Article

Cite This: ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX

Biomimetic Tooth Repair: Amelogenin-Derived Peptide Enables in

Vitro Remineralization of Human Enamel
Sami Dogan,† Hanson Fong,‡ Deniz T. Yucesoy,‡ Timothee Cousin,§ Carolyn Gresswell,‡ Sefa Dag,‡
Greg Huang,§ and Mehmet Sarikaya*,‡,∥,⊥
†
Department of Restorative Dentistry, ‡GEMSEC, Genetically Engineered Materials Science and Engineering Center and Department
of Materials Science and Engineering, §Department of Orthodontics, ∥Deparment of Chemical Engineering, ⊥Department of Oral
Health Sciences, University of Washington, Seattle, Washington 98195, United States
*
S Supporting Information

ABSTRACT: White spot lesions (WSL) and incipient caries

on enamel surfaces are the earliest clinical outcomes for
demineralization and caries. If left untreated, the caries can
progress and may cause complex restorative procedures or
even tooth extraction which destroys soft and hard tissue
architecture as a consequence of connective tissue and bone
loss. Current clinical practices are insufficient in treating dental
caries. A long-standing practical challenge associated with
demineralization related to dental diseases is incorporating a
functional mineral microlayer which is fully integrated into the
molecular structure of the tooth in repairing damaged enamel.
This study demonstrates that small peptide domains derived
from native protein amelogenin can be utilized to construct a mineral layer on damaged human enamel in vitro. Six groups were
prepared to carry out remineralization on artificially created lesions on enamel: (1) no treatment, (2) Ca2+ and PO43− only, (3)
1100 ppm fluoride (F), (4) 20 000 ppm F, (5) 1100 ppm F and peptide, and (6) peptide alone. While the 1100 ppm F sample
(indicative of common F content of toothpaste for homecare) did not deliver F to the thinly deposited mineral layer, high F test
sample (indicative of clinical varnish treatment) formed mainly CaF2 nanoparticles on the surface. Fluoride, however, was
deposited in the presence of the peptide, which also formed a thin mineral layer which was partially crystallized as fluorapatite.
Among the test groups, only the peptide-alone sample resulted in remineralization of fairly thick (10 μm) dense mineralized layer
containing HAp mineral, resembling the structure of the healthy enamel. The newly formed mineralized layer exhibited
integration with the underlying enamel as evident by cross-sectional imaging. The peptide-guided remineralization approach sets
the foundation for future development of biomimetic products and treatments for dental health care.
KEYWORDS: dental remineralization, white spot lesion, demineralization, bioinformatics, molecular biomimetics,
amelogenin-derived peptides

■ INTRODUCTION
Dental caries is a major public health problem and a highly
fluoride products remains the primary treatment modality for
caries prevention and remineralization, with major limitations
prevalent disease among the global population.1 Incipient caries regarding the efficacy of these products for the reversal or
and white spot lesions (WSL) as well as hypersensitivity are the prevention of dental caries. Fluoride delivery systems, therefore,
earliest clinical evidence of enamel demineralization and dental are not sufficient to overcome the high caries risk, especially in
caries.2 Caries forms when tooth enamel is exposed to acid the younger and elderly populations.15
produced by cariogenic bacteria. As a result, acid diffuses into Despite dental caries being a preventable infectious disease,16
surface enamel and dissolves hydroxyapatite (HAp) mineral. oral health promotion and prevention can fail due to many
Due to its nonregenerative nature, enamel is unable to heal and factors.17 The advanced cavitation of the carious lesion
repair itself postdemineralization.3,4 necessitates restoring the tooth with materials such as metals,
Traditionally, fluoride (F) has been used as the key agent in composite resins, and ceramics to replace the lost enamel or
prevention of caries. Fluoride functions primarily via topical even dentin. However, modern dental materials to repair
mechanisms.5−8 It is believed that fluoride forms a thin layer of cavitated carious lesions are not compatible with biological
new but harder mineral, namely fluorapatite (FAp), which is tissues at the lesion/restorative material interface mainly
incorporated into the existing HAp mineral on the tooth
surface.8 There is a trend of enhancing the remineralization Received: December 6, 2017
effect of fluoride with calcium and phosphate supplementation Accepted: March 9, 2018
in high risk individuals.9−14 Although controversial,15 the use of Published: March 9, 2018

© XXXX American Chemical Society A DOI: 10.1021/acsbiomaterials.7b00959

ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

because of their physical (crystallography, morphology,

property) and chemical differences (elemental compositions
■ MATERIALS AND METHODS
Sample Preparation and Test Groups. Extracted human molar
and phases) compared to the natural tooth structure.18 Even teeth with no visible white spot lesions, caries, or any other kind of
though treatment of early caries lesions by the application of restorations were collected from dental clinics around the King
various types of nanosized HAp or CaPO4 with or without F County area (WA, United States) and disinfected in 10% aqueous
has received considerable attention,19−22 their clinical validation bleach solutions. [Human Subject Division (University of Wash-
is still lacking.20,22 Low solubility of the calcium phosphates, ington) approved the application for nonidentifiable specimen use on
04/19/2016. It should be noted that for the extracted teeth, the
particularly in the presence of fluoride ions, is the main identification of the patients was not made and, therefore, approval
difficulty with the clinical application of remineralization. Using from the ethical committee was not needed.] Prior to the experiments,
biomimetic pathways, considerable attempts have been made to the teeth were cleaned to remove visible blood, gross debris, and soft
form a remineralized layer on the surface of enamel or dentin, connective tissue using a dental scaler under a light microscope.
or even cementum, to repair or reconstruct the lost mineral and Formation of Artificial Lesions. Enamel was demineralized to
hence restore the original structure and the resulting create artificial lesions to mimic white spot lesions and/or incipient
function.23−30 These studies included the use of full-length caries using a protocol modeled after previous work for creating
artificial lesions38−40 which is outlined in Figure 1. The teeth were
amelogenin,24,28 LRAP,29 peptides,26,27 dendrimers,30 and
physical chemistry approaches.25,30 These studies contributed
to the general knowledge of tooth surface remineralization; so
far, however, no clinical remineralization system has emerged to
promote biomimetic enamel subsurface remineralization in
vivo.
A protocol was recently developed in the authors’ lab to
identify peptide sequences from native proteins with the
potential to repair damaged dental tissues by biomimicking
HAp biomineralization.31 Using a newly developed bioinfor-
matics scoring matrix,32 specific peptide domains (each
containing 15−40 amino acids) within (180 amino acid-long)
amelogenin protein (rM180) were identified based on the
similarity with a set (155 sequences) of HAp-binding peptides
(HABPs) which were originally selected by 7-AA and 12-AA
phage display peptide libraries.33 Among these peptide
domains, referred to as amelogenin-derived peptides (ADPs),
a 22-amino acids long peptide ADP5 (see Supporting
Information) was shown to facilitate cell-free and fast formation
mineral layer on demineralized human root dentin. Dubbed as
cementomimetic layer, the newly formed mineralized coating
was found to be structurally and mechanically integrated into
the underlying dentin, resembling cementum with mechanical
and chemical durability providing biogenic surface for the PDL
cells to attach, grow, and proliferate.31 The ADPs in general and
ADP5 in particular epitomize the unique features of the natural
Figure 1. (a) White spot lesion was artificially created by exposing a
protein amelogenin, the key protein in enamel and cementum window on the tooth surface for demineralization. (b) Group 5 and 6
formation,34−36 especially the function of capturing the samples were exposed to shADP5 solution for 10 min at 37 °C. (c)
constituent ions, synthesizing the mineral, and controlling its Samples were then incubated in F/Ca2+/PO43− or Ca2+/PO43−
morphology on the surface and the root of the tooth. solutions for 1 h at 37 °C. (d) New mineral layer was characterized
Incorporating a functional and biomimetic mineral layer to structurally and mechanically.
the molecular structure of the tooth to repair damaged enamel
tissue has been a long-standing challenge.31,37 A better covered by lacquer leaving a 4 mm wide square window in the enamel
understanding of peptide-guided remineralization on human close to cemento−enamel junction. The exposed areas of 4 × 4 mm
tooth and, therefore, the ability to control the mineral layer enamel were treated with a cocktail of acetic acid/CaCl2/KH2PO4 for
properties, with no, low, or high-F content, has enormous 2 weeks to establish up to 200 μm deep artificially created
noncavitated WSL.38−40 Specifically, WSL were produced by daily
clinical implications to restore enamel and other dental hard cycling at 37 °C between demineralization and neutral solutions for 6
tissues. As a major step toward this overarching goal, the and 17.5 h, respectively. Demineralization solution was made of 1.2
objective of this study has been to develop an in vitro, cell-free, mM KH2PO4 (USB Corporation, Cleveland, OH, United States), 2.0
natural remineralization model on artificially induced enamel mM CaCl2 (Johnson Matthey Inc., Seabrook, NH, United States) and
lesions. Because of the unique biomineralization characteristic 75 mM acetic acid (EMD Chemicals, Savannah, GA, United States) in
and its short and simple sequence, shortened ADP5 (shADP5) deionized water at pH 4.2. Neutral solution was made of 0.9 mM
was used in this research as the active ingredient in solution for KH2PO4 (USB Corporation, Cleveland, OH, United States), 1.5 mM
CaCl2, 150 mM KCl (Johnson Matthey Inc., Seabrook, NH, United
the formation of mineralized layer. The work presented herein
States) in 20 mM Tris (EMD Chemicals, Savannah, GA, United
could eventually form the foundation of developing clinical States) buffered at pH 7.0. The demineralization solution was replaced
products (dental gels, toothpastes, and oral-health solutions) daily. Samples were then divided into control and test groups. Test
and treatments for the restoration of early stage cavities, e.g., groups were treated either with peptide, fluoride, or a combination of
incipient caries, white spot lesions, and hypersensitivity. both (Table 1).

B DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

Table 1. Experimental Test Groups and Mineralization for SEM involved cutting a notch on the back side using a low speed
Treatments saw (IsoMet, Buehler, Lake Bluff, IL, United States) before they were
subjected to WSL formation and remineralization as described above.
test groups treatment samples After the remineralization step was completed, specimens were rinsed
group 1: negative no treatment 5 with DI water, air-dried gently (<5 PSI), then carefully fractured into 2
control pieces along the notch. One of the fractured pieces was mounted on a
group 2: Ca2+and 2+
1 h, 4.80 mM Ca /2.89 mM PO4 3−
5 SEM stub with the mineralizing surface facing up for imaging the
O43− only surface morphology, and the second piece was mounted with the
group 3: low 1 h, 1100 ppm F (fluoridated toothpaste 5 cross-section facing up to show the thickness of the mineral layer.
concentration F concentrations), 4.80 mM Ca2+/2.89 mM Mounted specimens were then stored in vacuum for at least 2 h to
PO43− remove residual moisture, which were then sputter coated with 5 nm-
group 4: high 1 h, 20 000 ppm F (dental varnish 5 thick platinum (SPI-Sputter Module Coater, SPI Supplies, West
concentration F concentration), then 4.80 mM Ca2+/2.89 Chester, PA, United States). SEM characterization was performed
mM PO43−
using an FEI Sirion microscope (Sirion, FEI, Hillboro, OR, United
group 5: shADP5 (i) 10 min, 0.80 mM peptide 5
with low States) operating at 10 keV acceleration voltage. The chemical
(ii) 1 h, 1100 ppm F + 4.80 mM Ca2+/2.89 composition was measured by an onboard energy dispersive X-ray
concentration F mM PO43−
spectroscopy (EDXS) system (X-MaxN Si drift detector with
group 6: with (i) 10 min, 0.80 mM peptide 5
peptide, shADP5 AZtecEnergy software package, Oxford Instruments, Abingdon,
(ii) 1 h, 4.80 mM Ca2+/2.89 mM PO43− Oxfordshire, UK). The measurements for each group were pooled
from five specimens per group. The average values and standard
Peptide Design and Synthesis. The peptide shADP5, shortened deviations were calculated and expressed as the mean ± standard error.
ADP5, was generated using a procedure that was developed for Possible mineral phases were deduced from these EDXS measure-
designing protein-derived peptides, as described previously31 (see also ments. It should be noted that given the energy resolution of the
short synopsis of the procedure in Supporting Information). The instrument and topographical variations, precision of the EDXS
peptide (Table 2) was synthesized by using an automated solid-phase measurements was >2%.
Structural Characterization by Transmission Microscopy
Table 2. Molecular Characteristics of the Peptide shADP5a (TEM). After remineralization steps were completed, TEM samples
were collected by carefully shaving off the topmost surface of the
remineralized layer from the artificially created white spot lesion using
a clean razor blade. The shaved particles were suspended in 100%
ethanol, and the suspension was drop-casted onto a carbon coated
TEM grid, which was then vacuum-dried before TEM characterization.
TEM bright field imaging (BF) imaging and selected areas diffraction
were carried out using an FEI Tecnai (FEI, Hillboro, OR, United
a States) operating at 200 keV.
Color coding of amino acids is according to ref 41.
Mechanical Properties Characterization. Similar to SEM
specimen preparation, tooth samples were notched from the back of
tooth before remineralization, then fractured along the notch. The
synthesizer (CS336X; CS-Bio, Menlo Park, CA, United States)
specimens were then mounted in a room temperature-cure epoxy, and
through Fmoc-chemistry. In this procedure, in the reaction vessel,
the cross-section of the fracture was polished to 0.1 μm finish using
the Wang resin (Novabiochem, West Chester, PA, United States), was
diamond lapping films (Allied High Tech Products Inc., Rancho
treated with 20% piperidine in DMF to remove the preloaded Fmoc
Dominguez, CA, United States). Nanoindentation measurements were
group. Next, the incoming side chain protected amino acid was
made using a Triboindentor nanoindentation system (Hysitron Inc.,
activated with HBTU (Sigma-Aldrich, St Louis, MO, United States) in
Minneapolis, MN, United States) in air. Hardness (H) and elastic
dimethylformamide (DMF, Sigma-Aldrich) and then transferred into
modulus (Er) were determined by the software accompanying the
the vessel where it was incubated with the resin for 45 min. After the
nanoindentation unit (see Supporting Information).42−44 To obtain
resin was washed with DMF, this protocol was applied for the addition
the values that were not indentation volume dependent, maximum
of each of the next amino acids, and synthesis reaction was monitored
indentation depth for all measurements kept at 120 ± 10 nm. All
by UV absorbance at 301 nm. Following synthesis, the resulting resin-
reported H and Er values were averaged over 20 measurements.
bound peptides were cleaved and the side-chain deprotected using
In addition to nanoindentation measurements, microhardness
reagent-K [TFA:thioanisole:H2O:phenol:ethanedithiol (87.5:5:5:2.5),
testing was also performed on the surface to quantitatively assess
Sigma-Aldrich] and precipitated by cold ether. Crude peptides were
the mechanical properties of the mineral layers with larger areas and
purified by RP-HPLC with up to >98% purity (Gemini 10u C18 110A
volumes including the underlying enamel as a composite. Vicker’s
column). The sequence of the peptides was confirmed by a MALDI-
microhardness was performed at room temperature using Vicker’s
TOF mass spectrometry with reflectron (RETOF-MS) on an Autoflex
indenter on a Wilson Hardness Tukon 1202 microhardness tester at
II (Bruker Daltonics, Billerica, MA, United States).
10 kg applied load (Illinois Tool Works, Lake Bluff, IL). At least 20
Remineralization Protocol. Prior to remineralization, samples
measurements per group were recorded for obtaining an average and
requiring peptide treatments (groups 5 and 6) were incubated in 50
statistical analysis.

■
μL of 0.8 mM peptide dissolved in 50 mM Tris buffer solution (TBS)
(pH 7.4) for 10 min at 37 °C. Next, treatment samples were placed
into 800 μL of 50 mM TBS containing Ca2+/PO43− (groups 2 and 5) RESULTS
or Ca2+/PO43−/F− (groups 3, 4, and 6) at concentrations as listed in The incubation of samples in demineralization cocktail exposed
Table 1 for 1 h at 37 °C and then rinsed with deionized (DI) water, enamel rods on the surface of the samples before the
dried by forced air, and stored at room temperature until remineralization treatment was undertaken, as shown in
characterization. group 1, negative control (Figures 2a and b). Elemental
Sample Characterization by SEM and EDXS Analyses:
Imaging and Elemental Composition. After remineralization compositional analysis of the surface by EDXS gives a ratio of
experiments were completed, secondary electron imaging (SEI) in Ca2+/PO43− 1.56 ± 0.12 (Figure 2c). As seen in the cross-
the scanning electron microscope (SEM) was used to characterize the sectional view of (Figure 2d), well-aligned enamel rods of ∼3
surface morphology and to show the thickness of newly formed μm diameter extend to the exposed surface where they display
mineral layer in cross sections where applicable. Specimen preparation HAp crystallites constituting the rods. After 1 h of exposure to
C DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

concentration of the most commonly used toothpaste available

over the counter for daily home care.5−7 The analysis of the
SEM images suggests nonuniformly deposited layer with a fine
(<1 μm) roughness (Figures 3a and b). A detailed analysis of

Figure 2. Face-on (a and b) and edge-on (d) SEM images and EDXS
analyses (c) of group 1. Face-on (e and f) and edge-on (h) SEM
images and EDXS analysis (g) of group 2: Ca2+ and PO43− only. Insets
in b and f show enamel rods and HAp crystallites exposed on the
surface of damaged enamel as a result of demineralization. The inset Figure 3. Face-on (a and b) and edge-on (d) SEM images and EDXS
panels are 1 × 1 μm. analysis (c) of group 3, 1100 ppm F− + Ca2+/PO43−. Face-on (e and f)
and edge-on (h) SEM images and EDXS analysis (g) of group 4,
20 000 ppm F + Ca2+/PO43−. Insets in b and f show loosely packed
Ca2+/PO43− solution, no substantial remineralization was nanospherical particles (of diameter ∼20−30 nm) as a result of F
observed on the samples in group 2. Considering that the deposition. The inset panels are 1 × 1 μm. Wide arrows in panels d
imprints of enamel rods remained visible as shallow depressions and h indicate the boundary between the new layer and original tooth
on the enamel surface (delineated with circles, Figures 2e and surface.
f), any possible deposit of solid material, possibly the result of
Ca2+ and PO43− ions reacting to form an amorphous deposit,
remained extremely thin. In fact, a very thin (≪1 μm) layer is
barely visible in the SEM image of the cross-sectioned sample the surface structure, e.g., at higher magnification image in
shown in Figure 2h. Elemental analysis of the surface by EDXS Figure 3b, reveals fine nanoparticles of diameter 20−50 nm.
gives a ratio of Ca2+/PO43− 1.45 ± 0.04, possibly indicating a The cross-sectioned samples reveal a new layer with a thickness
mixed mineral composition (Figure 2g; also see Table 3). of about 1 μm covering the surface of enamel in the lesion
In group 3 (low concentration fluoride), 1100 ppm F was (Figure 3d). The elemental composition analysis from the
applied in the presence of Ca2+ and PO43− ions. The surface revealed prominent peaks of OKα, PKα, and CaLα as well
concentration of 1100 ppm fluoride corresponds to the as a small peak corresponding to FKα. The Ca/F ratio gives a

Table 3. Elemental Composition Analyses of the Remineralization Test Groups by EDXS

remineralized layer possible mineral

test group Ca/P Ca/F Ca/O formed
1: control 1.56 ± 0.12 0.51 ± 0.11 only HAp
2: ions only 1.45 ± 0.04 0.28 ± 0.02 amorphous Ca−P transition phase
3: low F 1.69 ± 0.05 36.99 ± 2.9 0.74 ± 0.15 Ca−P−F transition phase + CaF2
4: high F 5.62 ± 0.96 0.49 ± 0.11 1.01 ± 0.16 mainly CaF2
5: low F + peptide 1.60 ± 0.11 8.74 ± 0.78 0.34 ± 0.08 FAp + some CaF2
6: peptide only 1.54 ± 0.12 0.56 ± 0.13 only HAp

D DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

values of more than 30, while the Ca/P ratio is close to 1.70
(Figure 3c).
In group 4 (high concentration fluoride), 20 000 ppm
fluoride (concentrations of most commonly used dental
varnishes) applied with Ca2+ and PO43− ions.5−7 The analysis
of the SEM images recorded from this treatment displayed
significantly different surface topography, structures, and
elemental composition as compared to the samples in the
previous groups of nonfluoride or low concentration F
treatment. Although at low magnifications (Figure 3e) the
surface appears fairly smooth, higher magnification (Figure 3f)
revealed small spherical particles of 100−200 nm diameter
covering the overall surface (indicated by arrow in Figure 3f).
The secondary electron images, recorded with the SEM, from
the cross-sectioned samples reveal an about a micrometer-thick
new layer on the surface of the teeth (Figure 3h). The EDXS
spectra acquired from the surface gives a high concentration of
FKα peak, the most prominent among all the peaks in the
spectra from this group of samples (Figure 3g). The
quantitative analysis of the spectra from the samples prepared
in this group exhibited the Ca/F ratio of 0.49 (Table 3).
In group 5 (the peptide with low concentration fluoride),
shADP5 was applied with 1100 ppm fluoride along with Ca2+
and PO43−. The microstructure of the samples displays fairly
smooth surface with about 1−2 μm thickness (Figures 4a−d).
Enamel rod imprints remained visible in the lower magnifica-
tion image (Figure 4a). Higher magnification image of the
sample surface, however, exhibits two different surface
morphologies (see insets in Figure 4b): somewhat loosely
deposited nanoparticles of 50−100 nm diameter and dense Figure 4. Face-on (a and b) and edge-on (d) SEM images and EDXS
analyses (c) of group 5, shADP5 + 1100 ppm F− + Ca2+/PO43−. Insets
structure composed of rod-like nanoparticles of few tens of in b show loosely crystallized regions of accumulated 100 nm diameter
nanometers in diameters with the diameter/length aspect ratio spherical nanoparticles on the surface. Face-on (e and f) and edge-on
of 1/5. Elemental analysis of the samples from this group (h) SEM images and EDXS analysis (g) of group 6, shADP5 + Ca2+/
revealed fairly noticeable FKα peak in addition to highly PO43−. Inset f displays a highly uniform, plate-like HAp crystallites
prominent CaKα and PKα peaks with the elemental ratio of Ca/ within newly formed (h) mineral layer in shADP5 + Ca2+/PO43−
F, 8.7 (Figure 4c). treatment. The inset panels are 1 × 1 μm. Wide arrows in panels d and
In group 6 (shADP5 + Ca2+/PO43−), the SEM images in h indicate the boundary between the new layer and original tooth
Figures 4e and f give a continuous layer of plate-like crystals surface.
growing from the surface of the underlying enamel lesion when
the surface is exposed to aqueous peptide plus Ca2+/PO43−. diameter were observed (Figures 5d−f). On the peptide
Compared to the negative control (group 1) or low treatment group (group 6), particles (in the shown projection)
concentration fluoride treatment (group 3), the enamel rod of HAp were found (Figures 5g−i), possibly corresponding to
imprints in the face-on images are no longer visible, indicating plate-shape mineral. It is noted that groups 2, 3, and 5 had
that the new mineral layer is thick enough to mask the structural characteristics similar to those of group 1 in that the
previously exposed enamel rods (Figures 4e and f). The cross- mineral had morphological characteristics similar to those of
sectional image in Figure 4h shows a 10 μm-thick continuous enamel crystallites (data not shown here, but given in the
remineralized layer with fairly smooth surface topography. Supporting Information). In all cases, it was challenging but not
Elemental analysis of the samples from this group revealed impossible (as demonstrated in Figure 5 above) to differentiate
prominent CaKα and PKα peaks with a ratio of 1.54 ± 0.12; this the newly formed crystallites from the HAp crystallites present
is close to ideal ionic ratio of 1.6 in HAp composition (Figure in the underlying enamel.
4g; Table 3). Mechanical properties of the mineralized layers were
To further analyze the structural characteristics of the determined using two different tests. The microhardness test
mineral layers, imaging and diffraction analyses were also was carried out using a Vicker’s indenter loading on the
carried out on all samples by using transmission electron mineralized tooth surface. The hardness for the negative
microscopy (Figure 5 and Figure S2, see Supporting control group (group 1) was 128 ± 8 HV10, which was used as
Information). The TEM samples were prepared by gently the baseline figure representing microhardness of the surface of
shaving fragments off the surface of tooth specimens. Group 1 bare, artificially created WSL against which other experimental
received no mineralization treatment, and the enamel fragments groups compared. As further reference, the microhardness tests
were analyzed. As shown in Figures 5a−c, textured, elongated were also conducted on healthy enamel and dentin, away from
HAp crystals of 30−50 nm were encountered, typical of the demineralized surface (Table 4). The values for groups 2−5
prismatic rods constituting the enamel rods in healthy enamel ranged between 130 ± 12 HV10 to 134 ± 12 HV10. Unequal
tissue. In the case of high concentration F treatment (group 4), variance t test between group 1 and each of other groups
generally round particles CaF2 in the range of 100−250 nm in revealed no statistically significant difference (p > 0.05). The
E DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

Figure 5. TEM bright field images and corresponding selected diffraction patterns for no-treatment negative control (group 1) (a−c), showing high-
aspect ratio rod-like HAp crystallites; high concentration F treatment (group 4) (d−f), exhibiting CaF2 particles; and peptide-treatment (group 6)
(g−i), showing plate-like HAp crystallite formation.

Table 4. Vicker’s Microhardness of All Experimental Groups, from the no-treatment negative control and each of the groups
n ≥ 20 2 through 5 in both hardness and reduced elastic modulus, p >
0.05 in all cases. However, the average hardness and elastic
test group hardness (HV10, MPa) STDEV (MPa)
modulus for group 6 were higher than those of the no-
group 1: negative control 128.1 8.0 treatment samples with hardness of 2.23 ± 0.23 GPa vs 2.10 ±
group 2: Ca/PO4 only 130.1 11.7 0.26 GPa, p = 0.02 and elastic modulus of 58.6 ± 4.7 GPa vs
group 3: low conc F 130.6 12.6 55.1 ± 4.3 GPa, p = 0.02. Not surprisingly, the healthy enamel
group 4: high conc F 131.8 12.9 and dentin had, respectively, higher and lower values of both
group 5: shADP5 + low conc F 133.5 12.4 hardness and elastic moduli compared to the experimental
group 6: shADP5 141.1 7.8 groups involving remineralization. In conclusion, the mechan-
healthy enamel 290.6 20.1 ical properties (H and E) are higher than those of dentin but
healthy dentin 63.1 3.0 lower than those of the healthy enamel.

microhardness values of group 6 had slightly higher average

value of 141 ± 8 HV10. Unequal variance t test against group 1
■ DISCUSSION
A natural, cell-free, biomimetic model was developed to
revealed significant difference with p ≪ 0.01. The results remineralize artificially induced lesions on human enamel
indicate that the microhardness values of group 6 as well as the using a 15-amino acid long amelogenin-derived peptide,
rest of experimental groups fell between that of enamel and shADP5, along with properly tuned ionic concentrations of
dentin. Ca2+/PO43− in vitro in the presence and absence of low and
Nanomechanical tests were also conducted using nano- high fluoride content that were chosen based on the values in
indentation that not only provided hardness (H) but also elastic the frequently used current dental treatments. Significant
modulus (Er) values, and the tests were carried out in spatially differences were encountered among the surface characteristics
selected regions as the test facilitates scanned surface images. of the samples from six different groups of tests. The surface of
The nanoindentation tests of all the experimental groups were the artificially demineralized enamel displayed enamel rods, and
conducted on samples in cross-sectional geometry, i.e., indentor up to 3-μm diameter depressions exposed on the surface
direction being parallel to the surface (as opposed to vertical in appeared (Figures 2a and b) with a roughness of approximately
microhardness tests) (Figure S1a). The results (Figure 6) 1 μm. Another prominent feature was the fine structure of the
reveal a trend similar to the those of the microhardness data; no individual rod-shaped HAp crystallites of a few tens of
significant differences were encountered among the samples nanometers in thickness that constitute the enamel rods.
F DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

incorporated into the newly formed surface layer. The

concentration in the newly formed layer may imply that
majority of F was not delivered to the desired mineralization
site on the tooth surface in the peptide-free samples. The
presence of nanoparticles deposited on the surface, therefore,
may be due to F reacting with the excess Na in the buffer
solution forming NaF.
In group 4, the new layer is predominantly composed of
aggregated spherical nanoparticles. Considering the ideal Ca/F
ratio of 0.5 in CaF2 (Table S1, Supporting Information, and
Figures 5d−f, TEM results), the spherical particles are likely to
be CaF2. Another major difference in this group was the value
of CaKα/PKα ratio being >5.0. Even considering that some of
the Ca ions might be confined to CaF2, this ratio still indicates
unusually high concentration of Ca trapped in the newly
formed surface layer. Calcium fluoride is a highly stable
compound, likely to form under the experimental conditions
used in this study. The clinical products that contain high F
concentration are designed for forming apatite in the
demineralized product. However, as shown here, the mineral
formed is most likely calcium fluoride. In summary, neither of
the F samples exhibited the ideal ratio of Ca/P of 1.6 for HAp
(Table 3), meaning that F alone was not incorporated into the
enamel or remineralized layer. These findings are consistent
Figure 6. (a) Atomic force microscopy images of the surfaces of the with previous studies that reported CaF2 formation under
mineralized layers in samples from group 2 (left), where there is no home-use (low-F) and clinical (high-F) products.45−47 It has
apparent mineral layer on the lesion, and group 6 (right), showing a been suggested that CaF2 may serve as a reservoir of fluoride,
clear boundary between the lesion and newly formed mineral layer presumably, to be incorporated later into the structure of
(arrows). (b) Hardness (left) and elastic modulus (right) of the enamel.47−49
experimental groups used here were measured by nanoindentation, n > In groups 5 and 6, the peptide was used as the precursor in
20. the remineralization procedure. In group 5, shADP5 was
applied in the presence of 1100 ppm fluoride, resulting in the
When ionic precursors are used alone (group 2), the formation of the spherical particles as aggregates as opposed to
treatment resulted in a thin (<1 μm) layer with a highly porous being widely disseminated on the tooth surface compared to
morphology on the surface. The composition was off- the same F concentration without peptide in group 3. Enamel
stoichiometric for that of HAp, with Ca/P ratio of <1.5, rod imprints remained visible in the lower magnification image
possibly due to the formation of calcium-phosphate transition (Figures 4a and b), although these were less prominent than
phases (Table 3 and Table S1). The effect of F on mineral those seen in the no-treatment samples (Figures 2a and b). The
formation was examined under 2 different fluoride ion resulting mineralized structure presented two morphologies:
conditions, 1100 ppm in group 3 and 20 000 ppm in group clusters of 50−100 nm diameter spherical nanoparticles accrued
4, specifically chosen to mimic the F concentrations of the usual nonuniformly on the surface, and a structure primarily
over-the-counter toothpaste and clinical fluoride varnish, composed of highly dense nanorods. There was a considerably
respectively. The surface structures of the teeth in these two more prominent F peak compared to the no-peptide samples,
treatments revealed different morphologies. First, F treatments with an overall Ca/F ratio of 8.74 ± 0.78. This explains that
resulted in aggregates of nanoparticles in group 3, and a thin there was considerably more F in the mineral formed with the
mineralized layer of about 1 μm in group 4. The layers were peptide containing low-fluoride treatment compared to the
composed of nanoparticles which were about an order of samples with low-fluoride only (Ca/F = 36.99 ± 2.9). If the
magnitude smaller in group 3 than in group 4 samples, ∼20 nm presence of FAp is considered, i.e., corresponding to the dense
versus ∼200 nm, respectively. The application of F in dental nanorods, the ideal ratio of Ca/F turns out to be 5.0 (Table 3),
care products primarily focuses on remineralization, aided by F then the rest of the fluoride in the mineral layer could be
or incorporation of F into the existing HAp structure, desirably accounted for the formation of NaF nanoparticles. Considering
forming FAp. In this respect, the results of the elemental that the observed Ca/P ratio reflects either HAp or FAp
analyses obtained from the fluoride-treated surfaces are quite stoichiometry (Table 3), the new mineral was formed on the
intriguing. The group 3 samples (low-F) presented hardly any teeth surface by partially incorporating F in the presence of
F-peaks in the EDXS spectra, giving Ca/F ratio of almost 40. peptide. The presence of F could be explained in at least two
The elemental composition analysis of the sample surface ways: either the fluoride was incorporated into the newly
displayed relatively strong peaks of OKα, PKα, and CaLα as well forming HAp mineral replacing OH partially, or both HAp and
as a minor peak corresponding to FKα (NaKα and ClLα peaks are FAp were formed on the surface. In both cases, the effect of
from the treatment solutions). Possible sources of the low F peptide appears to be necessary to incorporate fluoride into the
concentration might be explained either as being due to the structure because in the absence of shADP5, very little or no F
formation of a very thin newly deposited layer in the signals was found in the remineralized layer. The significance here is
mainly originating from the underlying healthy enamel, as that dental formulations used in the current treatments, e.g.,
determined EDXS, or due to the very low amount of F pastes, gels, varnish, and solutions, could contain peptide as a
G DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

means to effectively carry F to the surface and help incorporate Detailed description of the design of amelogenin-derived
it into the remineralized HAp on the tooth surface. peptides; a list of calcium, phosphate, fluoride, and
In group 6, shADP5 peptide in addition to calcium and oxygen containing minerals with the calculated elemental
phosphate precursors were used which remineralized a thick composition ratios; TEM studies of the samples from all
(>10 μm) layer composed of crystal morphology specific to six test groups; and comparison of microhardness testing
HAp among the calcium phosphate polymorphs. Considering and nanoindentation plus a table of hardness and elastic
that he observed Ca/P ratio is 1.54, close to that of ideal HAp modulus values from nanoindentation tests (PDF)

■
composition, it is concluded that the peptide was capable to
catalyze a newly mineralized layer composed of HAp crystallites
(Figures 5g−i). As evident from SEM images (Figures 4e−f), AUTHOR INFORMATION
the roughness of remineralized surface was reduced (<300 nm) Corresponding Author
compared to the demineralized surface in group 1 (Figures 2a, *Phone: 206-543-0724; Fax: 206-543-3100; E-mail: sarikaya@
b, and d). u.washington.edu.
The loading in the microhardness tests were carried out in
ORCID
the direction relevant to functional dentition loading while
using the lowest possible Vicker’s indentation load to maximize Mehmet Sarikaya: 0000-0003-3856-6360
the contribution from the mineralization layer. Even with the Notes
lowest load, however, the indentor likely penetrated through The authors declare no competing financial interest.

■
the thin mineralized layers in the samples from groups 2−5 as
no discernible measurements were determined among these ACKNOWLEDGMENTS
groups compared to the no-treatment negative control group.
The SEM observations supported these findings as the mineral This research was mainly supported (D.T.Y., H.F., M.S.) by
layers in groups 2−5 were significantly thinner (∼1−2 μm or WA-LSDF (Washington State Life Sciences Discovery Funds)
less) and discontinuous. The remineralized layer, however, was and also by NSF through MGI/DMREF program (DMR-
thicker in peptide-guided remineralized group, reflected in 1629071) at GEMSEC, Genetically Engineered Materials
higher microhardness values, although slight, compared to the Science and Engineering Center (D.T.Y., M.S.). S.D., G.H.,
negative control group. and T.C. were supported by Spencer Fund (Restorative

■
Dentistry, University of Washington); D.T.Y. was also
CONCLUSIONS supported by TC-Education Ministry Fund, and S.D. by
The present in vitro study demonstrated that crystalline mineral NIDCR T32. The work was carried out at the GEMSEC-SECF,
a member of Materials Facilities Network of MRSEC.

■
layer is formed on an artificially created lesion on human
enamel in the presence of Ca2+ and PO43− ions under
physiologically viable conditions by using shADP5, a 15-AA REFERENCES
long amelogenin-derived peptide. This study also showed that (1) Selwitz, R. H.; Ismail, A. I.; Pitts, N. B. Dental caries. Lancet 2007,
the presence of biomineralizing peptide also facilitated the 369, 51−59.
delivery to tooth surface and incorporation of fluoride ions into (2) Gorelick, L.; Geiger, A. M.; Gwinnett, A. J. Incidence of white
the remineralized layer even at low F concentrations, providing spot formation after bonding and banding. Am. J. Orthod. 1982, 81,
an opportunity for dental health products to incorporate both 93−98.
(3) Featherstone, J. D. The science and practice of caries prevention.
elements in potential clinical and everyday dental product
J. Am. Dent. Assoc., JADA 2000, 131, 887−899.
formulations. (4) Huang, Z.; Newcomb, C. J.; Bringas, P., Jr; Stupp, S. I.; Snead, M.
Establishing the scientific foundations for remineralization L. Biological synthesis of tooth enamel instructed by an artificial
has a high potential to empower practicing dentists to address matrix. Biomaterials 2010, 31, 9202−9211.
caries-related dental problems due to enamel demineralization, (5) American Dental Association Council on Scientific, A.
the origin of most dental ailments. Although remineralization Professionally applied topical fluoride: Evidence−based clinical
on human teeth in vitro using peptides on artificially damaged recommendations. J. Am. Dent. Assoc., JADA 2006, 137, 1151−1159.
enamel discussed herein provides guidance toward addressing (6) Ammari, A. B.; Bloch-Zupan, A.; Ashley, P. F. Systematic review
this challenge, the procedures developed need to be further of studies comparing the anti-caries efficacy of children’s toothpaste
optimized toward clinical applications. Further research containing 600 ppm of fluoride or less with high fluoride toothpastes
of 1,000 ppm or above. Caries Res. 2003, 37, 85−92.
concerning the repair of enamel defects is necessary to achieve
(7) Petersson, L. G.; Twetman, S.; Dahlgren, H.; Norlund, A.; Holm,
an easy-to-apply, fast-growing enamel-like biomineralized tissue A. K.; Nordenram, G.; Lagerlof, F.; Soder, B.; Kallestal, C.; Mejare, I.;
for biomimetic tooth repair. Future studies (underway in the Axelsson, S.; Lingstrom, P. Professional fluoride varnish treatment for
authors’ laboratories) include implementing this advanced caries control: a systematic review of clinical trials. Acta Odontol. Scand.
remineralization technology under in vivo conditions by 2004, 62, 170−176.
utilizing clinically applicable peptide delivery systems (e.g., (8) Cate, J. M. Review on fluoride, with special emphasis on calcium
gels or pastes with or without F) and applying tests to the fluoride mechanisms in caries prevention. Eur. J. Oral Sci. 1997, 105,
mineralized layer to ensure its mechanical and chemical 461−465.
durability and adherence to the underlying tooth structure. (9) Al-Mullahi, A. M.; Toumba, K. J. Effect of slow-release fluoride

■
devices and casein phosphopeptide/amorphous calcium phosphate
ASSOCIATED CONTENT nanocomplexes on enamel remineralization in vitro. Caries Res. 2010,
44, 364−371.
*
S Supporting Information
(10) Krithikadatta, J.; Fredrick, C.; Abarajithan, M.; Kandaswamy, D.
The Supporting Information is available free of charge on the Remineralisation of occlusal white spot lesion with a combination of
ACS Publications website at DOI: 10.1021/acsbiomater- 10% CPP-ACP and 0.2% sodium fluoride evaluated using Diagnodent:
ials.7b00959. a pilot study. Oral Health Prev Dent 2013, 11, 191−196.

H DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX
ACS Biomaterials Science & Engineering Article

(11) Meyer-Lueckel, H.; Wierichs, R. J.; Schellwien, T.; Paris, S. (32) Oren, E. E.; Tamerler, C.; Sahin, D.; Hnilova, M.; Seker, U. O.;
Remineralizing efficacy of a CPP-ACP cream on enamel caries lesions Sarikaya, M.; Samudrala, R. A novel knowledge-based approach to
in situ. Caries Res. 2015, 49, 56−62. design inorganic-binding peptides. Bioinformatics 2007, 23, 2816−
(12) Rølla, G.; Saxegaard, E. Critical Evaluation of the Composition 2822.
and Use of Topical Fluorides, with Emphasis on the Role of Calcium (33) Gungormus, M.; Fong, H.; Kim, I. W.; Evans, J. S.; Tamerler, C.;
Fluoride in Caries Inhibition. J. Dent. Res. 1990, 69, 780−785. Sarikaya, M. Regulation of in vitro calcium phosphate mineralization
(13) Sitthisettapong, T.; Phantumvanit, P.; Huebner, C.; Derouen, T. by combinatorially selected hydroxyapatite-binding peptides. Bioma-
Effect of CPP-ACP paste on dental caries in primary teeth: a cromolecules 2008, 9, 966−973.
randomized trial. J. Dent. Res. 2012, 91, 847−852. (34) Moradian-Oldak, J.; Paine, M. L.; Lei, Y. P.; Fincham, A. G.;
(14) Vanichvatana, S.; Auychai, P. Efficacy of two calcium phosphate Snead, M. L. Self-Assembly Properties of Recombinant Engineered
pastes on the remineralization of artificial caries: a randomized Amelogenin Proteins Analyzed by Dynamic Light Scattering and
controlled double-blind in situ study. Int. J. Oral Sci. 2013, 5, 224−228. Atomic Force Microscopy. J. Struct. Biol. 2000, 131, 27−37.
(15) Wong, M. C.; Clarkson, J.; Glenny, A. M.; Lo, E. C.; Marinho, (35) Paine, M. L.; White, S. N.; Luo, W.; Fong, H.; Sarikaya, M.;
V. C.; Tsang, B. W.; Walsh, T.; Worthington, H. V. Cochrane reviews Snead, M. L. Regulated gene expression dictates enamel structure and
on the benefits/risks of fluoride toothpastes. J. Dent. Res. 2011, 90, tooth function. Matrix Biol. 2001, 20, 273−292.
573−579. (36) Snead, M. L.; Lau, E. C.; Zeichner-David, M.; Fincham, A. G.;
(16) Balakrishnan, M.; Simmonds, R. S.; Tagg, J. R. Dental caries is a Woo, S. L.; Slavkin, H. C. DNA sequence for cloned cDNA for murine
preventable infectious disease. Aust. Dent. J. 2000, 45, 235−245. amelogenin reveal the amino acid sequence for enamel-specific
(17) Featherstone, J. D. The caries balance: the basis for caries protein. Biochem. Biophys. Res. Commun. 1985, 129, 812−818.
management by risk assessment. Oral Health Prev Dent 2004, 2, 259− (37) Han, S.; Fan, Y.; Zhou, Z.; Tu, H.; Li, D.; Lv, X.; Ding, L.;
Zhang, L. Promotion of enamel caries remineralization by an
264.
amelogenin-derived peptide in a rat model. Arch. Oral Biol. 2017,
(18) Yamagishi, K.; Onuma, K.; Suzuki, T.; Okada, F.; Tagami, J.;
73, 66−71.
Otsuki, M.; Senawangse, P. Materials chemistry: a synthetic enamel for
(38) Hoppenbrouwers, P. M.; Driessens, F. C. The effect of lactic
rapid tooth repair. Nature 2005, 433, 819.
and acetic acid on the formation of artificial caries lesions. J. Dent. Res.
(19) Huang, S.; Gao, S.; Cheng, L.; Yu, H. Remineralization potential
1988, 67, 1466−1467.
of nano-hydroxyapatite on initial enamel lesions: an in vitro study. (39) Lynch, R. J. M.; ten Cate, J. M. The Effect of Lesion
Caries Res. 2011, 45, 460−468. Characteristics at Baseline on Subsequent De- and Remineralisation
(20) Ingram, G. S.; Silverstone, L. M. A chemical and histological Behaviour. Caries Res. 2006, 40, 530−535.
study of artificial caries in human dental enamel in vitro. Caries Res. (40) Naumova, E. A.; Niemann, N.; Aretz, L.; Arnold, W. H. Effects
1981, 15, 393−398. of different amine fluoride concentrations on enamel remineralization.
(21) Koulourides, T.; Cueto, H.; Pigman, W. Rehardening of J. Dent. 2012, 40, 750−755.
softened enamel surfaces of human teeth by solutions of calcium (41) Lesk, A. Introduction to Bioinformatics; Oxford University Press:
phosphates. Nature 1961, 189, 226−227. Oxford, 2013.
(22) ten Cate, J. M.; Jongebloed, W. L.; Arenas, J. Remineralization (42) Fong, H.; Sarikaya, M.; White, S. N.; Snead, M. L. Nano-
of artificial enamel lesions in vitro. IV. Influence of fluorides and mechanical properties profiles across dentin-enamel junction of human
diphosphonates on short- and long-term reimineralization. Caries Res. incisor teeth. Mater. Sci. Eng., C 1999, 7, 119−128.
1981, 15, 60−69. (43) Gallagher, R. R.; Balooch, M.; Balooch, G.; Wilson, R. S.;
(23) Bleek, K.; Taubert, A. New developments in polymer-controlled, Marshall, S. J.; Marshall, G. W. Coupled Nanomechanical and Raman
bioinspired calcium phosphate mineralization from aqueous solution. Microspectroscopic Investigation of Human Third Molar DEJ. J. Dent
Acta Biomater. 2013, 9, 6283−6321. Biomech 2010, 1, 256903.
(24) Fan, Y.; Sun, Z.; Moradian-Oldak, J. Controlled remineralization (44) Oliver, W. C.; Pharr, G. M. An improved technique for
of enamel in the presence of amelogenin and fluoride. Biomaterials determining hardness and elastic modulus using load and displacement
2009, 30, 478−483. sensing indentation experiments. J. Mater. Res. 1992, 7, 1564−1583.
(25) Fowler, C. E.; Li, M.; Mann, S.; Margolis, H. C. Influence of (45) Harding, A. M.; Zero, D. T.; Featherstone, J. D.; McCormack, S.
surfactant assembly on the formation of calcium phosphate materials - M.; Shields, C. P.; Proskin, H. M. Calcium fluoride formation on
A model for dental enamel formation. J. Mater. Chem. 2005, 15, 3317− sound enamel using fluoride solutions with and without lactate. Caries
3325. Res. 1994, 28 (1), 1−8.
(26) Kirkham, J.; Firth, A.; Vernals, D.; Boden, N.; Robinson, C.; (46) Petzold, M. The influence of different fluoride compounds and
Shore, R. C.; Brookes, S. J.; Aggeli, A. Self-assembling peptide scaffolds treatment conditions on dental enamel: a descriptive in vitro study of
promote enamel remineralization. J. Dent. Res. 2007, 86, 426−430. the CaF(2) precipitation and microstructure. Caries Res. 2001, 35, 45−
(27) Li, Q. L.; Ning, T. Y.; Cao, Y.; Zhang, W. B.; Mei, M. L.; Chu, 51.
C. H. A novel self-assembled oligopeptide amphiphile for biomimetic (47) Cate, J. M. Review on fluoride, with special emphasis on calcium
mineralization of enamel. BMC Biotechnol. 2014, 14, 32. fluoride mechanisms in caries prevention. Eur. J. Oral Sci. 1997, 105 (5
(28) Ruan, Q.; Zhang, Y.; Yang, X.; Nutt, S.; Moradian-Oldak, J. An Pt 2), 461−465.
amelogenin-chitosan matrix promotes assembly of an enamel-like layer (48) Tenuta, L. M.; Cerezetti, R. V.; Del Bel Cury, A. A.; Tabchoury,
with a dense interface. Acta Biomater. 2013, 9, 7289−7297. C. P.; Cury, J. A. Fluoride release from CaF2 and enamel
(29) Shafiei, F.; Hossein, B. G.; Farajollahi, M. M.; Fathollah, M.; demineralization. J. Dent. Res. 2008, 87 (11), 1032−1036.
Marjan, B.; Tahereh, J. K. Leucine-rich amelogenin peptide (LRAP) as (49) Tenuta, L. M.; Zamataro, C. B.; Del Bel Cury, A. A.; Tabchoury,
a surface primer for biomimetic remineralization of superficial enamel C. P.; Cury, J. A. Mechanism of fluoride dentifrice effect on enamel
defects: An in vitro study. Scanning 2015, 37, 179−185. demineralization. Caries Res. 2009, 43 (4), 278−285.
(30) Wu, D.; Yang, J.; Li, J.; Chen, L.; Tang, B.; Chen, X.; Wu, W.; Li,
J. Hydroxyapatite-anchored dendrimer for in situ remineralization of
human tooth enamel. Biomaterials 2013, 34, 5036−5047.
(31) Gungormus, M.; Oren, E. E.; Horst, J. A.; Fong, H.; Hnilova,
M.; Somerman, M. J.; Snead, M. L.; Samudrala, R.; Tamerler, C.;
Sarikaya, M. Cementomimetics-constructing a cementum-like bio-
mineralized microlayer via amelogenin-derived peptides. Int. J. Oral Sci.
2012, 4, 69−77.

I DOI: 10.1021/acsbiomaterials.7b00959
ACS Biomater. Sci. Eng. XXXX, XXX, XXX−XXX