127

Inflammatory Mediator Response as a

Potential Risk Marker for Periodontal
Diseases in Insulin-Dependent Diabetes
Mellitus Patients
Giovanni E. Salvi,*' Behnaz Yaida* John G. Collins* Bettye H. Jones* Franees
W. Smith,* Roland R. Arnold,* and Steven Offenbacher*

The gingival crevicular fluid (GGF) and monocytic secretion of Prostaglandin E2

(PGE2) and interleukin lß (IL-lß) were measured in a group of 39 insulin-dependent
diabetes mellitus (IDDM) patients and 64 systemically healthy individuals. Diabetics
were divided into Group A (gingivitis or mild periodontal disease) and Group (mod-
erate or severe periodontal disease). Diabetics had significantly higher GCF levels of
both PGE2 and IL-lß as compared to non-diabetic controls who were matched with
regard to periodontal disease severity (P < 0.00001 and 0.0005, respectively).
=

Within the diabetic population, the GCF levels of these inflammatory mediators were
almost 2-fold higher in Group as compared to Group A (P 0.01, =
0.006,
=

respectively for GCF-PGE2 and IL-lß). Furthermore, diabetics as a group had a sig-

nificantly higher monocytic PGE2 and IL-lß production in response to various con-
centrations of both Escherichia coli and Porphyromonas gingivalis lipopolysaccharide
(LPS) as compared to non-diabetic patients with adult Periodontitis ( 0.0001). LPS
=

dose-response curves demonstrated that monocytes from Group diabetics produced

approximately 3 times more PGE2 than Group A monocytes; however, there was no
significant difference in monocytic IL-lß secretion within the IDDM patients. The
levels of GCF or monocytic mediators did not correlate with age, race, or glycosylated
hemoglobin (HbA,c) levels. Our data suggest that the high GCF and monocytic se-
cretion of PGE2 and IL-lß in IDDM patients may be a consequence of a systemic
response trait and that the presence of Gram-negative infections such as periodontal
diseases may interact synergistically to yield high local levels of these mediators and
a more severe periodontal condition. J Periodontol 1997;68:127-135.

Key Words: Gingival crevicular fluid/analysis; diabetes mellitus; Prostaglandin E2;

interleukin lß.

Insulin-dependent diabetes mellitus (IDDM) is generally
associated with an increased risk for bacterial infections
including oral infections such as periodontal diseases.1-2
Periodontal diseases are widespread anaerobic Gram-neg-
ative infections that lead to the destruction of the bone
and connective tissues supporting the teeth.3 Epidemio-
logie studies by Haber et al.4 and Grossi et al.5 have un-
equivocally established diabetes as a risk factor for Per-
iodontitis with an attributable risk odds ratio of 2.1 to 3.0.
Traditionally, diabetics have been considered more sus-
*University of North Carolina, Dental Research Center, Department of
Periodontics, Chapel Hill, NC.
'University of Berne, School of Dental Medicine, Berne, Switzerland.
ceptible to severe periodontal disease due to metabolic
and host response imbalances that result in impaired
wound healing, microangiopathy, and diminished salivary
flow.6-8 More recently, defects in neutrophil Chemotaxis,
phagocytosis, and bacterial killing which have been re-
ported in diabetics910 have been cited as possible mech-
anisms for the increased severity of periodontal disease
associated with IDDM. Neutrophils serve an important
protective function in the periodontium and defects in
neutrophil function dramatically increase the risk for se-
vere and early-onset forms of Periodontitis.1112
Recent reviews13"15 have provided additional insights
into the pathophysiological consequences of chronic hy-
perglycemia. Advanced glycation end products (AGEs)

128 MEDIATOR RESPONSE AND PERIODONTAL DISEASE IN IDDM PATIENTS
arenon-enzymatically glycated proteins and lipids which
can form spontaneously in healthy individuals. However,
a chronically elevated glucose level leads to accelerated
AGE formation and accumulation, especially in vessel
walls. Both monocytes and endothelial cells possess spe-
cific receptors for AGEs called RAGEs16 which upon ac-
tivation upregulate cytokine expression and induce oxi-
dative stress.17 This would lead to an enhanced inflam-
matory responsiveness trait and may therefore be a pos-
sible candidate mechanism to explain why diabetics have
a more severe expression of periodontal disease.
Another potential complication of hyperglycemia re-
sults from the glycation of collagen.1819 This may lead to
an increased thickness of the basement membrane due to
the glycation of type IV collagen and resultant poor per-
fusion enhancing oxidative stress and reducing metabolic
waste removal within the tissues. Glycation of the con-
stitutive type I collagen in gingiva, periodontal ligament,
and bone also results in increased collagen cross-linking
which serves to reduce solubility and metabolic turnover.
Thus, diabetes-associated alterations in connective tissue
metabolism may contribute to the severity of periodontal
disease expression.20
An altered monocyte cytokine secretory phenotype has
been suggested to be common in IDDM patients.21 Mono-
cytes isolated from certain diabetic patients have been
reported to express an exaggerated inflammatory response
to a given bacterial challenge, as evidenced by the ex-
cessive production of inflammatory mediators such as
PGE2, IL-lß, and tumor necrosis factor alpha (TNF-a) in
response to lipopolysaccharide (LPS).22-24 This monocytic
2- to 3-fold increase in mediator secretion as a response
to Gram-negative LPS (a M0+ trait) has been suggested
to be regulated by genes in the HLA-DR region.24 The
extensive polymorphism in the HLA-D region results in
genetically distinct interindividual differences in antigen-
specific immune responsiveness and in cytokine re-
sponses,25 which has been suggested to lead to differential
susceptibility to autoimmune disorders including rheu-
matoid arthritis and IDDM. The diabetic susceptibility
genes are currently believed to map in or around the HLA-
DR3I4 and, more recently, the HLA-DQ region.26-27 These
associations lead to the suggestion that the LPS and im-
mune response genes in the HLA-DR region that control
monocyte inflammatory responses to bacterial challenge
may be linked to diabetes susceptibility genes. Logically,
one might expect that IDDM patients with periodontal
disease may have a more severe disease expression, if the
M0+ trait is present. For example, genetic polymorphisms
associated with TNF-a hypersécrétion have been dem-
onstrated to predispose children affected with cerebral
malaria to lethal outcomes28 and to increase morbidity in
patients affected with mucocutaneous leishmaniasis.29
There is growing evidence that increased levels of
PGE2 and IL-lß in the gingival crevicular fluid can serve
J Periodontol
February 1997
as markers of periodontal disease severity and can reflect
periodontal disease activity.3031 Hypothetically, the pres-
ence of the M0+ trait in an IDDM patient may result in
excessively elevated GCF levels of PGE2 and IL-lß and
increased disease severity. To test this hypothesis we ex-
amined the levels of crevicular fluid inflammatory me-
diators and the monocytic secretion of these mediators to
identify potential phenotypic patterns for Periodontitis
susceptibility within a group of IDDM patients with and
without severe periodontal disease. IDDM patients, with
or without severe Periodontitis, have been matched on the
basis of glycosylated hemoglobin levels to control for the
effect of metabolic abnormalities on the expression of
periodontal disease severity. Both gingival crevicular flu-
id levels and monocytic secretion of PGE2 and IL-lß are
elevated 2- to 3-fold in diabetics, relative to non-diabetic
Periodontitis controls. Thus, our data suggest that diabetic
patients have exaggerated inflammatory responses as
compared to systemically healthy individuals. Further-
more, within the diabetic population, severe Periodontitis
was associated with a 2-fold increase in these inflamma-
tory mediator responses, as compared to the HbA,c
matched patients without severe periodontal disease.
MATERIALS AND METHODS
Patient Selection and Clinical Assessment
Patients were volunteers selected from the School of Den-
tistry and Memorial Hospitals and associated clinics from
the University of North Carolina at Chapel Hill. All di-
abetic patients were insulin-dependent. Patients were ex-
cluded if they had a history of antibiotic therapy within
the preceding 6 months or non-steroidal anti-inflamma-
tory drug therapy within the preceding 3 months. Patients
with other systemic conditions which are known to mod-
ify periodontal disease expression, such as pregnancy or
Crohn 's disease, were excluded. The blood values of gly-
cosylated hemoglobin were obtained from the diabetolo-
gist within 3 months of the clinical examination and sam-
ple collection. Unfortunately, information regarding onset
and duration of diabetes mellitus was not obtained. Thir-
ty-nine IDDM patients participated in the study; 43 non-
diabetic patients with various degrees of adult Periodon-
titis and 21 periodontally healthy individuals served as
controls. Informed consent was obtained from all sub-
jects. All 103 patients received a full-mouth periodontal
examination including assessment of bleeding on probing
(BOP) and probing depth. Clinical attachment level
(CAL) measurements using the cemento-enamel junction
(CEJ) as a reference point were obtained for Periodontitis
patients. Probing measurements were performed at 6 sites
per tooth using a computer-interfaced controlled force (25
g) probe.* Bleeding on probing was expressed as the per-
*Florida Probe Corporation, Gainesville, FL.
Volume 68
Number 2
cent of sites which bled after probing. Probing depth and
clinical attachment level measurements were expressed as
mean mm per site averaging all sites (6/tooth) over all
teeth.
Measurement of Gingival Crevicular Fluid, PGE2,
and IL-lß Levels
Crevicular fluid samples were collected from all patients
using paper strips8 at 8 sites per patient, two in each quad-
rant. The interdental areas between the two most posterior
teeth in each quadrant were selected for sampling. GCF
collection and determination of GCF-PGE2 concentrations
were performed by radioimmunoassay as described pre-
viously.32 GCF strips were placed at the entrance of the
pocket until visibly moistened. The volume of each paper
strip was determined using a calibrated Periotron 6000.11
Each strip was eluted and independently analyzed for
PGE2 or IL-lß and. expressed as a GCF concentration.
IL-lß levels in GCF were determined by enzyme
linked immunoabsorbent assay (ELISA) as per manufac-
turer's instructions.11 The amount of IL-lß was determined
by interpolation from standard curve plotting log IL-lß
versus cumulative normal proportion bound. As with the
PGE2 determinations, the mean GCF-IL-lß concentration
for each patient represents the average of 4 sites.
Monocyte Isolation and Culturing
Peripheral blood monocytes were isolated, cultured and
stimulated with LPS using methods described by M0lvig
et al.25 Approximately 40 ml of whole blood was obtained
from each patient by standard venipuncture in heparinized
tubes. Peripheral blood monocytes were isolated from a
single donor by layering one ml of cold Ficoll-hypaque*
polysucrose (5.7 g/dl) and sodium diatrizoate (9.0 g/dl)
onto 3 ml of resolving media 8% (w/v) Ficoll 400, 5.65%
(w/v) sodium diatrizoate and 11.3% (w/v) meglumine dia-
trizoate** in 15 ml centrifuge tubes. Five ml of whole
blood was gently layered over the Ficoll-hypaque** and
centrifuged at 300 X g for 30 minutes. Platelets and
mononuclear cells were collected with a sterile Pasteur
pipette and washed twice with PBS by centrifugation at
300 X g for 10 minutes to remove non-adherent platelets.
The cell pellet was resuspended at a final concentration
of 10 X 106 cells/ml in RPMI 1640 supplemented with
10% pooled AB serum** and 1% penicillin/streptomycin.
Monocytes were allowed to adhere to 24-well culture
plates (0.5 ml/well) for 2 to 3 hours at 37°C and 5% C02.
Non-adherent cells were removed with 2 gentle washings
§Periotron ProFlow, Inc., Amityville, NY.
"Periopaper, ProFlow, Inc.
'Cistron Biotechnology, Pinebrook, NJ.
Histopaque 1077, Sigma Chemical Co., St. Louis, MO.
**Mono-poly Resolving Media, Flow Labs, Irvine, CA.
1
'Sigma Chemical Co., St. Louis, MO.
«Serotype 0127:B8, Sigma Chemical Co., St. Louis, MO.
SALVI, YALDA, COLLINS, ET AL. 129
with PBS. Adherent monocytes were removed with
0.01M EDTA-PBS, pH 7.4 at 4°C, and replated at 108
cells/ml. Aliquots of supplemented RPMI containing var-
ious concentrations of Porphyromonas gingivalis A 7436
(P. gingivalis) LPS (0, 0.003, 0.03, 0.3, and 3.0 pg/ml)
or Escherichia coli LPS (0, 0.03, 0.3, 3.0 µg/ml) were
added to each culture well. After overnight stimulation
(12 to 18 hours), the supernatants were collected and fro-
zen at 80°C until analysis. Representative cell cultures
—
were assayed for total DNA content of remaining adher-
ent cells using the fluorescent DAPI technique described
by Brunk et al.33 There were no significant differences in
cell number or final plating density among patient groups
or culturing conditions (CV 12 to 15%). For this rea-
=
son, mediator levels are expressed as final concentrations
within the culture media.
LPS Preparation
P. gingivalis LPS was isolated using a hot phenol/water
extraction method followed by CsCl isopycnic density
gradient centrifugation using methods described by Kas-
per.34 Briefly, bacteria grown in Schaedler's broth for 48
hours were washed and resuspended in PBS (approxi-
mately 2 ml in 15 ml Pyrex centrifuge tubes). The sus-
pension was heated to 68°C in a water bath and an equal
volume of 90% aqueous solution of phenol added. This
suspension was vortexed every 2 to 3 minutes for 15 min-
utes replacing the suspension in the 68°C water bath be-
tween mixings. After the last mixing, the phenol-water
suspension was chilled on ice, then centrifuged at 4°C at
1,000 X g for 10 minutes to separate the phenol and water
phases. The aqueous phase containing the LPS was care-
fully transferred to another tube. The extraction procedure
was repeated 2 more times and the aqueous phase from
each extraction pooled and dialyzed against distilled wa-
ter to remove the phenol. The extracted LPS and capsule
material were further purified on CsCl isopycnic density
gradients. CsCl (2.8 g) was added to 4.8 ml aqueous sam-
ples containing LPS. Gradients were established by ultra-
centrifugation (20,000 X g at 4°C for 48 hours). Fractions
(200 µ each) were collected from the bottom of the gra-
dient and the density of each fraction determined by re-
fractive index. The LPS containing fractions were iden-
tified by the limulus amoebocyte lysate reaction. Hot phe-
nol/water extracted E. coli LPS was purchased** and add-
ed directly to media to final concentrations without further
purification.
Measurement of Monocytic Secretion of PGE2 and
IL-lß
PGE2 concentrations in cell supernatants were assayed as
described previously by radioimmunoassay32 following
purification by organic solvent extraction and reverse
phase preparative chromatography.§§ IL-lß concentrations
5§Sep-pak, Millipore Corp., Milford, MA.
 J Periodontol
130 MEDIATOR RESPONSE AND PERIODONTAL DISEASE IN IDDM PATIENTS February 1997

Table 1A. Characteristics of the Systemically Healthy Subjects
(Patients are divided based upon the severity of their periodontal
disease. Periodontally healthy subjects display no clinical or radio-
graphic evidence of periodontal disease. AAP Type I and II group
represents individuals with gingivitis or early Periodontitis; AAP
Table IB. Characteristics of Diabetic Patient Population (Diabetics
are divided based upon the severity of their periodontal disease.
Group A represents individuals with gingivitis or early Periodontitis;
Group includes diabetics with moderate to severe periodontal dis-
ease.) .

Type III and IV group includes patients with moderate to severe

periodontal disease. The normal HbAlc level for systemically healthy Group A Group
subjects ranges from 5.0 to 7.5%.) (AAP Type (AAP Type
III) III-IV) Total
AAP Case Type 10 29 39
III III-IV Age (mean ± SEM) 37.7 ± 5.2 52.6 ± 2.6 48.8 ± 2.5
Age (range) 22-75 35-81 22-81
21 22 21
20.4 ± 0.05 50.2 ± 4.3 HbAIC
Age (mean ± SEM) 34.4 ± 4.5
(mean ± SEM) 9.92 ± 1.0 10.69 ± 0.36 10.49 ± 0.37
Age (range) 20-21 23-67 33-73 Caucasian/African-
Caucasian/African- 16/13 23/16
American 7/3
American 21/0 16/6 12/9 Female/male 8/2 17/12 25/14
Female/male 12/9 18/4 7/14
Probing depths
Probing depth (mean ± SEM) 2.71 ±0.11 3.33 ±0.11 3.17 ± 0.09
(mean ± SEM) 2.53 ± 0.14 3.31 ± 0.21
CAL
(mean ± SEM) 2.58 ± 0.10 4.27 ± 0.20 3.83 ± 0.19
% sites BOP 21.9 33.2 30.3 ± 2.95
in supernatants were determined by ELISA using rhlL-
lß as standard.1111 PGE2 and IL-lß concentrations in mono-
cyte supernatants were assayed in duplicate at each LPS A patients were diagnosed as having gingivitis and early
concentration. periodontal disease (AAP classification types I and II).
Group patients had moderate to severe periodontal dis-
Data Analysis ease (AAP classification types III and IV). AAP types I
Gingival crevicular fluid mediator levels were pooled to and II as well as types III and IV were pooled to form
form a patient mean and monocyte mediator values groups A and respectively, since the differences in the
pooled at each LPS concentration for each patient. Patient GCF-PGE2 and GCF-IL-lß levels between types I and II
values were pooled to form group means using a pooled and between types III and IV were very small and not
estimate of standard error correcting for non-equal vari- statistically significant. The mean age of the entire IDDM
ance. Differences in group means were tested using Mann
population was 48.8 years; however, the mean age of
Whitney U test or non-paired r-test for GCF mediator Group patients was significantly greater than that of
levels, mean age, HbAlc levels, and the clinical measure- Group A patients (52.6 ± 2.6 vs. 37.7 ± 5.2 years) at
ments. An alpha level of less than 0.05 was considered =
0.01. This is not surprising since the positive relation-
statistically significant. In all instances, the patient was ship between increasing severity of periodontal disease
the unit of observation, not the number of sites. Differ- and age is well established. However, age was not sig-
ences between groups in proportion to race or gender
nificantly associated with any of the other outcome vari-
were determined using the Fisher exact test. Differences ables. There were more Caucasians and more females in
in dose-response curves were determined using repeated this sample IDDM population. However, this difference
measures 2-way analysis of variance (ANOVA). If sig- in proportions is not statistically different between groups
nificant dose or group effects were seen by repeated mea- (Fisher exact test). Comparing mild periodontal disease
sures ANOVA, then non-paired i-test or the Mann Whit- to more severe forms, Group A to Group patients, there
ney U test was used to test for differences among groups is no significant difference in the mean HbA,c between
at specific doses. Since age was significantly different be- the 2 groups. As can be seen, most patients had relatively
tween diabetics and non-diabetics, we used age-adjusted high HbA,c levels, reflecting moderate to poor control.
mediator levels for statistical testing. Gender or race ad- Comparing Group A to Group B, there is no significant
justments were not performed since these were quite sim- difference in mean percent of sites which exhibit bleeding
ilar between comparison groups. on probing. As expected, there is a significant difference

(P < 0.05) in the clinical attachment level and probing

RESULTS
Tables 1A and IB summarize the biographical data as
well as the periodontal disease status of the 64 systemi-
cally healthy control subjects and the 39 diabetic patients,
respectively. Diabetics were divided into two groups (A
and B) on the basis of periodontal disease severity. Group
depths between the 2 groups, with those in the severe
Periodontitis group having more attachment loss.
GCF Mediators
The PGE2 and IL-lß levels present within the GCF
mean
are shown in Table 2. Patients are grouped into 5 cate-

gories based upon systemic and periodontal health. Non-

"Cistron Biotechnology, Pinebrook, NJ. diabetic patients free of periodontal disease have the low-
Volume 68
Number 2 SALVI, YALDA, COLLINS, ET AL. 131

Table 2. GCF-PGE2 and IL-lß Levels Compared Between Non-Diabetic and IDDM Patients (Dia-
betics have significantly higher levels of both mediators relative to non-diabetics. Within the diabetic
group, there are significantly higher levels of GCF-IL-lß in Group patients compared to Group A.
Non-diabetics with periodontal disease also had significantly higher GCF-PGE2 and IL-lß levels com-
pared to periodontally healthy non-diabetics.)
Periodontal Mean GCF-PGE, Mean GCF-IL-lß
Diagnosis ng/ml ng/ml
Systemically healthy
Health 21 20.5 ± 7.6 16.5 ± 9.3
I/II (gingivitis/mild Periodontitis) 22 43.3 ± 8.1 114.3 ±49.3
III/IV (moderate/severe Periodontitis) 21 42.9 ± 11.9 261.2 ± 70.9
Total I-IV Periodontitis patients 43 43.1 ± 7.05 186.0 ± 43.8
IDDM
I/II (Group A) gingivitis/mild Periodontitis) 9 183.2 ± 76.0 451.2 ± 57.8
III/IV (Group B) (moderate/severe Periodontitis) 25 319.1 ± 62.2 788.0 ± 85.8
Total I-IV Periodontitis patients 34 283.2 ± 9.7 698.9 ± 68.5

400 •IDDM Group A (AAP Type l-ll)

est GCF-PGE2 and GCF-IL-lß levels at 20.5 and 16.5
ng/ml, respectively. Systemically healthy individuals with [
k
IDDM Group (AAP Type lll-IV)
Non-diabetic Adult Periodontitis
gingivitis/mild Periodontitis show a significant elevation
in both mediators relative to healthy patients free of peri- 300
c
odontal infection. There is a 2.1 fold elevation in GCF-
111
PGE2 (20.5 vs. 43.3 ng/ml at < 0.05) and a dramatic o
6.9 fold increase in GCF-IL-lß at < 0.001. In non- . 200
o
diabetics, the GCF-PGE2 level in moderate/severe Perio- 4-«
>.
dontitis patients is essentially the same as the gingivitis/ o
o
c 100
mild Periodontitis group. However, the GCF-IL-lß levels o
are elevated in this more severe form of periodontal dis-
ease in non-diabetics (261.2 ± 70.9 ng/ml vs. 114.3 ±
49.3 ng/ml at < 0.05). In contrast, the IDDM patients
as a group display significantly elevated GCF levels of 0 ^ 0.001 ÖÖ1 oTÍ LO
both PGE2 and IL-lß. The elevation of GCF-PGE2 is 13.8 P. gingivalis LPS fag/ml)
fold over health and 6.6 fold over non-diabetic patients
with moderate to severe disease (P < 0.00001). The se- Figure 1. Monocyte PGE2 responses to increasing concentrations o/P.
gingivalis LPS shown in both groups of diabetics and non-diabetics with
verity of periodontal disease comparing the IDDM pa- adult Periodontitis. The PGE2 response is significantly lower in non-
tients to the moderate/severe Periodontitis non-diabetic diabetic patients relative to both groups of diabetics regardless of their
group was not significantly different (data not shown). periodontal status. There is a significant difference in the monocytic
However, as one examines the difference in GCF-PGE2 PGE, secretion between Group A and diabetics.
and IL-lß levels between Group A and Group IDDM
patients, there is a statistically significant increase in both for Group patients (P 0.01). The Group elevation
=

PGE2 and IL-lß levels in Group patients. Comparing in GCF-IL-lß relative to Group A is also statistically sig-
the Group A patients to the non-diabetic periodontal dis- nificant at 0.006. It is interesting to note that the
=

ease matched group, there is a significant 4.2 fold ele-
vation in GCF-PGE2 (P 0.003) and 3.9 fold increase
=
in GCF-IL-lß ( 0.0005). Thus, the diabetic state is
= dramatic increase in the levels of both GCF-PGE2 and
associated with a 4-fold increase in the GCF inflamma-
tory mediator level in patients with mild periodontal con-
ditions. The increase in Group IDDM patients relative
to the non-diabetic, periodontally-diseased matched pa-
tients is substantially greater. In IDDM patients there is a
7.4 fold increase in GCF-PGE2 levels and a 3.0 fold in-
crease in GCF-IL-lß, relative to non-diabetic, periodon-
tally-diseased matched individuals at 0.0002 and
=
< 0.0001, respectively. As one compares the GCF-PGE2
levels between Group A (183.2 ng/ml) and Group
(319.1 ng/ml), the GCF-PGE2 level is significantly higher
IL-lß/PGE2 ratio is 2.4 for both Group A and Group
patients. The data clearly suggest that IDDM results in a
IL-lß, irrespective of periodontal status. Furthermore,
within the diabetic population there is a significant 1.75
fold increase in both of these GCF inflammatory media-
tors in patients with more severe periodontal disease.
Monocyte Data
Figures 1 and 2 show the mean monocytic PGE2 and
IL-lß production and the standard error at each LPS con-
centration in response to various concentrations of P. gin-
givalis LPS. Figure 1 illustrates the mean dose-response
curves for Group A, Group and non-diabetic moderate
 J Periodontol
132 MEDIATOR RESPONSE AND PERIODONT AL DISEASE IN IDDM PATIENTS February 1997

value, but produce approximately 3-fold more PGE2 than

monocytes from Group A patients. The diabetic state ap-
pears to be associated with enhanced secretion of
an

monocytic PGE2 compared as non-diabetic

0 /
0.001 0.01 0.1 1.0
P. gingivalis LPS (jig/ml)
Figure 2. Monocyte IL-lß responses to increasing concentrations / .
gingivalis LPS shown for both groups of diabetics and non-diabetics
with adult Periodontitis. The IL-lß response is significantly higher in
diabetics relative to non-diabetics. There is no significant difference in
IL-lß secretion as a function of LPS concentration, comparing Group
A to Group patients.
to severe Periodontitis patients which serve as a Group
control. Dose responses were completed for 39 IDDM
patients and 21 non-diabetic adult Periodontitis patients.
Figure 1 shows the monocyte release of PGE2 in response
to P. gingivalis LPS pooling all patients within each
group at each LPS concentration. At zero LPS concentra-
tion the basal secretion of PGE2 is significantly higher
comparing either the pooled diabetic group (98.5 ± 1.94
ng/ml, data not shown) or Group (104.0 ng/ml) to the
non-diabetic group (57.2 ng/ml) at 0.0001. There is
=
to adult Per-
iodontitis patients. Within the IDDM patients there is a
3-fold elevation in monocytic PGE2 release in Group
patients as compared to Group A patients who have mild-
er periodontal disease (Table IB). This suggests that there
is heterogeneity in the monocytic response within the
IDDM patient population which is independent of the lev-
el of metabolic control.
As illustrated in Figure 2, the monocytic release of
IL-lß in IDDM patients is significantly elevated through-
out the LPS dose-response relative to the non-diabetic
patients. However, in contrast to the monocytic PGE2 re-
sponse, there is no significant difference in the IL-lß se-
cretion as a function of LPS concentration, comparing
Group A to Group patients. At zero LPS the basal se-
cretion of IL-lß is similar for Group A and Group
monocytes. At the lowest LPS dosage (0.003 µg/ml)
monocytes from Group patients appear to secrete great-
er amounts of IL-lß, but this difference is lost at con-
centrations above 0.03 µg/ml. Although there is a ten-
dency for Group monocytes to secrete more IL-lß at
various LPS concentrations, this difference was not sta-
tistically significant.
Monocytic dose-responses were also obtained using E.
coli LPS for a subset of 19 IDDM patients. The PGE2
and IL-lß levels at various LPS concentrations appear in
Table 3. The differences between Group A and Group

no difference in the basal secretion of PGE2 comparing

the two IDDM groups. In the non-diabetic patients the
dose response curve is relatively flat with an increase in
PGE2 secretion occurring as a sharp increase between 0.3
and 3.0 µg/ml P. gingivalis LPS. There is a statistically
significant difference in the level of PGE2 released as a
function of LPS concentration among all three groups, as
determined by repeated measures 2-way ANOVA. Mono-
cytes from Group A patients show a significant increase
in PGE2 secretion at low dosages of LPS (0.003 µg/ml)
and reach a maximum secretion of 169.0 ng PGE2/ml.
Group monocytes also have a similar half-maximum
monocytic PGE2 and IL-lß responses closely parallel that
observed with P. gingivalis LPS. Group monocytes se-
crete significantly more PGE2 and IL-lß, as compared to
monocytes from Group A patients. As cited above, the
elevation in Group IL-lß secretion in response to P.
gingivalis LPS did not achieve statistical significance
compared to Group A. However, using E. coli LPS, the
differences in IL-lß secretion were statistically significant
between these two groups (P < 0.05). There were no
significant differences in the potency of the two LPS
preparations, as both stimulated the secretion of similar
amounts of PGE2 and IL-lß in both groups of patients.

Table 3. Monocyte Dose-Responses to Various E. coli LPS Concentrations in a Subset of IDDM

Patients (n = 19) (ng/ml; mean ± SEM)

Group A Group B*
E. coli LPS
Concentra-
(AAP Type I II) (AAP Type III-IV)
tion ^g/ml) _

PGE, IL-lß PGE, IL-lß

0.0 51.0 ± 32.0 0.9 ± 0.27 65.0 ± 24.0 8.6 ± 3.7
0.03 93.0 ± 65.0 24.9 ± 14.8 278.0 ± 65.0 70.4 ± 13.7
0.3 179.0 ± 102.0 48.4 ± 7.4 223.0 ± 50.0 125.1 ± 42.2
3.0 128.0 ± 46.0 62.2 ± 7.9 314.0 + 72.0 104.6 ± 28.1

*Group patients secrete significantly more PGE2 and IL-lß compared to Group A patients (P < 0.05).
Volume 68
Number 2
DISCUSSION
Our data indicate that the levels of PGE2 and IL-lß in
GCF of IDDM patients are extraordinarily high. These
levels exceed all previous reports of PGE2 and IL-lß in
GCF of non-diabetic patients30 31 and when compared to
more aggressive forms of periodontal disease such as ear-
ly-onset and refractory disease.35 The observation that
IDDM patients with relatively mild periodontal problems
had excessively high levels of inflammatory mediators
was unanticipated. These high levels have previously
been shown to be highly correlated with active progres-
sive periodontal lesions in non-diabetic healthy individ-
uals30 and were considered to be inconsistent with peri-
odontal health. This suggests that in the diabetic there
must be compensatory anti-inflammatory and tissue pro-
tective responses which are preventing rampant destruc-
tion and can apparently maintain the patient in a stable,
albeit upregulated, state of equilibrium. The increase in
collagen cross-links which occur in diabetics may render
the collagen less susceptible to proteolysis. However, the
inflammatory response with regard to IL-lß and PGE2 in
the IDDM patients is exaggerated and may therefore be
a possible explanation for the increased prevalence of se-
vere periodontal disease in this population. We have also
recently reported an elevated monocytic secretion of
TNFa in an IDDM population compared to a systemically
healthy adult Periodontitis group.36
The underlying cause of this monocytic hypersecretory
inflammatory trait is presently unknown. However, the
microbial flora would not seem to be a factor in this re-
sponse, since the periodontal pathogens in the diabetic
oral flora are not unique to the diabetics and are essen-
tially equivalent to that found in non-diabetic adult Per-
iodontitis patients.37-38 None of the other variables mea-
sured in this study, such as HbAlc, age, or race was found
to be correlated with the amount of IL-lß or PGE2 se-
cretion in either diabetics or systemically healthy Perio-
dontitis subjects.
Furthermore, there was also no significant correlation
between diabetic metabolic control (HbAlc) and the se-
verity of periodontal disease. This observation is in agree-
ment with other studies which have failed to find a rela-
tionship between blood glucose levels and periodontal
disease severity.39-41 Some of the previous reports have
observed more severe periodontal disease in the black
population;42 however, there were no significant differ-
ences in any of the outcome variables when subjects were
compared based upon race in this sample population of
IDDM patients. Metabolic abnormalities may be an im-
portant contributing factor to the observed hyperinflam-
matory monocytic trait (M0+). However, we did not see
an association between the level of monocytic mediator
secretion and degree of metabolic control as reflected by
the level of glycosylated hemoglobin (HbAIC). Regretta-
bly, data were not collected with respect to onset, dura-
SALVI, YALDA, COLLINS, ET AL. 133
tion, and overall history of metabolic control. Since there
was a statistically significant difference in age between
Group A and diabetics, this may not exclude the pos-
sibility that in Group patients there has been a longer
exposure to a hyperglycémie state. This longer exposure
may contribute to a prolonged AGE accumulation and to
a hyperresponsive monocytic phenotype. It appears likely
that chronic hyperglycemia would lead to an accelerated
AGE accumulation and upregulation of the monocyte
phenotype in diabetics as has been previously demonstrat-
ed for IL-1 and TNFa by Vlassara et al.43
A new finding in this study is the observation that those
IDDM patients with the highest inflammatory response
were in Group patients who had more moderate/severe
periodontal disease. We have previously reported a sim-
ilar, pattern in TNFa monocytic secretion in IDDM pa-
tients with or without severe periodontal diseases.36
Therefore, the exaggerated monocytic response may be a
phenotypic marker for susceptibility to Gram-negative in-
fections in diabetics. Although this trait has not been
characterized in the general population as increasing risk
for diabetes, it may be a marker for susceptibility to
Gram-negative infections such as periodontal disease. It
appears likely that the high GCF levels of PGE2 and
IL-lß in the IDDM patients are a consequence of a sys-
temic monocytic response trait and that the presence of
severe periodontal disease is associated with the more ex-
cessive expression of this trait. Other investigators have
suggested that in IDDM patients pleomorphic variations
in the promoter region of TNFa and IL-lß genes can be
associated with low normal or hypersecretory monocytic
phenotypes.23-44 This raises the question as to whether
there is an underlying cytokine genotype which is asso-
ciated with the M0+ trait and disease susceptibility.
Previous investigations regarding monocytic IL-lß se-
cretion in IDDM patients have been equivocal. It appears
to be clearly elevated in newly diagnosed IDDM children,
but more variable in adults.24-45-46 However, none of these
experiments has controlled for IL-lß gene polymorphism
or the presence of periodontal infection. Thus, in the con-
text of our present experiments, these variables may fur-
ther explain the elevated IL-ß secretion seen in our IDDM
patient population. Similarly, elevated circulating levels
of interferon y (IFN- ) have been reported in IDDM pa-
tients.47-49 IFN--y can clearly upregulate monocytic PGE2
and IL-lß secretion in response to LPS. It is possible that
periodontal infection may upregulate IFN--y levels in
these patients and contribute to the exaggerated mono-
cytic and GCF inflammatory responses. Furthermore, el-
evated GCF levels of IL-la may, in part, be due to in-
creased plasma levels of this cytokine as reported by Hus-
sain et al.50 However, recent reports by Prabhu et al.51
have failed to demonstrate significant elevations of plas-
ma levels of several cytokines in adult Periodontitis pa-
tients.

134 MEDIATOR RESPONSE AND PERIODONTAL DISEASE IN IDDM PATIENTS
Further experiments on the relationships among the na-
ture of the inflammatory response, specific gene poly-
morphism and circulating cytokine assessments should
help clarify the role of these observed exaggerated in-
flammatory responses and periodontal disease in the di-
abetic.
Acknowledgments
This work was supported by the Swiss National Science
Foundation, by NIH grants HD26652, DE00325 and
DE10519, and by the Swiss Society of Periodontology.
REFERENCES
1. Katz PR Wirthlin MR Jr, Szpunar SM, Selby JV, Sepe SJ, Showstack
JA. Epidemiology and prevention of periodontal disease in individ-
uals with diabetes. Diabetes Care 1991;14:375-385.
2. Yalda B, Offenbacher S, Collins JG. Diabetes as a modifier of peri-
odontal disease expression. Periodontol 2000 1994;6:37-49.
3. Slots J. Bacterial specificity in adult Periodontitis. A summary of
recent work. J Clin Periodontol 1986;13:912-917.
4. Haber J, Wattes J, Crowley R. Assessment of diabetes as a risk
factor for Periodontitis. J Dent Res 1991;70(Spec. Issue):414(Abstr.
1187).
5. Grossi SG, Zambón JJ, Norderyd CM, et al. Microbiological risk
indicators for periodontal disease, i Dent Res 1993;72(Spec. Issue):
206(Abstr. 818).
6. Broadley KN, Aquino AM, Hicks B, et al. The diabetic rat as an
impaired wound healing model: stimulatory effects of transforming
growth factor-beta and basic fibroblast growth factor. Biotechnol-
Ther 1989-90; 1:55-68.
7. Hove KA, Stallard RE. Diabetes and the periodontal patient. J Per-
iodontol 1970;41:713-718.
8.
Lamey PJ, Darwazeh AMG, Fisher BM. Oral disorders associated
with diabetes mellitus. Diabet Med 1992;9:410-416.
9. Cutler CW, Eke R Arnold RR, Van Dyke TE. Defective neutrophil
function in an insulin-dependent diabetic patient. A case report. J
Periodontol 1991;62:394-401.
10. Manoucher-Pour M, Spagnuolo PJ, Rodman HM, Bissada NF. Com-
parison of neutrophil chemotactic response in diabetic patients with
mild and severe periodontal disease. J Periodontol 1981;52:410-
414.
11. Van Dyke TE, Horoszewicz HU, Cianciola LJ, Genco RJ. Neutro-
phil Chemotaxis dysfunction in human Periodontitis. Infect Immun
1980;27:124-132.
12. Lavine WS, Maderaso EG, Stolman J, et al. Impaired neutrophil
Chemotaxis in patients with juvenile and rapidly progressive Perio-
dontitis, i Periodont Res 1979;14:10-19.
13. Schmidt AM, Hori O, Brett J, Yan SD, Wautier JL, Stern D. Cellular
receptors for advanced glycation end products. Implications for in-
duction of oxidant stress and cellular dysfunction in the pathogenesis
of vascular lesions. Arterioscler Thromb 1994;14:1521-1528.
14. Vlassara H. Recent progress on the biologic and clinical significance
of advanced glycosylation end products. J Lab Clin Med 1994;124:
19-30.
15. Brownlee M. Glycation and diabetic complications. Diabetes
1994;43:836-841.
16. Neeper M, Schmidt AM, Brett J, et al. Cloning and expression of
RAGE: a cell surface receptor for advanced glycation end products
of proteins. J Biol Chem 1992;267:14998-15004.
17. Yan SD, Schmidt AM, Anderson GM, et al. Enhanced cellular ox-
idant stress by the interaction of advanced glycation end products
with their receptors/binding proteins. J Biol Chem 1994;269:9889-
'9897.
J Periodontol
February 1997
18. Dyer D, Dunn J, Thorpe S, et al. Accumulation of Maillard reaction
products in skin collagen in diabetes and aging. J Clin Invest
1993;91:2463-2469.
19. McCance D, Dyer D, Dunn J, et al. Maillard reaction products and
their relation to complications in insulin-dependent diabetes melli-
tus. J Clin Invest 1993;91:2470-2478.
20. Yu Z, Ramamurthy NS, Leung M, Chang KM, McNamara TF, Go-
lub LM. Chemically-modified tetracycline normalizes collagen me-
tabolism in diabetic rats: a dose-response study. J Periodont Res
1993;28:420-428.
21. M0lvig J. A model of the pathogenesis of insulin-dependent diabetes
mellitus. Dan Med Bull 1992;39:509-541.
22. Nerup J, Mandrup-Poulsen T, M0lvig J. The HLA-IDDM associa-
tion: implications for etiology and pathogenesis of IDDM. Diabetes
Metab Rev 1987;3:779-802.
23. Pociot F, M0lvig J, Wogensen L, Worsaae H, Nerup J. A Taq 1
polymorphism in the human interleukin-1 ß (IL-lß) gene correlates
with IL-lß secretion in vitro. Eur J Clin Invest 1991;22:396-402.
24. Santamaria R Gehrz RC, Bryan MK, Barbosa JJ. Involvement of
class II MHC molecules in the LPS-induction of IL-l/TNF secre-
tions by human monocytes. J Immunol 1989;143:913-922.
25. M0lvig J, Baek L, Christensen R et al. Endotoxin-stimulated human
monocyte secretion of interleukin 1, tumor necrosis factor alpha and
Prostaglandin E, shows stable interindividual differences. Scand J
Immunol 1988;27:706-716.
26. Todd JA. Genetic control of autoimmunity in type 1 diabetes. Im-
munol Today 1990;11:122-129.
27. Leslie RDG, Lazarus NR, Vergani D. Etiology of insulin dependent
diabetes mellitus. Br Med Bull 1989;45:58-72.
28. McGuire W, Hill AVS, Allsopp CEM, Greenwood BM, Kwiat-
kowski D. Variation in the TNFct promoter region associated with
susceptibility to cerebral malaria. Nature 1994;371:508-511.
29. Cabrera M, Shaw MA, Sharpies C, et al. Polymorphism in TNF
genes associated with mucocutaneous leishmaniasis. J Exp Med
1995;182:1259-1264.
30. Offenbacher S, Odie BM, Van Dyke TE. The use of crevicular fluid
Prostaglandin E, levels as a predictor of periodontal attachment loss.
J Periodont Res 1986;21:101-112.
31. Massada MR Persson R, Kenney JS, Lee SW, Page RC, Allison AC.
Measurement of interleukin-la and -lß in gingival crevicular fluid:
Implications for the pathogenesis of periodontal disease. J Periodont
Res 1990;24:156-163.
32. Offenbacher S, Fair DH, Goodson JM. Measurements of Prosta-
glandin E, in crevicular fluid. J Clin Periodontol 1981;8:359-367.
33. Brunk CF, Jones KC, James T. Assay for nanogram quantities of
NDA in cellular homogenates. Analytic Biochem 1979;92:497-500.
34. Kasper DL. Chemical and biological characterization of the lipo-
polysaccharide of Bacteroides fragilis subspecies fragilis. J Infect
Dis 1976;134:59-66.
35. Offenbacher S, Collins JG, Yalda B, Haradon G. Role of Prosta-
glandins in high risk Periodontitis patients. In: Genco R, Hamada S,
Lehner T, McGhee J, Mergenhagen S, eds. Molecular Pathogenesis
of Periodontal Disease. Washington DC: ASM Press; 1994;203-
214.
36. Salvi GE, Collins JG, Yalda B, Arnold RR, Lang NR Offenbacher
S. Monocytic TNFa secretion patterns in IDDM patients with peri-
odontal diseases. J Clin Periodontol 1997 (in press).
37. Mandel] RL, DiRienzo J, Kent R, Joshipura K, Haber J. Microbi-
ology of healthy and diseased periodontal sites in poorly-controlled
insulin-dependent diabetics. J Periodontol 1992;63:274-279.
38. Zambón JJ, Reynolds H, Fisher JG, Shlossman M, Dunford R, Gen-
co RJ. Microbiological and immunological studies of adult Perio-
dontitis in patients with non-insulin-dependent diabetes mellitus. J
Periodontol 1988;59:23-31.
39. Barnett ML, Baker RL, Yancey JM, MacMillan DR, Kotoyan M.
Absence of Periodontitis in a population of insulin-dependent dia-
betes mellitus patients. J Periodontol 1984;55:402-405.
Volume 68
Number 2
40. Sastrowijoto SH, van der Veiden U, van Steenbergen TJ, et al. Im-
proved metabolic control, clinical periodontal status and subgingival
microbiology in insulin-dependent diabetes mellitus. A prospective
study. J Clin Periodontol 1990;17:233-242.
41. De Pommereau V, Dargent PC, Robert JJ, Brion M. Periodontal
status in insulin-dependent diabetic adolescents. J Clin Periodontol
1992;19:628-632.
42. Burmeister JA, Best AM, Palcanis KG, Caine FA, Ranney RR. Lo-
calized juvenile Periodontitis and generalized severe Periodontitis.
Clinical findings. J Clin Periodontol 1981;11:181-192.
43. Vlassara H, Brownlee M, Manogue K, Dinarello CA, Pasagian A.
Cachectin/TNF and IL-1 induced by glucose-modified proteins:role
in normal tissue remodeling. Science 1988;240:1546-1548.
44. Pociot F, M0lvig J, Wogensen L, et al. A tumour necrosis factor beta
gene polymorphism in relation to monokine secretion and insulin-
dependent diabetes mellitus. Scand J Immunol 1991;33:37-49.
45. Luger A, Schernthaner G, Urbanski A, Luger TA. Cytokine pro-
duction in patients with newly diagnosed insulin-dependent (type 1)
diabetes mellitus. Eur J Clin Invest 1988;18:233-236.
46. M0lvig J, Pociot F, Baek L, et al. Monocyte function in IDDM
patients and healthy, individuals. Scand J Immunol 1990;32:297-
311.
SALVI, YALDA, COLLINS, ET AL. 135
47. Tovo PA, Cerutti F, Palomba E, Salomone C, Pugliese A. Evidence
of circulating interferon-7 in newly diagnosed diabetic children.
Ada Ped Scand 1984;73:785-788.
48. Scuderi F, Muscettola M, DiLeo MAS, et al. High levels of serum
interferon-gamma do not decrease with cyclosporin treatment in type
1 (insulin-dependent) diabetes. Diabetologia 1989;32:540A(Abstr.
454).
49. Lorini R, Montagna D, Lanfranchi A. Alterations of in vitro inter-
leukin 1 and 2 in diabetic children. Eur J Pediatr 1989; 148:732-
734.
50. Hussain MJ, Peakman M, Leslie RDG, Viberti GC, Watkins PJ,
Vergani D. Circulating levels of interleukin-la, tumour necrosis fac-
tor-a and interferon- are high in newly-diagnosed type 1 diabetes.
Diabetologia 1990;33(Suppl.):A25(Abstr. 55).
51. Prabhu A, Michalowicz BS, Mathur A. Detection of local and sys-
temic cytokines in adult Periodontitis. J Periodontol 1996;67:515-
522.
Send reprint requests to: Dr. Giovanni E. Salvi, University of North
Carolina at Chapel Hill, School of Dentistry, Dental Research Center,
Department of Periodontics, CB #7455, Room 222, Chapel Hill, NC
27599-7455.
Accepted for publication July 2, 1996.