Reactive pocket epithelium in untreatedchronic periodontal disease: possiblederivation from developmental remnants ofthe enamel organ and root sheath

Neil Hunter1

Brent Nicholls1

Mukesh Srivastava1

Cheryl C. Chapple1

Hans F. Zoellner2

John R. Gibbins3

1Institute of Dental Research, Sydney,2Faculty of Dentistry and

3Department of Pathology, University ofSydney, Sydney, Australia

Correspondence to:Neil HunterInstitute of Dental Research, United DentalHospital, 2 Chalmers Street, Surry Hils 2010,NSW, Australia

Accepted for publication April 3, 2000

Copyright C Munksgaard 2001J Oral Pathol Med . ISSN 0904-2512

Printed in Denmark . All rights reserved

178

Abstract: The pathological lining epithelium of destructive peri-odontitis was studied by analysis of the expression of intermedi-ate filament proteins in biopsies of untreated advanced peri-odontitis. The cytokeratin (CK) pair 8/18 characteristic of simpleepithelia was expressed consistently in a distribution patternconfined to the reactive pocket epithelium. The pattern of CK8/18 expression was complex with two broad presentations evi-dent. In two-thirds of the advanced disease biopsies, the entirepathological lining epithelium was strongly reactive for both CK8and CK18. In the remainder, the more superficial lining epithel-ium was mixed with foci of reactive and unreactive cells, with thedeeper epithelium uniformly reactive. Only occasional highlylocalised reactivity for the simple keratins (CK8/18) was found inthe lining epithelia of biopsies from minimally inflamed peri-odontal tissues. The pathological lining epithelium of advancedperiodontitis was further characterised by the co-expression inbasal layers of CK14, and of CK13 but not CK4, which are charac-teristic of suprabasal layers of stratified squamous epithelia.Cytokeratin 17, a marker of high turnover and migrating epithelialcells was extremely variable with no clear association betweenexpression pattern and location of the epithelium or disease sta-tus. There was no reactivity for CK10/11 typical of cornifyingcells nor of vimentin, the characteristic intermediate filament ofmesenchymal cells. The intermediate filament protein profile ofthe reactive lining epithelium was indistinguishable from the re-active epithelium present in three of five biopsies of periapicalgranulomas containing hyperplastic epithelium from activation ofthe developmental remnants of Hertwig’s sheath, known as thecell rests of Malassez. The data reported are compatible with acontribution by remnants of developmental epithelium, includingthe reduced enamel epithelium and the cell rests of Malassez,to the reactive lining epithelium of the subgingival pocket in thepathogenesis of chronic periodontitis.

Key words: cytokeratins; developmental remnants; peri-odontitis; reactive epithelium

J Oral Pathol Med 2001: 30: 178–86

Accumulation of microbial plaque in the vicinity of the gingival

sulcus induces an inflammatory response that is confined to the

Pocket epithelium in chronic periodontitis

marginal gingiva and in which the epithelial attachment to the tooth

remains intact. In some gingivitis sites, breakdown of the epithelial

attachment and apical re-positioning along the tooth root creates a

cleft or pocket characteristic of periodontitis. This is assumed to

occur by migration of the cells of the attachment, analogous to the

behaviour of epithelium in wound healing (1). Further, synthesis of

metalloproteinase XIII by the lining epithelium has been described

(2), with implication for the destructive potential of this tissue. Al-

tered barrier function of the pathological lining epithelium could

critically influence the ingress of microbial products into the tissues

and, therefore, regulate the pathogenesis of destructive peri-

odontitis.

The prevailing understanding of the nature of the epithelial

attachment is that it arises by a fusion of the reduced enamel epi-

thelium (REE), adherent to the erupting tooth, with the epithelium

of the oral mucosa during the process of eruption (1). It has been

proposed that the turnover associated with maturation could pro-

duce a gradual replacement of the reduced enamel epithelium com-

ponent of the attachment, so that the epithelial attachment is pro-

gressively replaced by the oral epithelium to become a product of

the oral epithelium (3). Developmental remnants derived from Hert-

wig’s root sheath epithelium persist in the periodontal ligament as

an attenuated network or as isolated clusters of epithelial cells,

termed epithelial cell rests of Malassez, after development of the

root is completed (4). The relation of this network to the gingival

attachment is controversial, but some studies have indicated conti-

nuity between the residual clusters in the periodontal ligament and

the epithelial attachment in the gingival sulcus (5–7).

The cell rests of Malassez within the periodontal ligament have

been implicated in a number of orodental pathologies, including

dental cysts arising from periapical granulomas of teeth (8). Spouge

(9) postulated that these cell rests could also contribute to the reac-

tive epithelial lining of the periodontal pocket in periodontitis. Test-

ing this hypothesis requires techniques capable of discriminating

the epithelial phenotype and, in this context, the expression of the

cytokeratin intermediate filament proteins provides a powerful tool.

Cytokeratins are a group of intermediate filament proteins that

form part of the intracellular cytoskeleton in epithelial cells. These

proteins are subdivided into the acidic type I keratins (CK10 to

CK20) and the neutral-basic type II keratins (CK1 to CK9) that form

heterodimers characteristic of the tissue type and the position of

cells within the strata of stratified squamous epithelia (10). CK8/18

form a primitive cytokeratin pair characteristic of simple epithelia

in the adult and are expressed in many epithelia during embryonic

development. They are not a feature of stratified squamous epithelia

in the adult, and it is only in severe dysplasia and neoplasia that

the keratinocytes are found to express these proteins (11, 12). The

179J Oral Pathol Med 30: 178–86

CK5/14 heteropolymer pair is characteristic of the basal layers of

stratified epithelia, while CK4/13 is characteristic of suprabasal cells

and CKs 1, 2, 10, 11 are typical of cornified envelopes (10). Cytoker-

atins 6, 16 and 17 are characteristic of high turnover or migrating

epithelium (13). While the analysis of cytokeratin expression in oral

epithelia related to neoplasia has been a subject of considerable

activity, there have been relatively few attempts to study the pheno-

type of the lining epithelium of the pocket pathognomonic of

chronic periodontitis. Furthermore, in the majority of studies of hu-

man periodontal disease, tissue specimens have been obtained from

lesions that have been treated therapeutically in some way, such as

scaling or root planing. Patterns of change detected in tissues fol-

lowing such therapeutic interventions may not reflect the natural

history of the disease process accurately, and the present study was

designed specifically to address this problem by ensuring that all

tissues examined were removed prior to initiation of any treatment.

Material and methods

Tissues

Gingival biopsies were obtained with informed consent from adults

selected from patients attending the United Dental Hospital of Syd-

ney for tooth extractions. None of these patients had received peri-

odontal treatment of any kind and no relevant pre-existing medical

or drug histories, including smoking, were cited. Full depth biopsies

were obtained from the buccal aspect of premolars without restora-

tions. Biopsies from 40 patients yielded sections that were satisfac-

tory in relation to tissue preservation and orientation. Minimally

inflamed tissues had a probing depth of the gingival sulcus of 3

mm or less and were considered to be clinically healthy, but all of

the biopsies showed some evidence of leukocyte infiltration. Sixteen

biopsies were obtained from eight women and eight men, with an

average age of 52 years, (range 30 to 72 years) and probing depths

#3 mm with no bleeding on probing. One biopsy from a man par-

tially involved the interdental region and one biopsy was taken from

a site with gingival recession equivalent to 3 mm of loss of attach-

ment to the tooth. Twenty-four biopsies of advanced periodontitis

were obtained from seven women and seventeen men, with an aver-

age age of 59 years (range 49–69 years) and probing depths 6–9

mm. One biopsy from a woman and three from men partially in-

volved the interdental region and four biopsies from men were taken

from sites with 4–8 mm of recession.

Histologically, all tissues in the advanced periodontitis group

showed significant pathology and all showed extensive accumu-

lations of plasma cells in foci, as described previously (14). In ad-

Hunter et al.

dition, five periapical granulomas were obtained from five adult

patients. These chronic inflammatory lesions resulted from exten-

sion of pulpal infection as a consequence of dental caries. Histolog-

ically, these lesions frequently contain activated epithelia arising

from developmental remnants of Hertwig’s root sheath epithelium

(15). All five lesions contained proliferating epithelium associated

with an intense infiltration, predominantly of plasma cells.

Evaluation

For each biopsy, the tissue was trimmed until a representative block

face was achieved and then every tenth section was stained with

haematoxylin and eosin (H-E) to provide a reference library. The

parameters were studied in at least one bracket of sequential sec-

tions and, where possible, two brackets separated as widely as

possible. Sections were coded and examined blind by a single inves-

tigator with frequent reference to a standard. The distribution of

reacting cells was mapped in relation to an intensity of staining

score of 1 to 4 and significance of differences analysed by a Mann-

Whitney U test. The general structural features were recorded to-

gether with the distribution, nature and intensity of the inflamma-

tory infiltrate. The epithelium was divided into three regions for

minimally inflamed tissues, comprising oral epithelium, gingival

crest and gingival sulcus and attachment. A fourth region compris-

ing the reactive pocket epithelium was present in sections of lesions

of advanced periodontitis.

Immunolabelling

Tissues were snap-frozen in isopentane cooled in liquid nitrogen,

and 6 mm sections were subsequently prepared at a cryostat tem-

perature of ª12æC to ª20æC. Slides with sequential sections were

sealed in slide boxes and stored at ª70æC until required.

Sections were fixed in ice-cold methanol for 3 min, washed in

Dulbecco phosphate-buffered saline (PBS) and then blocked for 1 h

in 20% horse serum in PBS. Murine monoclonal antibodies were

obtained from Zymed Laboratories Inc. (San Francisco, CA, USA)

[anti-cytokeratin (CK) 8-clone DE-K13 and anti-CK18-clone DC-10],

Sigma BioSciences (St. Louis, MO, USA) (anti CK-18-clone CY-90,

anti-CK14-clone CKB1, anti-CK13-clone KS-1A3, anti-CK 4-clone

6B10, anti-CK17-clone CK-E3, anti-CK10,11-clone K8.60) and Dako

Australia (Sydney, Australia) (anti-vimentin-clone Vim 3B4). The

primary antibodies were diluted in 10% foetal calf serum in PBS

and incubated on the sections for 1 h at room temperature in a

humid chamber. Slides were washed five times for 10 min each in

PBS prior to application of the second-stage antibody.

Goat anti-mouse antibody (Dako P447) conjugated with horse-

180 J Oral Pathol Med 30: 178–86

radish peroxidase was incubated in a humid chamber at room tem-

perature for 1 h, followed by five washes of 10 min in PBS, and

subsequently detected and visualised with diaminobenzidine (metal-

enhanced reagent from Pierce Ltd) before counterstaining with

haematoxylin.

Fig. 1. Outline diagram showing predominant cytokeratin profiles in theepithelia of minimally inflamed and advanced periodontitis specimens in re-lation to structural changes. OE5oral epithelium, GC5gingival crest, S/A5

sulcus/attachment zone, PE5pocket epithelium.

Pocket epithelium in chronic periodontitis

Fig. 2. Patterns of immunolabelling for cytokeratins. a-d) Immunolabelling for cytokeratin 8. a) Pocket wall of advanced periodontitis lesion showinguniform staining of the reactive pocket epithelium (arrow). b) Pocket wall of another lesion showing focal staining in the coronal aspect of the reactivepocket epithelium. c) Reactive epithelium associated with a periapical granuloma showing strong reaction for CK8. d) Attachment zone of a biopsy ofminimally inflamed gingival tissue sample showing scattered reactivity confined to basal epithelial cells (arrow). e) Pocket epithelium (PE) showing reactivityfor CK17 (arrow) and underlying inflammatory infiltrate (II). f) As for e, but showing pattern of staining for CK4.

Endogenous peroxidase labelling was blocked by pre-incubating

slides in 3% H2O2 for 1 min. Minimally inflamed specimens were

prepared and treated alongside the advanced disease tissues. Nega-

181J Oral Pathol Med 30: 178–86

tive controls were incubated without the primary antibody. Ad-

ditional controls consisted of replacement of the primary antibody

with an isotype-matched irrelevant murine antibody (Dako).

Hunter et al.

Results

The structural changes characteristic of the developing lesion of

chronic periodontitis are shown diagrammatically in Fig. 1. The

minimally inflamed gingival biopsies obtained from 16 subjects

demonstrated similar histological features. The oral and sulcular

epithelium were keratinised, with a broad spinous layer and wide,

regular rete-pegs extending into the underlying connective tissue.

The epithelium was supported by dense fibrous connective tissue,

consisting of bundles of collagen fibres, fibroblasts, blood vessels,

and occasional inflammatory cells. By contrast, tissue biopsies from

patients with advanced periodontal disease demonstrated a sulcular

lining epithelium with tortuous strands permeating the oedematous

subjacent tissue. In some areas, the epithelium was only a few cells

wide. Beneath this epithelium, large numbers of chronic inflamma-

tory cells had replaced the dense fibrous connective tissue of the

periodontium. The predominant cell type was the plasma cell. Nu-

merous blood vessels were prominent. In addition, the original sul-

cular epithelium was oedematous, while the epithelium on the gingi-

val (oral) surface was frequently acanthotic.

Evaluation of antibodies

The antibody to CK8 gave intense specific staining of glandular

epithelial cells in frozen sections of pancreas. It also produced speci-

fic staining of epithelial cells in three periapical granulomas. It did

not stain stratified squamous epithelium. The antibody to CK18

from Zymed behaved identically but the anti-CK18 from Sigma at

the recommended dilution produced staining of the spinous layer

of stratified squamous epithelium. Selective, specific staining was

produced by extensive dilution of this antibody preparation. The

other antibodies used in this study produced the expected pattern

of reactivity in control gingival tissue. In each case the protocol was

optimised for intense, specific staining (see Fig. 2). In each case

studied, examination of a second bracket of sections produced a

very similar result. Parallel controls treated with irrelevant anti-

bodies of the same isotype were completely clear.

Staining patterns

The predominant CK profiles are illustrated diagrammatically in

Fig. 1.

Minimally inflamed tissues

Reactivity for CK8/18 was observed only in scattered basal cells in

some biopsies and in occasional foci of superficial cells in the region

182 J Oral Pathol Med 30: 178–86

of the epithelial attachment (Table 1 and Fig. 2). In contrast, CK14

and CK13 were distributed extensively throughout all regions, but

not all strata of the epithelial lining. Cytokeratins 4 and 17 were

more localised in highly individual patterns (Fig 2). Cytokeratins

10/11 were expressed in suprabasal layers of the oral epithelium in

relation to terminal differentiation. Vimentin was expressed exten-

sively by connective tissue cells but never by epithelial cells.

Advanced periodontitis

There was a pronounced strong reactivity of CK8 confined to the

pocket epithelium in 24/24 biopsies studied and for CK18 in 21/22

Fig. 3. Staining intensity patterns for cytokeratins 8 and 18 in the epithelialattachment of minimally inflamed gingival tissues and the reactive pocketepithelium of chronic periodontitis. A) Cytokeratin 8, showing profiles forminimally inflamed gingival tissues (open bars) and periodontitis (filledbars). B) Cytokeratin 18, showing profiles for minimally inflamed gingivaltissues (open bars) and periodontitis (filled bars).

Pocket epithelium in chronic periodontitis

Tab

le1.

Patt

erns

ofex

pres

sion

ofcy

toke

ratin

prot

eins

ingi

ngiv

alep

ithel

ialc

ells

and

reac

tive

epith

eliu

m

CK

8C

K18

CK

14

CK

13

CK

4C

K1

7C

K1

0/1

1

Bio

psie

sR

eact

ivity

OE

GC

S/A

PEO

EG

CS

/APE

OE

GC

S/A

PEO

EG

CS

/APE

OE

GC

S/A

PEO

EG

CS

/APE

OE

GC

S/A

PE

Non

e1

61

61

016

16

12

22

47

55

22

13

Min

imal

lyTr

ace

64

43

27

66

33

1in

flam

edB

asal

13

12

13

15

14

16

(n5

16)

Sup

raba

sal

16

16

16

14

15

16

43

42

18

82

Foca

l1

67

64

53

3Al

lcel

ls1

110

12

12

Non

e2

42

42

41

24

24

24

11

37

11

10

65

15

10

23

24

Adva

nced

Trac

e3

51

01

31

17

61

87

7pe

riodo

ntiti

sB

asal

713

714

54

12

(n5

24)

Sup

raba

sal

112

15

91

18

96

14

19

73

73

Foca

l1

11

14

38

12

45

41

Allc

ells

22

23

12

915

22

514

17

23

11

Non

e2

23

5Pe

riapi

cal

Trac

egr

anul

oma

Bas

al(n

55)

Sup

raba

sal

Foca

l2

5Al

lcel

ls3

35

5

OE5

oral

epith

elia

lasp

ect;

GC

5gi

ngiv

alcr

est;

S/A

5su

lcus

/att

achm

ent

zone

;PE

5po

cket

epith

eliu

mfo

rad

vanc

edpe

riodo

ntiti

sbi

opsi

es.

183J Oral Pathol Med 30: 178–86

biopsies studied (Table 1 and Fig. 2). Both the extent and the inten-

sity of reaction for CK8/CK18 were enhanced in advanced lesions.

The Mann-Whitney U test was used to determine the statistical

significance of differences in the relative distribution of labelling

intensity for both CK8 and CK18 between minimally inflamed and

periodontitis specimens. There was significantly less CK8 and CK18

labelling in minimally inflamed specimens as compared with peri-

odontitis specimens (P,0.001), as shown in Fig. 3. For both

markers, in relation to the apical-coronal dimension, 14 biopsies

studied showed uniform strong staining of all layers throughout the

length, whereas in the remainder the coronal aspect was more vari-

able (Fig. 2). A similar pattern of expression was noted for CKs 14

and 13 whereas, by contrast, the majority of biopsies showed only

trace reactivity in the pocket epithelium for CKs 4, 17 and 10/11.

This general pattern of reactivity was also found in 3/5 biopsies of

periapical granulomas that contained epithelium derived from the

cell rests of Malassez. None of the advanced disease or granuloma

biopsies displayed reactivity for vimentin within epithelial cells.

Discussion

The findings of this study have implications for understanding the

origin, structure and function of the reactive pocket epithelium in

the development of chronic periodontal disease. The presentation of

consistent and extensive reactivity for the simple cytokeratins CK8

and CK18 in a pattern restricted to the reactive pocket epithelium

of untreated chronic periodontitis lesions provokes the question of

the origin, structure and function of this tissue. In this context, the

pocket epithelium was demonstrated to be poorly equipped to serve

as a barrier to infection (16). Markedly decreased expression of spe-

cific epithelial differentiation products, E-cadherin, gap junction

communication channels and involucrin, as well as disruption of

filamentous actin, support this contention (16). Cytokeratins 8 and

18 are not a feature of the keratinocytes of differentiated stratified

squamous epithelia. This is exemplified by the extensive analysis

of the respiratory tract (17) and urogenital tract (18) for expression

of these proteins. Cytokeratins 8 and 18 were restricted to pseudo-

stratified columnar and transitional epithelia with a sharp demar-

cation at the borders, with stratified squamous epithelia uniformly

unreactive for these simple keratins. Knowledge relating to the

sources of pocket epithelial cells becomes important in the dissec-

tion of the critical interface between clinically reversible gingivitis

and progressive, destructive periodontitis when events within and

adjacent to the epithelial attachment are pivotal. Feghali-Assaly et

al. (19) used immunohistochemistry on frozen sections and two-di-

Hunter et al.

mensional (2-D) electrophoresis of microdissected samples to study

the cytokeratin profile of the gingival attachment in erupting teeth.

While the epithelium of the attachment apparatus contained both

CK13 and CK4, CK8/18 were not detected in agreement with other

reported findings (20). This confirmed numerous prior studies, in-

cluding that of Bosch et al. (21) who used immunohistochemistry,

2-D electrophoresis of microdissected tissues, in situ hybridisation

and Northern blot analysis to probe cytokeratin profiles of normal

and inflamed gingivae from undefined surgical specimens. While

CK19 expression was reported as markedly increased in inflamed

gingival tissues, there was no evidence for CK8/18 expression in

health or disease, with the exception of putative Merkel cells. These

occurred as scattered basal cells in most biopsies, and they were

notably diminished in inflamed tissues (21).

These cumulative data are discordant with the findings on gingi-

val attachment microdissected from teeth with a sound periodontal

attachment and, in the case of the reduced enamel epithelium, from

unerupted third molars (1). A low frequency of CK8/18-positive cells

was reported only in cells cultured from the epithelial attachment.

Of note, a high proportion of these cells in culture were positive

for vimentin, indicating a possible loss of epithelial differentiation

characteristics under tissue culture conditions. It was also reported

that the junctional epithelium (JE) and reduced enamel epithelium

(REE) contained a low percentage of cells reactive for CKs 13 and

4, but a very high percentage of positive cells was noted for CK6/

16 in the sulcular epithelium, JE and REE. Similar data were re-

ported for intermediate filament protein expression in cells cultured

from the epithelial attachment of healthy gingiva (22). It was noted,

however, that CK8/18 were not detectable in situ within the junc-

tional epithelium in contrast to an earlier study that reported a

positive finding (23).

In one of the few studies of the cytokeratin profile of inflamed

periodontal tissues, Ouhayoun et al. (24) reported findings based on

immunohistochemistry and 2-D electrophoresis of tissue biopsies

from teeth with associated periodontitis. The pocket epithelium was

characterised by a strong reaction for CK13 in all layers and a

strong reaction for CK4 in suprabasal layers. Scattered cells were

reactive for CK19. Cytokeratins 8/18, 14 and 17 were not studied.

Expression of CK8 and CK18 was, however, addressed in a study

by Mackenzie et al. (25) who reported scattered reactivity of apical

pocket epithelial cells in surgical material. This may reflect rem-

nants of deep pocket epithelium after initial preparation of the sites

prior to surgery.

In situ hybridisation analysis indicated that mRNAs for CKs 7,

8 and 18 were frequently present in the basal layers of normal oral

epithelia, but the corresponding proteins were not detected, indi-

cating downregulation at the level of translation (26). This con-

184 J Oral Pathol Med 30: 178–86

trasted with the expression of these cytokeratin proteins in severe

dysplasia and neoplasia (27).

In a recent report on an assortment of biopsy samples, Pritlove-

Carson et al. (28) noted that cytokeratin profiles in gingival epithelia

were markedly influenced by underlying inflammatory reactions.

Cytokeratin 19 was found to be a consistent marker of the epithelial

attachment and the pocket epithelium while CK8/18 were more vari-

able, depending on the antibodies employed and the intensity of the

inflammatory response. These researchers (28) also found that CK8/

18 were expressed in the stratified epithelia adjacent to the attach-

ment in a pattern related to inflammation. This latter finding ap-

pears to be novel and was not detected in our study where we

did not examine tissues affected by clinically recognised gingivitis

without evidence of periodontitis defined by failure of the epithelial

attachment. A possible explanation for the data of Pritlove-Carson

et al. is that the reduced enamel epithelial component of the attach-

ment had undergone a selective hyperplasia related to the inflam-

matory response in gingivitis, with the expression of CK8/18 re-

flecting the developmental origin of this tissue. Expression of these

cytokeratins at distal sites in the deep periodontal pocket of ad-

vanced disease would be consistent with downgrowth of this hyper-

plastic component of the attachment. A contribution to the epithel-

ium lining the pocket from epithelial remnants persisting within the

connective tissues of the periodontal ligament, the rests of Malassez,

that have a similar developmental origin and also express CK8/18

must also be considered. An additional factor in this study was a

history of tobacco use by the subjects. On this basis, expression of

simple keratins related to dysplastic changes cannot be excluded.

Data compatible with hyperplasia of the REE in response to in-

flammation was also observed in our study in relation to focal stain-

ing of the attachment in minimally inflamed tissues. Using a separ-

ate set of antibodies from those used by Pritlove-Carson et al. (28),

we observed consistent strong reactivity for CK8/18 within pocket

epithelia of periodontitis lesions. Variability of staining in the more

exposed regions of some biopsies of advanced diseased tissue could

reflect replacement of the original attachment cells with cells of

gingival phenotype, as a consequence of preferential migration of

gingival cells during repair of microwounds in the lining of the

pocket. There was, however, no consistent pattern of reactivity for

CK17 to support this interpretation directly.

While there is abundant evidence to indicate cytokeratin profiles

of stratified squamous epithelia are reactive to inflammatory stim-

uli, this does not appear to involve the expression of the simple

keratins CK8 and CK18. Thus, the evidence for expression of CK8

and CK18 by epithelial cells lining inflamed sinus tracts arising

from periapical lesions (29) could be explained by contributions

from activated developmental remnants. In a different context, the

Pocket epithelium in chronic periodontitis

appearance of small foci of CK8 and CK18 reactive cells in the

attachment zone of the epithelial abutment to dental implants (30)

could also be explained by activation of remnants of the dental

lamina that have persisted in the edentulous area.

Study of cytokeratin profiles in rests of Malassez and periapical

lesions indicated that epithelial proliferation in periapical lesions

was associated with increased expression of CK14 (a marker for

stratifying epithelia), new expression of CKs 4 and 13 (differen-

tiation markers for non-cornifying epithelia), and variable low levels

of CK8/18 (markers of simple epithelia) (26). This represents an un-

usual epithelial phenotype, indicating a phenotypic change from the

resting cell rests of Malassez defined by Peters et al. (31). In contrast

to the situation noted frequently in radicular cysts (26), the develop-

mental remnants that comprise the major part of the reactive lining

epithelium of periodontitis, as indicated by our findings in this

study, appear not to have committed to a differentiation program

other than the widespread expression of CK13.

Taken together, the results reported in the present paper are

interpreted as showing that the reactive lining epithelium of the

subgingival pocket in untreated chronic periodontitis continues to

express an intermediate filament profile that reflects an origin from

developmental epithelium of the enamel organ and cell rests of Mal-

assez.

References

1. Gao Z, MacKenzie IC. Patterns of phenotypic expression of human junc-tional, gingival and reduced enamel epithelia in vivo and in vitro. Epi-thelial Cell Biol 1992; 1: 156–67.

2. Uitto VJ, Airola K, Vaalamo M, et al. Collagenase-3 (matrix metallopro-teinase-13) expression is induced in oral mucosal epithelium duringchronic inflammation. Am J Pathol 1998; 152: 1489–99.

3. Pitaru S, McCulloch CA, Narayanan SA. Cellular origins and differen-tiation control mechanisms during periodontal development and woundhealing. J Periodontal Res 1994; 29: 81–94.

4. Thomas HF. Root formation. Int J Dev Biol 1995; 39: 231–7.5. Cutress TW, Crigger M. Cell rests of the sheep periodontium. N Z Dent

J 1974; 70: 39–49.6. Grant DA, Bernick S. A possible continuity between epithelial rests and

epithelial attachment in miniature swine. J Periodontol 1969; 40: 87–95.7. Spouge JD. A new look at the rests of Malassez. A review of their em-

bryological origin, anatomy, and possible role in periodontal health anddisease. J Periodontol 1980; 51: 437–44.

8. Browne RM, Cutress TW, Crigger M. The pathogenesis of odontogeniccysts: a review. J Oral Pathol 1975; 4: 31–46.

9. Spouge JD. The rests of Malassez and chronic marginal periodontitis. JClin Periodontol 1984; 11: 340–7.

10. Moll R, Franke WW, Schiller DL, Geiger B, Krepler R. The catalog ofhuman cytokeratins: patterns of expression in normal epithelia, tumorsand cultured cells. Cell 1982; 31: 11–24.

185J Oral Pathol Med 30: 178–86

11. Morgan PR, Leigh IM, Purkis PE, Gardner ID, Van Muijen GNP, Lane EB.Site variation in keratin expression in human oral epithelia – an immuno-cytochemical study of individual keratins. Epithelia 1987; 1: 31–43.

12. Lindberg K, Rheinwald JG. Three distinct keratinocyte subtypes iden-tified in human oral epithelium by their patterns of keratin expressionin culture and in xenografts. Differentiation 1990; 45: 230–41.

13. Paladini RD, Takahashi K, Bravo NS, Coulombe PA. Onset of reepi-thelialization after skin injury correlates with a reorganization of keratinfilaments in wound edge keratinocytes: defining a potential role for kera-tin 16. J Cell Biol 1996; 132: 381–97.

14. Pinchback JS, Taylor BA, Gibbins JR, Hunter N. Microvascular angiopa-thy in advanced periodontal disease. J Pathol 1996; 179: 204–9.

15. Summers L, Papadimitriou J. The nature of epithelial proliferation inapical granulomas. J Oral Pathol 1975; 4: 324–9.

16. Ye P, Chapple CC, Kumar K, Hunter N. Expression patterns of E-cadher-in, involucrin and connexin gap junction proteins in the lining epitheliaof inflamed gingiva. J Pathol (to be published).

17. Nagle RB, Moll R, Weidaur H, Nemetschek H, Franke WW. Differentpatterns of cytokeratin expression in the normal epithelia of the upperrespiratory tract. Differentiation 1985; 30: 130–40.

18. Schaafsma HE, Ramaekers FCS, van Muijen GNP, Ooms ECM, RuiterDJ. Distribution of cytokeratin polypeptides in epithelia of the adulthuman urinary tract. Histochemistry 1989; 91: 151–59.

19. Feghali-Assaly M, Sawaf MH, Serres G, Forest N, Ouhayoun JP. Cytoker-atin profile of the junctional epithelium in partially erupted teeth. J Peri-odontal Res 1994; 29: 185–95.

20. Morgan PR, Johnson NW, Leigh IM, Lane EB. Structure of gingival epi-thelium as revealed by monoclonal antibodies to keratins. In: Lehner T,Cimasoni G, eds. The borderland between caries and periodontal disease,Vol. III. Geneva: Medicine et Hygiene, 1986; 48–59.

21. Bosch FX, Ouhayoun JP, Bader BL, et al. Extensive changes in cytoker-atin expression patterns in pathologically affected human gingiva. Vir-chows Arch 1989; 58: 59–77.

22. Matsuyama T, Izumi Y, Sueda T. Culture and characterization of humanjunctional epithelial cells. J Periodontol 1997; 68: 229–39.

23. MacKenzie IC, Rittman G, Gao Z, Leigh I, Lane EB. Patterns of cytoker-atin expression in human gingival epithelia. J Periodontal Res 1991; 26:468–78.

24. Ouhayoun JP, Goffaux JC, Sawaf MH, Shabana AH, Collin C, Forest N.Changes in cytokeratin expression in gingiva during inflammation. JPeriodontal Res 1990; 25: 283–92.

25. MacKenzie IC, Gao Z. Patterns of cytokeratin expression in the epitheliaof inflamed human gingiva and periodontal pockets. J Periodontal Res1993; 28: 49–59.

26. Gao Z, MacKenzie IC, Williams DM, Cruchley AT, Leigh I, Lane EB.Patterns of keratin-expression in rests of Malassez and periapicallesions. J Oral Pathol 1988; 17: 178–85.

27. Su L, Morgan PR, Lane EB. Protein and mRNA expression of simpleepithelial keratins in normal, dysplastic, and malignant oral epithelia.Am J Pathol 1994; 145: 1349–57.

28. Pritlove-Carson S, Charlesworth S, Morgan PR, Palmer RM. Cytokeratinphenotypes at the dento-gingival junction in relative health and in-flammation, in smokers and nonsmokers. Oral Dis 1997; 3: 19–24.

29. Gao Z, MacKenzie IC, Pan S, Shi J. Epithelial lining of sinus tracts associ-ated with periapical disease: an immunocytochemical study using mono-clonal antibodies to keratins. J Oral Pathol Med 1991; 20: 228–33.

30. MacKenzie IC, Tonetti MS. Formation of normal gingival epithelialphenotypes around osseo-integrated oral implants in humans. J Peri-odontol 1995; 66: 933–43.

Hunter et al.

31. Peters BH, Peters JM, Kuhn C, Zoller J, Franke WW. Maintenance ofcell-type-specific cytoskeletal character in epithelial cells out of epi-thelial context: cytokeratins and other cytoskeletal proteins in therests of Malassez of the periodontal ligament. Differentiation 1995;59: 113–26.

186 J Oral Pathol Med 30: 178–86

Acknowledgements

This study was supported by the National Health and Medical ResearchCouncil of Australia and the Dental Board of NSW.