Effects of alkaline phosphatase and its inhibitor levamisole on the modulation of androgen metabolism by nicotine and minocycline in human gingival and oral periosteal fibroblasts

Archives of Oral Biology (2003) 48, 69Ð76

Effects of alkaline phosphatase and its inhibitor levamisole on the modulation of androgen metabolism by nicotine and minocycline in human gingival and oral periosteal ®broblasts M. Soory*, A. Suchak Division of Periodontology, Guy's King's and St. Thomas' Dental Institute, King's Dental Hospital, Caldecot Road, London SE5 9RW, UK Accepted 8 August 2002

KEYWORDS Steroid hormones; Nicotine; Minocycline; Healing

Summary The aim of this investigation was to establish the implications of nicotine, minocycline, alkaline phosphatase (AP) and its inhibitor levamisole (L) on tissue turnover in human gingival and periosteal ®broblasts (HGF, HPF) using [14 C]-testosterone as substrate. Monolayer cultures of HGF and HPF established from four patients were incubated in duplicate with serial and optimal concentrations of nicotine and minocycline, alone and in combination, for 24 h in Eagle's MEM, with the substrate [14 C]-testosterone. Further experiments were carried out on HPF only, to investigate the effects of alkaline phosphatase (AP) and its inhibitor levamisole (L) on the metabolism of [14 C]-testosterone, followed by the effects of L on the modulatory actions of nicotine. The cell-conditioned medium was then solvent-extracted, analysed and quanti®ed for steroid metabolites using a radioisotope scanner. At low concentrations, nicotine stimulated the synthesis of the physiologically active androgen 5a-dihydrotestosterone (DHT) from [14 C]-testosterone, with inhibition at higher concentrations (n ˆ 4; P < 0:01). Minocycline stimulated the synthesis of DHT, with decreased yields in the presence of nicotine (n ˆ 4; P < 0:01), but greater than with nicotine alone. Alkaline phosphatase signi®cantly enhanced the synthesis of androgen metabolites by HPF (n ˆ 4; P < 0:01), with inhibition in response to L alone and in combination with AP, to less than control values (n ˆ 4; P < 0:01). L also caused further inhibition in the yields of androgen metabolites when incubated with nicotine, implying that some of the inhibitory effects of nicotine could be due to inhibition of AP activity. Conclusion: This investigation has shown that nicotine can inhibit the formation of matrix-stimulatory steroid metabolites in ®broblasts, partly due to inhibition of AP activity. Minocycline is a useful adjunct, in reducing the inhibition of androgen metabolism caused by nicotine. ß 2003 Elsevier Science Ltd. All rights reserved.

Introduction Abbreviations: MEM, minimal essential medium; TLC, thinlayer chromatography  Corresponding author. Fax: ‡44-20-7346-3185. E-mail address: [email protected] (M. Soory).

This is an investigation of the effects of nicotine and minocycline on the formation of anabolic androgen metabolites by ®broblasts and modulation of their effects by alkaline phosphatase and its inhibitor

0003±9969/03/$ Ð see front matter ß 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0003-9969(02)00157-7

70 levamisole. There is a strong positive correlation between tobacco use and increased incidence and severity of periodontal disease.1 Wound healing may be in¯uenced by tobacco products and nicotine is one of the main components isolated from tobacco. In view of the anabolic effects of circulating androgens on healing and repair2 it is relevant to study the modulatory effects of nicotine on androgen metabolism in cultured gingival and oral periosteal ®broblasts. While smaller concentrations of nicotine to which light smokers are exposed can enhance ®broblast attachment and growth, higher levels of nicotine can be damaging. Fibroblast adherence and growth may be differentially affected by the presence of nicotine, dependent on the concentrations and duration of exposure.3 Other studies,4 have shown that 0.075% nicotine inhibits proliferation and ®bronectin and type 1 collagen production in human gingival ®broblasts. These contradictory results might be a re¯ection of the nicotine concentrations used in these studies and of heterogeneity among sub-populations of ®broblasts. This indicates a need for further clari®cation of the effects of a range of concentrations of nicotine on gingival and oral periosteal ®broblasts. In the investigation, gingival and periosteal ®broblasts derived from chronically in¯amed sites were used to demonstrate the effects of nicotine, minocycline, alkaline phosphatase and its inhibitor on androgen metabolism as an index of healing responses. In the context of the periodontal patient in whom nicotine can affect healing responses, adjunctive antimicrobials have been used to complement mechanical therapy in aggressive or recalcitrant cases.5,6 It is pertinent to investigate the effects of the tetracycline antibiotic minocycline on the modulation of androgen metabolism by nicotine in ®broblasts, since the tetracyclines stimulate matrix synthesis in connective tissue and bone7 and also inhibit matrix metallo proteinases,8 in addition to their anti-microbial actions. The stimulatory effects of tetracyclines on tissue matrices may be attributed to up-regulation of alkaline phosphatase activity, thus enhancing the expression of the mature osteoblast phenotype.9 Sub-populations of ®broblasts engaged in in¯ammatory repair express increased alkaline phosphatase activity.10 Induction of alkaline phosphatase activity has been demonstrated in periosteal ®broblasts, in response to demineralised bone matrix.10 The speci®c alkaline phosphatase inhibitor levamisole has been shown to reduce the expression of collagen, osteocalcin and bone alkaline phosphatase in a rat osteoblastic cellculture model. In this context it is relevant to study the effect of alkaline phosphatase and its inhibitor

M. Soory, A. Suchak levamisole on androgen metabolism in order to identify a physiological link between androgen metabolites and matrix turnover. The physiologically active androgen 5a-dihydrotestosterone is formed from testosterone and is implicated in stimulating matrix-forming cells. The rate of periosteal bone formation has been shown to be signi®cantly elevated in response to DHT alone and in combination with oestrogen.2 The expression of androgen receptors has been detected in a high proportion of periodontal and gingival tissue and in ®broblasts derived from the same source.11 This ®nding suggests that matrix turnover in gingiva is characteristic of androgen target tissue and that steroid hormone metabolic pro®les could serve as indices of healing responses in gingiva. In view of the association between smoking, the incidence and severity of periodontal disease and response to treatment, the aim of this investigation was to establish steroid hormone pro®les in the presence or absence of nicotine and minocycline in gingival and periosteal ®broblasts in culture, using the radiolabelled androgen substrate [14 C]testosterone. The effects of alkaline phosphatase and its inhibitor levamisole on testosterone metabolism were also investigated.

Materials and methods Authentic steroids were obtained from Sigma Chemical Co. (Poole, Dorset, UK). The radioisotope [14 C]-testosterone (speci®c radioactivity 2.15 GBq/ mmol) was obtained from Amersham International, Amersham, UK. Solvents for thin-layer chromatography (TLC; benzene, acetone), ethyl acetate for extraction of metabolites, chloroform to redissolve the dried extract and TLC plates (precoated silica gel Kieselgel 60) were obtained from Merck Ltd., Dagenham, Essex, UK. The incubation medium was Eagle's MEM with L-glutamine, antibiotic/antimycotic solution and sodium bicarbonate which were all provided by Gibco Ltd., Paisley, Scotland. Minocycline used in the incubations was obtained from the pharmacy at GKT School of Medicine & Dentistry (King's Campus). For the derivatisation of steroids, O-(penta¯uorobenzyl)-hydroxylamine-HCl (Florox reagent) and N , O - bis ( trimethylsilyl ) tri¯uoroacetamide were obtained from Pierce and Warriner (UK Ltd., Chester). Nicotine, alkaline phosphatase and levamisole were purchased from Sigma Chemical Company.

Cell-cultures Chronically in¯amed gingival tissues were obtained from four periodontal patients aged 30Ð50 years,

Effects of alkaline phosphatase undergoing periodontal surgical procedures at King's Dental Hospital, London, UK. Patient consent and approval from the local Ethics Committee were obtained. They had all completed the initial phase of periodontal treatment consisting of scaling and root planing. There was no bleeding on probing, prior to the surgical procedures for the isolation of gingival tissues from periodontal pockets of 6Ð 8 mm depth. The experiments were conducted with four individual gingival ®broblast cell-lines derived from these patients; they were performed in duplicate, each with its own controls, in the absence of testing agents. The data are presented for a sample number of four, since there were no signi®cant differences between males and females, for control incubations. Similarly, small samples of periosteal tissue were isolated from four other patients, during periosteal fenestration of vertically advanced ¯aps, for regenerative therapy. The tissue was obtained from the surface of the bone, at the base of split-thickness mucoperiosteal ¯aps.12 All patients used in this study for the donation of tissue samples were non-smokers. Metabolism of testosterone in in¯amed gingivae or ®broblasts from an in¯amed source is equally prevalent amongst males and females, whereas in tissues from a healthy source, testosterone is metabolised predominantly in males.13Ð16 This study aimed to investigate the responses of cells from chronically in¯amed tissues, to the hormones and agents studied, in an attempt to simulate an in vivo environment. The gingival and periosteal tissues isolated were minced into 1 mm3 fragments and established in primary culture in 25 cm2 tissue culture ¯asks. This has been described previously with photographs of the ®broblast-like cell type isolated.17 Serial passaging of the primary cultures was carried out by partial digestion with 0.25% trypsin solution. Fibroblasts of the fourth to ninth passage in monolayer culture were used for the experiments. Each test incubation was compared with individual controls from the same source, subjected to the same conditions and in the absence of testing agents; this eliminates any possible effect of media components on the results. The cell-lines were used individually in duplicate, with no pooling of samples. The data are presented as means for the four cell-lines used.

Experimental design Cells (2:2  106 ) from a fully con¯uent 25 cm2 ¯ask were divided among 24 wells of a multiwell dish in Eagle's MEM for each of the four lines of human gingival and periosteal ®broblasts investigated. The

71 cells were allowed to become fully con¯uent, to avoid mitotic effects. Duplicate incubations were performed with human gingival/periosteal ®broblasts, using [14 C]-testosterone (0.925 kBq/ml) as substrate and testing agents as described below, in Eagle's MEM for 24 h. Each cell-line had its own duplicate controls incubated with the substrate [14 C]-testosterone, in the absence of testing agents. Incubations were performed for 24 h, allowing adequate time for equilibration of testing agents15,16 and optimising yields of androgen metabolites. The trypan blue exclusion test con®rmed cell viability at all concentrations of nicotine studied. The metabolism of [14 C]-testosterone as substrate by gingival and periosteal ®broblasts, in response to serial concentrations of nicotine and an optimal concentration of minocycline, alone and in combination. Duplicate incubations of the four human gingival and periosteal cell-lines were performed for 24 h with [14 C]-testosterone, nicotine at serial concentrations of 1, 5, 30, 50 and 100 mg/ml or minocycline at 20 mg/ml, as established previously,18 and combined incubations of nicotine (at the above concentrations) with minocycline. The metabolism of [14 C]-testosterone by human periosteal ®broblasts in response to alkaline phosphatase and its speci®c inhibitor levamisole, alone and in combination; effects of levamisole on the modulatory effects of nicotine. Four human periosteal ®broblast cell-lines were incubated with the substrate [14 C]-testosterone as described above, serial concentrations of alkaline phosphatase and an optimal inhibitory concentration of levamisole (30 mg/ml) established previously.16 Similar experiments were performed with nicotine and levamisole, alone and in combination. At the end of a 24 h incubation period, in a humidi®ed incubator at 37 8C, the medium was removed from the wells. Solvent extraction of the medium was performed with ethyl acetate after the addition of unlabelled carrier steroid metabolites. Procedural losses during solvent extraction were determined in a separate set of experiments, using known amounts of radiolabelled androgen which were solvent-extracted and subsequently quanti®ed.15 The mean of sixteen repeated procedures was used to make a correction for losses during this investigation.

Detection and characterisation of androgen metabolites The extracts were evaporated to dryness in a vortex evaporator (Gyrovap, VA Howe Ltd., Banbury,

72

M. Soory, A. Suchak

Oxon, UK), redissolved in 100 ml chloroform and spotted on TLC plates. TLC was performed in a benzene acetone solvent system (4:1 (v/v)) for the separation of metabolites isolated, and quanti®ed using a radioisotope scanner. The metabolites were tentatively identi®ed using the mobility of unlabelled standards added to the samples and disclosing them in iodine. Further con®rmation of the identity of steroid metabolites was established by carrying out gas chromatography/mass spectrometry (GCÐMS).

Statistical analysis

Characterisation of 5a-dihydrotestosterone by GC±MS

Results

As 5a-dihydrotestosterone is the most signi®cant biologically active testosterone metabolite in stimulating ®broblast synthetic activity, it was important to con®rm its identity. Under the conditions described above, several incubations were performed with unlabelled testosterone (10 6 M/l). After extraction, the identity of 5a-dihydrotestosterone as a metabolite in the dried extracts was con®rmed by gas chromatographyÐmass spectrometry (courtesy of Professor A.I. Mallet, St. Thomas' Hospital, London, UK), after derivatisation to penta¯uorobenzyloxime trimethyl silyl ether (PFBO/TMS). The derivatised biological material had a molecular ion (557) and mass spectral fragmentation pattern identical to those of the authentic PFBO/TMS ether of 5a-dihydrotestosterone, but at lower levels, due to smaller concentrations of the steroid. These procedures have been described in detail, showing the ionic fragmentation pattern in graphic form.19

Mean values were obtained for each of the metabolites isolated from incubations of four gingival and periosteal ®broblast cell-lines in duplicate. Since individual controls were set up for each cell-line, the data are representative of the number of cell-lines studied. Mean values and standard deviations from the mean are shown in the Figures. One-way ANOVA was used for signi®cance testing.

The metabolism of [14 C]-testosterone by human gingival ®broblasts in response to serial concentrations of nicotine (N) and an optimal concentration of minocycline (M), alone and in combination (Fig. 1). Human gingival ®broblasts metabolised [14 C]-testosterone mainly to the diols, 5a-dihydrotestosterone and 4-androstenedione. Low concentrations of nicotine at 1 and 5 mg/ml stimulated the formation of DHT by 12%, with 15Ð25% inhibition at N50,100 (Fig. 1; n ˆ 4; P < 0:01). There was 61% stimulation of 5a-dihydrotestosterone synthesis in response to minocycline at 20 mg/ml, maintaining fairly high yields with 35% stimulation at M ‡ N1,5, decreasing to 18% stimulation at M ‡ N30 (n ˆ 4; P < 0:01), with values similar to those of controls at M ‡ N50,100. The yields of 4-androstenedione decreased by 18Ð32% at N30,50,100 (n ˆ 4; P < 0:01), with a marginal increase in response to minocycline and values similar to those of N alone, for the combinations. There were signi®cant reductions in the formation of

Figure 1 The metabolism of [14 C]-testosterone as substrate by human gingival ®broblasts, in response to serial concentrations of nicotine (N) and the optimal concentration of minocycline (M), alone and in combination. Duplicate incubations of nicotine (N) at serial concentrations of 1, 5, 30, 50 and 100 mg/ml, minocycline (M) at 20 mg/ml and their combinations were performed with four human gingival ®broblast cell-lines for 24 h in MEM and [14 C]-testosterone. The medium was solvent-extracted for metabolites, separated by TLC and quanti®ed using a radioisotope scanner. C: Control, in the absence of testing agent. Mean values and standard deviations are shown; n ˆ 4. This applies to all ®gures.

Effects of alkaline phosphatase

73

Figure 2 The metabolism of [14 C]-testosterone as substrate by human periosteal ®broblasts, in response to serial concentrations of nicotine (N) and an optimal concentration of minocycline (M), alone and in combination. Duplicate incubations of nicotine (N) at serial concentrations of 1, 5, 30, 50 and 100 mg/ml, minocycline (M) at 20 mg/ml (as established previously) and combined incubations of nicotine (N) (at the above concentrations) with minocycline (M) (M ‡ N) were performed with four human periosteal ®broblast cell-lines for 24 h in MEM and [14 C]-testosterone. The metabolites were isolated and quanti®ed as described in Fig. 1.

diols, from 26 to 30% at N30,50,100 and a 47% increase in response to M (n ˆ 4; P < 0:01). The combinations of M ‡ N1,5 showed 16% increases in the diols, while combinations with higher nicotine concentrations showed values similar to those of N alone. The metabolism of [14 C]-testosterone by human periosteal ®broblasts, in response to serial concentrations of nicotine (N) and the optimal concentration of minocycline (M), alone and in combination (Fig. 2). The formation of 5a-dihydrotestosterone was increased by 63% and 57% in response to N1 and N5 (Fig. 2; n ˆ 4; P < 0:01), reaching control values at N30,50 and a little less than controls at N100. The formation of 4-androstenedione was decreased by 23Ð38% at N30,50,100 mg/ml (n ˆ 4; P < 0:01). There was 93% stimulation of 5a-dihydrotestosterone synthesis in response to minocycline at M20, which remained fairly high, with 47% stimulation at M ‡ N1 ,

27% stimulation with other combinations, decreasing to 17% stimulation at M ‡ N100 (n ˆ 4; P < 0:01). The yields of 4-androstenedione remained reduced by 23% over controls, in response to minocycline and in response to the combinations with nicotine. The formation of the diols was reduced from 35 pmol/ml at baseline to 14.7 pmol/ml in response to N100, resulting in more than a two-fold reduction, with similar values for the combinations M ‡ N30,50,100 (n ˆ 4; P < 0:01). The metabolism of [14 C]-testosterone by human periosteal ®broblasts in response to alkaline phosphatase and its speci®c inhibitor levamisole, alone and in combination (Fig. 3). Alkaline phosphatase stimulated 5a-dihydrotestosterone synthesis by 30Ð40% at concentrations of 1Ð5 mg/ml (n ˆ 4; P < 0:01). Levamisole reduced 5a-dihydrotestosterone yields by 47% ( n ˆ 4; P < 0:01). The combinations of alkaline phosphatase

Figure 3 The metabolism of [14 C]-testosterone by human periosteal ®broblasts, in response to alkaline phosphatase (AP) and its speci®c inhibitor levamisole (L) alone and in combination. Serial concentrations of AP (1Ð5 mg/ml), and an optimal concentration of levamisole (L, 30 mg/ml), were incubated alone and in combination with four cell-lines of human periosteal ®broblasts in Eagle's MEM, using 14 C-testosterone as substrate. After 24 h, the incubates were analysed for steroid metabolites as described in Fig. 1.

74

M. Soory, A. Suchak

Figure 4 Effects of levamisole (L) on the modulation of androgen metabolism by nicotine (N). Serial concentrations of nicotine (N), an optimal concentration of levamisole (L), and their combinations were incubated with four cell-lines of human periosteal ®broblasts in Eagle's MEM and 14 C-testosterone as substrate for 24 h. The incubate was subsequently solvent-extracted for metabolites and quanti®ed as described previously.

with levamisole showed 28 and 22% inhibition at the range of concentrations of alkaline phosphatase studied (n ˆ 4; P < 0:01). The yields of 4-androstenedione showed similar trends in response to alkaline phosphatase with a 28% increase over controls in response to 4 mg/ml and 30% inhibition in response to levamisole (Fig. 3; n ˆ 4; P < 0:01). The combinations alkaline phosphatase ‡ levamisole caused 10Ð 20% inhibition over controls (n ˆ 4; P < 0:01), being less pronounced than values for levamisole alone. Effects of levamisole on the modulation of androgen metabolism by nicotine (N) in human periosteal ®broblasts (Fig. 4). At 1 mg/ml, nicotine enhanced 5a-dihydrotestosterone synthesis by 44% decreasing to control values at 30Ð100 mg/ml. Levamisole reduced the yields of 5a-dihydrotestosterone by 30% over controls. The combination of N1 ‡ levamisole caused a 30% reduction in 5a-dihydrotestosterone over N1 alone, with further 10% reductions in combination with N30,50,100 (n ˆ 4; P < 0:01). The formation of 4-androstenedione was reduced by 45Ð53% at N30Ð100 (n ˆ 4; P < 0:01), over controls and by 50% in response to levamisole, with a further 40% reduction in combination with N50,100. The diols were signi®cantly reduced by 35% in response to N30Ð100 and 56% by levamisole, compared to control values (n ˆ 4; P < 0:01). When compared with levamisole alone, there was further inhibition of 40% in the yields of diols in the combinations of N50 ‡ levamisole and N100 ‡ levamisole (n ˆ 4; P < 0:01).

Discussion The results of this investigation show that nicotine at low concentrations enhanced the expression of

5a-reductase activity, resulting in the formation of 5a-dihydrotestosterone from testosterone. There was a gradual decline in the yields of 5a-dihydrotestosterone to control values and some inhibition at higher concentrations of 30, 50 and 100 mg/ml, indicating decreased anabolic activity in ®broblasts. There was inhibition of 4-androstenedione synthesis at nicotine concentrations of 30Ð100 mg/ml, due to inhibition of. 17b-hydroxysteroid dehydrogenase activity, responsible for the conversion of testosterone to 4-androstenedione. Although, the anabolic potential of ®broblasts may be enhanced in response to low levels of nicotine via 5a-reductase expression, the formation of 4-androstenedione was reduced considerably. This metabolite can contribute to androgenic actions in tissues and also result in inter-conversion to testosterone and subsequently to 5a-dihydrotestosterone. Decreased yields of 4androstenedione due to inhibition of 17b-hydroxysteroid dehydrogenase activity, can result in reduced repair potential within the periodontium in response to nicotine, similar to the effects of impaired 5adihydrotestosterone synthesis at higher concentrations of nicotine. Although common trends were observed, there was some variation amongst periosteal ®broblasts, in the yields of 5a-dihydrotestosterone in response to low concentrations of nicotine and of 4-androstenedione in response to higher concentrations of nicotine. There was some heterogeneity between gingival and periosteal ®broblasts, in the metabolic yields of diols and their response to minocycline. There were also some differences in the effects of minocycline on 5a-dihydrotestosterone yields in the two cell types. The formation of the diols was signi®cantly inhibited by nicotine at higher concentrations, in both human gingival and periosteal ®broblasts. As

Effects of alkaline phosphatase the diols are useful markers of androgenicity in peripheral tissues,20 they too can stimulate connective tissue and bone matrix synthesis and signi®cant reduction in their yields in response to nicotine could affect matrix turnover by gingival and periosteal ®broblasts. All concentrations of nicotine used were compatible with the levels isolated in cultured gingival ®broblasts exposed to nicotine.21 Other workers have shown that low concentrations of nicotine can enhance ®broblast adhesion, proliferation and matrix synthesis, while higher concentrations appear to be detrimental.3 This may be re¯ected in the dose dependent androgen response of human gingival and oral periosteal ®broblasts to nicotine, in this investigation. The ®ndings of our investigation demonstrate the dichotomous effects of nicotine. This may be due to interaction between ligand dependent and independent pathways, for stimulation of androgen metabolising enzymes and the release of inhibitory substances at higher concentrations of nicotine, which interfere with androgen metabolism. It has been suggested by other workers that smoking has a direct effect on osteoblasts and is therefore implicated as a risk factor for bone resorption, attributed to the inhibition of bone metabolism by nicotine.22 The results of our investigation with regard to the inhibition of 5a-reductase activity by nicotine, resulting in reduced yields of 5a-dihydrotestosterone in periosteal ®broblasts, could be another mechanism for bone loss in smokers. Our previous investigation using oral periosteal ®broblasts demonstrated that periosteum is a target tissue for androgen metabolism, with speci®c inhibition of 5a-reductase activity in response to the 5a-reductase type 2 enzyme inhibitor ®nasteride.12 This diminished the yields of the anabolic androgen DHT, speci®cally associated with the 5a-reductase type 2 enzyme activity. Minocycline enhanced the synthesis of 5a-dihydrotestosterone signi®cantly, which can contribute to anabolic effects in the healing periodontium, during adjunctive therapy. In the current investigation, the extent of inhibition of androgen metabolism caused by nicotine was not as pronounced when it was combined with minocycline. This suggests that in refractory cases of periodontal disease associated with smoking, adjunctive treatment with minocycline may have a positive effect on periodontal healing. Signi®cantly high yields of DHT in response to alkaline phosphatase, implies its involvement in the formation of androgen metabolites. Inhibition of 5a-dihydrotestosterone synthesis by the speci®c alkaline phosphatase inhibitor levamisole and moderation in combination with alkaline phosphatase,

75 con®rms an alkaline phosphatase mediated mechanism for androgen metabolism in ®broblasts with implecations on matrix synthesis. The production of physiologically active androgen metabolites by oral periosteal ®broblasts could contribute to matrix formation in these cells, in response to minocycline. The inhibitory effects of nicotine discussed above, may be partly due to inhibition of alkaline phosphatase activity. Levamisole signi®cantly reduced the yields of androgen metabolites, while higher concentrations of nicotine in combination with levamisole reduced these yields even further. Modulation of alkaline phosphatase activity is a possible explanation for some of the inhibitory actions of nicotine on ®broblasts, which could affect tissue turnover. This investigation has demonstrated the effects of nicotine, minocycline, alkaline phosphatase and its inhibitor levamisole on androgen metabolism in gingival and periosteal ®broblasts. The androgen metabolic pro®le in response to these agents alone and in combination is suggestive of a possible mechanism for wound healing in smokers, in the context of adjunctive treatment with minocycline. The concentrations of nicotine used in this study are similar to the levels detected in cultured gingival ®broblasts after nicotine exposure.21 The lower concentrations used enhanced some metabolic activity in human gingival and periosteal ®broblasts, while higher concentrations inhibited the formation of metabolites implicated in stimulating matrix synthesis. The inhibition caused by nicotine was not as great in the presence of minocycline, which suggests that this antibiotic may be bene®cial in the adjunctive treatment of patients with refractory periodontal disease. The increased incidence of periodontal disease and decreased healing responses in smokers may partly explained by the ®ndings of this investigation.

Acknowledgements We wish to thank Dr. I. Tavares for usage of the radioisotope scanner at the Rayne Institute, Guy's King's and St. Thomas' School of Medicine, King's campus, London, UK.

References 1. Ismail AI, Morrison EC, Burt BA, Caffese RG, Kavanagh MT. Natural history of periodontal disease in adults: ®ndings from the Tecumseh periodontal disease study, 1959Ð1989. J Dent Res 1990;69:430Ð435. 2. Coxam V, Bowman BM, Meecham M, Roth CM, Miller MA, Miller SC. Effects of dihydrotestosterone alone and in combination with oestrogen on bone mineral density, bone

76

3.

4. 5. 6.

7. 8.

9.

10.

11.

12.

M. Soory, A. Suchak growth and formation rates in ovariectomised rats. Bone 1996;19:107Ð114. Peacock ME, Sutherland DE, Schuster GS, Brennan WA, O'Neal RB, Strong SL, Van Dyke TE. The effect of nicotine on reproduction and attachment of human gingival ®broblasts in vitro. J Periodontol 1993;64:658Ð665. Tipton DA, Dabbous MK. Effects of nicotine on proliferation and extracellular matrix production of human gingival ®broblasts in vitro. J Periodontol 1995;66:1056Ð1064. Genco RJ. Using antimicrobial agents to manage periodontal diseases. J Am Dent Assoc 1991;122:30Ð38. Litch JM, Encarnaciou M, Chen S, Leonard J, Burkoth TL. Use of the polymeric matrix as internal standard for quanti®cation of in vivo delivery of tetracycline HCl from Actisite tetracycline ®bre during periodontal treatment. J Periodont Res 1996;31:540Ð544. Schneir M, Ramamurthy NS, Golub L. Minocycline treatment of diabetic rats normalises skin collagen production and mass. Matrix Coll Rel Res 1990;10:112Ð123. Ramamurthy NS, Rifkin BR, Greenwald RA, Xu J-w, Liu Y, Turner G, Golub LM, Vernillo AT. Inhibition of matrix metalloproteinase-mediated periodontal bone loss in rats: a comparison of 6 chemically modi®ed tetracyclines. J Periodontol 2002;73:726Ð734. Sasaki T, Kaneko H, Ramamurthy NS, Golub LM. Tetracycline administration restores osteoblast structure and function during experimental diabetes. Anat Records 1991;231: 25Ð34. Groeneveld MC, Everts V, Beertsen W. Formation of a®brillar acellular cementum-like layers induced by alkaline phosphatase activity from periodontal ligament explants maintained in vitro. J Dent Res 1994;73:1588Ð1592. Parkar MH, Newman HN, Olsen I. Polymerase chain reaction analysis of oestrogen and androgen receptor expression in human gingival and periodontal tissue. Arch Oral Biol 1996; 41:979Ð983. Soory M, Tilakaratne A. The effect of minocycline on the metabolism of androgens by human oral periosteal ®bro-

13. 14. 15. 16.

17. 18. 19.

20.

21. 22.

blasts and its inhibition by ®nasteride. Arch Oral Biol 2000; 45:257Ð265. Rappaport SC, Vittek J, Altman K, Gordon GG, Southren AL. Sex differences in the metabolism of androgens by human gingiva. J Dent Res 1976;55:1371. Sooriyamoorthy M, Gower DB. Phenytoin stimulation of 5areductase activity in in¯amed gingival ®broblasts. Med Sci Res 1989;17:989Ð990. Soory M, Gower DB. The in¯uence of prostaglandins on steroid conversions by human gingival ®broblasts. J Periodont Res 1998;33:439Ð447. Tilakaratne A, Soory M. Effects of the anti-androgen ®nasteride on 5a-reduction of androgens in the presence of progesterone in human gingival ®broblasts modulatory actions of the alkaline phosphatase inhibitor levamisole. J Periodont Res 2000;35:179Ð185. Kasasa SC, Soory M. The effect of PDGF, TGF-b and IGF in combination on androgen metabolism by ®broblasts. J Clin Periodontol 1998;25:640Ð646. Soory M, Virdi H. Effects of the anti-androgen ®nasteride on 5a-reductase activity in human gingival ®broblasts in response to minocycline. J Clin Periodontol 1998;25:67Ð73. Soory M. Bacterial steroidogenesis by periodontal pathogens and the effect of bacterial enzymes on steroid conversions by human gingival ®broblasts. J Periodont Res 1995;30: 124Ð131. Minut GJ, Cambie M, Lanzoni M, Marubini E, Secreto G. Urinary 5alpha-androstanediol and 5beta-androstanediol measurement by gas chromatography after solid phase extraction and high performance liquid chromatography. Int J Biol Markers 1999;14:154Ð159. Hanes PJ, Schuster GE, Lubas S. Binding, uptake and release of nicotine by human gingival ®broblasts. J Periodontol 1991;62:147Ð152. Williams AR, Weiss NS, Ure CL, Ballard J, Daling JR. Effect of weight, smoking and oestrogen use on the risk of hip and forearm in post menopausal women. Obstet Gynecol 1982; 60:695Ð699.