Central Role of Pyrophosphate in Acellular Cementum
Brian L. Foster1*, Kanako J. Nagatomo2, Francisco H. Nociti Jr.1,3, Hanson Fong4, Daisy Dunn2,
Anne B. Tran1, Wei Wang5, Sonoko Narisawa5, Jose Luis Milla ´n5, Martha J. Somerman1
1Laboratory of Oral Connective Tissue Biology, National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), National Institutes of Health (NIH), Bethesda,
Maryland, United States of America, 2Department of Periodontics, University of Washington School of Dentistry, Seattle, Washington, United States of America, 3Division
of Periodontics, School of Dentistry at Piracicaba, State University of Campinas, Piracicaba, Sa ˜o Paulo, Brazil, 4Materials Science and Engineering, University of
Washington, Seattle, Washington, United States of America, 5Sanford Children’s Health Research Center, Sanford-Burnham Medical Research Institute, La Jolla, California,
United States of America
Background: Inorganic pyrophosphate (PPi) is a physiologic inhibitor of hydroxyapatite mineral precipitation involved in
regulating mineralized tissue development and pathologic calcification. Local levels of PPiare controlled by antagonistic
functions of factors that decrease PPiand promote mineralization (tissue-nonspecific alkaline phosphatase, Alpl/TNAP), and
those that increase local PPi and restrict mineralization (progressive ankylosis protein, ANK; ectonucleotide
pyrophosphatase phosphodiesterase-1, NPP1). The cementum enveloping the tooth root is essential for tooth function
by providing attachment to the surrounding bone via the nonmineralized periodontal ligament. At present, the
developmental regulation of cementum remains poorly understood, hampering efforts for regeneration. To elucidate the
role of PPiin cementum formation, we analyzed root development in knock-out (2/2) mice featuring PPidysregulation.
Results: Excess PPi in the Alpl2/2mouse inhibited cementum formation, causing root detachment consistent with
premature tooth loss in the human condition hypophosphatasia, though cementoblast phenotype was unperturbed.
Deficient PPiin both Ank and Enpp12/2mice significantly increased cementum apposition and overall thickness more than
12-fold vs. controls, while dentin and cellular cementum were unaltered. Though PPiregulators are widely expressed,
cementoblasts selectively expressed greater ANK and NPP1 along the root surface, and dramatically increased ANK or NPP1
in models of reduced PPioutput, in compensatory fashion. In vitro mechanistic studies confirmed that under low PPi
mineralizing conditions, cementoblasts increased Ank (5-fold) and Enpp1 (20-fold), while increasing PPi inhibited
mineralization and associated increases in Ank and Enpp1 mRNA.
Conclusions: Results from these studies demonstrate a novel developmental regulation of acellular cementum, wherein
cementoblasts tune cementogenesis by modulating local levels of PPi, directing and regulating mineral apposition. These
findings underscore developmental differences in acellular versus cellular cementum, and suggest new approaches for
Citation: Foster BL, Nagatomo KJ, Nociti FH Jr, Fong H, Dunn D, et al. (2012) Central Role of Pyrophosphate in Acellular Cementum Formation. PLoS ONE 7(6):
Editor: Songtao Shi, University of Southern California, United States of America
Received March 30, 2012; Accepted May 9, 2012; Published June 4, 2012
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for
any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: This research was supported in part by the Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases
(NIAMS) of the National Institutes of Health (NIH). Grants R01DE15109 (MJS), R01 AR47908 and R01 DE12889 (JLM) were received from NIH. The funders had no
role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com
The mineralized tissues of the teeth and skeleton are subject to
homeostasis of inorganic phosphate (Pi) for normal development
and maintenance . The hydroxyapatite (HAP) deposited to
mineralize these hard tissues is a compound of Pi and ionic
calcium. Pyrophosphate (PPi), composed of two molecules of Pi,
functions as a pivotal regulator of physiological mineralization and
pathologic calcification by acting as a potent inhibitor of HAP
crystal precipitation [2–5]. Though the potential for PPito inhibit
biological mineralization is clear from in vitro experiments, the in
vivo role and regulation of PPihas been more difficult to elucidate.
Through study of the heritable conditions such as hypophospha-
tasia (HPP), spontaneous mutations, and directed gene ablations in
mouse models, the key regulators of PPihave been identified, and
their roles in shaping mineralized tissues have been partially
defined. As measurement of PPiin vivo at mineralization fronts is
not possible, the analysis of cellular proteins that manufacture,
transport, or degrade PPihas served to clarify the mechanisms for
PPimodulation, in conjunction with in vitro experiments.
Local tissue concentrations of PPiare controlled by a number of
regulatory enzymes and transporters. Tissue nonspecific alkaline
phosphatase (TNAP) is an ectoenzyme capable of hydrolyzing PPi
and providing Pi. TNAP is expressed by mineralizing cells of
bones and teeth, and is critical for proper skeletal mineralization
[5,7]. Hydrolysis by alkaline phosphatase activity (ALP) thus
PLoS ONE | www.plosone.org1June 2012 | Volume 7 | Issue 6 | e38393
provides a mechanism for clearance of PPi, allowing mineraliza-
tion to proceed. Loss of function mutations in the TNAP gene Alpl
cause hypophosphatasia (HPP), a disease marked by poor bone
mineralization, rickets, and osteomalacia, as well as tooth
phenotypes [8,9]. Ablation of the homologous mouse gene Alpl
(formerly Akp2) produces a phenotype consistent with increased
PPi and mineralization disorders of infantile HPP [10,11].
Conversely, two factors have been identified which increase local
PPiin tissues. The progressive ankylosis gene (Ank; Ankh in humans)
encodes a multipass transmembrane protein that regulates
transport of intracellular PPito the extracellular space [12–14].
Ectonucleotide pyrophosphatase phosphodiesterase 1 (NPP1;
encoded by the Enpp1 gene) also works to increase extracellular
PPiby hydrolysis of nucleotide triphosphates . PPiremoval by
ALP activity thus antagonizes provision of PPiby ANK and NPP1,
thereby creating a concerted regulation of Pi and PPi levels
(Figure 1), and ultimately, mineralization [16,17].
The cementum covering the tooth root provides attachment for
the tooth proper to surrounding alveolar bone, via the non-
mineralized periodontal ligament (PDL) [18–20]. Cementum was
first linked to PPi metabolism by the condition HPP, where
premature tooth exfoliation was discovered to result from
developmental cementum aplasia or hypoplasia, and thus poor
periodontal attachment [7,21,22]. Intriguingly, studies to date
suggest the acellular cementum (acellular extrinsic fiber cemen-
tum, AEFC) of the cervical portion of the root is severely affected
by PPidysregulation, while the apically located cellular cementum
(cellular intrinsic fiber cementum, CIFC) is unaffected, or much
less so [21,23]. Proper cementum formation is critical for
dentoalveolar function, though cementogenesis remains poorly
understood in terms of associated cells and regulatory factors
involved. This is especially true in regard to differences between
the acellular and cellular varieties, and how cementum differs
developmentally from other hard tissues, bone and dentin [18,19].
To address how the process of cementogenesis is shaped by PPi
metabolism, a set of studies was designed that employed in vivo
transgenic mouse models featuring disrupted PPiregulation, as
well as in vitro approaches using a cementoblast cell line for further
In order to develop a comprehensive understanding of how PPi
regulates tooth root development, we performed a detailed
histological study of developing first mandibular molars and
incisors of mice harboring homozygous knock-out (2/2) of Alpl
(high PPi), Ank, or Enpp1 (low PPi), compared to age-matched
homozygous wild-type (+/+) controls. Days were selected to capture
developmental time points of interest during molar root formation,
i.e., during acellular cementogenesis (14 days postnatal, dpn), at
completion of the root and following cellular cementogenesis
(26 dpn), and after more than a month in occlusion (60 dpn).
Alpl2/2mice were limited to a maximum age of 21 dpn because of
shortened lifespan. Morphological observations on H&E stained
sections were paired with in situ hybridization (ISH) and
immunohistochemistry (IHC) for selected mineralized tissue-
Acellular cementogenesis requires diminution of
In the infantile form of HPP, the skeleton is properly
mineralized at birth, but postnatal skeletogenesis is compromised
. Alpl2/2mice phenocopy aspects of infantile HPP, where loss
of TNAP was previously reported to have little effect on bone until
postnatal day 6 [10,24]. At 14 dpn, the majority of alveolar and
mandibular bone in Alpl2/2mice was well developed, though
signs of hyperosteoidosis were noted in the bone adjacent to the
molar root (Figure 2A and B). In Alpl+/+molars, acellular
cementum (AEFC) covered the root dentin as a thin and uniform
basophilic layer. Alpl2/2molars were marked by disruption of
acellular cementum, visible as reduction of the basophilic layer
(cementum aplasia or severe hypoplasia) and direct contact of PDL
cells and tissues with dentin. By 21 dpn this cementum defect was
sometimes associated with tearing at the PDL-AEFC interface,
suggesting poor integration of Sharpey’s fibers at the root surface
(not seen at the PDL-bone interface) (Figure 2C and D) and
consistent with HPP case reports observing premature tooth
exfoliation. This is not likely to be a processing artifact, as
infiltrating cells were present in the tear zone. These results agree
with AEFC disruption described in this Alpl2/2model , as
well as a different TNAP loss-of-function mouse .
To further investigate the mechanism for the cementum defect
in Alpl2/2mice, IHC was performed for two cementum markers,
extracellular matrix (ECM) proteins bone sialoprotein (BSP) and
osteopontin (OPN), which are present at high concentrations in
acellular cementum of controls (Figure 2E and G). Both BSP and
OPN immune localization were disrupted on the Alpl2/2root
surface (Figure 2F and H), compared to the strong, even staining
on Alpl+/+controls. Scanning electron microscopy (SEM) provided
improved resolution to explore the root surface. While Alpl+/+
molars displayed a cementum layer on the root dentin surface, this
layer was absent in the Alpl2/2molar (Figure 3). The disruption of
cementum initiation and concomitant lack of BSP and OPN
localization supports the hypothesis that high PPiin Alpl2/2is
acting to inhibit cementogenesis and HAP apposition on the root
Attenuation of pyrophosphate increases acellular
Both Ank and Enpp12/2mice are deficient in extracellular PPi,
though by different mechanisms. In molars of both null mice at
Figure 1. Pyrophosphate homeostasis in the extracellular
space. Inorganic phosphate (Pi) is a component of mineral hydroxy-
apatite (HAP), while pyrophosphate (PPi) is a potent inhibitor of HAP
crystal precipitation and growth. The enzyme tissue nonspecific alkaline
phosphatase (TNAP) hydrolyzes PPi to release ionic Pi, creating
conditions conducive for mineralization. Local PPiis increased by the
functions of the progressive ankylosis protein (ANK) and ectonucleotide
pyrophosphatase phosphodiesterase 1 (NPP1), which act to keep the
mineralization process in check.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org2 June 2012 | Volume 7 | Issue 6 | e38393
14 dpn, the developing cervical cementum was expanded (hyper-
cementosis) compared to Ank and Enpp1+/+controls (Figure 4A–
C). At the completion of root development at 26 dpn, both Ank
and Enpp12/2molars featured a nearly identical cementum
phenotype where cervical cementum width was expanded several
fold over controls (Figure 4D–F). This thick cervical cementum
included numerous cell inclusions in the matrix, in a region that is
typically acellular type cementum (AEFC). Intriguingly, for both
homozygous knock-out models, apical cementum (CIFC) was not
morphologically different from controls (Figure 4G–I), PDL space
remained unmineralized, and dentin was not altered compared to
Ank and Enpp1+/+mice.
The incisor in the mouse is divided into a (labial) crown analogue
featuring enamel, and a (lingual) root analogue featuring strictly
AEFC type cementum. Notably, histological changes in Ank and
Enpp12/2incisors paralleled those in molars, featuring expanded
cementum (Figure 4J–L). Sagittal sections of the mandible allowed
observation of all three molars. Loss of Ank affected all molars
similarly, with thickened cementum evident on all root surfaces
compared to controls (Figure 4M–R). The fact that acellular
cementum on all murine teeth was similarly affected by reduced PPi
supports this as a central molecular regulator of cementogenesis
which is not tooth- or stage-specific in its influence.
Both Ank and Enpp12/2mice featured a hypercementosis
phenotype, indicating both PPiregulators function in controlling
cementum formation. Comparative analysis between Ank and
Enpp12/2and their respective controls was accomplished by
measuring the growth rate of cervical cementum over time.
During early root formation between 14 and 26 dpn, Ank and
Enpp12/2molars featured at least 10-fold greater cementogenesis
compared to controls (Figure 5A). Ank and Enpp12/2cementum
continued to increase at a rate of 0.2–0.7 mm/day from 26 to
60 dpn, while over the same period, controls featured tightly
controlled apposition, growing at the much slower pace of 0.01–
While cementum was dramatically affected by loss of ANK or
NPP1, dramatic changes in other tissues were not observed.
Histomorphometry at age 26 dpn was performed to measure
cross-sectional widths to determine if PDL and alveolar bone were
affected. Cementum was significantly increased in both null
models, with Ank2/2at 14-fold and Enpp12/2at more than 13-
fold the width of age-matched controls (Figure 5B). A direct
comparison of the two homozygous knock-out models revealed
that Ank2/2featured slightly, but significantly, thicker cementum
at the age sampled. Histomorphometry confirmed that PDL space
was maintained in both null models, even significantly larger in
Ank2/2vs.+/+, despite exuberant cementogenesis. Alveolar bone
on the lingual aspect tended towards reduced cross sectional
dimension in both Ank and Enpp2/2models, though the effect was
not statistically significant as measured here. Tartrate resistant acid
phosphatase (TRAP) staining confirmed increased numbers of
osteoclast-like cells (TRAP positive, multinucleated) on the bone
surface adjacent to the tooth root in Ank2/2molars . A
modeling/remodeling of bone away from the root provides a
mechanism for maintenance of the PDL in the face of expanding
One of the key functional characteristics of the cervical
cementum is the extrinsic nature of the collagen fibers, which
serve to anchor the tooth to surrounding alveolar bone. Picrosirius
red staining in association with polarized light microscopy was
used to visualize the birefringent collagen fibers of the periodontia
. The thick cementum of Ank and Enpp12/2molars featured a
high concentration of extrinsic collagen fibers, which were
continuous with the fibers in the PDL proper (Figure 6B and D).
As this thick cementum in the null molars features dense extrinsic
collagen fibers, but also contains numerous cell inclusions, it could
properly be labeled cellular extrinsic fiber cementum (CEFC), a
form of cementum not typical for cervical molar roots, and
furthermore, not previously described in the cementum family.
Importantly, the observation of an ongoing, progressive apposition
on the root surfaces of Ank and Enpp12/2mice confirms this is
thickening of the normally present extrinsic fiber cementum, and is
not likely to be a different type of ectopic calcification on the root
surface. As a comparison, Alpl2/2molars were examined, and
confirmed tearing at the root-PDL interface, osteoid invasion of
the PDL space, and poorly organized and sparsely embedded
collagen fibers at the cervical root (Figure 6F).
Cementum, bone, and dentin are also characterized by their
extracellular matrix (ECM) protein composition, and these ECM
proteins contribute to crystal growth and regulation, and affect
mechanical properties of these tissues. Because of the dramatic
changes in cementum apposition, we investigated the ECM profile
Figure 2. Acellular cementogenesis requires diminution of
pyrophosphate. The Alpl+/+control first molar root at (A) 14 dpn and
(C) 21 dpn, shows a normal periodontal architecture with a continuous
layer of basophilic cementum (c) covering the root dentin (d) surface. In
Alpl2/2molars, ablation of TNAP resulted in (B) hyperosteoidosis (*) and
loss of the acellular cementum layer, and (D) a weak cementum-PDL
interface, manifested by tearing (#). (E–H) Disrupted localization of
cementum markers bone sialoprotein (BSP) and osteopontin (OPN)
compared to control supported histological observations of cementum
hypoplasia in 14 dpn Alpl2/2mouse molars. Abbreviations: d=dentin;
c=acellular cementum; p=periodontal ligament; b=bone. Scale
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org3 June 2012 | Volume 7 | Issue 6 | e38393
in PPideficient mice. In the low PPienvironment of the Ank and
Enpp12/2mice, the thick cervical cementum was marked by
increased OPN and dentin matrix protein 1 (DMP1), proteins of
the SIBLING family (Figure 7A–F and G–L). OPN staining
strongly labeled control acellular cementum, and was intensely
expressed in the corresponding Ank and Enpp12/2cervical
cementum and associated cementoblast cells. DMP1, a marker
for osteocytes, odontoblasts, and cementocytes, was present at low
or undetectable levels in acellular cementum in controls, in
contrast to intense localization in expanded Ank and Enpp12/2
cementum. OPN and DMP1 levels were not changed in Ank or
Enpp12/2apical cementum, as well as in other dentoalveolar
locations. The source of the increased OPN and DMP1 protein
was confirmed, as cementoblast gene expression for both Opn and
Dmp1 mRNA was increased in Ank and Enpp2/2mice (Figure 8A–
C and D–F). OPN and DMP1 expression changes were not
observed in other cell populations in the dentoalveolar complex in
these mice. Another characteristic marker for cementum, BSP,
was present in control and null cementum (Figure 7M–R), and
where protein concentration was diluted in the larger cementum
volume of the Ank and Enpp12/2mice, mRNA levels in
cementoblasts were unaltered (Figure 8G–I).
Thus, increased cementogenesis in Ank and Enpp12/2teeth was
linked to increased OPN and DMP1 specifically in cervical
cementum. It is notable that OPN was increased in cementum as a
result of reduced extracellular PPi. This change is opposite to the
decreased OPN that has been documented in osteoblasts and
articular locations in mice lacking ANK or NPP1 [14,16].
Cementoblasts express pyrophosphate regulators in a
time and space restricted manner
Acellular cementum was shown to be exceptionally sensitive to
regulation by PPi; with increased PPi(as in Alpl2/2mice) AEFC
was severely inhibited, and under reduced PPiconditions (as in Ank
and Enpp12/2mice) cementum thickness increased significantly, a
trend not reflected in other dental hard tissues. In order to
understand the sensitivity of acellular cementum to PPimetabo-
lism, we mapped the expression of TNAP, ANK, and NPP1
during tooth root formation. We also assayed these factors in all of
the null models to determine if there were compensatory or
antagonistic expression changes that would contribute to pheno-
types under PPidysregulation.
TNAP was widely expressed during molar root formation, most
strongly in mineralizing osteoblasts, odontoblasts, and cemento-
blasts (Figure 9A). As previously reported, TNAP was also strongly
localized to the PDL region [28,29]. TNAP localization was not
altered in developing Ank and Enpp12/2molars (Figure 9B and C).
We previously reported wide expression of ANK gene and
protein in the tooth and supporting tissues , paralleling
previous findings that ANK is expressed in several tissues system-
wide . Using a refined immunohistochemistry technique,
which allowed more sensitive identification of differential ANK
protein localization, we discovered that after acellular cementum
formed, ANK was labeled most intensely in cementoblasts lining
the molar and incisor roots (Figure 9D). Developmental localiza-
tion of NPP1 protein was similar to that of ANK, with most
intense staining found in cementoblasts (Figure 9G). Both ANK
and NPP1 stained weakly in other cells, including PDL cells,
osteoblasts, and odontoblasts. Immunolocalization revealed com-
pensatory up-regulation, where NPP1 was increased in Ank2/2
and ANK was increased in Enpp12/2(Figure 9F and H). Most
interestingly, the observed increase was found only in cemento-
blasts, and not in other cell populations of the dentoalveolar
region. These data suggested that ANK and NPP1 were
differentially expressed by cementoblasts and employed to tightly
regulate PPiand developmental cementum apposition. However,
Figure 3. Lack of acellular cementum on Alpl2/2molar root surfaces. Backcattered SEM was employed to explore the cervical root surface
(white boxes) in (A) Alpl+/+control and (B) Alpl2/2first molars. At higher magnification, the acellular cementum layer (white arrows) in the (C) control
molar can be distinguished by contrast differences due to slightly lower mineralization than underlying dentin (d). (D) No acellular cementum layer
was apparent in the cervical region of the Alpl2/2molar. Abbreviations: d=dentin; c=acellular cementum; p=periodontal ligament.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org4 June 2012 | Volume 7 | Issue 6 | e38393
it still remained unclear by what mechanism PPiwas controlling
cementum apposition and ECM composition.
Pyrophosphate controls mineralization and coupled
gene expression in cementoblast cultures
PPiregulators ANK and NPP1 were preferentially expressed by
cementoblasts after initiation of cementogenesis, and their
expression was modulated under conditions of low extracellular
PPi and increased apposition. Expression levels of cementum
ECM proteins OPN and DMP1 were also responsive to PPi
deficiency, reflecting the altered homeostasis of Pi/PPiratio or
increased cementum apposition in Ank and Enpp12/2. These data
together suggested that cementoblasts associated with AEFC
regulate PPias a means to tightly control the process of apposition
and related gene expression. In vitro experiments were performed
to determine how these genes were regulated during mineral
formation, and what potential role PPiplayed in their regulation.
Because of the technical obstacles in isolating and identifying
Figure 4. Attenuation of pyrophosphate increases acellular cementum. The cervical cementum (c) is a thin, acellular layer in Ank; Enpp1+/+
control molars at (A) 14 dpn and (D) 26 dpn, while the (G) apical cementum is thicker and contains cementocytes. Knock-out of either Ank or Enpp1
results in expanded cervical cementum compared to control, visible by 14 dpn (B and C), and progressively thicker by 26 dpn (E and F). In contrast,
the apical cementum in Ank and Enpp12/2molars (H and I) was not different from+/+control. (J–L) Acellular cementum of the incisor lingual root
analog was similarly expanded in Ank and Enpp12/2vs. control. (M–R) Hypercementosis resulting from loss of ANK was confirmed on all three
mandibular molars. Abbreviations: d=dentin; c=acellular cementum; p=periodontal ligament; b=bone. Scale bar for A–L represents 200 mm, and
for M–R represents 400 mm.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org5 June 2012 | Volume 7 | Issue 6 | e38393
primary cementoblasts, we opted to use an immortalized
cementoblast cell line (OCCM.30) and modulate exogenously
added PPi. OCCM.30 cells were cultured in control media or
mineralization media where 5 mM b-glycerophosphate (BGP) was
added. BGP served as an organic Pisource, mimicking similar
sources in vivo and commonly used for in vitro mineralization
experiments [30–33]. Cells receiving control media lacking BGP
failed to mineralize during the course of the experiment. While
cells cultured with BGP produced mineral nodules by day 6, with
increased staining and calcium incorporation at day 8 (Figure 10A
Cells were introduced to exogenous PPi to create culture
conditions of low (10 mM) and high (100 mM) PPi. The lower dose
of 10 mM PPidid not affect mineralization, while the higher dose
of 100 mM was confirmed as an inhibitor of mineral nodule
formation under these conditions. While PPiis an inhibitor of
HAP crystal precipitation, it has also been reported to have cell
signaling effects in osteoblasts [4,14,16]. Neither dose of PPi
affected OCCM.30 cell proliferation, viability, or collagen
synthesis compared to controls (Figure 11), therefore these
processes were not indirectly affecting mineralization. Cemento-
blast ALP enzyme activity was uniform across treatments and
times, and added 100 mM PPidid not appreciably affect ALP
(Figure 10C), indicating the effect of PPion mineralization was not
by inhibition of TNAP. An enzymatic assay for 59-nucleotide
phosphodiesteraseI and nucleotide
(NTPPPH) activity demonstrated significantly increased NPP1
function with mineralization at days 4, 6, and 8, while 100 mM PPi
brought activity back to basal levels of non-mineralizing cultures
PPiassociated and cementoblast marker genes were assayed by
quantitative PCR. Under non-mineralizing conditions, Ank, Enpp1,
Figure 5. Increased cementum apposition in Ank and Enpp12/2teeth. (A) During early root formation between 14 and 26 dpn, both Ank and
Enpp12/2molars featured at least 10-fold greater cementogenesis compared to controls. From 26 to 60 dpn, Ank and Enpp12/2cementum
continued to increase at a rate of 0.2–0.7 mm/day, while Ank and Enpp1+/+controls featured tightly controlled apposition at the pace of 0.01–
0.05 mm/day. (B) Histomorphometry confirmed Ank or Enpp12/2cervical cementum was significantly increased compared to controls, while PDL
width was maintained and alveolar bone thickness tended towards reduction. Values with the same letter were not significantly different, while
different letters indicate a statistically significant intergroup (genotype) difference (p,0.05) as tested by ANOVA followed by the Tukey test for direct
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org6 June 2012 | Volume 7 | Issue 6 | e38393
Opn, and Dmp1 did not change over the course of the experiment
(Figure 12). However, all four genes increased significantly under
mineralizing conditions at days 3 and 5, when mineral nodules
were forming. At day 3, when increases were most dramatic, Ank
increased almost 10-fold, Enpp1 increased 30-fold, Opn increased
more than 30-fold, and Dmp1 increased 140-fold in mineralizing
cultures compared to controls. These four genes also responded in
parallel fashion to PPi. While inclusion of 10 mM PPihad a mild
effect on gene expression compared to ascorbic acid (AA) + BGP
cultures (paralleling effects on mineralization), the higher dose of
100 mM PPisignificantly depressed Ank, Enpp1, Opn, and Dmp1
expression on day 3 compared to mineralizing cells. Cells receiving
the 100 mM dose also maintained significantly lower expression of
Ank, Enpp1, and Opn on day 5. By day 7, expression levels of the
four genes were low, and there were no differences between any of
the treatment conditions. Other cementoblast marker genes
assayed, including Alpl, Bsp, and Col1, did not show a coherent
pattern in response to addition of PPi. Notably, the increase in
Enpp1 gene expression associated with mineralization corresponds
to the increase in NTPPPHase activity recorded, and inclusion of
PPi decreased mineralization and correspondingly decreased
Enpp1 gene expression and NPP1 enzyme activity. In contrast,
PPidid not perturb cementoblast mineralization by affecting Alpl
expression or ALP activity.
These results showed that PPiregulated cementoblast mineral-
ization and associated gene expression, in vitro. In an additional
experiment of similar design, the addition of 100 mM PPiwas
discontinued in some wells midway through the experiment. Cells
with 100 mM PPi for the duration did not mineralize, while
cultures relieved of PPiinhibition at day 4 showed mineralization
by day 6, increased Ank, Enpp1, Opn, and Dmp1 by day 5,
coincident with mineralization (Figure 13). This experiment
demonstrated that even if PPiinhibited initiation of mineralization
for the first 4 days, its removal facilitated both mineralization and
concomitant gene expression. These results support expression of
Ank, Enpp1, Opn, and Dmp1 as being functionally coupled to matrix
mineralization, i.e. linked to changes in the mineralizing matrix.
Importantly, these results parallel in vivo observations, where ANK,
NPP1, OPN, and DMP1 were all increased by cementoblasts
under conditions of reduced PPi, i.e. Ank or Enpp1 ablation.
These studies aimed to define the regulatory role of PPiin tooth
root cementum development. We demonstrate here that PPiserves
as an essential regulator of tooth root acellular cementum
development, and a key determinant defining the hard-soft
interface between the cementum and PDL. Dysregulation of PPi
resulting from loss of any of the central PPicontrolling factors
explored here had profound consequences on development of
acellular extrinsic fiber cementum (AEFC), a tissue essential to
tooth attachment and function. To wit, loss of TNAP caused
severe underdevelopment or even absence of acellular cementum.
Loss of either ANK or ENPP1 resulted in loss of control of
cementum apposition, causing an exceptional hypercementosis.
Because these three factors, TNAP, ANK, and NPP1, primarily
adjust extracellular PPi, this strongly supports PPi as the key
mechanistic factor uniting the cementum phenotypes in all three of
these mouse models, prompting us to propose that PPiregulates
acellular cementum in a molecular ‘‘rheostat’’ fashion, i.e.
acellular cementum thickness relates inversely to PPiproduction.
Based on these collective data, we propose a model whereby PPi
plays a central and novel role in acellular cementum formation
(Figure 14). The periodontal region is extremely rich in ALP
activity (reducing local PPi) and thus a permissive milieu for
cementum formation on the root surface. In the course of normal
development, cementoblasts modulate PPito curb apposition (by
increasing PPivia ANK and NPP1) to maintain AEFC as a thin
tissue on the root surface. When one of these PPi factors is
removed from the equation, apposition cannot be fully regulated
and cementoblasts attempt to compensate by increasing expression
of its counterpart PPiregulator. In addition to directly controlling
cementum mineral apposition, these studies suggest PPiinfluences
ECM protein composition; in the face of rapid cementogenesis,
cementoblasts increased expression of OPN and DMP1. The
increase in OPN, a negative regulator of HAP crystal growth, may
be an additional mechanism cementoblasts employ to limit extent
of cementum apposition. In vitro experiments support this
interpretation of the role of PPi in controlling both mineral
accumulation and cementoblast expression profile. What emerges
is a portrait of acellular cementum as a mineralized tissue heavily
governed by regulation of the physical-chemical process of mineral
Figure 6. Progressive mineralization of extrinsic collagen fibers
in Ank and Enpp12/2cervical cementum. Picrosirius red staining
with polarized light microscopy was used to visualize birefringent
collagen fibers of periodontal tissues in mandibular first molar roots.
Histological sections of 60 dpn (A) control Ank; Enpp1+/+cut in a
horizontal plane and (C) coronal plane revealed high density of
embedded extrinsic fibers in the acellular cementum, where the high
degree of birefringence (intense coloration) makes visible the
organization and orientation of the major PDL collagen fibers.
Observation of (B) Ank2/2and (D) Enpp12/2expanded cervical
cementum (yellow dotted outline, flanked by white arrows) in the
same orientations revealed a similar high density of embedded fibers,
continuous from PDL through the cementum. (E) Control Alpl+/+molars
at 21 dpn cut in a coronal plane show an organized and attached PDL,
while conversely, (F) Alpl2/2molars exhibited tearing at the root-PDL
interface (#), osteoid invasion of the PDL space, and poorly organized
and sparsely embedded collagen fibers at the cervical root. Abbrevi-
ations: d=dentin; c=acellular cementum; p=periodontal ligament;
b=bone. Scale bar=100 mm.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org7June 2012 | Volume 7 | Issue 6 | e38393
precipitation, and the cementoblast as a cell capable of directing
PPimetabolism to promote and restrain cementogenesis.
On the role of pyrophosphate as a negative regulator of
In these studies, we confirm that loss of TNAP function in the
Alpl2/2mouse causes aplasia or severe hypoplasia of the acellular
cementum. This is in line with a previous report from this and
another model of TNAP loss-of-function [23,25], as well as reports
from human hypophosphatasia (HPP) subjects [21,34], who
harbor a mutation in the human homologue, Alpl . We extend
previous analyses of the Alpl2/2tooth cementum phenotype with
gene and protein assays. Cementoblasts express similar levels of
Bsp mRNA, while protein distribution of both BSP and OPN
appear disrupted. We interpret these results to mean that
cementoblast phenotype is maintained in the face of loss of
TNAP, but that disruption of AEFC synthesis prevents accretion
of BSP and OPN proteins on the root surface. Loss of OPN
protein under conditions where acellular cementum was inhibited
has been reported previously [25,35], and this observation makes
sense because BSP and OPN are both mineral-binding members
of the SIBLING family which play a role in the mineralization
process , and in the close relationship of AEFC cementogen-
esis with the act of mineralization.
We show strong TNAP localization in the developing root
region, and ALP activity has been reported to be strong in the
periodontium, with highest activity adjacent to the mineralizing
bone and developing cementum surfaces . Moreover, the same
Figure 7. Reduced pyrophosphate alters acellular cementum matrix composition. IHC was performed on Ank and Enpp1+/+(Control) and
2/2tissues at 26 dpn. OPN defines the acellular cementum layer in (A) wild-type cervical cementum, and is intensely localized to the thick AEFC in (B,
C) both2/2models. DMP1 did not label acellular cementum in (G)+/+controls, but was increased dramatically in the thickened cervical cementum of
(H, I) both2/2models. BSP was present in (M) control AEFC, as well as in (N, O) Ank and Enpp12/2AEFC in diluted concentrations. Localization of
OPN, DMP1, and BSP was not different in cellular cementum of null models vs. controls (D–F, J–L, and P–R). Abbreviations: d=dentin; c=acellular
cementum; p=periodontal ligament; b=bone. Scale bar=100 mm.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org8 June 2012 | Volume 7 | Issue 6 | e38393
Figure 8. Reduced pyrophosphate alters gene expression in cervical cementoblasts. Opn mRNA is markedly increased in root-lining
cementoblasts in both (B) Ank and (C) Enpp12/2, compared to (A) Ank and Enpp1+/+controls. Increased numbers of cells associated with the thick
cervical cementum express Dmp1 in (E) Ank and (F) Enpp12/2molars, compared to (D)+/+controls. Bsp gene expression was not different in
cementoblasts in (H) Ank and (I) Enpp12/2vs. (G)+/+controls. Black arrowheads indicate regions of positively stained cells. All panels are samples from
mice at 14 dpn. Abbreviations: d=dentin, c=(cervical) cementum; p=periodontal ligament; b=bone. Scale bar=100 mm.
Figure 9. Cementoblasts express pyrophosphate regulators in a time and space restricted manner. TNAP was expressed strongly in all
the periodontal tissues in (A) Ank; Enpp1+/+control as well as both (B, C) Ank and Enpp12/2models. Loss of ANK or NPP1 did not alter cementoblast
TNAP expression. ANK was localized selectively to cementoblasts in (D) control, and was increased when (F) Enpp1 was ablated. Like ANK, NPP1 was
found at selectively greater concentrations in cementoblasts in (G) controls, and was increased upon (H) Ank2/2. Specificity of antibody staining was
confirmed in null mice in (E) and (I). All panels are mandibular first molar teeth at 26 dpn. Abbreviations: d=dentin; c=acellular cementum;
p=periodontal ligament; b=bone. Cervical cementum is indicated by opposing black arrows. Scale bar=100 mm.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org9 June 2012 | Volume 7 | Issue 6 | e38393
study identified a significant correlation between measured TNAP
activity and acellular cementum thickness. The critical influence of
TNAP on cementum apposition is likely by clearance of
mineralization inhibitor PPi, rather than by providing local Pi
for HAP precipitation, for several reasons. First, circulating PPiis
in the micromolar range, while Piis much higher in the millimolar
range, so in the highly vascular periodontal region hydrolysis of
PPi is not likely to appreciably increase Pi available for
mineralization, though local, compartmentalized ionic dynamics
in vivo are difficult to predict. Secondly, mouse models of
hypophosphatemia described to date tend to feature bone, dentin,
and cellular cementum disorders, while acellular cementum is less
Figure 10. Pyrophosphate regulates cementoblast mineralization and nucleotide pyrophosphohydrolase (NTPPPH) activity, in
vitro. (A) By von Kossa staining, OCCM.30 cells cultured with 5 mM BGP produced mineral nodules by days 6 and 8, while cells receiving only AA did
not mineralize. The low dose of 10 mM PPidid not affect mineral nodule precipitation, however, the higher dose of 100 mM was a potent inhibitor of
mineral nodules. (B) Quantitative calcium assay performed on days 6 and 8 confirmed visual mineral nodule staining by von Kossa. (C) Relative ALP
enzyme activity was not affected by inhibition of mineralization by 100 mM PPi. (D) NTPPPHase activity was increased under mineralizing conditions,
but inclusion of 100 mM PPibrought activity back to basal levels of non-mineralizing cultures. Graphs show mean +/2 SD for n=3 samples.
Lowercase letters indicate treatment comparison at each time point, where different letters indicate a statistically significant intergroup difference.
Uppercase letters indicate comparisons over time in the same treatment group, where different letters indicate a statistically significant intragroup
difference. Values sharing the same uppercase or lowercase letter in were not significantly different. Means were compared by ANOVA (p,0.05)
followed by the Tukey test for direct pair-wise comparisons.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org 10June 2012 | Volume 7 | Issue 6 | e38393
affected [37,38]. Thirdly, in studies employing a PPianalog, 1-
hydroxyethylidene-1, 1-bisphosphonate (HEBP), it was found that
HEBP inhibited formation of acellular cementum entirely, while
cellular cementum and bone matrices were produced, but
remained unmineralized [35,39,40]. A parallel pattern emerged
when mineralization inhibitor matrix gla protein (MGP) was
ectopically expressed in bones and teeth; bone, dentin, and cellular
cementum matrices were produced yet remained unmineralized,
while AEFC was absent ; MGP and PPimay have parallel
functions as mineral regulators throughout the body. These studies
indicate that cementogenesis depends heavily on creation of a
physicochemical environment conducive for apposition, such as by
Diminished pyrophosphate relieves the negative
regulation on cementogenesis
Further evidence for PPias a central regulator of cementum
thickness was garnered from studying models with deficient PPi,
the Ank and Enpp12/2mice. In these mice, a progressive
thickening of AEFC was found during root development, which
corroborated previous findings in mice harboring suspected loss-
of-function mutations in these genes [42,43]. By completion of
root formation, these null models exhibited 12-fold or greater
AEFC vs. controls, with a significantly increased rate of apposition
over the developmental time period. Importantly, we have shown
this expanded cervical cementum shares the same mineral and
mechanical properties as WT controls [26,44,45]. This is strong
evidence that PPiis a key factor controlling acellular cementum
formation, for several reasons. Firstly, ANK and NPP1 are
membrane-bound proteins, which have been identified as primary
regulators of extracellular PPiconcentrations around mineralizing
cell types, as well as elsewhere in the body. However, they operate
by different mechanisms, with ANK affecting PPitransport and
NPP1 acting as an ectoenzyme, producing PPithrough catalysis of
trinucleosides. The common link in functions of both these
proteins is extracellular PPi production. That nearly identical
AEFC phenotypes result from ablation of either of these genes is
potent evidence for the indispensable role of PPiin influencing
acellular cementum formation. Though ANK and NPP1 share
similarity in function by increasing extracellular PPi, loss of NPP1
causes a more severe skeletal hypermineralization phenotype in
mice, a difference possibly related to inclusion of NPP1 in matrix
vesicles, whereas ANK was found to be absent in matrix vesicles
. It is intriguing then that loss of ANK or NPP1 had nearly
identical phenotypic results on acellular cementum, a tissue where
there is no clear role of matrix vesicles in mineralization.
Figure 11. Pyrophosphate does not affect cementoblast proliferation or collagen synthesis, in vitro. (A) Cell proliferation was assayed by
MTS assay where absorbance at 570 nm is proportional to the number of living cells in culture. No difference in OCCM.30 cementoblast cell number
was found between non-mineralizing (AA) and mineralizing (AA + BGP) treatments at concurrent time points, including with doses of 10 or 100 mM
PPi. (B) Picrosirius red dye was used to stain collagen deposited by cementoblasts at days 3, 5, and 7. (C) Quantification of the collagen-binding assay
did not identify any treatment differences for collagen deposition at any of the time points. For both (A) and (C), graphs show mean +/2 SD for n=3
samples, and no intergroup significant differences (at the same time point) were identified by one-way ANOVA and post-hoc Tukey analysis, for
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org11 June 2012 | Volume 7 | Issue 6 | e38393
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org12 June 2012 | Volume 7 | Issue 6 | e38393
Secondly, ANK and NPP1 are expressed in the dentoalveolar
region during tooth formation and cementogenesis. While both
ANK and NPP1 are widely expressed throughout the body, both
were found to be selectively more highly expressed in cemento-
blasts lining the tooth root. Also, a special importance for PPi
production in regulating cementum was indicated indirectly by
findings that human PDL tissue expresses significantly higher basal
levels of TNAP, ANK, and NPP1 than pulp [21,46]. This
hypothesis is supported by the finding that cementoblasts
dramatically increased either ANK or NPP1 expression in
response to loss of the other factor, likely an attempt to
compensate for lack of extracellular PPi output in these mice.
The nature of the interaction between ANK and NPP1 in tooth
formation is currently the subject of study in a series of double-
deficient mice. That expression of these PPiregulating factors is
enriched in tooth root and they are inducible in each others’
absence supports a central physiologic function for PPiin normal
control of cementogenesis. The possible involvement of other
complementary and antagonistic factors in PPihomeostasis in the
root region is an intriguing question currently being studied. One
candidate is CD73, a cell surface protein operating downstream of
NPP1 which may regulate Alpl expression, and that has been
linked to vascular calcification .
This essential role of PPi, however, seems to be limited to the
cervical acellular cementum. ANK and NPP1 were not as
consistently localized to regions of apical cementum, did not
exhibit compensatory up-regulation in the apical portion of knock-
out molars, and loss of ANK and NPP1 did not impact the
phenotype of CIFC. In this respect, the cellular cementum showed
a clear difference in developmental regulation from AEFC and
more similarity to alveolar bone. Similarly, loss of TNAP and the
resulting increased PPiaffected cellular cementum and bone in
On the influence of pyrophosphate metabolism on
cementum extracellular matrix composition
Reduced PPinot only resulted in more rapid AEFC apposition,
but also led to altered cementoblast gene expression and matrix
composition. OPN and DMP1, both mineral-regulating ECM
proteins from the SIBLING family , were increased at the
gene and protein level in Ank and Enpp1
Expression of Bsp, a key cementoblast marker and SIBLING
family member was unaltered in Ank and Enpp12/2teeth. BSP
immunostaining indicated a diffuse presence in the thick
cementum, likely diluted in relatively greater volume of mineral-
ized cementum. Bsp expression in cementoblasts in vitro was
unaffected during mineralization or its inhibition by PPi.
OPN is a multifunctional ECM protein and a marker for
cementum [48–50]. OPN has been shown in vitro to be an inhibitor
of hydroxyapatite mineral crystal growth [51,52], an observation
supported by study of the Opn2/2(Spp12/2) mouse , as well as
other models where increased OPN was found to disrupt skeletal
mineralization [16,54]. We suggest that increased expression of
OPN by Ank and Enpp2/2cementoblasts represents an additional
mechanism for control of apposition; like NPP1, an attempt at
normalization of the cementogenesis process. While previous
studies using osteoblasts have cited PPias a signal increasing Opn
expression [4,16], we found here that cementoblasts exposed to
PPi in vitro significantly reduced Opn expression during the
mineralizing phase of the experiment. The cementoblast reaction
to increase OPN in response to mineralization under low PPi
conditions, is opposite that of osteoblasts, which were found to
reduce OPN in Ank and Enpp12/2mice, contributing to the bone
pathology. The divergent response underscores the unique mineral
metabolism of cementum.
DMP1 is highly expressed in the osteocytes embedded in bone
matrix, and is associated with maintenance of the lacunar-
canalicular system of these cells [55–57]. DMP1 was increased
in osteocytes in loaded bone, perhaps functioning in the
mechanical response [58,59]. In the context of increased
cementum apposition in Ank and Enpp12/2mice, we hypothesize
that induction of Dmp1 gene expression reflects rapid apposition
and embedding of cervical root cementoblasts as cementocytes.
The cementoblasts that direct AEFC normally remain as lining
cells adjacent to this thin tissue. In vitro, mineralizing cementoblasts
also increased Dmp1 gene almost 140-fold, paralleling other studies
where DMP1 was induced in periodontal ligament cells in
mineralizing 3-dimensional gels .
The mechanism for altered Opn and Dmp1 gene and protein
expression is unknown, but under investigation. Based on in vivo
and in vitro data, we propose an ‘‘outside-in’’ type of matrix-cell
signaling mechanism whereby increased cementum apposition
switches on expression of Opn and Dmp1, as well as Ank and Enpp1.
Cementoblasts as pyrophosphate sensitive cells
Localization of PPi regulators over the course of tooth
development supported a role for PPiin modulating cementogen-
esis. The developing periodontal region shows strong immunolo-
calization of TNAP and high ALP activity, in effect producing a
highly pro-mineralization environment. These are favorable
circumstances for apposition of the cementum layer on the root
dentin surface. However, a question that arises is how cementum
may remain a thin and slow growing mineralized layer in such an
environment permissive for mineralization. ANK and NPP1
localized most strongly to the cementoblasts of the AEFC
following cementum formation, suggesting initiation of cemento-
genesis occurs under the influence of TNAP activity, but that after
cementum deposition (usually several mm in mice), cementoblasts
increase ANK and NPP1 to restrict further cementum apposition.
Thus, in a scenario where either ANK or NPP1 function is lost,
cementum apposition is not adequately controlled and the other is
up-regulated, along with increased OPN, in an attempt to regain
homeostasis of cementum.
In vitro experiments employing a cementoblast cell line provided
a mechanistic platform for probing these proposed roles of ANK,
NPP1, and PPiin cementoblast mineralization and gene expres-
Figure 12. Pyrophosphate regulates cementoblast mineralization-coupled gene expression. Mineralizing cultures (AA + BGP) increased
expression of Ank, Enpp1, Opn, and Dmp1 at days 3 and 5, concurrent with mineralization. The higher dose of 100 mM PPisignificantly depressed
expression of all four genes on day 3 compared to mineralizing cultures; Ank, Enpp1, and Opn suppression was maintained on day 5. The lower dose
of 10 mM PPishowed milder effects on all four genes at day 3. Differences were not maintained by day 7. Unlike expression of Ank, Enpp1, Opn, and
Dmp1, where PPiwas able to block mineralization-associated induction, additional markers Alpl, Bsp, and Col1 were not regulated in coordinated
fashion by mineralization or inclusion of either dose of PPi. Graphs show mean +/2 SD for n=3 samples. Lowercase letters indicate treatment
comparison at each time point, where different letters indicate a statistically significant intergroup difference. Uppercase letters indicate comparisons
over time in the same treatment group, where different letters indicate a statistically significant intragroup difference. Values sharing the same
uppercase or lowercase letter in were not significantly different. Means were compared by ANOVA (p,0.05) followed by the Tukey test for direct pair-
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org13June 2012 | Volume 7 | Issue 6 | e38393
Figure 13. Timing of pyrophosphate removal determines cementoblast mineralization and coordinated gene expression in vitro. (A)
By von Kossa staining, OCCM.30 cells cultured with 5 mM BGP produced mineral nodules by days 4, 6, and 8, while inclusion of 100 mM PPiinhibited
mineralization for the entire experiment. When PPiwas removed after 4 days, OCCM.30 cells began mineralizing the matrix by days 6 and 8. (B)
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org14 June 2012 | Volume 7 | Issue 6 | e38393
sion. The response for cementoblasts to increase ANK, NPP1, and
OPN in light of rapid apposition was supported by in vitro
experiments employing a cementoblast cell line. Results from the
studies described here demonstrated that gene expression of PPi
regulators Ank and Enpp1 and ECM proteins Opn and Dmp1 are
functionally coupled to the mineralization process, increasing
under mineralizing conditions and coincident with mineral nodule
formation. Under conditions of higher PPi, mineral apposition was
hindered, and expression of Ank, Enpp1 (and Opn and Dmp1) was
blocked in dose-response fashion. These in vitro data together with
Ank and Enpp12/2mouse phenotypes, as well as ANK and NPP1
tissue localization, support the hypothesis that PPimodulation is
employed by cementoblasts to guide the relative amount of
The question can be raised as to whether or not these
experiments were carried out within a relevant physiological
range of PPi. Circulating PPi in normal individuals has been
measured in the low mM range. We carried out experiments with
10 and 100 mM added PPi, and justified these doses based on the
following considerations. Firstly, our primary interest was to
determine if PPiat a dose that could inhibit mineralization would
also influence cementoblast gene expression. Using these cells
under these culture conditions, the lower dose of 10 mM PPiwas
insufficient in blocking mineralization in these cells. Mineralization
is affected by multiple factors including cellular activities, cell
culture media, Ca2+and Piavailability, as well as known and
unknown factors present in fetal bovine serum (FBS) supplemented
to the media; therefore, it is difficult to directly compare doses
across studies where one or more of these variables may differ,
especially cell type and time points examined [4,16]. Based on
these criteria, we employed the higher dose of 100 mM PPito
create high PPiconditions for cells. Secondly, while circulating PPi
levels are reported, it remains unknown how local, pericellular
concentrations of PPimay vary. Biomineralization is well known to
be a process dependent on compartmentalization, i.e. creation of
localized, protected regions conducive to mineral precipitation.
Therefore, we reasoned that cementoblasts, cells that show high in
vivo and in vitro expression of PPiregulators ANK and NPP1, could
potentially create localized, high PPi conditions to mediate
biomineralization-related activities. This hypothesis supported
using a higher exogenous dose of PPi in order to effectively
regulate OCCM.30-mediated mineralization in vitro.
Insights into acellular and cellular cementum
development and regeneration
In mammals, the fibrous connection of the tooth to the bony
socket is classified as a gomphosis, or fibrous joint, and is unique in
the body in that the periodontal ligament joins bone on one side to
a non-bone substance on the other side, in this case the tooth root
cementum . The gomphosis attachment is unique to mammals
and crocodilian reptiles, having developed from more ancient
forms of tooth attachment such as direct ankylosis to bone [62–
65]. This interposed ligament was made possible likely by a
combination of events such as alterations in the HERS during root
development, changes in the supporting bone (‘‘bone of attach-
ment’’), and the rise of a unique tissue, the cementum, as well as
diminution of mineralization in the region between bones and
teeth . The development of a mineral-free PDL region
between mineralized bones and teeth requires localized expression
of factors to establish a mineralization boundary at the hard-soft
tissue interface and continue to maintain PDL space throughout
the life of the tooth. In these studies, we demonstrate that
maintenance of that hard-soft interface at the tooth root surface
Quantitative calcium assay performed on days 4, 6, and 8 confirmed visual mineral nodule staining by von Kossa. (C) Mineralizing cultures (AA + BGP)
increased expression of Ank, Enpp1, Opn, and Dmp1 at day 3, concurrent with mineralization. Inclusion of 100 mM PPisignificantly depressed
expression of Ank, Enpp1, and Opn on day 3 compared to mineralizing cultures. Removal of PPion day 4 led to increased Ank, Enpp1, Opn, and Dmp1
on day 5, coincident with mineralization. Graphs in (B) and (C) show mean +/2 SD for n=3 samples. Lowercase letters indicate treatment comparison
at each time point, where different letters indicate a statistically significant intergroup difference. Uppercase letters indicate comparisons over time in
the same treatment group, where different letters indicate a statistically significant intragroup difference. Values sharing the same uppercase or
lowercase letter in were not significantly different. Means were compared by ANOVA (p,0.05) followed by the Tukey test for direct pair-wise
Figure 14. Model for the hypothesized role of in formation of
acellular cementum. Cementum apposition depends on precipita-
tion of calcium (Ca2+) and phosphate (Pi) ions on the root surface, and
pyrophosphate (PPi) acts as a potent inhibitor of hydroxyapatite crystal
precipitation. Local pericellular PPiconcentration is controlled primarily
by three cellular factors: Tissue nonspecific alkaline phosphatase (TNAP;
hydrolyzes PPi to Pi), progressive ankylosis protein (ANK; regulates
transport of PPi from the intracellular to extracellular space), and
ectonucleotide pyrophosphatase phosphodiesterase 1 (NPP1; produces
PPifrom hydrolysis of nucleotide triphosphates). Cementoblasts express
TNAP, ANK, and NPP1 in order to regulate local PPihomeostasis and
control the amount of cementum apposition. Cementoblasts are
responsive to apposition by some type of outside-in feedback
mechanism, and are capable of modulating expression of PPiregulators
as well as production of secreted matrix proteins such as OPN and
DMP1, which may influence mineralization or properties of the matrix.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org 15 June 2012 | Volume 7 | Issue 6 | e38393
depends on finely tuned PPihomeostasis. Intriguingly, unlike the
ectopic calcification found in the joints, the PDL remained
nonmineralized in the face of cementum expansion in Ank and
Enpp12/2mice, indicating the presence of additional factors that
inhibit mineralization across the PDL space. These likely include
multiple, redundant negative regulators of mineral growth, as well
as factors that indirectly prevent mineralization by influencing cell
These studies lend support to the idea that there are key
differences influencing acellular versus cellular cementum devel-
opment. Namely, acellular cementum is dependent on precise
modulation of local PPi, whereas cellular cementum is much less
sensitive to fluctuations in local PPi. The morphology, speed of
formation, and ECM protein composition of acellular cementum
were altered dramatically by disruption of local PPihomeostasis.
Conversely, cellular cementum remained unaffected in all of these
aspects. The primary cementoblasts of the AEFC adapted their
expression of ANK, NPP1, and OPN in attempts to compensate
for loss of control over cementum apposition, a response notably
absent in cementoblasts of cellular cementum, as well as other cells
of the dentoalveolar complex. However, when the regulatory
influence of PPiwas lessened, as in Ank and Enpp12/2mice, the
cementum of the cervical root grew rapidly, engulfed cells to
become cementocytes, and switched on DMP1, approximating
several aspects of apical cellular cementum. Thus, we propose that
strict regulation by PPiis one of the major differences between
acellular vs. cellular cementum types.
This finding provides insight into the origins of the two main
types of cementum, and also informs clinical regenerative
therapies. The profound influence of PPimetabolism on acellular
cementum development immediately suggests the concept of PPi
modulation for cementum regeneration. It is especially attractive
to consider such a novel approach when growth and differenti-
ation factors used to date have been limited in terms of true
cementum regeneration, PDL integration, and/or predictability
[68–70]. In a preliminary proof-of-principle study, we employed
the Ank2/2mouse, featuring deficient extracellular PPi and
increased cementogenesis, to analyze tissue repair and regenera-
tion in a periodontal fenestration model . Importantly, we
found that Ank2/2mice featured significantly greater new
cementum vs. controls, more organized mineral deposition on
the root surface in the defect areas, and recapitulated expression
patterns mapped during cementum development, including strong
OPN and DMP1 in the cementum matrix, and elevated NPP1 in
associated cementoblasts. Thus, in this pilot study in mice we
found that reduced local levels of PPi promoted increased
cementum regeneration. There has been concern voiced about
regenerated cementum being the cellular type in a majority of
studies [as summarized in ], and thus not optimal for PDL
attachment. Our findings support that the cellular or acellular
nature may be a reflection of the speed of formation, and that both
can support sufficient extrinsic PDL fiber insertion.
Materials and Methods
This study was performed in accordance with the recommen-
dations in the Guide for the Care and Use of Laboratory Animals
of the National Institutes of Health and AVMA Guidelines on
Euthanasia. The protocol for all animal studies was approved by
the Institutional Animal Care and Use Committee (IACUC),
University of Washington, Seattle, WA (Protocols 4010-01 and
Preparation and genotyping of mouse models was previously
described for Ank2/2[26,71], Alpl2/2(previously known as
Akp22/2), and Enpp12/2[10,14,16,54]. Ank and Alpl mice were
maintained on a mixed background of 129S1/SvImJ and C57BL/
6 strains, and Enpp1 mice were maintained on a mixed
background of C57BL/66129/SvTerJ strains. Mice were housed
in a specific pathogen free facility in 12 hr light-dark cycles with
access to water ad libitum. Ank and Enpp12/2mice were fed a
standard rodent diet, while Alpl litters were provided a vitamin B6
enforced diet to reduce seizures and prolong lifespan (TestDiet,
Richmond, IN). Heterozygote breeding pairs were employed to
prepare homozygote2/2mice and age-matched+/+controls at
specific ages during tooth development. Heterozygotes were
examined to determine any morphological tooth phenotype. Mice
were sacrificed by cervical dislocation and mandible tissues
harvested. At least three control (+/+) and null (2/2) mandibles
were examined for each age of interest.
Mouse mandibles were harvested and prepared for histology as
previously described . Briefly, mandibles were sagittally
hemisected and fixed in Bouin’s solution at 4uC overnight.
Hemi-mandibles were demineralized (for tissues post 8 dpn) in
acetic acid/formalin/sodium chloride (AFS) solution at 4uC.
Tissues were paraffin embedded after standard histological
processing. Five mm buccal-lingual (coronal) serial sections of the
first mandibular molar or longitudinal (sagittal) sections of hemi-
mandibles were prepared by rotary microtome and mounted on
charged glass slides. Slides were deparaffinized in xylene for
histological analyses, including hematoxylin and eosin (H&E)
staining used for morphological characterization.
Growth measurement and histomorphometry
Cervical cementum on the lingual aspect of the mesial root of
the mandibular first molar was measured at a fixed distance of
300 mm from the cementum-enamel junction (CEJ) in Ank and
Enpp1+/+and2/2sections from 14–60 dpn, using two central
sections from the set of serial sections. Static histomorphometry
was used to measure cervical cementum, PDL, and alveolar ridge
bone width on the lingual aspect of mesial roots of mandibular first
molars at the age 26 dpn. Calibrated measurements were made
using SPOT software (Diagnostic Instruments, Sterling Heights,
MI). ANOVA followed by the Tukey test for direct comparisons
was employed for statistical testing of histomorphometric mea-
surements (PASW (SPSS) Statistics software, version 19).
Picrosirius red stain for collagen in histological sections
Tissues processed for histology were stained with a picrosirius
red staining kit according to manufacturer directions (Polysciences,
Inc., Warrington, PA). Deparaffinized slides were immersed in
0.2% phosphomolybdic acid hydrate, rinsed in water, incubated in
direct red 80 for 60 min, then 0.01 N HCl solution for an
additional 2 min. Samples were rinsed in 75% ethanol for 45 sec,
then dehydrated in xylene, cleared, and mounted with coverslips.
Digital images were captured with an OptiPhot-2 microscope
(Nikon Instruments, Inc., Melville, NY) fitted with a light
polarizer, using an EOS 5D Mark II digital camera (Canon
U.S.A., Inc., Lake Success, NY).
Immunohistochemistry (IHC) was performed on histological
sections as previously described . Primary antibodies were
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org16 June 2012 | Volume 7 | Issue 6 | e38393
used with biotinylated secondary antibodies (Vectastain Elite
ABC, Vector Labs, Burlingame, CA) and color reactions were
developed to a red product using a 3-amino-9-ethylcarbazole
(AEC) substrate kit (Vector Labs). Positive controls included
normal mouse tissues and negative controls were performed in the
absence of primary antibody. Primary antibodies included:
monoclonal rat anti-human ALPL/TNAP (R&D Systems, Min-
neapolis, MN); rabbit anti-mouse progressive ankylosis protein
(ANK3) ; rabbit anti-mouse bone sialoprotein (BSP), (a gift
from Dr. Renny Franceschi, University of Michigan); rabbit anti-
rat dentin matrix protein-1 (DMP1) raised against an N-terminal
(90–111) portion of DMP1 (Takara, Shiga, Japan); polyclonal goat
anti-human NPP1 (Abcam, Cambridge, MA); and LF-175 rabbit
anti-mouse osteopontin (OPN) (Dr. Larry Fisher, NIDCR) .
ALPL and ANK3 staining was performed with an additional
unmasking step wherein slides were incubated overnight in 8.0 M
guanidine HCl (pH 8.0) solution. IHC for each target protein was
performed in sections from at least three (n=3) animals for each
age, with representative staining chosen for photographs shown in
In situ hybridization
In situ hybridization was performed using a non-radioactive in
situ hybridization (ISH) protocol employing a digoxigenin (DIG)-
labeled cRNA probe for genes of interest, as described previously
. Linearized probes were cleaned by phenol-chloroform
precipitation. Riboprobe synthesis was performed using a digox-
igenin–UTP-labeled kit (Roche Applied Science, Indianapolis,
IN). Probes were fractionated at 60uC and riboprobe concentra-
tions were checked by dot blot on Hybond N+nylon membrane
(GE Healthcare, Piscataway, NJ). Messenger RNAs were labeled
by incubation of deparaffinized sections with NBT/BCIP (Nitro
blue tetrazolium chloride/5-Bromo-4-chloro-3-indolyl phosphate,
toluidine salt). Probes used for ISH included: mouse Dmp1 plasmid
(provided by Dr. Ann George, Northwestern University) ;
mouse Opn and Bsp probes (provided by Dr. Marian Young, NIH/
NIDCR) . Negative controls included sense probes.
Scanning electron microscopy (SEM) analyses were performed
on hemi-mandibles from 20 dpn control and Alpl
previously described . Briefly, mandibles were sequentially
dehydrated in aqueous ethanol solutions and mounted in room-
temperature-cure epoxy (Allied High Tech Inc, Rancho Dom-
inguez, CA). Specimens were cut using a precision wafering saw
(Buehler Ltd, Lake Bluff, IL) to expose the mesial surface of the
first molar. The cut surface was then ground further distally to
expose the interior of the first molar using 600 then 1500 grit SiC
papers, followed by smoothening via ultramicrotoming with a 45u
angle diamond knife (Diatome, Inc., Hatfield, PA) fitted onto a
MT 6000-XL ultra-microtome (Bal-Tec RMC, Inc., Tucson, AZ).
Specimens were mounted on SEM stubs, sputter coated with 5 nm
of Pt for electron conductivity (SPI Supplies Inc, West Chester,
PA), and imaged by an JSM7000F (JEOL-USA, Inc., Peabody,
MA) SEM operating at 15 kV in backscattering mode.
Cell culture and in vitro assays
Isolation and characterization of OCCM.30 murine cemento-
blasts has been previously described [75,76]. Cells were grown in
Dulbecco’s Modified Eagle Medium (DMEM) with 10% v/v fetal
bovine serum (FBS), 2 mM L-glutamine, 100 U/ml penicillin, and
100 mg/ml streptomycin (all reagents from Invitrogen, Carlsbad,
CA). For gene expression and mineralization experiments,
OCCM.30 cells were plated in standard media as described
above, with media changed after 24 hrs to DMEM with 1% FBS
with 50 mg/ml ascorbic acid (AA). Media were changed every
48 hrs for the remainder of the experiment. Inclusion of organic
phosphate source b-glycerophosphate (bGP; 5 mM) was used to
create mineralizing conditions. Inorganic PPi(10 or 100 mM) was
added to assay effects on cell function. Both bGP and PPiwere
purchased from Sigma-Aldrich (St. Louis, MO). Cell culture
experiments were performed at least three times in triplicate with
representative results presented.
Cell proliferation was measured using a non-radioactive, MTS-
based assay, following manufacturer’s directions (CellTiter 96H
AQueousproliferation assay, Promega, Madison, WI). Absorbance
was measured at 570 nm, with reference reading at 750 nm.
Absorbance is proportional to the number of living cells in culture.
Production of collagen by cells in vitro was quantified by
picrosirius red staining, using methods modified from previous
reports [77,78]. Briefly, cells were rinsed with PBS and fixed in
Bouin’s solution for 1 hr at room temperature. The fixative was
removed and the plate rinsed several times in water to remove
excess Bouin’s solution. The collagenous matrix in plates was
stained by incubation with picrosirius red dye (Direct Red 80,
Polysciences, Inc., Warrington, PA) while gently shaking. Un-
bound dye was removed by rinsing several times with 0.01 N HCl.
Bound dye was removed by incubation and shaking with 0.1 N
NaOH for at least 1 hr. Picrosirius red was quantified by reading
the absorbance at 550 nm. Quantity of collagen was calibrated
against a standard curve created by plating and eluting known
concentrations of rat tail collagen.
Von Kossa staining for mineral nodule formation was
performed using standard procedures . Silver stain was
visualized as black, and stain intensity indicated the amount of
calcium phosphate precipitation in the cell matrix (silver ions react
with phosphate). Cell mineralization in vitro was quantitatively
assayed by measuring calcium deposits, using a method modified
from a previous report . To cell culture wells, 500 ml 0.5 N
HCl was added and plates were agitated for 60 min to dissolve
calcium-phosphate precipitations. Eluted calcium was measured
using a calcium assay (Genzyme Diagnostics, Farmingham, MA).
One ml of sample was added to 99 ml Arsenazo reagent and
absorbance was read at 650 nm. Standard curves were prepared
using a calcium stock solution.
A modified assay for measuring in vitro alkaline phosphatase
activity (ALP) was used . Briefly, cell cultures were rinsed with
PBS and incubated with 200 ml p-Nitrophenyl phosphate (PNPP,
Sigma) in the dark at ambient room temperature for 30 min. After
incubation, 10 ml supernatant for each condition was transferred
to a 96 well plate containing 90 ml of 3 N NaOH (stop solution)
per well. Absorbance was recorded at a wavelength of 405 nm.
An enzymatic assay was employed to measure the activity of
pyrophosphate-generating ectoenzymes (nucleoside triphosphate
pyrophosphohydrolase, NTPPPHase activity), based on a previ-
ously described procedure . Cells were rinsed with PBS and
incubated for 2 hrs with 2.0 ml of 1 mM thymidine 59
monophosphate p-nitrophenyl ester sodium (TMPNP) solution at
37uC and without CO2. After incubation, 20 ml supernatant for
each condition was transferred to a 96 well plate containing 80 ml
of 0.1 N NaOH (stop solution) per well. Absorbance was recorded
at a wavelength of 410 nm.
RNA isolation and real-time quantitative RT-PCR
Isolation of RNA, synthesis of cDNA, and performance of real-
time quantitative PCR was undertaken as previously described
. Total RNA from cells was isolated using the RNeasy Micro
kit (Qiagen, Valencia, CA) and cDNA was synthesized from
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org17 June 2012 | Volume 7 | Issue 6 | e38393
1.0 mg RNA (Transcriptor kit, Roche Applied Science). PCR
reactions were performed with DNA Master SYBR Green I kit
(Roche Applied Science) on the Roche Lightcycler 480 system
(Roche Diagnostics GmbH, Mannheim, Germany) using intron-
spanning primers (http://www.gene-expression-analysis.com/).
Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was em-
ployed as a housekeeping/reference gene for target gene
normalization and relative quantification with amplification
efficiency correction. Primer sequences used are listed in Table 1.
PCR product identification was performed by post-amplification
melting curve analysis. To detect intra- and intergroup gene
expression differences, we employed a one-way ANOVA with
post-hoc Tukey test, using PASW (SPSS) Statistics 19 software.
A portion of this research was completed when BLF, FHN, ABT, and MJS
were affiliated with the University of Washington School of Dentistry,
Department of Periodontics (Seattle, WA, USA). The authors would like to
thank Jirawan Wade for preparing histological sections, Casey Self for
assistance with polarized light microscopy, and Lay Soon for assistance
with histomorphometry. Thanks to Ann Rosenthal (Medical College of
Wisconsin, Milwaukee, WI) for assistance with the NTPPPH activity
Conceived and designed the experiments: BLF MJS. Performed the
experiments: BLF KJN HF DD ABT SN WW. Analyzed the data: BLF
MJS FHN JLM. Contributed reagents/materials/analysis tools: JLM.
Wrote the paper: BLF.
1. Foster BL, Tompkins KA, Rutherford RB, Zhang H, Chu EY, et al. (2008)
Phosphate: known and potential roles during development and regeneration of
teeth and supporting structures. Birth Defects Res C Embryo Today 84: 281–
2. Fleisch H, Bisaz S (1962) Mechanism of calcification: inhibitory role of
pyrophosphate. Nature 195: 911.
3. Meyer JL (1984) Can biological calcification occur in the presence of
pyrophosphate? Arch Biochem Biophys 231: 1–8.
4. Addison WN, Azari F, Sorensen ES, Kaartinen MT, McKee MD (2007)
Pyrophosphate inhibits mineralization of osteoblast cultures by binding to
mineral, up-regulating osteopontin, and inhibiting alkaline phosphatase activity.
J Biol Chem 282: 15872–15883.
5. Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G (2005) Unique
coexpression in osteoblasts of broadly expressed genes accounts for the spatial
restriction of ECM mineralization to bone. Genes Dev 19: 1093–1104.
6. Millan JL (2006) Mammalian Alkaline Phosphatases. Weinheim: Wiley-VCH
Verlag GmbH & Co. KGaA. 322 p.
7. Whyte MP (1994) Hypophosphatasia and the role of alkaline phosphatase in
skeletal mineralization. Endocr Rev 15: 439–461.
8. Whyte MP (2002) Hypophosphatasia Nature’s window to alkaline phosphatase
in man. In: Bilezikian JP, Raisz LG, Rodan GA, editors. Principles of Bone
Biology. 2nd ed. San Diego, CA: Academic Press. pp. 1229–1248.
9. Reibel A, Manie `re M, Clauss F, Droz D, Alembik Y, et al. (2009) Orodental
phenotype and genotype findings in all subtypes of hypophosphatasia.
Orphanet J Rare Dis 4: 6.
10. Narisawa S, Frohlander N, Millan JL (1997) Inactivation of two mouse alkaline
phosphatase genes and establishment of a model of infantile hypophosphatasia.
Dev Dyn 208: 432–446.
11. Milla ´n J, Narisawa S, Lemire I, Loisel T, Boileau G, et al. (2008) Enzyme
replacement therapy for murine hypophosphatasia. J Bone Miner Res 23: 777–
12. Ho AM, Johnson MD, Kingsley DM (2000) Role of the mouse ank gene in
control of tissue calcification and arthritis. Science 289: 265–270.
13. Gurley KA, Reimer RJ, Kingsley DM (2006) Biochemical and genetic analysis of
ANK in arthritis and bone disease. Am J Hum Genet 79: 1017–1029.
14. Johnson K, Goding J, Van Etten D, Sali A, Hu SI, et al. (2003) Linked
deficiencies in extracellular PP(i) and osteopontin mediate pathologic calcifica-
tion associated with defective PC-1 and ANK expression. J Bone Miner Res 18:
15. Johnson K, Moffa A, Chen Y, Pritzker K, Goding J, et al. (1999) Matrix vesicle
plasma cell membrane glycoprotein-1 regulates mineralization by murine
osteoblastic MC3T3 cells. J Bone Miner Res 14: 883–892.
16. Harmey D, Hessle L, Narisawa S, Johnson KA, Terkeltaub R, et al. (2004)
Concerted regulation of inorganic pyrophosphate and osteopontin by akp2,
enpp1, and ank: an integrated model of the pathogenesis of mineralization
disorders. Am J Pathol 164: 1199–1209.
17. Hessle L, Johnson K, Anderson H, Narisawa S, Sali A, et al. (2002) Tissue-
nonspecific alkaline phosphatase and plasma cell membrane glycoprotein-1 are
central antagonistic regulators of bone mineralization. Proc Natl Acad Sci U S A
18. Foster BL, Popowics TE, Fong HK, Somerman MJ (2007) Advances in defining
regulators of cementum development and periodontal regeneration. Curr Top
Dev Biol 78: 47–126.
19. Bosshardt D (2005) Are cementoblasts a subpopulation of osteoblasts or a unique
phenotype? J Dent Res 84: 390–406.
20. Diekwisch TG (2001) The developmental biology of cementum. Int J Dev Biol
21. van den Bos T, Handoko G, Niehof A, Ryan LM, Coburn SP, et al. (2005)
Cementum and dentin in hypophosphatasia. J Dent Res 84: 1021–1025.
22. Bruckner R, Rickles N, Porter D (1962) Hypophosphatasia with premature
shedding of teeth and aplasia of cementum. Oral Surg Oral Med Oral Pathol 15:
23. Beertsen W, VandenBos T, Everts V (1999) Root development in mice lacking
functional tissue non-specific alkaline phosphatase gene: inhibition of acellular
cementum formation. J Dent Res 78: 1221–1229.
24. Fedde KN, Blair L, Silverstein J, Coburn SP, Ryan LM, et al. (1999) Alkaline
phosphatase knock-out mice recapitulate the metabolic and skeletal defects of
infantile hypophosphatasia. J Bone Miner Res 14: 2015–2026.
25. McKee MD, Nakano Y, Masica DL, Gray JJ, Lemire I, et al. (2011) Enzyme
Replacement Therapy Prevents Dental Defects in a Model of Hypophospha-
tasia. J Dent Res.
26. Foster BL, Nagatomo KJ, Bamashmous SO, Tompkins KA, Fong H, et al.
(2011) The Progressive Ankylosis Protein Regulates Cementum Apposition and
Extracellular Matrix Composition. Cells Tissues Organs.
Table 1. Real time quantitative PCR primer sequences.
Gene SymbolGene Name Forward (59–39) Reverse 59–39
AlplTissue nonspecific alkaline phosphataseGGGGACATGCAGTATGAGTTGGCCTGGTAGTTGTTGTGAG11647
Ank Progressive ankylosis proteinGAATCAGTCGGCCCAT GTTCGCCAGTTTATTGCT 11732
BspBone sialoprotein GAGACGGCGATAGTTCCAGTGCCGCTAACTCAA15891
Col1 Collagen type 1 alpha 1CACCCCAGCCGCAAAGAGT CGGGCAGAAAGCACAGCACT12842
Dmp1 Dentin matrix protein 1GCGCGGATAAGGATGA GTCCCCGTGGCTACTC13406
Enpp1 Ectonucleotide pyrophosphatase
Gapdh Glyceraldehyde-3 phosphate
ACCACAGTCCATGCCATCAC TCCACCACCCTGTTGCTGTA 14433
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org 18 June 2012 | Volume 7 | Issue 6 | e38393
27. Junqueira L, Bignolas G, Brentani R (1979) Picrosirius staining plus polarization Download full-text
microscopy, a specific method for collagen detection in tissue sections.
Histochem J 11: 447–455.
28. van den Bos T, Beertsen W (1999) Alkaline phosphatase activity in human
periodontal ligament: age effect and relation to cementum growth rate.
J Periodontal Res 34: 1–6.
29. Groeneveld MC, Everts V, Beertsen W (1995) Alkaline phosphatase activity in
the periodontal ligament and gingiva of the rat molar: its relation to cementum
formation. J Dent Res 74: 1374–1381.
30. Tenenbaum HC (1981) Role of organic phosphate in mineralization of bone in
vitro. J Dent Res 60 Spec No C: 1586–1589.
31. Tenenbaum HC (1987) Levamisole and inorganic pyrophosphate inhibit beta-
glycerophosphate induced mineralization of bone formed in vitro. Bone Miner
32. Bellows CG, Heersche JN, Aubin JE (1992) Inorganic phosphate added
exogenously or released from beta-glycerophosphate initiates mineralization of
osteoid nodules in vitro. Bone Miner 17: 15–29.
33. Bellows CG, Aubin JE, Heersche JN (1991) Initiation and progression of
mineralization of bone nodules formed in vitro: the role of alkaline phosphatase
and organic phosphate. Bone Miner 14: 27–40.
34. Chapple IL (1993) Hypophosphatasia: dental aspects and mode of inheritance.
J Clin Periodontol 20: 615–622.
35. Jayawardena CK, Takahashi N, Watanabae E, Takano Y (2002) On the origin
of intrinsic matrix of acellular extrinsic fiber cementum: studies on growing
cementum pearls of normal and bisphosphonate-affected guinea pig molars.
Eur J Oral Sci 110: 261–269.
36. Fisher L, Fedarko N (2003) Six genes expressed in bones and teeth encode the
current members of the SIBLING family of proteins. Connect Tissue Res 44
Suppl 1: 33–40.
37. Fong H, Chu EY, Tompkins KA, Foster BL, Sitara D, et al. (2009) Aberrant
cementum phenotype associated with the hypophosphatemic hyp mouse.
J Periodontol 80: 1348–1354.
38. Ye L, Zhang S, Ke H, Bonewald LF, Feng JQ (2008) Periodontal breakdown in
the Dmp1 null mouse model of hypophosphatemic rickets. J Dent Res 87: 624–
39. Beertsen W, Niehof A, Everts V (1985) Effects of 1-hydroxyethylidene-1, 1-
bisphosphonate (HEBP) on the formation of dentin and the periodontal
attachment apparatus in the mouse. Am J Anat 174: 83–103.
40. Alatli-Kut I, Hultenby K, Hammarstrom L (1994) Disturbances of cementum
formation induced by single injection of 1-hydroxyethylidene-1,1-bisphospho-
nate (HEBP) in rats: light and scanning electron microscopic studies.
Scand J Dent Res 102: 260–268.
41. Kaipatur NR, Murshed M, McKee MD (2008) Matrix Gla protein inhibition of
tooth mineralization. J Dent Res 87: 839–844.
42. Nociti FH Jr, Berry JE, Foster BL, Gurley KA, Kingsley DM, et al. (2002)
Cementum: a phosphate-sensitive tissue. J Dent Res 81: 817–821.
43. Fong H, Foster BL, Sarikaya M, Somerman MJ (2009) Structure and
mechanical properties of Ank/Ank mutant mouse dental tissues-An animal
model for studying periodontal regeneration. Arch Oral Biol.
44. Fong H, Foster BL, Sarikaya M, Somerman MJ (2009) Structure and
mechanical properties of Ank/Ank mutant mouse dental tissues–an animal
model for studying periodontal regeneration. Arch Oral Biol 54: 570–576.
45. Rodrigues TL, Nagatomo KJ, Foster BL, Nociti FH, Somerman MJ (2011)
Modulation of phosphate/pyrophosphate metabolism to regenerate the
periodontium. A novel in vivo approach. J Periodontol.
46. Rodrigues TL, Foster BL, Silverio KG, Martins L, Casati MZ, et al. (2011)
Correction of Hypophosphatasia (Hpp) Associated Mineralization Deficiencies
In vitro by Phosphate/Pyrophosphate Modulation in Periodontal Ligament
Cells. J Periodontol.
47. St Hilaire C, Ziegler SG, Markello TC, Brusco A, Groden C, et al. (2011) NT5E
mutations and arterial calcifications. N Engl J Med 364: 432–442.
48. Bronckers A, Farach-Carson M, Van Waveren E, Butler W (1994) Immuno-
localization of osteopontin, osteocalcin, and dentin sialoprotein during dental
root formation and early cementogenesis in the rat. J Bone Miner Res 9: 833–
49. MacNeil R, Berry J, D’Errico J, Strayhorn C, Piotrowski B, et al. (1995) Role of
two mineral-associated adhesion molecules, osteopontin and bone sialoprotein,
during cementogenesis. Connect Tissue Res 33: 1–7.
50. McKee M, Zalzal S, Nanci A (1996) Extracellular matrix in tooth cementum
and mantle dentin: localization of osteopontin and other noncollagenous
proteins, plasma proteins, and glycoconjugates by electron microscopy. Anat
Rec 245: 293–312.
51. Boskey AL, Maresca M, Ullrich W, Doty SB, Butler WT, et al. (1993)
Osteopontin-hydroxyapatite interactions in vitro: inhibition of hydroxyapatite
formation and growth in a gelatin-gel. Bone Miner 22: 147–159.
52. Hunter GK, Kyle CL, Goldberg HA (1994) Modulation of crystal formation by
bone phosphoproteins: structural specificity of the osteopontin-mediated
inhibition of hydroxyapatite formation. Biochem J 300 (Pt 3): 723–728.
53. Boskey AL, Spevak L, Paschalis E, Doty SB, McKee MD (2002) Osteopontin
deficiency increases mineral content and mineral crystallinity in mouse bone.
Calcif Tissue Int 71: 145–154.
54. Harmey D, Johnson KA, Zelken J, Camacho NP, Hoylaerts MF, et al. (2006)
Elevated skeletal osteopontin levels contribute to the hypophosphatasia
phenotype in akp2(2/2) mice. J Bone Miner Res 21: 1377–1387.
55. Toyosawa S, Shintani S, Fujiwara T, Ooshima T, Sato A, et al. (2001) Dentin
matrix protein 1 is predominantly expressed in chicken and rat osteocytes but
not in osteoblasts. J Bone Miner Res 16: 2017–2026.
56. Feng J, Ward L, Liu S, Lu Y, Xie Y, et al. (2006) Loss of DMP1 causes rickets
and osteomalacia and identifies a role for osteocytes in mineral metabolism. Nat
Genet 38: 1310–1315.
57. Rios H, Ye L, Dusevich V, Eick D, Bonewald L, et al. (2005) DMP1 is essential
for osteocyte formation and function. J Musculoskelet Neuronal Interact 5: 325–
58. Yang W, Lu Y, Kalajzic I, Guo D, Harris M, et al. (2005) Dentin matrix protein
1 gene cis-regulation: use in osteocytes to characterize local responses to
mechanical loading in vitro and in vivo. J Biol Chem 280: 20680–20690.
59. Gluhak-Heinrich J, Ye L, Bonewald L, Feng J, MacDougall M, et al. (2003)
Mechanical loading stimulates dentin matrix protein 1 (DMP1) expression in
osteocytes in vivo. J Bone Miner Res 18: 807–817.
60. Berendsen A, Smit T, Schoenmaker T, Walboomers X, Harris S, et al. (2010)
Inorganic phosphate stimulates DMP1 expression in human periodontal
ligament fibroblasts embedded in three-dimensional collagen gels. Cells Tissues
Organs 192: 116–124.
61. Ho SP, Marshall SJ, Ryder MI, Marshall GW (2007) The tooth attachment
mechanism defined by structure, chemical composition and mechanical
properties of collagen fibers in the periodontium. Biomaterials 28: 5238–5245.
62. McIntosh JE, Anderton X, Flores-De-Jacoby L, Carlson DS, Shuler CF, et al.
(2002) Caiman periodontium as an intermediate between basal vertebrate
ankylosis-type attachment and mammalian ‘‘true’’ periodontium. Microsc Res
Tech 59: 449–459.
63. Abe T (2004) Observation on the ultrastructure of the cementum in Caiman
crocodilus. J Oral Biosci 46: 125–151.
64. Gaengler P (2007) Evolution of tooth attachment in lower tetrapods to
vertebrates. In: Teaford MF, Smith MM, Ferguson MWJ, editors. Development,
function, and evolution of teeth. New York: Cambridge University Press. pp.
65. Luan X, Walker C, Dangaria S, Ito Y, Druzinsky R, et al. (2009) The mosasaur
tooth attachment apparatus as paradigm for the evolution of the gnathostome
periodontium. Evol Dev 11: 247–259.
66. Luan X, Ito Y, Diekwisch TG (2006) Evolution and development of Hertwig’s
epithelial root sheath. Dev Dyn.
67. Marchesan JT, Scanlon CS, Soehren S, Matsuo M, Kapila YL (2011)
Implications of cultured periodontal ligament cells for the clinical and
experimental setting: A review. Arch Oral Biol.
68. Giannobile WV, Somerman MJ (2003) Growth and amelogenin-like factors in
periodontal wound healing. A systematic review. Ann Periodontol 8: 193–204.
69. Bosshardt DD, Sculean A (2009) Does periodontal tissue regeneration really
work? Periodontol 51: 208–219.
70. Esposito M, Grusovin MG, Papanikolaou N, Coulthard P, Worthington HV
(2009) Enamel matrix derivative (Emdogain(R)) for periodontal tissue regener-
ation in intrabony defects. Cochrane Database Syst Rev: CD003875.
71. Gurley KA, Chen H, Guenther C, Nguyen ET, Rountree RB, et al. (2006)
Mineral Formation in Joints Caused by Complete or Joint-Specific Loss of ANK
Function. J Bone Miner Res 21: 1238–1247.
72. Ogbureke KU, Fisher LW (2004) Expression of SIBLINGs and their partner
MMPs in salivary glands. J Dent Res 83: 664–670.
73. Hao J, Zou B, Narayanan K, George A (2004) Differential expression patterns of
the dentin matrix proteins during mineralized tissue formation. Bone 34: 921–
74. FISHER LW, STUBBS JT 3RD, YOUNG MF (1995) Antisera and cDNA probes to
human and certain animal model bone matrix noncollagenous proteins. Acta
Orthop Scand Suppl 266: 61–65.
75. D’Errico JA, Berry JE, Ouyang H, Strayhorn CL, Windle JJ, et al. (2000)
Employing a transgenic animal model to obtain cementoblasts in vitro.
J Periodontol 71: 63–72.
76. Berry JE, Zhao M, Jin Q, Foster BL, Viswanathan H, et al. (2003) Exploring the
origins of cementoblasts and their trigger factors. Connect Tissue Res 44 Suppl
77. Addison W, McKee M (2010) Inositol hexakisphosphate inhibits mineralization
of MC3T3-E1 osteoblast cultures. Bone 46: 1100–1107.
78. Tullberg-Reinert H, Jundt G (1999) In situ measurement of collagen synthesis by
human bone cells with a sirius red-based colorimetric microassay: effects of
transforming growth factor beta2 and ascorbic acid 2-phosphate. Histochem
Cell Biol 112: 271–276.
79. Lillie RD (1965) Histopathological Technique and Practical Histochemistry.
New York: McGraw Hill Book Company.
80. Osathanon T, Giachelli C, Somerman M (2009) Immobilization of alkaline
phosphatase on microporous nanofibrous fibrin scaffolds for bone tissue
engineering. Biomaterials 30: 4513–4521.
81. Rosenthal A, Cheung H, Ryan L (1991) Transforming growth factor beta 1
stimulates inorganic pyrophosphate elaboration by porcine cartilage. Arthritis
Rheum 34: 904–911.
Pyrophosphate and Acellular Cementum
PLoS ONE | www.plosone.org19 June 2012 | Volume 7 | Issue 6 | e38393