroblasts cultured fromhereditary gingival fibromatosis and normal gingiva.J Periodont Res 1998;33:469-475.

30. Thompson K, Holliday R. Genetic effects on the long-evity of cultured human fibroblasts. Enhanced growthpotential of cystic fibrosis cells. Gerontology 1983;29:97-101.

31. de Andrade CR, Cotrin P, Graner E, Almeida OP, Sauk JJ,Coletta RD. Transforming growth factor-β1 autocrine stim-ulation regulates fibroblast proliferation in hereditary gingi-val fibromatosis. J Periodontol 2001;72:1726-1733.

32. Clark D. Gingival fibromatosis and its related syndromes.A review. J Can Dent Assoc 1987;2:137-140.

33. Witkop CR. Heterogeneity in gingival fibromatosis. BirthDefects Orig Artic Ser 1971;7:210-221.

34. Hart TC, Pallos D, Bozzo L, et al. Evidence of geneticheterogeneity for hereditary gingival fibromatosis. J DentRes 2000;79:1758-1764.

35. Shapiro SD, Jorgenson RJ. Heterogeneity in genetic dis-orders that affect the orofacies. Birth Defects OriginalArtic Ser 1983;19:155-165.

36. Hart TC, Pallos D, Bowden DW, et al. Genetic linkage ofhereditary gingival fibromatosis to chromosome 2p21.Am J Hum Genet 1998;62:876-883.

37. Xiao S, Wang X, Qu B, et al. Refinement of the locus forautomosal domninant hereditary gingival fibromatosis(GINGF) to a 3.8-cM region on 2p21. Genomics 2000;68:247-252.

38. Xiao S, Bu L, Zhu L, et al. A new locus for hereditary gin-gival fibromatosis (GINGF2) maps to 5q13-q22. Genomics2001;74:180-185.

39. Fryns JP. Gingival fibromatosis and partial duplicationof the short arm of chromosome 2 (dup(2)(p13-p21).Ann Genet 1996:39:54-55.

40. Shashi V, Pallos D, Pettenati MJ, et al. Genetic heterogene-ity of gingival fibromatosis on chromosome 2p. J MedGenet 1999;36:683-686.

41. Macias-Flores MA, Garcia-Cruz D, Rivera H, et al. A newform of hypertrichosis inherited as an X-linked dominanttrait. Hum Genet 1984;66:66-70.

42. Hart TC, Zhang Y, Gorry MC, et al. A mutation in the SOS1gene causes hereditary gingival fibromatosis. Am J HumGenet 2002;70:943-954.

43. Noonan K, Beck TA, Holzmayer J, et al. Quantitativeanalysis of MDR1 (multidrug resistance) expression inhuman tumors by polymerase chain reaction. Proc NatlAcad Sci (USA) 1990;87:7160-7164.

44. Kelly K, Cochran B, Stiles C, Leder P. Cell-specific regu-lation of the c-myc gene by lymphocyte mitogens andplatelet-derived growth factor. Cell 1983;35:603-610.

45. Heufelder A, Bahn RS. Modulation of cellular functions inretro-orbital fibroblasts using antisense oligonucleotidestargeting the c-myc protooncogene. Invest Ophthalmol VisSci 1995;36:1420-1432.

46. Capeans C, Piñeiro A, Dominguez F, et al. A c-myc anti-sense oligonucleotide inhibits human retinal pigment cellproliferation. Exp Eye Res 1996;66:581-589.

47. Bondurant MC, Yamashita T, Muta K, Krantz SB, KouryMJ. C-myc expression affects proliferation but not termi-nal differentiation or survival of explanted erythroid pro-genitor cells. J Cell Physiol 1996;168:255-263.

48. Cerutti J, Trapasso F, Battaglia C, et al. Block of c-mycexpression by antisense oligonucleotides inhibits proli-feration of human thyroid carcinoma cell lines. Clin Can-cer Res 1996;2:119-126.

Correspondence: Dr. David A. Tipton, Dental Research Centerand Department of Periodontology, College of Dentistry,University of Tennessee Health Science Center, 894 Union Ave.,Memphis, TN 38163. Fax: 901/448-7860; e-mail: dtipton@utmem.edu.

Accepted for publication July 7, 2003.

3062.qxd 3/10/04 9:23 AM Page 369