J Clin Periodontol 2009; 36: 634641 doi: 10.1111/j.1600-051X.2009.01440.

Markers of bone destruction and formation and periodontitis in type 1 diabetes mellitus
Lappin DF, Eapen B, Robertson D, Young J, Hodge PJ. Markers of bone destruction and formation and periodontitis in type 1 diabetes mellitus. J Clin Periodontol 2009; 36: 634641. doi: 10.1111/j.1600-051X.2009.01440.x. Abstract Aim: To determine plasma concentrations of bone metabolism markers in type 1 diabetes mellitus patients and non-diabetic and to evaluate the inuence of periodontitis on biomarkers of bone formation in these patient groups. Methods: Plasma concentrations of receptor activator of nuclear factor-kB ligand (RANKL), osteoprotegerin (OPG), C-terminal telopeptide of type 1 collagen and osteocalcin were measured in type 1 diabetes mellitus patients (n 5 63) and nondiabetics (n 5 38) who were also subdivided on the basis of their periodontal status. Results: Diabetics had signicantly lower osteocalcin concentrations, lower RANKL to OPG ratios and higher OPG concentrations (as shown by other researchers) than non-diabetics. The ratio of RANKL to OPG was altered by the periodontal status. Osteocalcin had a negative correlation and OPG a positive correlation with the percentage of glycated haemoglobin in the blood. Conclusion: Because, osteocalcin, a biomarker of bone formation, is lower in patients with periodontitis and in patients with type 1 diabetes mellitus with and without periodontitis than in non-diabetics without periodontitis, this might indicate that diabetics are less able to replace bone lost during active bursts of periodontitis and explain the greater severity of disease seen in studies of patients with diabetes.

David F. Lappin, Bob Eapen, Douglas Robertson, Jenny Young and Penny J. Hodge
Infection and Immunity Group, Faculty of Medicine, Dental Hospital and School, University of Glasgow, Glasgow, UK

Key words: bone-markers; diabetes; periodontitis Accepted for publication 10 May 2009.

Periodontitis is recognized as a complication of diabetes. It is a chronic inammatory disease of the supporting tissues of the teeth, generally considered to be caused by infection with Gram-negative bacteria and characterized by gingival inammation and bone resorption (Page 1991). Bone resorption and bone formaConict of interests and source of funding statement

The authors declare that they have no conict of interests. The clinical study was funded by the Chief Scientist Ofce of the Scottish Government and the Medical Faculty University of Glasgow. Bob Eapen was seconded to the University of Glasgow Dental School from the University of Abertay Dundee, Dundee, Scotland and funds were provided by University of Abertay, Dundee to cover bench fees.

tion are processes that are coupled but can take place independently (Corral et al. 1998). The function of bone resorbing osteoclasts is regulated by interaction with broblasts in the ligament between the tooth and the bone (Kanzaki et al. 2002). These periodontal ligament broblasts are involved in both stimulatory and inhibitory processes. Receptor activator of nuclear factor-kB ligand (RANKL) interacts with receptors on the surface of osteoclasts to stimulate bone resorption. Osteoprotegerin (OPG) also produced by periodontal ligament broblasts (Kanzaki et al. 2002) blocks the activity of RANKL by binding to RANK. High levels of RANKL are expressed during root resorption of deciduous teeth and raised levels of OPG are expressed where no root resorption is normally

taking place, i.e. adjacent to permanent teeth (Fukushima et al. 2003). The expression of OPG in periodontal tissues is decreased in both moderate periodontitis and advanced periodontitis, while RANKL mRNA is increased in periodontitis at the advanced stage (Liu et al. 2002). Higher circulating levels of OPG are found in type 1 diabetics than in healthy individuals (Kim et al. 2005). Plasma OPG concentrations have been shown to be implicated in the regulation of vascular calcication, atherogenesis, coronary artery disease (Dhore et al. 2001, Jono et al. 2002) and associated with progression of coronary artery calcication in type 2 diabetes mellitus (Anand et al. 2007). An increased ratio of RANKL to OPG mRNA in periodontitis may determine local bone resorption. C-terminal

634

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

Bone markers and periodontitis in diabetes

telopeptide of type 1 collagen (ICTP) is the carboxyterminal telopeptide of type 1 collagen, the only collagen type found in mineralized bone (Risteli & Risteli 1993). ICTP is released into the circulation during bone resorption (Abildgaard et al. 1997). Increased amounts of this product in serum or plasma indicate an increase in osteoclast activity. Osteocalcin is a major non-collagenous protein of the bone matrix. Osteocalcin is produced by osteoblasts. Serum osteocalcin levels have been used as biomarkers for bone formation because this protein is released into the circulation during this process (Bullon et al. 2005). A number of studies have measured osteocalcin levels in relation to periodontitis and in diabetic patients (Alexopoulou et al. 2006) but there are no studies examining diabetic patients with periodontitis. Higher serum levels of osteocalcin were measured in rats, with experimentally induced periodontitis, treated with doxycycline and alendronate indicating higher bone formation (Buduneli et al. 2005). Bone loss seen in periodontitis is associated with a perturbation in the balance of RANKL and OPG and may be reected in alterations in osteocalcin and/or ICTP levels. Our hypothesis is that higher levels of OPG reported for type 1 diabetes mellitus patients modify the level of destruction seen in those patients also suffering from periodontitis. The aim of this study was to measure the circulating levels of the markers of bone metabolism in diabetics with and without periodontitis. We also aimed to determine whether the presence of periodontitis had an inuence on the plasma concentrations of markers of bone metabolism in diabetics and systemically healthy individuals.
Methods
Ethics and informed consent

635

and female Caucasians, who were generally healthy and had at least 20 teeth.
Exclusion criteria

Smoking within the past 5 years, pregnancy, immunosuppression, medication which predisposes to gingival hyperplasia, hormone replacement therapy; patients prescribed systemic antibiotics within the previous 6 weeks of the oral examination and patients who required prophylactic antibiotic cover before a periodontal examination.
Clinical measurements

Ethical approval was obtained from the Glasgow Royal Inrmary Ethics Committee. Informed consent was obtained from all the participants. Patients were told that they had the right to withdraw from the study at any time.
Inclusion criteria

The percentage of glycated haemoglobin (HbA1c%) was measured by high-performance liquid chromatography using an automated analyser Menarini HA8160 analyzer (Menarini Diagnostics Firenze, Italy) according to the method described by Schnedl et al. (2005). Sixty-three of the patients recruited were diabetics and 38 were non-diabetics. The patients underwent a detailed periodontal examination and a blood sample was taken into a tube containing an anticoagulant. A full-mouth six-point periodontal charting was successfully carried out for 99 of the 101 patients recording the following: gingival recession, clinical probing depth (CPD), clinical attachment loss (CAL), measured from the amelocemental junction, but only 96 of the 101 patients had complete records of bleeding on gentle probing. To be included in the periodontitis groups patients had to have a minimum of two sites with CPD andCAL of 44 mm. Non-periodontitis control subjects had no sites withCAL or CPD of 44 mm. Once the study population was divided according to these criteria, 13 had to be excluded from the dataset because they did not t into either the periodontitis or the non-peridontitis groups (two because the clinical data were not complete and 11 because they would not t either category). Two nondiabetics which did not have full-mouth bleeding on probing (BOP) recorded were retained because full-mouth records of CPD and CAL were available.
RANKL, OPG, ICTP and osteocalcin enzyme-linked immunosorbent assays (ELISAs)

ately before assay. The human soluble RANKL (hsRANKL) ELISA development kit (Peprotech EC, London, UK), the human OPG (hOPG) ELISA development kit (R & D Systems, Abingdon, UK) the UniQ ICTP EIA kit (Orion Dianostica, Espoo, Finland) and the osteocalcin immunoassay kit (BioSource Europe, Nivelles, Belgium) were purchased. ICTP assays were performed using a 1/3 dilution of plasma and osteocalcin assays were performed using a 1/5 dilution of plasma. The method used for each assay followed the manufacturers guidelines. The concentration of sRANKL and OPG in each of the samples was then determined by comparing the average sample optical density readings with the concentrations from the assay standard curve. The relevant assays for these proteins are capable of detecting both the free and RANKLOPG complex (Lappin et al. 2007). The relative plasma concentrations of OPG and sRANKL and osteocalcin and ICTP (nanograms per millilitre) were calculated.
Statistical analyses of the ELISA data

A pilot experiment, where OPG levels were measured and a twofold difference obtained, was utilized in statistical power calculations. With a power of 80% and an a 5 0.0125 the minimum number of patients required for the comparisons was 68. Non-parametric distribution-free statistical tests were used to analyse the data. The postulated null hypothesis was no statistically signicant change and the maximum number of comparisons was four. The relationship between the percentage of glycated haemoglobin in the blood (HbA1c%) and (RANKL), (OPG), (osteocalcin) and (ICTP) was determined using the Spearman correlation coefcient, correcting for multiple testing by applying a Bonferroni correction. The correlation analysis was also controlled for confounding variables such as gender, age and presence of periodontitis.

Results
Demographic and clinical parameters

All patients were non-smokers recruited from out-patient clinics at Glasgow Royal Inrmary, Stobhill Hospital, Glasgow, and Glasgow Dental Hospital and School. The age range of the subjects was 2256 years and included both male

The plasma was separated from the blood by centrifugation and stored in aliquots at 701C and thawed immedi-

The non-diabetic group consisted of 16 male and 22 female subjects and the diabetic group consisted of 30 male and 33 female patients. The low HbA1c group consisted of 12 male and 18 female patients and the high HbA1c group of 17

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

636

Lappin et al.

Table 1. Number of sites and teeth with clinical probing depths 44 mm, sites and teeth with clinical attachment loss 44 mm and proportion of sites bleeding on probing (mean standard deviation) Number of sites CPD44 mm Diabetic status (n) Non-diabetic (38)n(36) Low HbA1c ( 8.5%) (30)n o (29) High HbA1c (48.5%) (33)w (31) Periodontal status (n)z Non-diabetic control (19) Non-diabetic periodontitis (17) n (15) Diabetic (26) Diabetic periodontitis (26) 4.13 5.61 3.96 8.85 3.88 5.61 0.00 0.00 8.71 6.09 0.00 0.00 8.86 10.45 Number of teeth CPD44 mm 2.16 1.39 2.14 4.62 2.36 2.19 0.00 0.00 4.57 2.68 0.00 0.00 5.00 4.89 Number of sites CAL44 mm 2.82 6.00 2.89 6.13 4.33 5.99 0.00 0.00 6.16 7.84 0.00 0.00 9.05 7.70 Number of teeth CAL44 mm 1.39 2.57 1.82 3.51 2.39 2.40 0.00 0.00 2.39 2.68 0.00 0.00 5.10 3.63 Proportion of sites BOP 0.33 0.17 0.46 0.19 0.51 0.17 0.35 0.14 0.32 0.21 0.40 0.16 0.55 0.18

n Two of the non-diabetic (periodontitis) patients do not have full-mouth BOP data and one diabetic has this information missing, but full charting of CAL and CPD are available. w Two patients do not have full CAL or BOP data. z Two non-diabetics and eleven diabetics were omitted from the periodontal status analysis as they failed to meet the inclusion criteria. CPD, clinical probing depth; CAL, clinical attachment loss; BOP, bleeding on probing.

male and 16 female patients. The median ages of the patients in the non-diabetic and the diabetic groups (low-HbA1c and high-HbA1c groups) were very similar; non-diabetics 40 (range 2156) years, diabetics with low HbA1c% 34 (2053) years and diabetics with high HbA1c% 40 (2054) years. Clinical data are presented in Table 1. The number of sites and teeth with CPD and CAL44 mm and the number of sites BOP are indicated. The major differences between the diabetic patient groups and the healthy patient groups with periodontitis were as follows: The healthy patients appeared to have lower average levels of BOP than the diabetics, but contrary to what we expected the non-diabetics without periodontitis had a slightly higher level of BOP than the non-diabetics with periodontitis. The diabetics with low HbA1c% appeared to have lower numbers of teeth and sites with CAL44 mm and CPD44 mm than the diabetics with high HbA1c, but there was no statistically signicant difference in these parameters between these two groups. The KruskalWallis test results comparing the three groups were as follows: CPD44 mm, p 5 0.394; number of teeth with CPD44 mm, p 5 0.299; CAL4 4 mm, p 5 0.304 and number of teeth with CPD44 mm, p 5 0.171. Although, the number of teeth and sites with CPD44 mm was similar in both diabetic and non-diabetic periodontitis groups there were more sites and teeth with CAL44 mm in the diabetic periodontitis group. None of these differences were statistically signicant.

Fig. 1. The box plot with quartiles, showing outliers (open diamonds) and extreme values (asterisks), represents the plasma concentration of (a) RANKL, (b) osteoprotegerin, (c) osteocalcin and (d) ICTP in 38 patients without diabetes (non-diabetic) and 63 patients with type 1 diabetes mellitus (diabetic). ICTP, C-terminal telopeptide of type 1 collagen; RANKL, receptor activator of nuclear factor-kB ligand.
Plasma concentrations of RANKL, OPG, ICTP and osteocalcin

differences between the groups in RANKL and ICTP levels were not statistically signicant (Fig. 1).
Inuence of glycated haemoglobin on plasma RANKL, OPG, ICTP and osteocalcin

Diabetic patients had a higher concentration of OPG and a lower concentration of osteocalcin than non-diabetic patients. The MannWhitney U-test indicated that both ndings were statistically signicantly different (OPG, po0.001; osteocalcin, p 5 0.010). The

Diabetics were divided into two groups high and low depending on their mean

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

Bone markers and periodontitis in diabetes

637

Table 2. Comparison of median (Quartile 13) plasma osteocalcin, ICTP, RANKL, OPG levels and RANKL:OPG ratios in the patient groups Osteocalcin (ng/ml) Diabetic status (n) Non-diabetic (38) HbA1c ( 8.5%) (30) o High HbA1c (48.5%) (33) Periodontal status (n)w Non-diabetic (19) Non-diabetic Periodontitis (17) Diabetic (26) Diabetic Periodontitis (26)
n

ICTP (ng/ml)
n

RANKL (ng/ml) 0.83 (0.361.26) 0.66 (0.192.27) 0.66 (0.231.92) 0.95 (0.711.23) 0.51 (0.152.00) 0.63 (0.251.68) 0.95 (0.711.23)

OPG (ng/ml) 1.11 (0.651.55) 2.20 (1.543.02) 1.88 (1.313.11) 0.65 (0.611.15) 1.82 (1.372.72) 2.16 (1.282.89) 0.65 (0.611.15)

RANKL:OPG (mol:mol) 1.45 (0.572.83) 0.65 (0.241.11) 0.90 (0.261.44) 1.09 (0.461.70) 2.22 (2.023.06) 0.68 (0.211.91) 0.84 (0.341.14)

4.78 (3.206.24) 3.34 (2.045.55) 3.30 (2.525.81) 4.80 (3.225.86) 5.22 (4.286.50) 3.15 (2.265.40) 3.71 (2.265.57)

7.18 (4.6214.60) 5.40 (2.9810.44) 5.29 (2.717.82) 5.00 (3.596.98) 6.23 (4.1810.20) 5.32 (3.639.95) 5.00 (3.596.98)

Statistically signicant differences between the groups are shown by the linkages. Two non-diabetics and eleven diabetics were omitted from the periodontal status analysis as they failed to meet the inclusion criteria. Because the data were not normally distributed, the MannWhitney U-test was used to compare the differences between the patient groups and Bonferroni correction was applied to adjust the statistic for multiple comparisons. ICTP, C-terminal telopeptide of type 1 collagen; OPG, osteoprotegerin; RANKL, receptor activator of nuclear factor-kB ligand.

HbA1c% over the previous 2 years. The median HbA1c% (8.60%) was taken as the cut-off point and the two groups were compared with each other and the non-diabetic group (Table 2, Fig. 2). The results obtained for OPG indicated a statistically signicant difference between some of the groups according to the KruskalWallis test (Fig. 2b). When the three groups were compared and analysed post hoc with the MannWhitney U-test and a Bonferroni correction for multiple comparisons, statistically signicant differences were found for the following comparisons: non-diabetic group and low HbA1c group (po0.001) and the non-diabetic group and high HbA1c group (po0.001). Diabetics with high HbA1c% showed the lowest concentrations of osteocalcin (Fig. 2c; median 5 5.91 ng/ml). The results showed a statistically signicant difference between the groups according to the KruskalWallis test (p 5 0.006) When the three groups were analysed by the MannWhitney U-test with Bonferronis correction a statistically signicant difference in plasma osteocalcin concentrations was observed when comparing the non-diabetic group and the high HbA1c group (p 5 0.011) and between the low- and high-HbA1c groups (p 5 0.0163). There were no statistically signicant differences in RANKL or ICTP concentrations between any of the groups (p 5 0.649 and 0.092, respectively) (Fig. 2a and d).

Fig. 2. The box plot with quartiles, showing outliers open diamonds and extreme values (asterisks), represents the plasma concentration of (a) RANKL, (b) Osteoprotegerin, (c) Osteocalcin and (d) ICTP in 38 non-diabetic patients and 63 patients with diabetes split into two groups based on their HbA1c%: 30 with low HbA1c% (less than the median) and 33 with high HbA1c% (median HbA1c% and above). ICTP, C-terminal telopeptide of type 1 collagen; RANKL, receptor activator of nuclear factor-kB ligand.

There was a statistically signicant difference in the RANKL:OPG ratio (Table 2) between the non-diabetic and the diabetic patient groups according to the KruskalWallis test (p 5 0.0023). In

the post hoc analysis: both high- and lowHbA1c groups were statistically signicantly different from the non-diabetics (p 5 0.0015 and 0.0016, respectively), but not from each other (p 5 0.080).

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

638

Lappin et al.

Inuence of periodontitis on plasma RANKL, OPG, ICTP and osteocalcin

There was no difference in the extent of periodontitis between the non-diabetic and the diabetic patients. Thirteen patients (11 diabetics and 2 non-diabetics) were excluded from the rest of the study because of difculty in placing them in either the periodontitis group or the non-periodontitis control group. Therefore the groups consisted of diabetic non-periodontitis n 5 26; diabetic with periodontitis n 5 26; non-diabetic non-periodontitis n 5 19; and non-diabetic with periodontitis n 5 17. When the patients were allocated according to their disease status, diabetics with periodontitis had the highest plasma concentration of OPG (Table 2, Fig. 3b). The KruskalWallis test indicated that the results were statistically signicantly different (po0.001). The MannWhitney U-test with a Bonferroni correction indicated statistically signicant differences in plasma OPG (Fig. 3b) in the following comparisons: non-diabetics and diabetics without periodontitis (p 5 0.0193); non-diabetic patients with periodontitis and diabetics with periodontitis (po0.0001) and; non-diabetics without periodontitis and non-diabetics with periodontitis (p 5 0.0002). The highest osteocalcin levels were in the non-diabetics without periodontitis and the lowest in the non-diabetics with periodontitis but the levels in this non-diabetics without periodontitis were similar to those in both diabetic patient groups, i.e. the diabetic patients with and without periodontitis. There were statistically signicant differences in osteocalcin (Table 2, Fig. 3c) concentrations (p 5 0.0033) and the post hoc test showed statistically signicant differences between the non-diabetics without periodontitis and the non-diabetics with periodontitis groups (p 5 0.038); the non-diabetic group without periodontitis and diabetics without periodontitis (p 5 0.0082). There were no statistically signicant differences in RANKL or ICTP concentrations (Table 2, Fig. 3a and d). Statistically signicant differences in the ratio of RANKL:OPG (Table 2) were observed between the following patient groups: the nondiabetic patients without periodontitis and the non-diabetic patients with periodontitis (p 5 0.0018) and the diabetic patients with periodontitis and the nondiabetic patients with periodontitis (po0.0001).

Fig. 3. The box plot with quartiles, showing outliers (open diamonds) and extreme values (asterisks), represents the plasma concentration of (a) RANKL, (b) osteoprotegerin, (c) osteocalcin and (d) ICTP in 36 patients without type 1 diabetes mellitus (non-diabetic) and 52 patients with diabetes each split into two groups based on the presence of periodontitis. Seventeen of the non-diabetics had periodontitis and 19 did not and 26 diabetics had periodontitis and 26 did not. Two control patients and 11 diabetics were excluded from the analysis because they did not t the inclusion criteria to be included in either the periodontitis or the control (non-periodontitis) groups. ICTP, C-terminal telopeptide of type 1 collagen; RANKL, receptor activator of nuclear factor-kB ligand.
Correlations

The plasma concentrations of RANKL did not show any statistically signicant correlation with HbA1c%; Spearmans correlation coefcient (r) 5 0.059 (p 5 0.560). The plasma concentration of OPG indicated a statistically signicant positive correlation (rho 5 0.339, p 5 0.0005) with the HbA1c% in blood. The plasma concentration of osteocalcin indicated a statistically signicant negative correlation (rho 5 0.324; p 5 0.0018) with HbA1c%. These relationships remained statistically signicant when the data were controlled for the presence of periodontitis (HbA1c% and OPG, p 5 0.008 and HbA1c% and osteocalcin, p 5 0.009) the gender (p 5 0.008 and 0.009) or the age (p 5 0.006 and 0.004) of the patients. Osteocalcin concentrations also correlated with the extent of periodontitis (r 5 0.285, p 5 0.004) and appeared to be statistically signicant when controlling for the presence of diabetes (p 5 0.041), but this was before the data were corrected for multiple comparisons.

Plasma ICTP concentrations did not correlate with HbA1c% (r 5 0.146; p 5 0.147) or OPG levels, but there was a positive correlation between ICTP and RANKL (r 5 0.280; p 5 0.0050 and the RANKL:OPG ratio (r 5 0.323; p 5 0.0011). BOP correlated with the number of deep sites (po0.001) and with the HbA1c% (p 5 0.002).

Discussion

Activation of the transcription factor nuclear factor-kB following ligation of RANKL to the cell surface receptor RANK an event that promotes osteoclast formation (Hofbauer et al. 2000) and also induces the production of a cysteine proteinase, cathepsin K, in osteoclasts. Cathepsin K is involved in bone matrix solubilization (Crotti et al. 2003). OPG is a member of the TNF receptor family that is expressed by osteoblasts and other cell types, notably insulin secreting b cells of the pancreas (Schrader et al. 2007). OPG inhibits osteoclast formation

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

Bone markers and periodontitis in diabetes

by high afnity binding to RANKL and prevents RANKL from coupling with the RANK receptor (Hofbauer et al. 2000). This tends to result in reduced bone resorption. In periodontitis, increased concentrations of RANKL are found in diseased tissues, and the disruption to the balance between RANKL and OPG concentrations is associated with disease severity (Liu et al. 2002, Crotti et al. 2003, Garlet et al. 2006). Although diabetes has been established as a risk factor for periodontitis and the osteocalcin levels in serum, gingival tissues and saliva correlate to periodontal status in nondiabetics (Bullon et al. 2005), the osteocalcin concentrations in diabetics and their relationship to periodontitis have not been established. In this study in order to avoid possible confounding between the periodontitis and the non-periodontitis groups we dened the level of disease before the analysis, i.e. a diagnosis of periodontitis was conrmed by a minimum of two sites with 44 mm CAL and CPD. Therefore 13 patients with only one site of 44 mm CAL and/or CPD were excluded. Statistically signicant differences were not observed in RANKL concentrations, but in contrast statistically signicant differences in OPG and osteocalcin and RANKL:OPG ratios were observed. Serum osteocalcin is presently considered to be a valid marker of bone turnover when resorption and formation are coupled, and a specic marker of bone formation when formation and resorption are uncoupled (Giannobile et al. 2003). In this study both diabetic and periodontitis patients have lower levels of osteocalcin than non-diabetic patients without periodontitis, suggesting lower bone formation. Reduced osteocalcin levels in the presence of periodontitis have been reported before by other researchers (Bullon et al. 2005 and Buduneli et al. 2005) and the present data show a statistically signicant inverse relationship between the extent of periodontitis and the osteocalcin levels. The data also suggest that the failure to adequately control diabetes leads to a reduction in plasma osteocalcin and that the elevated blood glucose acts independently of the presence of other factors such as age gender or periodontitis to inuence the expression of osteocalcin. There are several mechanisms that might explain the greater incidence of periodontitis in diabetes: a greater susceptibility to infection as a result of diminished neutrophil function, the formation of advanced glycation end products which increase oxidative stress in the tissues; and the binding of advanced glycation end products to cell surface receptors which stimulates the increased production of inammatory cytokines and causes delayed wound healing. These mechanisms would induce periodontal destruction through osteoclastogenesis. However, there is also evidence that diabetes suppresses osteoblastogenesis after bacterial challenge and the net bone loss could be due to reduced bone formation because of apoptosis of bone lining cells (He et al. 2004). Advanced glycation end products have also been shown to induce apoptosis of osteoblasts in diabetics via the mitogen-activated protein (MAP) kinase and cytosolic apoptotic pathway (Alikhani et al. 2007). The reduced plasma osteocalcin concentrations in the diabetic patients in the present study could be explained by a reduction in the number of osteocalcin secreting osteoblasts. Numerous studies have established lower concentrations of OPG and lower expression of OPG in periodontitis patients compared with normal healthy individuals (Belibasakis et al. 2007, Bostanci et al. 2007, Lappin et al. 2007), but in contrast diabetes patients have higher levels of OPG than healthy individuals (Kim et al. 2005). The results of the present study conrmed that higher OPG concentrations are detected in diabetics, and indicated that there is a positive correlation between OPG and HbA1c%. It is particularly intriguing that the diabetic patients have higher circulating levels of OPG than healthy individuals because they have a greater severity of periodontitis than healthy subjects (Khader et al. 2006). OPG is considered to be a bone protective protein as it essentially reduces the activity of RANKL. High plasma levels of OPG are also linked to endothelial cell dysfunction (Secchiero et al. 2006) and the early onset of diabetes mellitus and have implications in the higher susceptibility of diabetics to coronary artery disease (Anand et al. 2007) and myocardial infarction (Avignon et al. 2007). The nding that the level of BOP was greater in the diabetics could also be partly explained if there was a dysfunction of the endothelial cells in the gingival blood vessel walls. Further studies are required to conrm this. We expected to see a signicantly higher level of BOP in the

639

periodontitis groups compared with those without periodontitis but in the non-diabetics this was not the case. This may be due to the poor level of oral hygiene seen in the general population in the United Kingdom (Kelly et al. 1998). Although the number of teeth and sites with 44 mm CPD were similar between the diabetic and non-diabetic periodontitis groups the diabetic periodontitis group had more teeth and sites affected by CAL44 mm. These differences were accounted for by a greater prevalence of gingival recession in the diabetics with periodontitis. Hyperglycaemia does not appear to be directly responsible for elevated plasma OPG in diabetics (Knudsen et al. 2007). However, our results suggest a strong relationship between HbA1c% and increased OPG concentrations acting independently of other factors such as age, gender or periodontitis. A question that has yet to be resolved is whether high levels of OPG are directly involved in the pathogenic process or are indicative of the process taking place. The pathogenesis of type 1 diabetes mellitus is caused by immunemediated pancreatic b cell destruction. Schrader et al. (2007) indicated that cytokine induced OPG production may protect pancreatic b cells from further damage and may be partially mediated through inhibition of p38 MAP kinase phosphorylation and suggested that OPG may be elevated to protect the pancreatic b cells. In diabetics increased levels of OPG should in theory reduce bone resorption; however, diabetic patients have been shown in a recent meta analysis to have a similar extent of periodontitis to non-diabetics, but more severe disease (Khader et al. 2006). Bone remodelling is a coupled process involving bone formation and bone resorption. As bone formation (indicated by osteocalcin levels) is lower in diabetic patients, regardless of periodontal status, cross-linked ICTP a specic biochemical marker of bone resorption was measured (Bacovsky et al. 2002). Although lower levels of ICTP were observed in the plasma of the diabetic patients there was no statistically signicant difference in ICTP levels between non-diabetic and diabetic patients. Increased levels of this bone marker have been observed in the GCF of periodontitis patients (Giannobile 1999, Giannobile et al. 2003). The current results would tend to agree since higher median plasma concentrations of

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

640

Lappin et al.
protegerin: a novel independent marker for silent myocardial ischemia in asymptomatic diabetic patients. Diabetes Care 30, 29342939. Bacovsky, J., Scudla, V., Vytrasova, M., Budikova, M. & Myslivecek, M. 2002 Monitoring of bone resorption and bone formation in multiple myeloma. Biomedical Papers 146, 5961. Belibasakis, N. G., Bostanci, N., Hashim, A., Johansson, A., Aduse-Opoku, J., Curtis, M. A. & Hughes, F. J. 2007 Regulation of RANKL and OPG gene expression in human gingival broblast and periodontal ligament cells by Porphyromonas gingivalis: a putative role of the Arg-gingipains. Microbial Pathogenesis 43, 4653. Bostanci, N., Ilgenli, T., Emingil, G., Afacan, B., Han, B., Toz, H., Atilla, G., Hughes, F. J. & Belibasakis, G. N. 2007 Gingival crevicular uid levels of RANKL and OPG in periodontal diseases: implications of their relative ratio. Journal of Clinical Periodontology 34, 370376. Bridges, M. J., Moochhala, S. H., Barbour, J. & Kelly, C. A. 2005 Inuence of diabetes on peripheral bone mineral density in men: a controlled study. Acta Diabetologica 42, 8286. Buduneli, E., Buduneli, N., Vardar-Senugul, S., Kardeseler, L., Atilla, G., Lappin, D. & Kinane, D. F. 2005 Systemic low dose doxycycline and alenodronate administration and serum Interleukin 1 Beta, osteocalcin and C-reactive protein levels in rats. Journal of Periodontology 76, 19271933. Bullon, P., Goberna, B., Guerrero, J. M., Segura, J. J., Parez-Cano, R. & MartinezSahuguillo, A. 2005 Serum, saliva and gingival crevicular uid osteocalcin: their relation to periodontal status and bone mineral density in postmenopausal women. Journal of Periodontology 76, 513519. Corral, D. A., Amling, M., Priemel, M., Loyer, E., Fuchs, S., Ducy, P., Baron, R. & Karsenty, G. (1998) Dissociation between bone resorption and bone formation in osteopenic transgenic mice. Proceeding of the National Academy of Sciences of the United States of America 95, 1383513840. Crotti, T., Smith, M. D., Hirsch, R., Soukoulis, S., Weedon, H., Capone, M., Ahern, M. J. & Haynes, D. (2003) Receptor activator NF kappa B ligand (RANKL) and osteoprotegerin (OPG) protein expression in periodontitis. Journal of Periodontal Research 38, 380387. Dhore, C. R., Cleutjens, J. P. M., Lutgens, E., Cleutjens, K. B. J. M., Geusens, P. P. M., Kitslaar, P. J. E. H. M., Tordoir, J. H. M., Spronk, H. M. H., Vermeer, C. & Daemen, M. J. A. P. (2001) Differential expression of bone matrix regulatory proteins in human atherosclerotic plaques. Arteriosclerosis Thrombosis Vascular Biology 21, 19982003. Fukushima, H., Kajiya, H., Takada, K., Okamoto, F. & Okabe, K. (2003) Expression and role of RANKL in periodontal ligament cells during physiological root-resorption in human deciduous teeth. European Journal of Oral Science 111, 346352. Garlet, G. P., Cardoso, C. R., Silva, T. A., Ferreira, B. R., Avila-Campos, M. J., Cunha, F. Q. & Silva, J. S. (2006) Cytokine pattern determines the progression of experimental periodontal disease induced by Actinobacillus actinomycetemcomitans through the modulation of MMPs, RANKL, and their physiological inhibitors. Oral Microbiology Immunology 21, 1220. Giannobile, W. V. 1999 C-telopeptide pyridinoline cross-links. Sensitive indicators of periodontal tissue destruction. Annals of the New York Academy of Science 878, 404412. Giannobile, W. V., Al-Shammari, K. F. & Sarment, D. P. 2003 Matrix molecules and growth factors as indicators of periodontal disease activity. Periodontology 2000 31, 125134. He, H., Liu, R., Desta, T., Leone, C., Gerstenfeld, L. C. & Graves, D. T. (2004) Diabetes causes decreased osteoclastogenesis, reduced bone formation, and enhanced apoptosis of osteoblastic cells in bacteria stimulated bone loss. Endocrinology 145, 447452. Hofbauer, L. C., Khosla, S., Dunstan, C. R., Lacey, D. L., Boyle, W. J. & Riggs, B. L. (2000) The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption. Journal of Bone Mineral Research 15, 212. Jono, S., Ikari, Y., Shioi, A., Mori, K., Miki, T., Hara, K. & Nishizawa, Y. (2002) Serum osteoprotegerin levels are associated with the presence and severity of coronary artery disease. Circulation 106, 11921194. Kanzaki, H., Chiba, M., Shimizu, Y. & Mitani, H. (2002) Periodontal ligament cells under mechanical stress induce osteoclastogenesis by receptor activator of nuclear factor kappaB ligand up-regulation via prostaglandin E2 synthesis. Journal of Bone Mineral Research 17, 210220. Kelly, M., Steele, J., Nuttall, N., Bradnock, G., Morris, J., Nunn, J., Pine, C., PiUs, N., Treasure, E. & White, D. (1998) Adult Dental Health Survey: Oral Health in the United Kingdom. The Stationery Ofce, London, UK. Khader, Y. S., Dauod, A. S., El-Qaderi, S. S., Alkafajei, A. & Batayha, W. Q. (2006) Periodontal status of diabetics compared with nondiabetics: a meta-analysis. Journal of Diabetes Complications 20, 5968. Kim, S. M., Lee, J., Ryu, O. H., Lee, K. W., Kim, H. Y., Seo, J. A., Kim, S. G., Kim, N. H., Baik, S. H., Choi, D. S. & Choi, K. M. (2005) Serum osteoprotegerin levels are associated with inammation and pulse wave velocity. Clinical Endocrinology 63, 594598. Knudsen, S. T., Jeppesen, P., Poulsen, P. L., Andersen, N. H., Bek, T., Schmitz, O., Mogensen, C. E. & Rasmussen, L. M. (2007) Plasma concentrations of Osteoprotegerin during normo-and hyperglycaemic clamping. Scandinavian Journal of Clinical and Laboratory Investigation 67, 135142. Lappin, D. F., Sherrabeh, S., Jenkins, W. M. M. & Macpherson, L. M. D. (2007) Effect of

ICTP were observed in the groups with periodontitis in this study but the differences failed to reach statistical signicance. This failure to see a statistically signicant difference may be due to measuring ICTP in plasma rather than GCF or to the lack of sufcient statistical power for determining differences in this particular marker. The lower levels of ICTP that were observed possibly reect the inuence of OPG at reducing bone resorption. In conclusion diabetics do not appear to be more susceptible to bone destruction, but may have a defect in bone formation. The reduced osteocalcin levels in type 1 diabetics suggests that these patients have a reduction in their intrinsic ability to replace bone, such as that which has been destroyed during acute bursts of periodontitis. This may make them more susceptible to progression of this disease. In support of this other researchers have not indicated a signicant change in bone density in diabetics but have indicated that bone formation may be lower in diabetic patients (Bridges et al. 2005, Oz et al. 2006). Future studies in type 1 diabetics should investigate markers of bone metabolism in combination with bone density measurements.

References
Abildgaard, N., Bentzen, S. M., Nielsen, J. L. & Heickendorff, L. (1997) Serum markers of bone metabolism in multiple myeloma: prognostic value of the carboxy-terminal telopeptide of type I collagen (ICTP). Nordic Myeloma Study Group (NMSG). British Journal of Haematology 96, 103110. Alexopoulou, O., Jamart, J. & Devogelaer, J. P. (2006) Bone density and markers of bone remodeling in type 1 diabetes patients. Diabetes Metabolism 32, 453458. Alikhani, M., Alikhani, Z., Boyd, C., Maclellan, C. M., Raptis, M., Liu, R., Pischon, N., Trackman, P. C., Gerstenfeld, L. & Graves, D. T. (2007) Advanced glycation end products stimulates osteoblast apoptosis via the MAP kinase and cytosolic apoptotic pathways. Bone 40, 345353. Anand, D. V., Lim, E., Darko, D., Bassett, P., Hopkins, D., Lipkin, D., Corder, R. & Lahiri, A. (2007) Determinants of progression of coronary artery calcication in type 2 diabetes role of glycemic control and inammatory/vascular calcication markers. Journal of the American College of Cardiology 50, 22182225. Avignon, A., Sultan, A., Piot, C., Mariano Goulart, D., Thuan Dit Dieudonne, J. F., Cristol, J. P. & Dupuy, A. M. (2007) Osteo-

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S

Bone markers and periodontitis in diabetes

smoking on serum RANKL and OPG in sex, age and clinically matched supportive therapy periodontitis patients. Journal of Clinical Periodontology 34, 271277. Liu, D., Xu, J. K., Figliomeni, L., Pavlos, N. J., Rogers, M., Tan, A., Price, P. & Zheng, M. H. (2002) Expression of RANKL and OPG mRNA in periodontal disease: possible involvement in bone destruction. International Journal of Molecular Medicine 11, 1722. Oz, S. G., Guven, G. S., Kilicarslan, A., Calik, N., Beyazit, Y. & Sozen, T. (2006) Evaluation of bone metabolism and bone mass in patients with type-2 diabetes mellitus. Journal of the National Medical Association 98, 15981604. Page, R. C. (1991) The role of inammatory mediators in the pathogenesis of periodontal disease. Journal of Clinical Research 26, 230242. Risteli, L. & Risteli, J. (1993) Biochemical markers of bone metabolism. Annals of Medicine 25, 385393. Schnedl, W. J., Lahousen, T., Wallner, S. J., Krause, R. & Lipp, R. W. (2005) Silent hemoglobin variants and determination of HbA(1c) with the high-resolution program of the HPLC HA-8160 hemoglobin analyzer. Clinical Biochemistry 38, 8891. Schrader, J., Rennekamp, W. & Niebergall, U. (2007) Cytokine-induced osteoprotegerin expression protects pancreatic beta cells through p38 mitogen-activated protein kinase signaling against cell death. Acta Diabetologia 50, 12431247. Secchiero, P., Corallini, F., Pandol, A., Consoli, A., Candido, R., Fabris, B., Celeghini, C., Capitani, S. & Zauli, G. (2006) An increased osteoprotegerin serum release characterizes the early onset of diabetes mellitus and may contribute to endothelial cell dysfunction. American Journal of Pathology 169, 22362244.

641

sites 44 mm Clinical Probing Depths or Clinical Attachment Loss attributable to periodontal disease) nor the periodontitis group (at least 2 sites with 44 mm Clinical attachment loss and increased Clinical Probing Depths): Study codes 4, 10, 18, 22, 32, 37, 42, 61, 66, 80, 102, 124, 131. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
Address: David F Lappin Infection and Immunity Group Faculty of Medicine Dental Hospital and School University of Glasgow 378 Sauchiehall Street Glasgow G2 3JZ UK E-mail: [email protected]

Supporting Information

Additional Supporting Information may be found in the online version of this article: Appendix S1. The following patients were excluded because they met neither the criteria for the control group (no cult to assess the impact of increased OPG levels on net bone turnover. Principal ndings: We conrmed that higher plasma OPG and lower RANKL:OPG ratios are found in diabetics and that the presence of periodontitis altered the RANKL:OPG ratio. Lower osteocalcin levels were found in diabetics and in periodontitis. However, there was no statistically signicant change in ICTP levels in the diabetic groups or in the groups with periodontitis.

Clinical Relevance

Scientic rationale for study: RANKL interacts with receptors (RANK) on the surface of osteoclasts to stimulate bone resorption. OPG decays RANKL reducing RANK activation. OPG levels in diabetics are reported to be high, therefore it should follow that the amount of bone loss in patients with periodontitis would be less in diabetics than non-diabetics. However without a measure of RANKL and an indicator of bone formation (osteocalcin) and destruction (ICTP) it is dif-

Practical implications: Biomarkers indicating bone formation are lower in type 1 diabetes mellitus patients. Inammatory signals that would tend to increase bone destruction are possibly counteracted by high OPG levels in diabetics. It therefore appears that overall bone turnover may be reduced in type 1 diabetes possibly due to reduced new bone formation as indicated by a reduction of plasma osteocalcin.

r 2009 John Wiley & Sons A/S Journal compilation r 2009 John Wiley & Sons A/S