J Periodontol • March 2005

443

* Department of Human Morphology-Interdisciplinary Laboratories of Advanced Technologies, Segrate, University of Milan, Milan, Italy.† Department of Human Morphology, University of Milan.‡ Department of Morphology and Embryology, Section of Histology and Embryology, University of Ferrara, Ferrara, Italy.§ Institute of Biomedical Techniques, Research National Center, University of Milan.

Depletion of Polyamines and Increase ofTransforming Growth Factor-ββ1, c-myc,Collagen-Type I, Matrix Metalloproteinase-1,and Metalloproteinase-2 mRNA in PrimaryHuman Gingival FibroblastsGiordano Stabellini,* Claudia Moscheni,* Nicoletta Gagliano,* Claudia Dellavia,† Carla Calastrini,‡Maria Elena Ferioli,§ and Magda Gioia*

Background: The polyamines spermidine, spermine, and putrescine are known to be deeply linked withgrowth processes, gene expression, and extracellular matrix synthesis. Their cellular content dependsprimarily on the activity of the enzyme ornithine decarboxylase. High levels of ornithine decarboxylase andpolyamines have been found in proliferative, inflammatory, and neoplastic pathologies of the oral cavity andin gingival fluid. Difluoromethylornithine (DFMO) selectively inhibits ornithine decarboxylase, thus depletingpolyamine content and preventing cell proliferation and synthesis activity. The aim of this study was toinvestigate whether DFMO treatment could modify the genes involved in cell proliferation and extracellularmatrix turnover.

Methods: Fibroblasts derived from non-inflamed gingiva were maintained in Dulbecco’s modified Eagle’smedium (DMEM) plus α-difluoromethylornithine for 4 days. At 0, 24, 48, 72, and 96 hours cell number wasassessed, polyamine levels were quantified with high performance liquid chromatography (HPLC) method,and transforming growth factor-β1 (TGF-β1), c-myc, matrix metalloproteinases (MMP)-1 and 2, collagen type I(COL-I) and tissue inhibitor of matrix metalloproteinases (TIMP)-1 were evaluated by reverse transcription-polymerase chain reaction (RT-PCR).

Results: Fibroblasts treated with DFMO significantly decreased cell proliferation, ornithine decarboxylaseactivity, and putrescine levels at all treatment times, spermidine after 72 and 96 hours, and spermine after96 hours of culture. Total polyamines decreased (P ≤0.01) at 96 hours after DFMO treatment, while c-myc,TGF-β1, MMP-1 and 2, COL-I mRNA significantly increased. Conversely, TIMP-1 did not show any signifi-cant change. The polyamines trend was not correlated to c-myc, TGF-β1, MMP-1 and -2, and TIMP-1 mRNAlevels. Transforming growth factor-β1 and c-myc mRNA expression were related and correlated to MMP-1and 2, COL-I and TIMP-1 mRNA trend after DFMO treatment.

Conclusions: Our data show that as the polyamine content decreases, TGF-β1, c-myc, MMP-1 and -2,and COL-I mRNA levels increase, therefore a negative regulatory role of the polyamines on the mRNAexpression could be suggested. J Periodontol 2005;76:443-449.

KEY WORDSComparison studies; genes, c-myc; growth factors, transforming; putrescine; spermidine; spermine.

4012.qxd 4/4/05 3:33 PM Page 443

444

Polyamine, TGF-ββ1, c-myc, COL-I, MMP-1, MMP-2 mRNA in Gingival Fibroblasts Volume 76 • Number 3

The polyamines (PA) such as putrescine (PUT),spermidine (SPD), and spermine (SPM) are nat-ural polycations that link negatively charged

DNA, RNA, proteins, phospholipids, and nucleosidetriphosphates. This ability may explain their physio-logical role in cell proliferation and differentiation.1,2

Variations among different PA, in the number and in thedistance of amino groups, provide for their functionalspecificity.2 Nevertheless it has been difficult to definetheir biological role, PA are required for transcriptionof the proto-oncogenes c-myc and c-fos; in particu-lar, SPD preferentially stimulates the transcription andthe expression of c-myc while PUT of c-fos.3 PA canplay a role in the transduction of signals between cellmembrane and nucleus;4 they may be involved in TGF-β signal transduction5 and are necessary for normalexpression of the TGF-β gene during cell migration.Their cellular content is closely regulated and is pri-marily dependent on the activity of enzyme ornithinedecarboxylase (ODC) which catalyzes the conversionof ornithine to PUT, the first step in PA biosynthesis.The gene encoding ODC is a proto-oncogene, tran-scriptional target for c-myc,6 and the ability of c-mycto promote cell cycle progression is due in part to tran-scriptional activation of the ODC gene.7 Difluo-romethylornithine (DFMO) is a competitive suicideenzyme inhibitor that selectively inhibits ODC, thusdepleting PA intracellular content and preventing cellproliferation, and up-regulates c-myc gene expression.8

Depletion of cellular PA by DFMO increases the expres-sion of the TGF-β type I receptor mRNA and protein,but it has no effect on the TGF-β type II receptorexpression.9 PA deficiency inhibits cell attachment toplastic and matrigel and the inhibition of the expressionof cell surface receptors for the binding to extracellularmatrix proteins may be part of the mechanism wherebyPA deficiency retards cell migration.10

By contrast, depletion of intracellular PA has no effecton post-transcriptional regulation c-myc and c-junmRNA levels in IEC-6 cells;11 DFMO treatment pre-vents increased expression of c-myc and c-jun proto-oncogenes at 4 days, but also significantly reducessteady-state levels of c-myc and c-jun mRNA between6 and 12 days.12

TGF-β stimulates PA through L-proline synthesis byinducing the genes that regulate the transport andmetabolism of L-arginine; in addition, TGF-β-stimu-lated collagen production is dependent on L-proline.13

TGF-β1 is a multifunctional cytokine involved in extra-cellular matrix (ECM) turnover, able to increase thesynthesis of ECM components, in particular of COL-I,and to inhibit their degradation by down-regulatingmetalloproteinase expression.

PA and TGF-β are able to increase the synthesis ofother ECM components such as glycosaminogly-cans.14,15 Moreover, the enzyme related to ECM degra-

dation, such as metalloproteinase, was shown to be spe-cific for TGF-β stimulation and independent of substratecomposition.16 Antibody against TGF-β1 reduces MMP-1 increase and MMP-2 expression decrease, whereasTIMP-1 expression was unaffected.17 Moreover, theincreased amounts of TGF-β isoforms isolated in hered-itary gingival fibromatosis fibroblasts confirm the role ofTGF-β in ECM turnover, since it stimulates the produc-tion of fibronectin and COL-I.18

The PA are produced also by the bacteria of micro-bial plaque; in fact they are present in gingival fluid ofdiseased periodontal sites.19 It has been demonstratedthat PA have an active role in oral cavity tumors20 andin inflamed periodontium,21 inhibit chemotaxis, anddysregulate the apoptosis of cells involved in the con-trol of periodontal bacteria such as human polymor-phonuclear leukocytes.19,22 The aim of this study wasto investigate the effects of PA depletion induced byDFMO, analyzing on human gingival fibroblasts thegenes involved in cell proliferation and collagen turn-over, and the possible correlation of their mRNA ex-pression with the PA content.

MATERIALS AND METHODSFibroblast Cell CulturesTissue specimens of non-inflamed periodontal gingivawere obtained from the premolar area during oral surgery(six females, 20 to 30 years old) after obtaining patients’informed consent. Each donor had undergone dentalprophylaxis and 3 weeks of intensive oral hygiene. Eachbiopsy was washed with 0.1 M Dulbecco’s phosphatebuffered saline (D-PBS) and immediately minced withsterile scissors. Tissue fragments were transferred to25 cm2 Nunc flasks and, after adherence, were supple-mented with 5 ml Dulbecco’s modified Eagle’s medium(DMEM)� containing 10% fetal bovine serum,¶ 100 UI/mlpenicillin,# 10 ng/ml streptomycin,# and 25 µg/mlamphotericin B.# Cultures were maintained in humidity-saturated atmosphere (5% CO2, 37°C) and routinelysubcultured after use of 0.1% trypsin 0.02% EDTA forcell release. Experiments were performed with fibroblastsbetween the fourth to sixth passages. Cultures were rou-tinely monitored by a phase-contrast inverted micro-scope** and fibroblast viability was measured by thecell ability to exclude trypan blue.23 The fibroblasts werefixed with 4% paraformaldehyde and 2% sucrose in 0.1MD-PBS, stained with hematoxylin and eosin, and observedby light microscopy.**

Fibroblast TreatmentAt confluence, the culture medium was replaced withcomplete MDM with 5 mM DFMO (DMEM as vehicle).

� Sigma Chemical Co., St. Louis, MO.¶ GIBCO, Paisley, U.K.# Serva, Heidelberg, Germany.** Leitz, Metzlar, Germany.

4012.qxd 4/4/05 3:33 PM Page 444

445

J Periodontol • March 2005 Stabellini, Moscheni, Gagliano, et al.

This dose inhibits the PA cyclewithout toxic phenomena.24 At0, 24, 48, 72, and 96 hours afterDFMO treatment (a generousgift from Centre de ResearchMerrell International, Strasburg,France), c-myc, TGF-β1, MMP-1, MMP-2, COL-I, and TIMP-1mRNA levels were determined;PUT, SPD, and SPM levels; ODCactivity; and cell number wereassessed. Fibroblasts incubatedwith complete medium withoutDFMO served as controls (CT).At established time intervals,fibroblasts were washed in D-PBS, trypsinized, harvested bycentrifugation (100× g, 5 min-utes) and resuspended in Tri-zol†† for RNA extraction, in 0.6.M perchloric acid for PA contentdetermination, or in 5 mM di-thiothreitol for ODC activityevaluation.

Measurement of Polyamine ContentFor polyamine analysis, extracts from cell samples wereprepared in 300 µl of 0.6 N perchloric acid by ultrason-ication and were centrifuged at 12,000 rpm for 20 minwith an Eppendorf microcentrifuge. Aliquots of the super-natant were used for the determination of polyaminesby HPLC‡‡ with fluorescence detection, according toLöser et al.25 with minor modifications. Polyamine lev-els were then DFMO. HPLC calculated based on exter-nal standard curves run within 24 hours of the samplechromatograms. The results were expressed as nmoles(106 cells)−1.

Measurement of Ornithine Decarboxylase ActivityFibroblasts were centrifuged at 50× g for 5 minutes at4°C, washed with PBS, and centrifuged at 50× g for10 minutes at 4°C. The cell pellets were extracted with100 µl of 50 mM Tris-HCl, at pH 7.5, 0.1 mM EDTA,5 mM dithiothreitol, 0.5 mM pyridoxal 5′-phosphate,and 0.5% Triton X-100 (lysis buffer). The cell lysateswere centrifuged at 7,000× g for 15 minutes at 4°C, andthe resulting supernatants were used immediately forODC assay and the determination of protein by theBradford method.26 ODC activity was measured ac-cording to the method of Stefanelli et al.27 Briefly,enzyme assays were carried out in a total volume of50 µl, 40 of which were of cell extract origin. The assaymixture (final concentration) contained 50 mM Tris-HCl, at pH 7.5, 0.1 mM EDTA, 5 mM dithiothreitol,0.05 mM pyridoxal 5′-phosphate, 0.5% Triton X-100,0.4 mM L-ornithine, and 0.1 µCi of 14C-ornithine (sp.

Act. 58 mCi mmol−1).§§ The assays were performed inglass tubes with a filter paper attached to the inside ofthe cap. The filter paper was moistened with 20 µl of1 M methylbenzethonium hydroxide in methanol totrap released 14CO2. The reaction was carried out for1 hour in a 37°C shaker bath; then it was stopped byinjecting 200 µl of 1 M citric acid through the rubbercap. After a further 30 minutes, the incubation filterswere cut out, and radioactivity was counted by scintil-lator counter.� � The results were expressed as nmolesof CO2 released h−1 mg−1protein.

Gene ExpressionTotal RNA was extracted from fibroblasts by Trizol. COL-I, MMP-1 and -2, TIMP-1, TGF-β1, and c-myc mRNAlevels were determined by RT-PCR. After DNase I diges-tion, cDNA from 1 µg of total RNA was reversed tran-scribed following manufacturer’s instructions in 20 µlfinal volume reaction mix containing 5 mM MgCl2, 1×RT buffer, 1 mM dNTPs, 1 µg/µl RNase inhibitor, 15 U/µlAMV transcriptase, and 0.5 µg random primers.¶¶ Theused primer sequences were obtained from publishedsequences and are shown in Table 1 and PCR protocolsamplification in Table 2. Amplification reactions wereperformed in a final volume of 25 µl containing 2.5 µlcDNA, 200 µM four deoxynucleotide triphosphates,

†† Invitrogen, San Giuliano, Milan, Italy.‡‡ LC pump series 410 B10; C18 Nova Pak column, Perkin-Elmer Analytic

Instruments, Monza, Italy.§§ Amersham, International, Buckinghamshire, U.K.�� Kontron Instruments, Milan, Italy.¶¶ Promega, Milan, Italy.

Table 1.

Synthetic Oligonucleotide Primers for RT-PCR

Length End-Product Primer Sequence (base pairs) (base pairs)

COL-I 5′-GGC GGC CAG GGC TCC GAC 18 3473′-AAT TCC TGG TCT GGG GCA CC 20

MMP-1 5′-GGT GAT GAA GCA GCC CAG 18 4373′-CAG TAG AAT GGG AGA GTC 18

MMP-2 5′-CCT CTC CAC TGC CTT CGA TAC ACC 24 1623′-AGC ATC TAT TCT TGG GCA CCG 21

TIMP-1 5′-AGT CAA CCA GAC CAC CTT ATA CCA 24 3863′-TTT CAG AGC CTT GGA GGA GCT GGT C 25

TGF-β1 5′-CAG AAA TAC AGC AAC AAT TCC TGG 24 1863′-TTG CAG TGT GTT ATC CCT GCT GTC 24

c-myc 5′-TCC AGC TTG TAC CTG CAG GAT CTG A 25 3383′-CCT CCA GCA GAA GGT GAT CCA GAC T 25

GAPDH 5′-ATT CCA TGG CAC CGT CAA GGC T 22 5713′-TCA GGT CCA CCA CTG ACA CGT T 22

4012.qxd 4/4/05 3:33 PM Page 445

TIMP-1, TGF-β1, and c-myc was determined relative tothat obtained for GAPDH.

Statistical AnalysisData are the mean ± SD of six duplicate experiments.Statistical analyses were performed using the Studentt test for unpaired data and P values <0.05 were con-sidered significant. Correlation analyses were per-formed between pairs of variables. Significance wasfixed at a probability of 1%.

RESULTSFibroblast CulturesLight microscopy observations of CT and DFMO treatedhuman gingival fibroblasts did not reveal any noticeablemorphological differences. Cell viability was notapparently influenced by DFMO administration. Cellnumber significantly increased for 48 hours; thereaftercells entered in a plateau phase which was maintaineduntil 72 hours, then decreased at 96 hours in bothDFMO-treated and CT. DFMO significantly decreasedcell proliferation (P ≤0.01) at 24, 48, and 72 hours, com-pared to CT (Table 3).

Ornithine Decarboxylase ActivityTime-course of the ODC activity during DFMO treatmentshowed that in primary cultured human gingival fibro-blasts, the enzyme significantly decreased at 6, 12, 24,and 48 hours after DFMO treatment, as compared toCT (Table 4).

Intracellular Polyamine LevelsIn DFMO-treated fibroblasts the intracellularPUT level decreased (P ≤0.01) at 24, 48, 72,and 96 hours compared to CT (Table 5). SPDcontent decreased (P ≤0.01) starting at 72and 96 hours, and SPM significantly decreasedafter 96 hours compared to CT (Table 5).Total PA significantly decreased (P ≤0.01)starting at 48 hours in DFMO-treated fibro-blasts compared to CT.

Gene Expression AnalysisIn Table 6, c-myc, TGF-β1, COL-I, MMP-1,MMP-2, and TIMP-1 mRNA display the sametrend, with a significant increase at 96 hoursin primary cultured human gingival fibro-blasts when compared to CT. After DFMOtreatment, c-myc gene expression decreased(P ≤0.05) at 72 hours and was upregulated(P ≤0.01) after 96 hours with a significant

increase when compared to CT. TGF-β1, COL-I, MMP-1, and MMP-2 displayed the same trend, while thechanges of TIMP-1 mRNA levels were not significant.

446

Polyamine, TGF-ββ1, c-myc, COL-I, MMP-1, MMP-2 mRNA in Gingival Fibroblasts Volume 76 • Number 3

Table 2.

RT-PCR Amplification Conditions

Number Gene Protocol Temperature Time of Cycles

COL-I Denaturation 94°C 1 minuteAnnealing 61°C 90 seconds 35Elongation 72°C 2 minutes

MMP-1 Denaturation 94°C 30 secondsAnnealing 53°C 1 minute 32Elongation 72°C 1 minute

+72°C 10 min to finalize extension

MMP-2 Denaturation 94°C 1 minuteAnnealing 60°C 2 minutes 30Elongation 72°C 3 minutes

TIMP-1 Denaturation 94°C 1 minuteAnnealing 60°C 2 minutes 30Elongation 72°C 3 minutes

TGF-β1 Denaturation 94°C 1 minuteAnnealing 60°C 90 seconds 35Elongation 72°C 2 minutes

c-myc Denaturation 94°C 30 seconds 35Annealing 65°C 1 minuteElongation 72°C 1 minute

GAPDH Denaturation 94°C 30 secondsAnnealing 62°C 1 minute 25Elongation 72°C 1 minute

+72°C 10 min to finalize extension

100 pmol each forward and reverse primers, 2.5 U TaqDNA polymerase.## The RT-PCR products (3 µl) wereresolved by electrophoresis in 1.5% agarose gels, stainedwith ethidium bromide and quantified by densitometricanalysis.*** The expression of COL-I, MMP-1, MMP-2,

Table 3.

Proliferation of Cultured Human Gingival FibroblastsAfter DFMO Treatment at Various Time Intervals(values [1 ×× 10−−3 cells/ml] are mean ±± SD of sixduplicate determinations)

Hours of Culture

Treatment 0 24 48 72 96 R

CT 250 ± 30 340 ± 37 675 ± 72 645 ± 67 290 ± 310.957*

T 250 ± 30 220 ± 27* 540 ± 61* 589 ± 63† 260 ± 33

* P ≤0.01 versus control.† P ≤0.05 versus control.

## Stratagene, La Jolla, CA.*** Image Pro-Plus, Gleichen, Germany.

4012.qxd 4/4/05 3:33 PM Page 446

correlation with COL-I andMMP-2 (Table 7).

DISCUSSIONIn cultured human gingival fi-broblasts, DFMO treatment pro-duced a decrease of fibroblastproliferation, ODC activity, andPUT content at all the timepoints; by contrast, SPD de-creased only from 72 to 96hours after DFMO adminis-tration, and SPM content de-creased only 96 hours aftertreatment. These data are in

agreement with previous studies.1,2 The gradual de-crease of the three PA reflects the trend of the PA cycle(PUT-SPD-SPM). PUT is a precursor of PA and a reduc-tion of ODC activity leads to loss of PUT formation, ren-dering the conversion of PUT to SPD the progressivelimiting step. This is demonstrated by the significantcorrelation between the trends of ODC activity andintracellular PUT levels. Ninety-six hours after DFMOtreatment, intracellular PUT, SPD, and SPM levels weresignificantly diminished, whereas c-myc, TGF-β1, MMP-1, MMP-2, and COL-I mRNA levels increased (P ≤0.01);TIMP-1 gene expression was not affected by DFMOadministration. In CT and DFMO-treated gingivalfibroblasts, the trend of the three PA was not positivelycorrelated to the expression of c-myc, TGF-β1, MMP-1, MMP-2, COL-I, and TIMP-1 mRNA, while in DFMO-treated fibroblasts c-myc, TGF-β1, MMP-1, MMP-2, andCOL-I gene expression was significantly related to thetrend of the respective CT. The absence of a signifi-cant correlation between the intracellular PA trend andthe TGF-β1 and c-myc gene expression in humangingival fibroblasts suggests that PA does not play adirect role in TGF-β1 and c-myc pathways. Moreover,the time course correspondence between the signifi-cant decrease of PUT, SPM, and SPD and the increaseof TGF-β1, c-myc, MMP-1, MMP-2, and COL-I mRNA,together with the correlation (P ≤0.01) of TGF-β1 andc-myc with the MMP-1, MMP-2, COL-I, and TIMP-1mRNA after DFMO treatment, suggests that the PA(although not directly acting) are negative regulatorsin the complex pathways controlling both the expres-sion of TGF-β1, c-myc, MMP-1, MMP-2, COL-I, andTIMP-1 mRNA and their relationship. These observa-tions are in agreement with data showing that the PAdepletion increases TGF-β/TGF-β1 receptor mRNA inIEC-9 cells9,11 and confirm that the PA may play aregulatory role in the expression of the TGF-β1

11,28

and c-myc,8,28 and contribute a similar role in c-myc,MMP-1, MMP-2, and COL-I mRNA expressions. Theabsence of any significant correlation between PA andc-myc gene expression is in contrast with reports that

CorrelationsThe changes of cell proliferation between DFMO-treatedcells and CT were significantly correlated, while thetrends of ODC, PA, TGF-β1, MMP-1, MMP-2, TIMP-1,and c-myc did not show any correlation. The inhibitionof ODC activity by DFMO was correlated to the decreaseof PUT cytoplasmatic levels (P ≤0.01), while no corre-lation was evident between the trends of ODC activity,PUT, SPD, and SPM content and c-myc, TGF-β1, MMP-1, MMP-2, COL-I, and TIMP-1 mRNA levels. C-myc andTGF-β1 mRNA levels were related (P ≤0.01) to MMP-1,MMP-2, COL-I, and TIMP-1 gene expression after DFMOtreatment, while in CT only c-myc showed a significant

447

J Periodontol • March 2005 Stabellini, Moscheni, Gagliano, et al.

Table 4.

Changes of ODC in CT and DFMO-Treated Human GingivalFibroblasts Over Time (values [nmol/h/mg proteins] are themean ±± SD for six duplicate determinations)

Hours of Culture

ODC 0 6 12 24 48

CT 0.438 ± 0.066 1.188 ± 0.154 3.625 ± 0.435 1.624 ± 0.244 4.361 ± 0.544

T 0.438 ± 0.066 0.198 ± 0.032* 0.150 ± 0.024* 0.120 ± 0.063* 0.130 ± 0.072*

* P ≤0.01 versus CT.

Table 5.

Intracellular PUT, SPD, and SPM Levelsin Primary Human Gingival FibroblastsTreated With DFMO (values [nmol/mgproteins] are the mean ±± SD of sixduplicate determinations)

Hours of Culture

Polyamine 24 48 72 96

PUTCT 0.68 ± 0.09* 0.86 ± 0.08* 0.55 ± 0.07* 0.42 ± 0.05*T 0.31 ± 0.04 0.35 ± 0.04 0.25 ± 0.03 0.19 ± 0.03

SPDCT 0.38 ± 0.05 0.60 ± 0.07 0.42 ± 0.05* 0.66 ± 0.09*T 0.38 ± 0.04 0.56 ± 0.06 0.29 ± 0.03 0.30 ± 0.03

SPMCT 1.39 ± 0.17 1.74 ± 0.19 1.45 ± 0.17 1.87 ± 0.21*T 1.41 ± 0.18 1.68 ± 0.22 1.30 ± 0.16 1.38 ± 0.15

Total PACT 2.44 ± 0.29 3.20 ± 0.38† 2.42 ± 0.31* 2.95 ± 0.41*T 2.10 ± 0.23 2.60 ± 0.22 1.83 ± 0.22 1.87 ± 0.25

* P ≤0.01 versus CT.† P ≤0.05 versus CT.

4012.qxd 4/4/05 3:33 PM Page 447

448

Polyamine, TGF-ββ1, c-myc, COL-I, MMP-1, MMP-2 mRNA in Gingival Fibroblasts Volume 76 • Number 3

Table 7.

Correlation Index Between c-myc, TGF-ββ1,COL-I, MMP-1, MMP-2, and TIMP-1 mRNALevels

c-myc TGF-β1

MMP-1CT 0.017 0.567T 0.959* 0.883*

MMP-2CT 0.974† 0.376T 0.956* 0.996*

COL-1CT 0.715† 0.193T 0.970* 0.975*

TIMP-1CT 0.464 0.149T 0.884* 0.721†

TGF-β1CT 0.433T 0.951*

* P ≤0.01.† P ≤0.05.

show they are related to gene expression3 and it sup-ports other findings suggesting a new role for PA asnegative regulators of c-myc expression.8,28 Moreover,ODC is necessary for triggering mitogen-activated pro-tein kinase cascade,29 and SPM provokes conforma-tional changes in the kinase;30 since TGF-β1 stimulatesprotein kinase,31,32 these substances could play a cen-tral role in the pathways linking PA to TGF-β1, c-myc,MMP-1, MMP-2 and COL-I gene expression.

These patterns of mRNA steady-state levels, togetherwith the relationship showing that TGF-β1 may favorECM changes, show that the ODC/PA system in humangingival fibroblasts plays an indirect role in the rela-tionship between TGF-β1, c-myc, MMP-1, MMP-2, andCOL-I gene expression, without interferring in TIMP-1mRNA synthesis.

ACKNOWLEDGMENTSThis study was supported by a Scientific and TechnicalResearch Fund (FIRST) grant, Milan, Italy.

REFERENCES1. Heby O. Role of polyamines in the control of cell prolif-

eration and differentiation. Differentiation 1981;19:1-20.2. Cohen SS. A Guide to the Polyamines. New York: Oxford

University Press; 1998;185-230.3. Tabib A, Bachrach U. Role of polyamines in mediating

malignant transformation and oncogene expression. IntJ Biochem Cell Biol 1999;31:1289-1295.

4. Tabib A, Bachrach U. Activation of proto-oncogenec-myc and c-fos by c-ras: Involvement of polyamines.Biochem Biophys Res Commun 1994;202:720-727.

5. Blachowski S, Motyl S, Grzelkowska K, Kasterka M,Orzechowski A, Interewicz B. Involvement of polyaminesin epidermal growth factor (EGF), transforming growthfactor (TGF)-α and β1 action on culture of L6 and fetalbovine myoblasts. Int J Biochem 1994;26:891-897.

6. Auvinen M, Paasinen A, Andersson LC, Hölttä E. Ornithinedecarboxylase activity is critical for cell transformation.Nature 1992;360:355-358.

7. Bello-Fernandez C, Packham G, Cleveland JL. Theornithine decarboxylase gene is a transcriptional targetof c-myc. Proc Natl Acad Sci (USA) 1993:90;7804-7808.

8. Frostesjö L, Heby O. Polyamines depletion up-regulatesc-myc expression, yet induces G1 arrest and terminaldifferentiation of F9 teratocarcinoma stem cells. J CellBiochem 1999;76:143-152.

9. Rao JN, Li L, Bass BL, Wang J-Y. Expression of theTGF-β receptor gene and sensitivity to growth inhibitionfollowing polyamine depletion. Am J Physiol Cell Physiol2000;279:C1034-C1044.

10. Santos MF, Viar MJ, McCormack SA, Johnson LR.Polyamines are important for attachment of IEC-6 cellsto extracellular matrix. Am J Physiol 1997;273:G175-G183.

Table 6.

c-myc, TGF-ββ1, COL-I, MMP-1, MMP-2, andTIMP-1 mRNA Levels in Primary GingivalFibroblasts Treated With DFMO (values,obtained after densitometric analysis, arethe mean ±± SD of six duplicateexperiments)

Hours of Culture

24 48 72 96

c-mycCT 1.13 ± 0.15 1.23 ± 0.16 1.29 ± 0.18 1.23 ± 0.17T 1.13 ± 0.17 1.21 ± 0.18 1.03 ± 0.13* 1.61 ± 0.20*

TGF-β1CT 1.24 ± 0.18 1.33 ± 0.20 1.50 ± 0.22 1.08 ± 0.17T 1.20 ± 0.24 1.36 ± 0.23 1.23 ± 0.20 1.97 ± 0.25†

COL-ICT 1.44 ± 0.22 1.47 ± 0.24 1.65 ± 0.23 1.54 ± 0.20T 1.30 ± 0.21 1.48 ± 0.19 1.38 ± 0.18 2.02 ± 0.24*

MMP-1CT 1.35 ± 0.18 1.32 ± 0.20 1.37 ± 0.20 1.40 ± 0.18T 1.39 ± 0.19 1.44 ± 0.24 1.18 ± 0.16 1.90 ± 0.22*

MMP-2CT 1.41 ± 0.28 1.48 ± 0.24 1.65 ± 0.23 1.60 ± 0.24T 1.37 ± 0.26 1.49 ± 0.27 1.35 ± 0.20 2.35 ± 0.32*

TIMP-1CT 1.07 ± 0.16 0.96 ± 0.17 1.21 ± 0.15 1.05 ± 0.14T 1.03 ± 0.15 1.12 ± 0.18 0.95 ± 0.13 1.19 ± 0.16

* P ≤0.05 versus CT.† P ≤0.01 versus CT.

4012.qxd 4/4/05 3:33 PM Page 448

449

J Periodontol • March 2005 Stabellini, Moscheni, Gagliano, et al.

11. Patel AR, Wang JY. Polyamines modulate transcriptionbut not post transcription of c-myc and c-jun in IEC-6cells. Am J Physiol 1997;273:C1020-C1029.

12. Wang JY, McCormack SA, Viar MJ, et al. Decreasedexpression of protoncogenes c-fos, c-myc, and c-jun fol-lowing polyamine depletion in IEC-6 cells. Am J Physiol1993;265:G331-G318.

13. Durante W, Lan Liao, Reyna SV, Peyton KJ, Schafer AI.Transforming growth factor-β1 stimulates L-Argininetransport and metabolism in vascular smooth musclecells. Role in polyamine and collagen synthesis. Circu-lation 2001;103:1121-1127.

14. Stabellini G, Fiocchi O, Evangelisti R, Pagliarini A,Caruso A, Altavilla G. Cyclosporin A, effect on cytoskele-ton and glycosaminolgycans in human gingival fibro-blasts. Immunohistochemical and biochemical evaluation.Eur J Bas Appl Histochem 1991;35:371-381.

15. Stabellini G, Carinci F, Bedani PL, et al. Cyclosporin Aand transforming growth factor beta modify the patternof extracellular glycosaminoglycans without causingcytoskeletal changes in human gingival fibroblasts.Transplantation 2002;73:1676-1679.

16. Richiert DM, Ireland ME. Matrix metalloproteinase secre-tion is stimulated by TGF-beta in cultures lens epithe-lial cells. Curr Eye Res 1999;19:269-275.

17. Coletta RD, Almeida OP, Reynolds MA, Sauk JJ. Alter-ation in expression of MMP-1 and MMP-2 but not TIMP-1and TIMP-2 in hereditary gingival fibromatosis is medi-ated by TGF-beta 1 autocrine stimulation. J Periodon-tal Res 1999;34:457-463.

18. Tipton DA, Dabbous MK. Autocrine transforming growthfactor beta stimulation of extracellular matrix produc-tion by fibroblasts from fibrotic human gingiva. J Perio-dontol 1998;69:609-619.

19. Walters JD, Miller TJ, Cario AC, Beck FM, Marucha PT.Polyamines found in gingival fluid inhibit chemotaxis byhuman polymorphonuclear leukocytes in vitro. J Perio-dontol 1995;66:274-278.

20. Mariggiò MA, Vinella A, Pasquetto N, Curci E, CassanoA, Fumarolo R. In vitro effects of polyamines on poly-morphonuclear cell apoptosis and implication in thepathogenesis of periodontal disease. ImmunopharmacolImmunotoxicol 2004;26:93-101.

21. Gallesio C, Colombatto S, Modica R. Free and acetylatedpolyamines as markers of oral cavity tumors. Oral SurgOral Med Oral Pathol 1994;77:167-171.

22. Ratasirayakorn W, Leone P, Leblebicioglu B, Walters JD.Polyamines found in the inflamed periodontium inhibitpriming and apoptosis in human polymorphonuclearleukocytes. J Periodontol 1999;70:179-184.

23. Patterson MK Jr. Measurement of growth and viability ofcells in culture. In: Jacob WB, Pastan IH, eds. Methodsin Enzymology, vol. 58. New York: Academic Press;1977:141.

24. Wallace HM, Fraser AV. Inhibitors of polyamine metab-olism. Amino Acids 2004;26:353-365.

25. Löser C, Wunderlich U, Fölsch UR. Reversed-phase liq-uid chromatographic separation and simultaneous fluo-rimetric detection of polyamines and their monoacetyl

derivatives in human and animal urine, serum and tissuesamples: An improved, rapid and sensitive method forroutine applications. J Chromatogr 1988;430:249-262.

26. Bradford MM. A rapid and sensitive method for the quan-titation of microgram quantitaties of proteins utilizingthe principle of protein-dye binding. Anal Biochem 1976;72:248-254.

27. Stefanelli C, Rossoni C, Ferrari F, Flamigni F, CaldareraCM. Ornithine decarboxylase and ornithine decarboxy-lase-inhibiting activity in rat thymocytes. Cell BiochemFunct 1992;10:243-250.

28. Liu L, Santora R, Rao JN, et al. Activation of TGF-beta-Smad signaling pathways following polyamine depletionin intestinal epithelial cells. Am J Physiol Gastrointest LiverPhysiol 2003;285:G1056-G1067.

29. Manni A, Wechter R, Gilmour S, Verderame MF, MaugerD, Demers LM. Ornithine decarboxylase over-expressionstimulates mitogen-activated protein kinase and anchor-age-independent growth of human breast epithelial cells.Int J Cancer 1997;70:175-180.

30. Leroy D, Heriche JK, Filhol O, Chambaz EM, Cochet C.Binding of polyamines to an autonomous domain of theregulatory subunit of protein kinase CK2 induces a con-formational changes in the holoenzyme. A proposed rolefor the kinase stimulation. J Biol Chem 1997;272:20820-20827.

31. Dowdy S, Mariani A, Janknecht R. HER2/Neu-and TAK1-mediated up-regulation of the transforming growth factorβ inhibition Smad7 via the ETS protein ER81. J BiolChem 2003;278:44377-44384.

32. Nicolás FJ, Hill CS. Attenuation of the TGF-β-Smad sig-naling pathway in pancreatic tumor cells confers resis-tance to TGF-β-induced growth arrest. Oncogene 2003;22:3698-3711.

Correspondence: Dr. Giordano Stabellini, Dipartimento diAnatomia Umana, Via Mangiagalli 31, 20133 Milan, Italy.Fax: 39-0258355387; e-mail giordano.stabellini@unimi.it.

Accepted for publication July 7, 2004.

4012.qxd 4/4/05 3:33 PM Page 449