EMDOGAIW REGULATION RAT PERIODONTIUMperiodontium, by Laura Chano. Degree: Master of Science,...
Transcript of EMDOGAIW REGULATION RAT PERIODONTIUMperiodontium, by Laura Chano. Degree: Master of Science,...
-
EMDOGAIW REGULATION OF CELLULAR DlFFERENTlATlON IN WOUNDEO RAT PERIODONTIUM
Laura Chano
A thesis submitted in wnfonnity with the requirernents for the degree of Master of Science Graduate Department of Dentistry
University of Toronto
@Copyright by Laura Chano 2001
-
National Library B*B of camda Bibliothèque nationale du Canada Acquisitions and Acquisitiorrs et Bibliographie Services services bibliographiques 395 Welington Street 395, rue Wdingtori OltawaON KlAON4 OiiawaON K 1 A W Caneda CaMda
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Libraty of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or se0 reproduire, prêter, disûibuer ou copies of this thesis in microfom, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de
reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extxacts fiom it Ni la thèse ni des extraits substantiels may be printed or othedse de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.
-
A bstract
EmdogairND regulation of cellular differentiation in wounded rat periodontium, by Laura Chano. Degree: Master of Science, Department of Periodontology, Faculty of Dentistry, University of Toronto, 2001.
EmdogainO is an enamel matrix derivative that may promote periodontal
regeneration by recapitulating critical events in tooth morphogenesis. I
hypothesized that EmdogainQY facilitates periodontal regeneration by promoting
the differentiation of cells required for the synthesis of periodontal ligament, bone
and cementum. Cell differentiation was examined in a rat periodontal window
wound model. Defects were filled with vehicle control or ErndogairNB (3 mg/ml or
30 mg/ml). Rats were sacrificed at 7, 14 and 21 days after wounding. Specimens
of periodontium were prepared for irnmunohistochemistry, morphometry and
radioautography. Rats treated with Emdogaim (30 mg/ml) showed widening of
the periodontal ligament at 7 days; by 14 and 21 days, periodontal ligament width
was restored to normal values for al1 groups. Emdogain exerted no effect on
cernentum thickness, bone volume, osteoid deposition rates, or extracellular
staining for osteopontin, bone sialoprotein or osteocalcin. Further, the percentage
of cells with intracellular staining for osteopontin, osteocalcin or bone sialoprotein
was unaffected by Emdogaim. Staining for a-smooth muscle actin
(myofibroblast marker) was abundant in the repopulating wound but was also
unaffected by EmdogaiMl. I conclude that EmdogaiM does not affect cell
differentiation or bone matrix protein synthesis in the repopulation response of
wounded rat molar periodontium .
-
Ackno wledgemenfs
I could not have accomplished this project without my supervisor, Dr. Chnstopher
McCulloch. I am grateful for the intelledual insight, encouragement, extreme
generosity and fnendship he has given me. I am also thankful for the technical
assistance of the laboratory staff, Hong Hong Chen and Balram Sukhu.
I would like to thank my family for al1 the attention, support,
encouragement, and love they gave me ovet the years. I am etemally grateful to
my mother for I tnily tealize and appreciate how much she has given and
supported me al1 through my Me. I also appreciate the encouragement that
Brenda gave me durhg these last three years and I deeply cherish the love and
support that James has given me.
-
Table of Contents
Abstract Acknowledgements Table List of Figures List of Abbreviations
ii iii vi vii viii
1. Literature Review 1
A. Penodontal diseases 1
B. Repair and regeneration of the periodontium 4
1. Wound healing in the periodontium 4
2. Cells in the periodontal wound healing response 7
3. Cellular markers of differentiation 10
4. Regulation of differentiation 12
5. Treatment approaches to regulate cellular differentiation 14
Cm Tooth and mot fornation 16
D. Emdogain 19
E- Mode1 systems and rationale 24
1. Non-human primates 26
2. Dogs
3. Rodents
II. Staternent of the problem 29
-
III. Materials and Methods
A. Wound mode1
B. Pmparation of implants and experimen fa1 design
C. Tissue preparation
D. Immunohistochemistry
E. Radioautography
F. Motphometric a n a l p s
G. Statistical analyses
IV. Results
A. Periodonfal ligament homeostasis
B. Cementum
C. Bone
D. Osteogenic difFerentiation
E. Fibmblast differentiation
F. Matrix formation
V. Discussion
A. Mode1
B. Mat& synthesis
C. DMemntiation
VI. Conclusions
VIL References
-
Table
Table 1. Matnx formation (appositional rate of osteoid and incorporation of
radiolabeled proline into nascent matrix proteins).
-
List of figures
Figure 1 A: Immunohistochemical staining for osteopontin
Figure 1 B: Morphometric analyses of periodontal ligament width
Figure 1 C: lmmunohistochemical staining for bone sialoprotein
Figure 1 D: Morphometric analyses of œmentum thickness
Figure 1 E: lmmunohistochemical staining for osteopontin
Figure 1 F: Morphometric analyses of nascent bone formation Figure 2A: lmmunohistochemical staining for intracellular osteopontin
Figure 28: Percentage of osteopontin cellsltotal cell count
Figure 2C: lmmunohistochemical staining for extracellular osteopontin
Figure 2D: Percentage of osteopontin stained matrix
Figure 3A: Immunohistochemical staining for intracellular bone sialoprotein
Figure 38: Percentage of bone sialoprotein cells/total cell wunt
Figure 3C: lmmunohistochemical staining for extracellular bone sialoprotein
Figure 3D: Percentage of bone sialoprotein stained rnatrix
Figure 4A: lmmunohistochemical staining for intracellular osteocalcin
Figure 48: Percentage of osteocalcin cellsRotal cell count
Figure 4C: lmmunohistochemical staining for extracellular osteocalcin
Figure 4D: Percentage of osteocalcin stained matrix
Figure SA: lmmunohistochemical staining for intracellular a-smooth muscle actin
Figure 58: Percentage of a-smooth muscle actin cellsltotal cell count
vii
-
Abbrevia fions
a-SMA
BMP
BSP
FGF
IGF
OC
OPN
PBS
PDGF
PL
TGF
Alpha-smooth muscle adin
Bone morphogenetic protein
Bone sialoprotein
Fibroblast growth factor
lnsulin growth factor
Osteocalcin
Osteopontin
Phosphate buffer saline
Platelet derived growth factor
Periodontal ligament
Transfonning growth fador
-
1. Literature Review
A. Penodontal diseases
Periodontal diseases are high prevalence infections of the periodontium that are
classified into two major groups- gingivitis and periodontitis (Page and
Schroeder, 1976). Gingivitis is a reversible inflammatory lesion of the
dentogingival junction that is not associated with bone loss (Lbe et al., 1965). In
wntrast, periodontitis is an ineversible destructive lesion (Page and Schroeder,
1976) of the connective tissue attachment and alveolar bone (Narayanan and
Page, 1983) that if left untreated can lead to tooth exfoliation.
Although root-borne, adherent bacterial biofilms are required for the
initiation of periodontal diseases, they are not sufficient alone to cause
progressive attachment loss (Offenbacher. 1996). Instead, epithelial and matrix
degradation are very likely attributable to disturbances in the balance between
host defense mechanisms and microbial assault. Accordingly, key factors in
disease progression indude both bacterial products (e.g. proteases, surface
membrane toxins) and host cell chernical mediators. Some of the host-derived
factors that are important in progressive periodontal tissue destruction include a
wide variety of pro-inflammatory molecules released by polymorphonuclear
leukocytes and macrophages and espedally the matrix metalloproteinases, zinc-
dependent endoproteinases which degrade collagen and other prominent
periodontal matrix proteins (Birkedal-Hansen et al., 1993). In bnef, the
development and progression of periodontitis is mnsidered to be a multifactorial
lesion involving interactions between the pathogenic wmponents of dental
-
biofilms, the vasculature, the innate and humoral immune systems, the epithelial
cells of the dentogingival jundion, stroma1 wnnective tissue cells and their
matrices (Offenbacher, 1996).
The reported prevalence (the fraction of a population exhibiting a
pathological condition at a specific time point) of periodontitis rnay Vary as a
result of the choice of measurement methods (Locker and Leake, 1993; Brown et
al., 1994). Nevertheless, a widely quoted estimate for the prevalence of
penodontitis in the United States is 36% (Brown et al., 1989). Notably, only a
fraction (i.e. 1 O-3O%) of those affected with periodontitis exhibit progressive
lesions and these susceptible individuals account for most of the disease burden
(Hirschfeld and Wasserrnan, 1978). As a result of the severity of the periodontitis
lesions in this susceptible group and because of intrinsic problems in treating
these infections, the dinical management of progressive penodontitis is a
substantial challenge to the dental health care system.
Numerous microbial species in dental biofilms have been implicated in the
initiation and progression of periodonti tis. However, only a relatively smal l
number have been examined in suffident detail to justify their inclusion as
putative periodontal pathogens. Some of the major pathogenic species include
Potphymnonas gingivalis, Actinobacillus actinomycetemcomitans and
Bactemides forsjdhus (Zambon, 1 996). In addition to the root-borne pathogen
load generated by these and other less virulent bacterial species, tome evidence
indicates that œmentum surfaces contaminated by periodontal pathogens
undergo morphological and biochernical changes that may contribute significantly
-
to the progression of the lesions (Aleo et al., 1974). Further, in the context of
periodontal wound healing, contaminated root œmentum may be a critical factor
in blocking reattachment of nascent collagen fibers to previously exposed root
surfaces since root-bound endotoxin can inhibit ceIl attachment in vitro (Aleo et
al., 1975).
Contamination of the root surface as well as the loss of potential growth
and differentiation factors from the cementum matrix are important obstacles for
periodontal repair and regeneration. Further, the mot surface is an avascular
structure and cannot directly contribute to the formation of blood vessels or to the
production of cells that are important for reattachment. Instead, the cells that
repopulate and attach to the exposed root surface must migrate from adjacent
vascularized wound edges. Hence, our understanding of root surface biology and
its role in cell recruitment and cell differentiation is pivotal in the regeneration of
perÏodontal tissues, particularîy since the repopulating cells in a periodontal
wound must migrate a considerable distance before they can contribute to
healing at the root surface. Notably, cells from contiguous endosteal spaces
(McCulloch et al., 1987) rnay be able to contribute to the repopulation of healing
periodontal wounds. However, we have limited ability to manipulate this
repopulation response favourably and predictably sinœ we do not know how to
selectively recniit or promote the differentiation of the cell types that are required
for nascent tissue synthesis (e.g. cementoblasts, osteoblasts, fibroblasts).
-
B. Repair and regeneration of the periodontium
1. Wound healing in the periodontium
Wound healing is the process by which an organism attempts to reconstitute
tissues damaged by injury and subsequently restores their function.
Reconstitution can be achieved either through regeneration or repair or both. In
regeneration the architecture and function of lost periodontal tissues are
completely renewed. When this is not accomplished, the damaged tissue is
repaired and the injured tissue is replaced with a fibrous scar or by cells and
attachment complexes that are not found in the original tissue structure (e.g. long
junctional epithelium). Repair is a biological process by which the continuity of
disrupted tissue is restored by new tissues that do not replicate the structure and
function of tissues destroyed by disease or injury (Arnerican Academy of
Periodontology, 1 992).
Following tissue injury, there is rapid formation of a blood clot and
stabilization by fibrin. Between 1 3 days following injury, polyrnorphonuclear
leukocytes and macrophages are reuuited to the wounded site. These cells
debride the wound site by phagocytosis, kill local microorganisms and facilitate
tissue remodeling by the release of matrix metalloproteinases and other neutral
proteases. The macrophage population also contributes an important regulatory
role since cytokines released by macrophages help to control local cell fundion.
Later on in wound healing (4-10 days), a provisional matrix is formed by
endothdial cells and fibroblasts that is the prewrsor for more mature periodontal
tissues. The granulation tissue f m e d at the wound site is colonized 4-10 days
-
after wounding by progenitor cells from the periodontal ligament and ccntiguous
endosteal spaces. The progenitors can undergo several rounds of ceIl division to
become periodontal ligament fibroblasts, cementoblasts and osteoblasts. In tum
these cells secrete the proteins of the specialized tissues of the periodontium
including periodontal ligament, cementum and bone. Finally, there is a long-terni
remodeling phase of the newly formed tissues that may require as long as 7
year.
The critical events in wound healing of many tissues and organs exhibit
several cornmon features and have been studied in depth using, most commonly,
experimental skin wound models. However, in cornparison to the relatively simple
wound healing responses that ocair in skin, healing of pefiodontium is
complicated by the diversity of the celi and tissue types as well as the inability of
the root surface to be vascularized. Notably, wound healing studies in the
-
periodontium demonstrate that the phenotype of the repopulating cells exert an
impact on the type of repair processes that occur subsequently (Nyman et al.,
1982). This finding and earlier work by Melcher (1 970) established the notion that
the tissue ongin of the cells impaded on the nature of the tissue that is ultimately
foned; this concept has provided the basis for the socalled guided tissue
regeneration procedure.
Early events in the wound healing process may have an impact on later
outcornes. Notably, Polson and Proye (1 983) extracted and re-implanted normal
teeth in squirrel monkeys after surgically denuding the coronal root surface of
connective tissue fibers and œmentum by root planing. Some test spedmens
were treated with topical application of citric acid; these teeth exhibited a different
response campared to control specirnens that were not acid-treated. In controis,
the epithelium migrated rapidly along the denuded root surface. Epithelium
reached the alveolar crest by 3 days and extended into the periodontal ligament
space ta the level of the denuded root by 21 days. In contrast, in roots treated
with citric acid, the epithelium did not migrate significantly along the denuded root
surfaces. At 1 and 3 days, infiammatory cells were embedded in a fibrin network
that was apparently attached to the root surface. At 7 and 21 days, the wound
site was repopulated with cannedive tissue cells and collagen fibers had
replaœd the fibrin. The authors speculated that formation of the collagen fiber
attachment to the root surface was preceded by a fibrin linkage. This linkage was
considered to be an important early event in the wound healing response and
underlines the importance of wound stabilization by fibrin. Further, the paper
-
highlights critical steps played by specific proteins (i.e. fibrin) in the wound
healing process.
2. Cells in the periodontal wound healing response
A very large vatiety of different cell types participate in periodontal wound healing
but only the synthetic cells of petiodontal tissue matrices will be described in any
detail here. Bnefly, the cells participating in wound healing originate from different
precursor populations. For example, polymorphonudear leukocytes are gtanular
leukocytes 7-9 Pm in diameter that are of hematogenous (myeloid) origin and
contribute important microbiocidal and phagocytic functions. Platelets are 2 4 pm
diarneter dis= devoid of a nudeus that are fomed from megakaryocytes and
play aitical roles in blood clotting. Some of the platelet granules also contain
important cytokines that regulate wound healing such as transforming growth
factor+ and platelet derived growth factor. The synthetic cells in periodontal
wounds are derivad from locally proliferating stroma1 precutsors and include
fibroblasts, cementoblasts and osteoblasts.
Fibroblasts are large, often flat, branching cells which appear fusifonn or
spindle-shaped in profile and exhibit an oval or elongated nucleus. They are
responsible for the formation and rernodelling of collagen fibers and are thought
to elaborate most, if not all, of the amorphous component of the rnatrix
(principally glycosaminoglycans). Osteoblasts have vanous shapes (i.e cuboidal
or squamous) depending on their synthetic activity. They exhibit a large nudeus
and are found on bone surfaces where matrix synthesis and minerakation are
extant. Cementoblasts are by definition, associated with the surfaces of newly
-
foming cementum and typically exhibit similar morphologies as osteoblasts.
However, 'restingn cernentoblasts are not squamous in shape but rather appear
to retract from the cementum margin and instead appear to be more fibroblastic
in appearance.
When considering the formative cells of the periodontal tissues, it is
worthwhile to consider their ability to divide and multiply since in wounds, cell
multiplication is critical for repopulation and reparative processes. In general,
mammalian cells may be classified into 3 types on the basis of their proliferative
potential (Leblond, 1 981 ). Renewal cell populations continuously divide and have
considerable division potential throughout postnatal Iife. For example, the crypt
and villus cells lining the gastrointestinal tract and the progenitor cells of blood
comprise renewal cell populations. Expanding cell populations are comprised of
cells that have limited division potential during the lifetime of the organism but
can divide extensively on demand, most notably after injury. Liver and kidney
cells are examples of expanding ceIl populations. Static or non-renewal cells are
cells which have exited the cell cycle, undergone differentiation, and do not
divide. Central nervous system neurons are examples of static cells; the deletion
of these cells by injury or loss of blood supply (Le. stroke) has important
implications for restoration of function.
From morphological and cell kinetic studies of mouse periodontal ligament
under conditions of both normal function and while undergoing regeneration,
several important properties of the formative cell population have k e n deduced.
First, the progenitor cells of periodontal ligament in paravascular zones exhibit
-
some of the features of stem cels of renewing populations. These cells
classically exhibit extensive proliferative capacity (Gould et al., 1977), are
responsive to control mechanisms (Gould et al., 1977, l98O), demonstrate self-
renewal capacity (Gould et a/., 1980) and are spanely distributed within the
proliferative compartment (McCulloch and Melcher, 1983a). Second, the progeny
of paravaswlar progenitor cells migrate to bon8 and cementum surfaces where
they differentiate into cementoblasts and osteoblasts (McCulloch and Melcher,
1983b; McCulloch et al., 1987). Third, the turnover of periodontal ligament cells is
relatively slow (McCulloch and Melcher 1983~). Collectively, these data suggest
that the formative cells of the periodontal ligament comprise as slowly renewing
cell population. Finally, it is possible that the relatively primitive cells immediately
surrounding blood vessels (Le. paravascular cells) are stem cells since they
exhibit many of the classical features of stem cells described above (McCulloch,
1 985).
In addition to providing cells for repopulation of the fibroblast population,
the periodontal ligament may also serve as a reservoir of precursor cells for
cementoblasts and osteoblasts. Cementoblasts are important for the apposition
of cementum that may be required in response to trauma and for the repair of
darnaged root surfaces (Beertsen and Everts, 1990). Periodontal ligament
fibroblasts may also have the ability to differentiate into cernentoblasts
responsible for the synthesis of acellular extrinsic fiber cementum (Beertsen and
Everts, 1990). Alveolar bone also undergoes continual remodeling under
physiological conditions and the progeny of paravascular fibroblastic cells in the
-
periodontal ligament c m migrate to the bone surface and there diffetentiate into
osteoblasts (McCulloch and Melcher, 1983a). Consequently, cells from the
periodontal ligament may be crucial in the formation and repair of cementum and
bone, processes that are important in maintaining the structural integrity of the
periodontium.
During tooth formation, cementoblasts and periodontal ligament fibroblasts
originate pnmarily from cells of the dental follicle proper and the cells of the
perifollicular mesenchyme respectively (Cho and Garant, 1988; 1989). Shortly
after the onset of cementogenesis, cementoblasts detach from the newly fonned
cementum surface and appear to join the fibroblast population in the periodontal
ligament. This suggests that both the dental follicle proper and perifollicular
mesenchymal cells contribute to the periodontal ligament cell pool. Thus the
periodontal ligament cell population is apparently a mixed cell population
containing precursors for cementogenic cells (among others). However, a more
definitive understanding of the wound healing potential of the formative cells of
the periodontium has (and will) rely on rnethods to measure unambiguously the
differentiation potential of the cells.
3. Cellular markers of d'ifferentiation
Molecular markers of cellular differentiation have been utilized in many studies
focusing on for example, blood formation (Till and McCulloch, 1980) and bone
formation (Bruder et al., 1990; Turksen et al., 1992). In mineralized tissue
formation several extracellular matrix macromolecules including osteopontin,
osteocalcin, and bone sialop rotein are expressed by differentiating osteogenic
-
cells. Measurement of these proteins has been used to identify discrete stages in
the formation of bone in vivo (Chen et al., 1992; Yoon et al., 1987) and for
studies of ceIl differentiation in periodontal tissues (Lekic et al.. 1996b).
Osteopontin and bone sialoprotein are major noncollagenous proteins
rscreted by osteoblastic cells and deposited into the bone matrix (Kasugai et al.,
1991 ; Nagata et a/. , 1 991 ). Both proteins are glycosylated, phosphorylated, and
sulfated and both have Arg-Gly-Arp (RGD) sequences that may provide cell
attachment motifs (Butler, 1989; Heinegard and Oldberg, 1989). However,
whereas bone sialoprotein is expressed almost exclusively by differentiated
mineralized tissue-forming cells (Bianco et al., 1991; Chen et al., t991a, 1992),
osteopontin is expressed by osteogenic and also nonosteogenic cells (Denhardt
and Guo, 1993). In osteogenesis, osteopontin mRNA is expressed during matrix
formation. In cornparison. the expression of bone sialoprotein coincides with
initial bone mineralization and is believed to be a nudeator of hydroxyapatite
crystal formation (Hunter and Goldberg, 1993). Since osteopontin and bone
sialoprotein are expressed in alveolar bone, cernentum and dentin (Chen et al.,
1991 a, 1993), the differential expression of these proteins can be used to study
the formation and repair of periodontal tissues and the contribution of periodontal
ligament cells to bone matrices that contain osteopontin and bone sialoprotein.
In camparison to osteopontin, bone sialoprotein (Kasugai et al., 1992) and
osteocalàn (Chen et al., 1992) are expressed at later stages of bone cell
differentiation and may be expressed also by cernentoblasts. Hence the
identification of these matrix proteins could suggest the presence of cells
-
committed to the osteogenic or cementogenic lineages. Osteocalcin is a y-
carboxyglutamic acid (Gla)-containing protein that is also a major
noncollagenous bone protein. It has been used for studies of later stages of bone
cell differentiation in rat tissues (McKee et al., 1993). Collectively, osteopontin,
bone sialoprotein and osteocalàn c m provide useful markers for bone and
possibly cementum cell differentiation.
a-smooth muscle actin is an actin isoform that has been extensively
studied in the differentiation of fibroblasts into myofibroblasts and their role as
contractile cells (Desmouliere et al., 1993). The a-smooth muscle actin isoform is
a well-descfibed fundional marker for a subpopulation of contractile periodontal
fibroblasts (Arora and McCulloch, 1994) and its assessrnent may facilitate the
identification of specific, periodontal cell subpopulations that are important in
rnatrix contraction and wound remodelling (Arora et al., 1999).
4. Regulation of differentiation
A wide variety of molecules participate in the regulatory processes required for
periodontal regeneration. Based on their mechanism of action, two important
classes of regulatory molecules are worth considering in the context of
periodontal regeneration and how exogenous application of these factors rnay be
exploited to facilitate regeneration: 1) growth factors and other inflammatory
mediators, including cytokines, lymphokines, and chemokines; 2) adhesian
molecules and matrix components such as fibronectin, laminin, collagens,
proteoglycans, and hyaluronan. The first group of molecules can regulate the
migration, proliferation and differentiation of cells during inflammation and wound
-
repair. Adhesion molecules localize cells at required sites and may be specific for
certain cell types or may be non-specific in their interactions. Matnx components
also provide important adhesive functions and are needed for the structural and
physiologic integrity of new tissues as well as for regulating cell differentiation.
These molecules may originate from the circulation or may be produœd locally
by cells residing in the tissue matrix.
Polypeptide growth factors are naturally ocuirring biological mediators
that orchestrate critical cellular events involved in regenerative processes sudi
as cell proliferation, chernotaxis, differentiation and matrix synthesis (Matsuda et
al., 1992). The growth factors found in bone matrices indude transfoming growth
factor-B (TGFS), insulin-like growth factors I and II (IGF-I and II), platelet-derived
growth factor (PDGF), acidic and basic growth factors (a- and b-FGF) and bone
morphogenetic proteins (Graves and Cochran, 1994). The prirnary cellular
sources of PDGF, FGF, TGF-f3 and IGF are platelets, macrophages and
osteoblasts (Giannobile, 1996). These factors are also stored in bone matrix and
may be released during bone remodelling, thus helping to couple bone formation
to resorption (Linkhart et al., 1996). In bone, osteoblasts are important target
cells of these growth factors although sorne of these cytokines can induce
periodontal ligament cells to proliferate as well (Graves and Cochran, 1994).
There are abundant epidemial growth factor (EGF)-binding sites on
differentiating perifollicular mesenchyme cells as well as on mature periodontal
ligament fibroblasts exhibiting synthetic adivity (Cho et al., 1991 ). This finding
suggests that EGF plays an important role in cell differentiation as well as during
-
the active synthetic activity of mature cells. Conversely, low binding of EGF to
dental follicle cells, precementoblasts and cementoblasts indicates that EGF
probably has little or no effect on cementoblast difFerentiation. Consequently, the
EGF-receptor may act to negatively regulate the differentiation of periodontal
ligament fibroblasts into mineralized tissue-forming cells (Cho et al., 1991 ).
5. Treatment approaches to regulate cellular differentiation
As noted above, vanous wound healing events and cellular activities assouated
with healing are regulated by polypeptide growth factors. Several exogenously
applied growth factors have been utilized to treat naturally and experimentally-
induœd periodontal defects in anirnals (Graves and Cochran, 1994). In beagle
dogs with naturally occurring periodontitis, Lynch et al. (1 989) demonstrated that
the combination of PDGF with IGF-1 stimulated regeneration of the periodontium,
possibly through its efFed on mesenchymal cells. Indeed, the formation of new
cementum-like deposits and alveolar bone was present in growth factor-treated
sites but not in mt ro ls receiving surgery and placebo gel.
Giannobile et al. (1 996) compared the effects of platelet-denved growth
factor-BB and insulin-like growth factor-1, individually and in combination, on
periodontal regeneration in Cynomolgus monkeys. Ligature-induced periodontitis
was initiated and after periodontal lesions were established, surgery was
performed, and either gel (i.0. vehicle control), or gel containing PDGF-BB, IGF-I
or bath was applied to the exposed root surfaces. They found that IGF-l alone, at
the dose tested, did not significantly alter periodontal healing. In contrast, PDGF-
86 alone strongly stimulated the formation of new attachment. In addition, the
-
PDGF-BBIIGF-I combination resulted in significant increases in new attachment
and production of new bone in the osseous defects 4 and 12 weeks post-
surgically. Further, in a study performed in monkeys, Rutherford et al. (1993)
demonstrated that a combination of PDGF and dexamethasone could also
promote regeneration.
Howell et al. (1997) perfomed a clinical trial to evaluate the therapeutic
effect of a combination of recombinant human PDGF-BB and recombinant
human IGF-l in patients with periodontitis. Subjects were treated in a splitmouth
design. The test sites received the local application of the dnig in one of two
doses while the control sites were either treated surgically only or received a
vehicle gel. Reentry procedures were perfomed 6 to 9 months post-surgically.
Patients treated with the higher dose of the wmbined dnig protocol
demonstrated statistically significant increases in alveolar bone formation. Sirnilar
results were found for furcation defeds.
Collectively, these studies suggest that the topical application of growth
factors has promise in the treatment of periodontitis but further studies are
needed to characterite the mechanism of action, the kinetics of dnig degradation
and dnig release, and to more carefully demonstrate the reliability and efficacy of
the therapeutic maneuver. While the topical application of these factors may
ultimately show promise clinically, there is uncertainty as to whether the cells
involved in wound healing in adult periodontal tissues are actually capable of
regeneration. Thus an important question is whether wound healing in aduît
tissues recapitulates the events of root and periodontal ligament formation that
-
occur during tooth development. Cognizant of these issues, in the next section I
will describe the salient events of tooth formation that relate to the development
and formation of the periodontal ligament and alveolar bone.
C. Tooth and mot fornation
In development, the organization of cells into tissues is accomplished in part
through morphogenesis; cellular diversification is achieved through the process
of differentiation. These two processes, in addition to growth and reproduction.
comprise critical components of mammalian development and are central to
organogenesis. After the 3 germ layers (ectoderm, mesodenn and endoderrn)
are formed through primary induction, organogenesis is initiated. The cells of the
germ layers interact and rearrange thernselves into specialized organs such as
the limbs, eyes and teeth. The formation of these and other complex organs
depends on sequential and reciprocal interactions between the epithelial and
mesenchyrnal tissue, a proœss called secondary induction.
Teeth arise as a result of secondary induction between the oral epithelium
and its adjacent dental mesenchyme (Lumsden, 1988; Thesleff and Sharpe,
1997). The oral epithelium originates from the stomodeal or pharyngeal region of
the developing embryo (Thesleff and Sharpe, 1997). The mesenchymal cells
underîying the epithelium of the first brachial arches (where the future rnaxillary
and mandibular processes reside) are of neural crest origin (Osumi-Yamashita et
al., 1994). Following migration of neural crest cells to the first brachial ara, a
group of neural crest cells interact with the overîying oral epithelium and cause
an invagination of a band of epithelial cells to form the dental lamina. Indeed, the
-
first distinct morphological feature indicative of tooth development is the
formation of the dental lamina and the localized thickening of the dental
epitheliurn. As the dental lamina expands, the epithelium invaginates into the
underiying mesenchyme, forming an epithelial bud (the bud stage of tooth
development). The mesenchymal cells condense around the bud and fonn the
dental papilla which gives rise to both the dental pulp and dentin-secreting
odontoblasts (Peters and Balling, 1 999). Notably, the proteoglycans syndewn-1
and tenascin, two important cell adhesion molecules, are involved in
mesenchymal cell condensation (Thesleff et ai., 1995). At the cap stage of
development, the enamel knot appears within the epithelial cornpartment mi le
the bel1 stage is characterizad by rapid cell proliferation and dental crown
formation (Kettunen and Thesleff, 1998).
The multiple redprocal interactions between oral epithelium and its
adjacent mesenchyme dernonstrate that the inductive and wmpetence
properties of each tissue are temporally and spatially restricted (Dassule and
McMahon, 1998). This has been shown in tissue recombination experiments in
which a non-dental, neural crest-derived mesenchyme from a mouse could be
induœd by an oral epithelium isolated from the mandibular arch. Apparently, the
odontogenic potential resides in the epithelial layer at this early stage of tooth
development. At the onset of tooth morphogenesis, odontogenic signaling events
originate fiom the dental epithelium (Lumsden, 1 988). The instructive signaling
center reverts to the enamel knot of the dental epithelium (Peters and Balling,
1999).
-
The inner enamel epithelium plays an essential role in the differentiation of
odontoblasts of the crown. For the development of root odontoblasts, the inner
epithelial cells of the root sheath are required. The detachment of the inner
epithelial cells from the dentin surface and the fragmentation of the underlying
basal lamina may result from the active migration of preœmentoblasts toward the
dentin surface and their penetration between the dentin surface and the inner
epithelial cells during their differentiation into cementoblasts (Cho and Garant,
1988). The disruption of the root sheath exposes the newly deposited dentin
matrix to the follicular connective tissue. a possible requirement for cementoblast
differentiation. The earliest sign of precementoblast differentiation is the
projection of cell processes from their leading edge toward and into the space
previously occupied by the root sheath. Cementoblast differentiation involves
directed cell migration of precementoblasts toward the dentin matrix causing
disruption of the epithelial root sheath and the eventual contact of newly
differentiated cementoblasts with root dentin. Upon contact with the dentin
surface, these cells exhibit the appearance of fully differentiated cementoblasts
involved in forrning acellular extrinsic fiber cementum. Differentiation and
migration of precementoblasts from the dental follide toward the dentinal surface
are probably initiated by a chemoattractant andhr promoted by an adhesion
gradient within the local extracellular dentinal matrix or from the inner basement
membrane of the root sheath (Cho and Garant. 1988). Soon after œmentum
deposition, newly differentiated cementoblasts detach and move away from the
newly fomed cementum, which implies that the signal for their chemotaxis is
-
transient. They then become a part of the periodontal ligament fibroblast
population (Cho and Garant, 1989).
The formation of acellular cementum involves an initial phase of directed
cell migration, followed by attachment and cementum matrix deposition during
which the cementoblast morphotype is deariy expressed. It then detaches and
assumes a fibroblast-like morphotype (Cho and Garant, 1989). The
mesenchymal cells of the dental follide undergo direded migration toward the
dentin and the perifolliwlar mesenchymal cells differentiate primarily into the
periodontal ligament fibroblasts. Thus both the dental follicle proper and
perifallicular mesenchyme contribute to the pool of periodontal ligament cells
(Cho and Garant, 1989).
Wth this brief background of tooth and periodontal ligament development
in mind, there has been considerable interest in applying principles of tooth
development to periodontal wound healing. Notably, the work of Lars
Hammarstrom has been pivotal in developing a biological basis for topical
application of enameldenved proteins that may be able to facilitate periodontal
wound healing. These enamelderived proteins are now commercially available
under the trade-name " E m d o g a i ~ .
D. Emdogalm
Prior to reviewing the putative role of EmdogainQD in periodontal regeneration, I
will first provide some of the relevant background that led to its dinical
development. Alrnost 20 years ago, it was suggested that the formation of new
periodontal attachment may be promoted selectively by 'guiding" periodontal
-
ligament cells into periodontal wounds (Nyman et al., 1982). It has yet to be
shown definitively that the cell exclusion methods described by Nyman and
colleagues in any way guide either tissue formation or cell migratory behaviour.
Indeed, while it has been suggested that the repopulation of wounded
periodontium with cells onginating from the periodontal ligament may result in
enhanced healing and functional repair (Boyko et al., 1981 ; Nyman et al., 1981;
Egelberg, 1987). lineage studies in which the origin of repopulating cells was
definitively demonstrated have not been conducted. Further, although it has been
suggested that only the cells located within the periodontal ligament have the
ability to regenerate the tissues of the periodontal attachment apparatus (Melcher
1976), it is equally likely that cells from contiguous endosteal spaces (McCulloch
et al., 1987) can also contribute to the repopulation of healing periodontal
wounds.
The metabolic behaviour of cells that originate from both PL and endosteal
spaces is modulated by factors (e.g.. chemokines and cytokines) that regulate
their reauitment from progenitors and may affect the outcornes of wound healing
(see above Section B). In this context, many factors including extracellular matrix
components (mg. collagen or enamel matrix proteins) may contribute to
enhanced regeneration. In this context, the major proteins of the enamel matrix
are known as arnelogenins. They are expressed at the apical end of the foming
toot (Lindskog 1 982a,b; Lindskog and Hammarstrorn, 1982) and comprise -90%
of the enamel matrix. The remaining 10% indude proline-rich enamel proteins
(Fukae and Tanabe, 1 987), tuftelin (Deutsch et al., 1 991 ), tuft proteins (Robinson
-
et al., 1975), various serum proteins and at least one salivary protein (Brookes et
al., 1995).
During root formation, Hertwig's epithelial root sheath, a derivative of the
inner cells of the enamel organ, induces the mesenchymal cells of the dental
papilla to fom the mantle predentin before it disintegrates and detaches from the
root surface. The mesenchymal cells exposed to the newly fomed dentin are
believed to induce cementogenesis (Cho and Garant, 1988). However, an
exposed dentin surface is thought to be an insuffident stimulus for œmentoblast
differentiation. Instead, it has been proposed that enamel matrix proteins may be
involved in the formation of acellular cementum during nascent root development
(Slavkin, 1976; Lindskog 1 982a, b; Slavkin et al., 1989, Hammarstrom, 1997). In
fact, there is increasing evidence that the inner epithelial cells of Hertwig's
epithelial root sheath may have the potential to produœ and secrete these
enamel-like proteins during root formation (Slavkin. 1976; Lindskog 1982a,b,
Slavkin et al., 1989, Hammarstrbm, 1 997). Further, during the initial formation of
the enamel matrix. mesenchyrnal cells exposed to these proteins differentiate
into cementoblasts. In addition, a non-cellular cementum-like tissue is formed on
the surface of enamel matrix when it is exposed to these mesenchymal cells
(Hammarstrom, 1997).
Slavkin et al. (1989) showed that acellular cementum contains proteins
that are immunologically related to proteins present in the enamel matrix. This
finding suggests that acellular œmentum is a secretory product of epithelium and
that it can only be fomed during tooth development. Formation of acellular
-
cernentum is unlikely during the adult Iife of mammals as the epithelial cells that
may regulate its formation (aside from the rests of Malasset) are no longer
present in the adult.
Alterations to œmentum structure and biochemical composition are
central factors in the periodontal disease proœss (Aleo et al., 1974; Robinson
1 975). Consequently, these changes impose significant limitations on
regenerative patential and impact on the management of the disease and
repairlregeneration treatment protocols. Hence, the application of enamel matrix
proteins to the exposed root surfaces could promote wound healing and the
regeneration of the periodontal ligament, cementum and alveolar bone. Based on
these observations, it has been hypothesized that enamel proteins play a
stimulatory role in the formation of cementum (Slavkin 1976; Hammarstrom et al.,
1 997).
Enamel matrix proteins have been relatively well conserved during
evolution (Slavkin and Diekwisch, 1996) and there appears to be a high degree
of amino acid similarity between the sequences of porcine and human enamel
proteins (Brookes, 1995). Arnelogenin is an enamel protein expressed at the
apical end of the forming root and is present in the area where cementogenesis
is initiated. An acellular cementum-like tissue is formed on the surface of enarnel
matrix when it is exposed to mesenchymal œfls of the dental follide. Thus
enamel matrix proteins are possibly involved in the development of œmentum.
With this in mind, enamel matrix derivatives have been developed as a dinical
treatment to promote periodontal regeneration. The commercial product,
-
EmdogairvID, has been described as a resorbable material and consists of
hydrophobic enamel matrix proteins bebnging to the amelogenin family extracted
from the enamel of developing teeth in porcine embryos (Heijl et al., 1997).
Emdogaim is supplied as freeze-âried enamel matrix proteins in a viscous
carrier, propylene glycol alginate. The carrier is a propylene glycol ester of alginic
acid (Gestrelius et al., 1997a).
The mechanism of action of enamel matrix derivative is not known in detail
but conceivably, the derivative mimics the role of enamel proteins in
cementogenesis during nascent root development. It appears that the temporary
deposition of enamel matrix proteins ont0 a root surface is an essential step
preceding the reformation of acellular cementum and that the formation of
periodontal ligament and alveolar bone is dependent on formation of acellular
cementum (HammarstrBm, 1997). The use of enamel proteins as an adjund in
periodontal surgery could possibly provide a 'natural" extracellular matrix for
recolonization of previously diseased mot surfaces by cells expressing a
cementoblastic phenotype.
Experimental studies have shown that aœllular cementum is fomed when
mesenchymal cells of the dental follicle are exposed to endogenous or
exogenous enamel matrix (Hammarstrom, 1997) so cementwn proteins cwld
modulate the biological activities of periodontal ligament and possibly gingival
fibroblasts. Indeed, Gestrelius et al. (1 997b) have demonstrated that periodontal
ligament cells show increased proliferation and mineralized nodule-formation in
the presence of these enamel proteins. Further, enamel matrix proteins may also
-
affect bacterial colonization of the root surfaces as the physico-chemical
properties of the environment (e.g. hydrophobicity) may modulate bacterial
adherence. Regardless of the favorable dinical results obtained in periodontal
therapy with EmdogainQB (Pontoriero et al., 1 999; Heden et al., 1 999; Sculean et
al., 1999), the biological effects and its mechanism of action are still unclear.
E. Mode1 sysfems and rationale
Elucidation of the critical regulatory factors in periodontal wound healing will likely
be predicated on the use of appropnate model systems, perhaps combining in
vitm as well as in vivo methods. In vitm cell culture analyses of periodontal cells
provide simplified approaches for understanding basic molecular mechanisms in
regenerative processes but without the interference of multiple cell types and
confounding in vivo factors (e.g. bacterial contamination). However, conclusions
from in vitro investigations may be incomplete since they cannot recapitulate the
events involved in regeneration and the cornplex intercellular communication
systems that may exist between the different types of periodontal œlls
(McCulloch, 1993). The major, and by far the most wmmonly used alternatives
are in vivo periodontal wound healing models.
A wide variety of different animal models have been used to facilitate
study of human periodontitis and its response to regenerative proœdures. The
most commonly used models employ rodents, dogs and non-human primates
(Page and Schroeder, 1982). The ability to closely replicate the periodontal
lesions of man and ultimately their utility as models for the study of human
periodontitis is a critical factor in model seledion. There are no aarently used
-
animal models of periodontitis that perfectly replicate human periodontal lesions;
indeed, it is uncertain if a perfect model will ever be found. Significant differences
between humans and animals in diets, oral habits, masticatory patterns, life-
span, tissue destruction pathways, tissue morphology, host defense mechanisms
and genetic traits underline the essential validity of this statement. Accordingly,
the diversity among animal species in susœptibility, progression, and
morphological features of periodontitis necessitate that the animal model used for
the research project be selected with great care and awareness of their
limitations.
The utilization of an in vivo model of the healing periodontium is often
complicated by the presence of bacteria and other soluble factors in the oral
cavity (e-g. salivary proteins, crevicular fluid enzymes). Consequently, the ability
to sequester experimentally these confounding variables can facilitate studies of
periodontal wound healing and simpfify the interpretation of biological outwmes.
A second and altical determinant is whether the healing wound can be
influenced by the protocol under test and whether experimentall y-induced
variations can be measured reliably. For example, critical size defeds, which do
not heal spontaneously during the lifetime of the animal, are very useful in
identifying the factors that promote wound healing since it is known that without
the experimental intervention, healing does not progress. Unfortunately, there are
no available models for critical sire defects in rodent periodontiurn. Indeed, King
et al. (1997) showed that despite variations in size, essentially al1 fenestration-
type defects created in rat molar periodontium will heal spontaneowly. As a
-
complete review of al1 animal models is beyond the swpe of this thesis, I will
briefly review three cornmonly used types of model systems that have been
employed in periodontal wound healing research.
1. Non-human primates
Monkeys, baboons and other non-human primates are models that
morphologically, most closely resemble the human dentition and periodontium.
With appropriate application of silk ligatures, the kinetics and microbiology of
disease progression can also be modeled reasonably well (Holt et al., 1988).
However, experiments using nonhuman primates are expansive, sample sizes
are generally small and ethical review boards very closely monitor the
experimental designs of non-human primate experiments. The position of
monkeys within the evolutionary hierarchy of animals and their dose physical
resemblance to man didates a careful consideration of experimental design, the
potential for animal suffering and the reasons for using them in experiments.
2. Dogs
Although they exhibit anatomical, topographical and physiological differences of
the periodontium campared to humans (Page and Schroeder, 1982). dogs
provide an excellent animal model to study gingivitis and other periadontal
diseases. Infiammatory lesions are readily induced by soft diets and several
breeds (0.g. beagles) spontaneously exhibit progressive periodontitis lesions with
increased age. The gingival lesion extends apically through the junctional
epithelium and the kinetics of increasing crevice depth are similar (but not
identical) to the deepening periodontal pocket in man and nonhuman primates
-
(Schroeder et al., 1975). Experimental periodontal regeneration in dogs has
produced results that are often difficult to interpret since some of the lesions heal
spontaneously, most notably in furcation defects (Bogle et al., 1 983). Further,
high purchase and maintenance costs are important considerations in the ability
of many investigators to obtain a substantial sample size.
3. Rodents
The teeth and periodontal tissues of rodents, such as mice, rats and hamsters,
undergo marked physiological changes throughout their relatively short life span
(Vignery and Baron, 1980; McCulloch and Melcher, 1983a). Even in teeth of
limited eruption such as the molars, any pathological changes must be
interpreted in the context of dynamic tissue remodelling that includes rapid matrix
turnover in bone, cementum and periodontal ligament. Another important
difference between rodents and man is that the sulwlar epithelium of rats (Page
and Sdiroeder, 1982) including the gen-free rat (Yamasaki et ai., 1979) is
keratinized compared to the non-keratinized sulwler epithelium of humans. This
feature may impact on the formation of pocket epithelium and the apical
extension of inflammatory cell infiltrates.
In spite of limitations related to variations in morphology and spontaneity
of healing in rats compared to humans, the rat periodontal window wound model
developed by Melcher (1 970) and further refined by Gould et al. (1977) and Lekic
et al. (1996b) provides an excellent system to study cell repopulation and cell
differentiation in the absence of oral bacteria and epithelial downgrowth. The
window wound model can be easily standardized and in the hands of
-
experienœd operators can provide relatively reproducible data with respect to
wound site, configuration and stability (Gould et al., 1 980; Lekic et al., 1996b).
While the rat model has several advantages, there are also several
drawbacks. First, the absence of bacterial biofilm formation perhaps over-
simplifies the wound healing environment since this important factor is eliminated
fmrn the model. Further, the window wound heals spontaneously over time, even
without therapeutic intervention. Nevertheless, the predictability and reliability of
the model facilitates studies of cell proliferation and cell differentiation in
response to various implanted materials on periodontal regeneration (Nguyen et
al., 1 997; King et al., 1 997, Rajshankar et al., 1 998). In this study, I have fowsed
on the impact of Emdogaina~ on cell differentiation in the repopulation response in
the wounded rat periodontal ligament. Consequently, the shortcomings of the
model described above (i.e. lack of bacterial biofilm formation, spontaneous
healing over time) do not significantly impact on the central outcomes of the
study.
-
II. Statement of the problem
Periodontal diseases are high prevalence infections that cause the destruction of
connective tissue, the loss of fibrous attachment and the resorption of alveolar
bone and cementum. If untreated, these infections can lead eventually to tooth
loss. Currently, most treatment approaches for the management of periodontitis
focus on the elimination of bacterial infection and the stabilization of the marginal
lesion. However, frequent consequences of periodontitis and of surgical
periodontal treatment include elongated dinical crowns and root exposure,
reduced periodontium and increased sensitivity to thermal stimuli. Accordingly,
despite the successes of conventional treatment, the ultimate goals of
periodontal therapy including the regeneration of connective tissue, the formation
of cementum and bone, and the attachment of new connective tissue fibers into
previously exposed root cementum (Egelberg, 1987, Aukhil et al., 1990, Polson,
1986) remain elusive.
One possible approach to enhance periodontal regeneration is to mirnic
the processes that take place dwing the development of the root and periodontal
tissues. Notably, Emdogaim is an enamel matrix derivative that may be able to
promote hard and soft tissue regeneration on the basis of its presumptive ability
to recapitulate critical events in tooth morphogenesis (Hammarstrom et al.,
1997). Indeed, some limited and very preliminary studies in the rat periodontal
window wound model have demonstrated that Emdogaim may produœ a large
increase in the volume of nascent bone and cementum matrices as early as one
week after wounding. Other data indicate that EmdogairS may greatly improve
-
the rate and nature of the regenerative process (Heijl et al., 1 997). Nevertheless,
separate studies have dernonstrated equivocal results following the topical
application of this agent (Sculean et al., 1 999. 2001 ).
Evidently, a deeper understanding of how Emdogaiw rnay promote
regeneration is essential for developing an improved biological basis for its use in
periodontal therapy, and for optirniring protocols that may lead to favorable and
predictable outwmes. Currently, the effects of Emdogaim on the differentiation
of cells in regenerating periodontal tissues are not known and the molecular
mechanisms by which this agent may prornote wound healing are poorly
understood. My hypothesis is that Emdogaiw facilitates regeneration of
periodontal tissues by promoting the differentiation of cells that are required for
the synthesis of new periodontal ligament, bone and cementum matrices.
To test my hypothesis, I have framed the following objectives to
investigate the efFed of Emdogaiw on the periodontium using the rat periodontal
window wound rnodel:
1-Ta compare the arnount of new bone and cementum as well as the
width of the periodontal ligament space following EmdogaiW or vehide
treatment of periodontal defects.
2-To study cellular differentiation following wounding and topical
application of Emdogaiw using intracellular and extracellular expression of
osteopontin, bon8 sialoprotein, osteocalcin, and a-smooth muscle actin as
markers of cellular differentiation.
-
III. Materials and Methods
Twenty-seven male CBL Wstar rats (90-115 g) were obtained from Charles
River mlmington, MA). The vehicle control (propylene glycol alginate) and
Emdogaim (stock concentration of 30 mglml) were obtained from BlORA AB
(Malmci, Sweden). '~pro l ine (specific activity=24 Cilmmol) at a final injected
concentration of 1 pCilg body weight was provided by Mandel Sdentific (Guelph,
ON). Mouse monoclonal antibodies to osteopontin (OPN; clone # MPIIIBI O) and
bone sialoprotein (BSP; Clone # WVIDl [9C5]) were obtained from the
Hybridoma Bank, Johns Hopkins University, Baltimore, MD. Mouse monoclonal
antibody to a-smooth muscle actin was obtained from Sigma Chemical (Clone #
lA4; Oakville, ON). Polydonal rabbit antibody to osteocatcin (OC) was obtained
from Dr. William Butler (University of Texas, Houston, Texas). Kodak nudear
track liquid emulsion (NTB-2) for radioautography was obtained from Kodak
(Eastman Kodak Co., Rochester, NY).
A. Wound model
The periodontal window wound model originally described by Melcher (1970) and
modified later by Gould et al. (1977) and Lekic et al. (1 996a,b) provides a
repopulating wound in which periodontal ligament cells are recruited to
regenerate spontaneously alveolar bone, cernenturn and periodontal ligament
(Le. non&tical size defect). The model facilitates studies of periodontal cellular
differentiation since a relatively well-syndwonized cohort of connective tissue
œfls proliferates and subsequently differentiates during the repopulation
-
response. Rats were caged in pairs in a room with a 12-hour darWlight cycle and
provided with food and water ad libitum. A total of 27 rats were included in these
experiments. Animals were anesthetized with Halothane (1.3%) and nitrous
0xide:oxygen (2:l). An incision -1 cm in length was made through the skin
overlying the incisor trunk. The posterior masseter muscle was identified and
retracted to locate the mental nerve that was dissected free and the anterior
fibers of the masseter muscle were incised at their point of insertion into the
mandible to expose the underlying bone. A rnodified end-aitting bur (0.6 mm in
diameter) driven by a slow speed dental hand-piece was used to drill a hole with
a final diameter of -0.8-1.0 mm through the alveolar bone over the mesiobuccal
root of the mandibular first molar. The hole was located mid-way between the
gingival rnargin and the mental nerve and was -1 mm posterior to the anterior
edge of the mandible. This hole extended to the most lateral surface of the
periodontal ligament but did not actually penetrate the soft tissue of the
periodontal ligament. W~th the aid of a dissecting microscope (Wild M3Z 10X)
and a 27 gauge needlq the periodontal ligament was extirpated dom to the level
of cernentum and the wound was cleared of debris with saline and a wet gauze.
Before the wound site was closed, either the vehide (control) or EmdogainQ54 at a
concentration of 3 mglml or 30 mglml (see below), was placed into the defect but
without overfiowing the defed. For al1 animals, wounds were performed on both
the left and the right mandibular first molars. Finally, the tissues were closed with
4 0 V i q î intempted sutures that resorbed spontaneously.
-
B. Preparation of implants and experimen fa1 aldesign
EmdogaiM was distributed in increments of -0.3 g each. The propylene glycol
alginate was prepared as 10 ml or 7 0 0 ml aliquots. An hour before the surgical
procedure was begun, the enamel matrix derivative was mixed with either 10 ml
or 100 ml of the propylene glycol alginate to yield an Emdogaiw preparation with
a concentration of 30 mglml or 3 mglml, respedively. The wound sites were
either: 1) implanted with the propylene glycol alginate (vehicle contrd); or 2)
implanted with EmdogainO at a concentration of 30 mglml; or 3) implanted with
EmdogairVB at a concentration of 3 mglml.
To study the effects of EmdogainQ3 on wound healing over time, three
animals (6 sides) for each experimental condition (vehicle control and
EmdogaiM3 at a concentration of 30 mglml or 3 rnglml) were sacrificed by CO2
asphyxiation at 7, 14 and 21 days following surgery. These time periods were
chosen on the basis of previous experiments (Lekic et al., 1996) showing that
these time periods correspond to the early proliferative stage of healing, to the
matrix formation stage of healing and finally to the completion of healing. One
and three days prior to sacrifice, one rat belonging to each of the three
experimental groups and from each of the three time periods studied, was
injected intraperitoneally with '~groline diluted with saline to produce a final
injectable concentration of 1 pCi/g body weight in a total volume of 2 ml. Injection
of the '~grol ine at 2 time periods facilitated assessrnent of the rate of matrix
synthesis in the healing tissues.
-
C. Tissue preparation
Following sacrifice, the mandible was removed, cleared of any attached soft
tissues and tnmmed at the mid-incisor and at the third rnolar region. Tissues
were fixed in 4% paraformaldehyde in phosphate buffer saline (PBS) at pH 7.4
for 24 hours at 4OC, demineralized for 24 hours in 0.2N HCI at rcom temperature
and washed in PBS at pH 7.4 for 24 hours, also at room temperature. Thereafter,
the specirnens were dehydrated by washing in graded ethanol solutions, cleared
in toluene and embedded in paraffin. Sections of -5 pm in thickness were cut
transversally to the longitudinal axis of the tooth with a Leitz microtome (mode1
#1512). Sections that were located closest to the drill site were stored on trays
and every fifteenth section was stained with toluidine Mue to determine the exact
location of the wound site. Sections in the middle of the wound sites were
attached ont0 Superfrost Plus@ slides (Fisher Scientific, Toronto, ON) and used
for immunohistochemical (Chen et al., 1991 b), morphornetric and
radioautographic analyses (Lekic et al., 1 996c).
D. lmmunohistochemistry
For immunohistochernical analyses, the sections were dewaxed in xylene (4 X
1.5 min) and rehydrated in a series of graded ethanol solutions (2 X 1.5 min in
100% ethanol, 1 X 2 min in 95% ethanol and 1 X 2 min in 70% ethanol). Slides
were washed in water (i X 5 min) as well as in PBS at pH 7.4 (2 X 5min). The
slides were subsequently incubated at room temperature with 3% H202 in
methanot for 30 min and protected from light. This step enabled inactivation of
-
endogenous peroxidase activity. Sections were washed with PBS at pH 7.4 (2 X
5 min).
To decrease non-specific background staining, sections were incubated
with a serumcontaining casein blocking solution (Rajshankar et al., 1998) for
one hour in a moist chamber, at room temperature. When staining for
osteopontin, bone sialoprotein and a-smooth muscle actin, mouse and horse pre-
immune sera were used at a dilution of 0.1% (vlv) in the blocking solution. For
osteocalcin staining, rabbit and goat pre-immune sera were used at a dilution of
0.02% (vfv) in the blocking solution (Vector, Burlingame. CA).
Sections were subsequently incubated at roorn temperature with one of
the following prirnary antibodies: mouse monoclonal antiosteopontin (1 : 1 Oûû),
mouse monoclonal anti-bone sialoprotein (1:500), rnouse monoclonal antia-
smooth muscle actin (1 :100) or rabbit polydonal antiosteocalcin antibody (150)
diluted with antibody diluting bMer (DAKO Diagnostics Laboratories,
Mississauga, ON) for 1.5 hours in a moist chamber. The sections were washed
with PBS at pH 7.4 (3 X 5 min) and incubated with secondary antibody diluted
with antibody diluting buffer (DAKO Diagnostics Laboratories. Mississauga, ON)
for 30 min in a moist chamber, also at room temperature. The secondary
antibody was biotinylated horse anti-mouse (1:2W) for osteopontin and bone
sialoprotein, and the same antibody was used at a dilution of 1 :A00 for staining of
a-smooth muscle actin. For osteocalcin, biotinylated goat anti-rabbit (1 50) was
used. Sections were washed with PBS at pH 7.4 (2 X Smin) and incubated at
room temperature with streptavidin horseradish peroxidase (PK-6100,
-
Vectastain) for 30 min, in a moist chamber. The sections were washed with PBS
at pH 7.4 (2 X 5 min) and incubated with diaminobenzidine (DAB) (SK-4100,
Vector, Burlingame, CA) for 15 min. The color reaction was stopped by gentle
rinsing with water ovemight. Finally, the sections were counter-stained with
hematoxylin and eosin, mounted in Penount, coverslipped and examined with a
Bioquant image analyzer (see below).
E. Radioautography
Radioautographs were prepared from the specimens labeled with 'H-proline
using the dipping method (Lekic et al., 1996~). First, slides were dewaxed in
xylene (4 X 1.5 min.) and rehydrated in a series of graded ethanol solutions (2 X
1.5 min in 10O0h ethanol, 1 X 2 min in 9S0h ethanol and 1 X 2 min in 70%
ethanol). The sections were air-dried for about one hour and dipped in Kodak
NTB-2 (Eastman Kodak Co., Rochester, NY), plaœd in light-proteded, dry boxes
and exposed for 2 weeks at 4OC. Following exposure, slides were developed for
6 minutes in D-19 developer (Eastman Kodak). This reaction was stopped by
placing the slides into 30°h ethanol for 30 sec and fixed for 5 min. The slides
were washed with water and stained with hematoxylin and eosin through the
emulsion.
F. MorphomeMc ana&ses
For specimens stained with hematoxylin and eosin and in sections
immunostained for osteopontin, several measurements of tissue structure were
-
made to assess the effect of Emdogaim on healing. Sections were examined
with a Leica Orthoplan microscope equipped with a drawing tube and analyzed
with a computerized morphornetry program (Bioquant). From each wound block,
three sections separated by at least 100 pm were examined. In each section, two
square sampling gnds of -40,000 were superimposed over the wound
compartment. First, the reversal line in the bone at the ait edges of the wound
was digitized in osteopontin-immunostained sections. This gave an estimate of
the original wound margins in the alveolar bone overlying the wound site. The
areas of new bone produœd within the defed were traced. These two areas
were then used to wmpute the percentage of new bone area fomed in the
wound. Next, the width of the periodontal ligament of wounded and contralateral
unwounded sites was measured and a ratio computed. This measurement
estimates whether possible growth of bone into the periodontal ligament space
(and hence loss of periodontal ligament homeostasis) has occurred. Finally, the
thickness of cementum on wounded and unwounded sides was deterrnined and
a ratio cornputed. The various measurements obtained from the wound side were
compared to the unwounded contralateral side of the rwt.
Analyses of immunostaining for intracellular Golgilendoplasmic retiwlum
proteins (OPN, BSP, OC or aSMA-stained cells) and for extracellular matrix
proteins (OPN, BSP or OC) were conducted in the regenerating bone
compartment of the wound by digital rnorphometric analyses (Lekic et al.,
l996a, b, 1 997) and by stereological procedures (McCulloch et ai., 1990).
Measurement of these proteins has been used to identify disaete stages in the
-
formation of bone in vivo (Chen et al., 1992; Yoon et al., 1 987) and for studies of
cell differentiation in periodontal tissues (Lekic et al., 1996b). The identification of
cells with intracellular immunostaining for these proteins can suggest the
presence of cells committed to the osteogenic or cementogenic lineages (Chen
et al., 1992.1993; McKee et al., 1 993) or to the fibroblastic lineage (for a-smooth
muscle actin; Arora and McCulloch, 1994). An intraocular grid system with 100
squares was superimposed over the wounded site to facilitate counting. Cells
with intracellular staining present within the gnd were wunted to provide
quantitative estimates of cell differentiation within the repopulating zone of the
PL. Total cell counts present within this grid were also obtained. These two cell
counts were then used to compute the percentage of cells with intracellular
staining for each marker present in the wound. Second, the presenœ or absence
of staining in the matrix for OPN, BSP or OC within each of the 100 squares of
the grid was counted and expressed as a %. This datum gives an estimate of the
volume density of nascent bone matrix produced by osteogenic cells. Previous
detailed threedimensional analyses of mineralized tissues have shown that
counting the number of cells that stain for a specific protein within a coherent grid
system allows the abundance of the cells present in a tissue to be estimated from
a 24imensional tissue section (McCulloch et al., 1990). Further, application of a
coherent grid system for estimating the relative abundance of staining for a
specific protein from a tissue section also gives an estimate of the perewitage of
the tissue volume that contains the protein (Rajshankar et al., 1998).
-
Analyses of radioautographs allowed measurement of matrix protein
synthesis rates by labeling nascent bone and cementum matrices (McCulloch
and Heersche 1988). In transverse sections, the distance between the two
labels, measured by morphometry, gives an estimate of matrix appositional rates
and of the extent of healing in the cementum and bone compartments. Grain
counts pet 1000 prd were also cornputed to provide an estimate of the
incorporation of labeled proline into matrix and these counts were adjusted by
subtracting background counts obtained over mature dentine.
G. Statistical analyses
Two-factor analysis of variance was performed to evaluate the differences
between the three dnig treatment groups over time after wounding. Further,
possible interactions between treatrnent and time were tested. The SAS system
(Cary, NC) was used to analyze the data. Analyses were wnducted for
measurements of tissue structure, immunostaining assessments and matrix
protein synthesis rates in the different sites examined. The data from each rat
(n=6) were pooled to obtain an animal-specific mean value. These animal means
were considered as single measures that were representative of each animal.
The results were reported as meanzstandard error of the mean where n-3
animals for the experimental groups and treated controls. Individual differences
between groups and time of sacrifice for the various experimental outcornes were
assessed by Tukey's test. A value of pc0.05 was considered statistically
significant.
-
IV. Results
A. ?%riodonial ligament homeostasis
The possible growth of bone into the periodontal ligament space and henœ the
loss of periodontal ligament homeostasis was assessed by measuring the width
of the periodontal ligament of both wounded and unwounded sides and by
computing the ratio of these widths. A ratio of greater than one indicates that the
periodontal ligament width of the wounded side is wider than that of the
unwounded side and may suggest ingrowth of osteogenic cells. At 7 days, rats
treated with either the vehicle control or Erndogairm at 3 mglml exhibited a
slightly wider periodontal ligament space than unwounded sites (Figure 1 A,B).
By 14 and 21 days, this ratio was very close to unity, indicating that homeostasis
of the periodontal ligament was restored. In uintrast, defects treated with
Emdogaiw ai 30 mglml showed widening of the periodontal ligament width at 7
days (p
-
Emôogmn 3 me(ml Emdogain 30 mglm1
Vehicle umtrol m Emdogain 3 mgiml rn EmQgain 30 mglml - --
74
m m (da- ' F
-
Figure 1.
Morphometric analyses.
A) lmmunohistochemical staining for osteopontin in vehicle control-treated
defect. Rat sacrificed at 14 days after wounding. Cementum (C), periodontal
ligament (PL) and bone (B). Note growth of bone into wound defect and
restoration of periodontal ligament width. Magnification 100X. 6 ) Histogram
showing ratio of periodontal ligament width of experimental (wounded)
defectslcantrol (non-wounded) sides. Defects were treated with vehicle control,
or EmdogaimB at 3 or 30 mglml. Rats were sacrifiœd at 7, 14 and 21 days after
wounding. Data are meanskSEM of ratios (n=6). *-pcO.001 wmpared to vehicle
control and to Emdogaim (3 mglml) at 7 days. C) Imrnunohistochemical staining
for bone sialoprotein on non-wounded side at 7 days after wounding. Cementum
(C), periodontal ligament (PL) and bone (B). Note pmminent staining for bone
sialoprotein in cementum. Magnification 4WX. D) Histogram showing ratio of
cementum thickness of experimental (wounded) defecWcontrol (non-wounded)
sides. Defects were treated with vehicle control, or EmdogaiM 3 or 30 mgtml.
Rats were sacrificed at days 7, 14 and 21. Data are meanstSEM (n=6).
*-p
-
B. Cernentum
The effect of Emdogairm on cementurn formation was deterrnined by measuring
the thickness of cementum on the wounded and unwounded sides and by
wmputing a ratio of these thicknesses. mus, the cementum thickness on the
wounded side was normalized to the relatively constant width of the cementum
on the unwounded side within each section. A ratio of less than one indicated
that the œmentum thickness on the wounded side is less than that of the
unwounded side and reflects loss of cementum at early time periods after
healing. This morphometric approach obviates systematic errors that may arise
because of section orientations deviating from a perpendicular orientation to the
longitudinal axis of the tooth. During wounding and extirpation of the periodontal
ligament, the cementum was scored with the drill and resorptian of cementum
ocairred as part of the healing process. Consequently, at 7 days, the ratio of the
cementum thickness was well below one for al1 groups. indicating that œmentum
formation was limited (Figure 1 C,D). As expected in normal periodontal wound
healing, the cementum thickness increased neady 2-fold between 7 and 14 days
(p
-
Figure 2.
Analyses of immunostaining for intracellular osteopontin and extracellular
matrix osteopontin.
A) Defect treated with Emdogaim (3 mglml) in rat sacfificecl at 7 days; cell with
intracellular staining for osteopontin (arrows) adjacent to cut bone margin of
wound defect. Magnification 400X. 6) Histogram showing perœntage of cells in
wound site staining for osteopontin/total cell count. Defects were treated with
vehicie control, or Erndogaiw at 3 or 30 mglml. Rats were sacrificed at 7, 14 and
21 days. Data are meanskSEM (n=6). C) Vehicfe control-treated defect in rat
sacrificed at 21 days showing extracellular staining for osteopontin (arrows) at
reversal lines. Magnification 250X. D) Histogram showing stereological analysis
of osteopontin staining in matrix. Defects were treated wWi vehicle control, or
Emdogaiw at 3 or 30 mglml. Rats were sacrificed at days 7, 14 and 21. Data
are means2SEM (n=6). - peO.01 comparing matrix staining for bone sialoprotein between 7 and 14 days.
-
while much lower effects were due to Emdogaim and even less to interanimal
variation. The coefficient of variation of the rnodel was 40%.
C. Bone
The area of new bone area fomed within the wound area was assessed by
digital morphometry of osteopontin-stained sections. Osteopontin-stained
reversal lines in the bone were used to identify the cut bone margin and
consequently, served to locate the original wound edges. Only limited bone
formation ocairred by 7 days after wounding in al1 3 groups (Figure 1 E,F). By 14
days, there was a dramatic IO-fold increase in bone area (p0.4)
at any of the three time periods. As shown by ANOVA, virtually al1 of the variation
in the bone measurements was attributable to time-dependent effects of healing
while there was virtually no effect because of Emdogaim, inter-animal variation
or interactions between time and dnig. The CO-efficient of variation of the ANOVA
model was 4 4 O h .
O. Osteogenic dHemntiaüon
lrnmunostaining for intracellular Golgilendoplasmic reticulum proteins and for
extracellular matrix proteins was conducted in the bone compartment of the
wound (Lekic et al., 1996a,b, 1997). The percentage of cells with intracellular
staining for osteopontin, a m a r k of early osteogenic differentiation, was sparse
at al1 time periods after wounding (-2%) and there was no difference in the % of
-
- . - . . - . - - - - -- . -. - . - - - - -- - . - - - - - - --- 0 Vehicle coritroi - 6 . œ EmdoQoin 3 mgiml c I Emdogain 30 mdml 1
Timm (dry.)
- -- O Vehicb cmtrd
100 r Emdogain 3 mgiml Y 0 . Emcbgain 30 mgiml - - - - -- - --
-
Figure 3.
Analyses of immunostaining for intracellular bone sialoprotein and
extracellular matrix bone sialoprotein.
A) Defect treated with Emdogaim at 30 mglml in rat sacrificed at 14 days after
wounding. Cell with intracellular staining for bone sialoprotein (arrow) in middle of
PL. Magnification 400X. B) Histogram of percentage of cells in wound site
staining for bone sialoproteinltotal cell count. Defects were treated with vehide
control, or Emdogaim at 3 or 30 mglml. Rats were sacrificed at 7, 14 and 21
days after wounding. Data are meanszSEM (n=6). *-pc0.05 wmparing defects
treated with Erndogaiw 3 and 30 mglml at 14 and 21 days after wounding. C)
Defect treated with Erndogaim at 30 mglml in rat sacrificed at 14 days after
wounding showing extracellular staining for bone sialoprotein (arrows) in nascent
bone matrix. Magnification 250X. D) Histogram of stereological analysis of bone
sialoprotein staining in matrix. Defects were treated with vehicle wntrol, or
Emdogaiw at 3 or 30 mglml. Rats were sacrificed at 7, 14 and 21 days after
wounding. Data are meanstSEM (n=6). -- pcO.01 comparing rnatrix staining for bone sialoprotein between 7 and 14 days; *-pe0.05 comparing defects treated
with Emdogaim 3 and 30 mglml at 7 and 21 days.
-
celfs with intracellular osteopontin staining between the three treatment groups at
any of the three time points (Figure 2 48). Stereological methods were also
used to estimate the volume density of nascent matrix in the wound site that was
imrnunostained for osteopontin. As expected. the time after wounding strongly
affected the expression of osteopontin in the matrix and there was a 5-fold
increase of osteopontin-stained matrix between 7-14 days (pc0.01) but there was
no difference between Emdogain treatments. 8y 21 days, -80% of th8 matrix
was stained with osteopontin. However. Emdogaim. even at 30 mglml, exerted
no detectable effect on extracellular osteopontin staining at al1 three time points
(Figure 2 D).
A small % of cells (1-4%) exhibited intracellular staining for bone
sialoprotein. a marker of the minerakation stage of osteogenic differentiation. At
14 and 21 days. there were apparent differences between the vehicle control-
treated defeds and the defeds treated with Emdogaiiwp at 3 and 30 mglml
(Figure 3 A,B) but these d-ifferences were small and of marginal statistical
significance (p=0.05). For extracellular bone sialoprotein staining in nascent
matrix, there was a nearly 4-fold increase of staining between 7-1 4 days (p
-
-. - - - - -. - - . - - -
6 . )Emdogain 3 mglml m Emdonein 30 mdml
-
Figure 4.
Analyses of immunostaining for intracellular osteocalcin and extracellular
matrix osteocalcin.
A) Defect treated with Emdogaim at 3 mglml; rat sacrificed at 21 days showing
an osteocyte with intracellular staining for osteocalcin (arrow). Magnification
400X. B) Histogram of percentage of cells in wound site staining for
osteocalcinltotal cell count. Defects were treated with vehide control or
Emdogain@ at 3 or 30 mglml. Rats were sacrifiiced at 7, 14 and 21 days after
wounding. Data are means5SEM (n=6). *-p'-0.01 comparing defects treated with
vehicle control or Emdogaiw (30 mglml) at day 7. C) Defect in vehicle control-
treated rat saaificed at 21 days depicting extracellular staining for osteocalcin
(arrows) in wound defect. Magnification 250X. D) Histogram showing
stereological analysis of osteocalcin staining in matrix within the wound site.
Defects were treated with vehicle control or Emdogaim at 3 or 30 mglml. Rats
were sacrificed at 7, 14 and 21 days after wounding. Data are meanszSEM
(n=6). l p 4 . 0 1 comparing defects treated with EmdogainQB (30 mglml) at day 7.
-
for osteocalcin was small (0.53%) for al1 sampling periods (Figure 4 A,B) and
most of these stained cells appeared to be osteocytes. There were some
variations in the % of osteocatcin-positive cells in the different groups at the
various time points of sampling. At 7 days after wounding, there was a very low
% of immunostained cells in the Emdogaim 3 mglml group (p
-
Time (da*
-
Figure 5.
Analyses of immunostaining for intracellular a-smooth muscle actin.
A) Defect treated with Emdogaim (3 mglml) in rat sacrificed at 14 days after
wounding. Cementum (C) and bone (B). Fibroblasts (F) and blood vessels (BV)
stain for a-smooth muscle actin. Magnification 4ûOX. B) Histogram of percentage
of cells in wound site staining for a-smooth muscle actiriltotal cell count. Defects
were treated with vehicle cuntrol or Emdogaim at 3 or 30 mglml. Rats were
sacrificed at 7, 14 and 21 days after wounding. Data are rneanszSEM (n=6). ' pe0.05 camparing intracellular staining for a-smooth muscle actin between 7 and
21 days after wounding.
-
significantly by 21 days after wounding for al1 groups (