The role of EphB4/ephrinB2 signaling in
root repair after orthodontically-induced
root resorption
Tiancheng Li,a Han Wang,a Ruojing Liu,a Xin Wang,b Li Huang,a Zuping Wu,a Xing Yin,a Shujuan Zou,a
and Peipei Duana
Chengdu and Beijing, China
Introduction: This study aimed to investigate the effect of EphB4/ephrinB2 signaling on orthodontically-induced
root resorption repair and the possible molecular mechanism behind it. Methods: Seventy-two 6-week-old male
Wistar rats were randomly divided into 3 groups: blank control group, physiological regeneration group (PHY),
and EphB4 inhibitor local injection group (INH). A root repair model was built on experimental rats of the PHY and
INH groups. The animals in the INH groups received a daily periodontal local injection of EphB4 inhibitor NVP￾BHG712, whereas the blank control group and PHY groups received only the vehicle. Results: Histologic stain￾ing and microcomputed tomography analysis showed that root regeneration was inhibited in the INH group
compared with the PHY group with a greater number of osteoclasts. Immunohistochemical staining showed
active EphB4/ephrinB2 signaling activities during root regeneration. The cementogenesis-related factors
cementum attachment protein, alkaline phosphatase, osteopontin, and runt-related transcription factor 2, and
osteoclastic-related factors RANKL and osteoprotegerin were affected by regulated EphB4/ephrinB2
signaling. Conclusions: These findings demonstrated that the EphB4/ephrinB2 signaling might be a promising
therapeutic target for novel therapeutic approaches to reduce orthodontically-induced root resorption through
enhancement of cementogenesis. (Am J Orthod Dentofacial Orthop 2021;159:e217-e232)
One of the most common undesirable complications
of orthodontic treatment is orthodontically￾induced root resorption (OIRR).1 Characterized by
the loss of apical root material, such as mineralized
cementum and dentin, this clinical problem can be
affected by the type of tooth movement, force magnitude,
and duration of treatment.2 If handled improperly, OIRR
may occasionally lead to severe root shortening that
threatens the health of the tooth and even the stability
of the treatment results. Therefore, investigating the un￾derlying mechanisms of OIRR and developing novel ther￾apeutic approaches to accelerate its regeneration has been
a popular research subject in recent years.
As a specialized calcified substance covering the root
of a tooth, cementum is the part of the periodontium
that attaches the teeth to the alveolar bone by anchoring
the periodontal ligament (PDL). Under physiological
conditions, cementum is fundamental in maintaining
periodontal homeostasis and long-term stability of
teeth, protecting root from resorption caused by infec￾tion, trauma, or orthodontic treatment.3,4 Its ability to
maintain its shape and repair during tooth movement
is considered an important biological foundation of or￾thodontic treatment.5 Meanwhile, external root resorp￾tion tends to begin with the resorption of cementum
when subjected to undesirable local orthodontic
force.6,7 Therefore, many studies recently have focused
on cementum regeneration under a pathologic state.
Among them, some have shown that resorption craters
might be restored through stimulation of cementogene￾sis.8-10 This suggests that promoting cementogenesis
could be a practical way of solving OIRR.
a
State Key Laboratory of Oral Diseases, National Clinical Research Center for Oral
Diseases, Department of Orthodontics, West China School and Hospital of Sto￾matology, Sichuan University, Chengdu, China.
b
Oral Diagnosis and Treatment Center, Aviation General Hospital, China Medical
University, Beijing, China.
All authors have completed and submitted the ICMJE Form for Disclosure of
Potential Conflicts of Interest, and none were reported.
This work is supported by the National Natural Science Foundation of China
(grant no. 81701005), Science and Technology Department, Sichuan, China
(grant nos. 2018JY0139 and 2019JDRC0099), and China Postdoctoral Science
Foundation (grant no. 2018M640929).
Address correspondence to: Peipei Duan, State Key Laboratory of Oral Diseases,
National Clinical Research Center for Oral Diseases, Department of Orthodontics,
West China School and Hospital of Stomatology, Sichuan University, No. 14, 3rd
section of Renmin South Road, Chengdu, 610041, China; e-mail, duanp@scu.
edu.cn.
Submitted, December 2019; revised and accepted, July 2020.
0889-5406/$36.00

2020 by the American Association of Orthodontists. All rights reserved.
ORIGINAL ARTICLE
Eph receptors belong to a subfamily of receptor
tyrosine kinases activated by ligands called Eph
receptor-interacting proteins (ephrins).11 Ephs and eph￾rins are both divided into 2 A and B groups. Generally,
EphA receptors (EphA1-A8, A10) interact with ephrinA
(ephrinA1-A5), whereas EphB receptors (EphB1-B6)
interact with ephrinB ligands (ephrinB1-B3).12 The
Eph/ephrin signaling participates in a wide spectrum of
developmental processes, including spatial organization
of different cell populations, axon guidance, formation
of synaptic connections between neurons, and blood
vessel remodeling.13 Its cross-regulation with other
communication pathways guarantees the proper func￾tion of the adult body, including the skeletal sys￾tem.11,14,15 In terms of bone biology, the bidirectional
EphB4/ephrinB2 signaling is found to be particularly
important in the maintenance of bone homeostasis.
The reverse signaling through ephrinB2 into osteoclast
precursors suppresses osteoclast differentiation by inhib￾iting the osteoclastogenic cellular oncogene-Fos (c-Fos)-
nuclear factor of activated T cells c1 (NFATc1) cascade,
whereas forward signaling through EphB4 into osteo￾blasts enhances osteogenic differentiation, and overex￾pression of EphB4 in osteoblasts increases bone mass
in transgenic mice.16,17
Recent studies suggested that EphB4/ephrinB2 pathway
might also play a vital role in the remodeling of alveolar
bone.18-20 One study reported that transgenic expression
of ephrinB2 in PDL stem cells could promote osteogenic
differentiation via stimulation of the phosphorylation of
ephrinB2 and EphB4.19 During orthodontic tooth move￾ment, upregulation of ephrinB2-EphB4 signaling between
fibroblasts within the PDLs and osteoblasts of the alveolar
bone might contribute to osteogenesis at tension sites.18
Stimulated with ephrinB2, osteoblasts increased their
osteoblastogenic genes runt-related transcription factor 2
(Runx2) and alkaline phosphatase (ALP) expression
and showed functional signs of osteoblastic differentia￾tion.18 In compression areas, EphB4 and ephrinB2 expres￾sion, was significantly decreased, which contributed to
alveolar bone resorption through regulating transcription
factors like Runx2 and Sp7.20 Although much research
progress has been made on the regulatory role of EphB4/
ephrinB2 signaling in alveolar bone remodeling, studies
are lacking on its biological functions on other oral and
maxillofacial tissues.
Root resorption repair is a process in which dynamic
cementum restoration and periodontal tissue remodel￾ing take place. Considering the similar biochemical
composition and molecular biological properties be￾tween cementum and bone,21-23 it is reasonable to
hypothesize that the EphB4/ephrinB2 pathway, as a
crucial regulator, could also involve in cementum
formation during root regeneration. Applying in vivo
approaches, we aimed to elucidate the effect of
EphB4/ephrinB2 signaling on OIRR and the possible
molecular mechanism behind it.
MATERIAL AND METHODS
All experimental procedures were approved by the
Ethics Committee of West China Hospital of Stomatol￾ogy. The number of experimental animals used in this
study was determined on the basis of the sample size
calculation using the resource equation approach ac￾cording to a previous study.24 For group comparison us￾ing 1-way analysis of variance, 6 was an appropriate
number of animals per group, and the total sample
size was calculated as 72 for this study. Accordingly,
72 6-week-old male Wistar rats weighing 200 6 10 g
were obtained from the university’s experimental animal
center. They were kept in plastic cages with a standard
12-hour light-and-dark cycle and fed with a soft diet
and water ad libitum.
After 2 days of acclimatization, an orthodontic
appliance was randomly applied on 48 experimental
animals, and 100 g of heavy force was exerted to
create OIRR using an orthodontic elastic closed-coil
spring (Grikin Advanced Materials, Beijing, China)
fixed between the maxillary left first molar and the in￾cisors.25,26 Same orthodontic appliance was applied
on the rest 24 rats without activation to serve as a
blank control group (CON). After 2 weeks, coil springs
were removed, and orthodontic wires were fixed be￾tween the maxillary left first molars and the incisors
passively to maintain the results of tooth movement.
At this point, the molar roots of experimental rats
started to regenerate from OIRR. Six rats of each
group were killed immediately by an overdose of
pentobarbital to investigate the original state right af￾ter root resorption (day 0).
Experimental animals were randomly divided into 2
groups of 24 animals: physiological regeneration group
(PHY) and EphB4 inhibitor local injection group (INH).
The animals in the INH groups received daily periodontal
local injection of 20 mL EphB4 inhibitor NVP-BHG712
(A8683; APExBIO, Houston, Tex) on the attached
gingiva on both the buccal and palatal side of the maxil￾lary left first molar with a concentration of 4 mmol/L. The
CON group and PHY group received only the vehicle
(0.1% acetic acid) of the same volume. During the next
28 days of root repair period, animals of each group
received injections daily until they were killed randomly
after 7, 14, and 28 days (n 5 6), and alveolar bone blocks
that included the maxillary left first molar were har￾vested. The regions of interest consisted of the distal
e218 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
buccal root of the molar and the adjacent PDLs. All
experiment procedures and the intraoral picture of the
animal models were presented in Figure 1, A and B.
Then, all the samples were scanned using the high￾resolution microcomputed tomography 50 system
(Scanco Medical, Bruttisellen, Switzerland) with a voxel €
resolution of 10 mm, passing through a 3-dimensional
(3D) Gaussian filter (mean, 1.2; filter support, 1). Mimics
21.0 software was employed to reconstruct the 3D model
of the maxillary left first molar (Fig 1, C). To evaluate the
degree and extent of root resorption, we separated the
distal buccal root of each sample, and the total volume
of the resorption pits on the mesial surface of the distal
buccal root was calculated. The calculation followed the
method of the 3D Convex Hull Algorithm.27,28 Consid￾ering the surface of the teeth was roughly a smooth
convex curve initially, Mimics 21.0 was used to draw
the assumed surface line around the resorption lacunae
Fig 1. Experimental design of animal study and methods of microCT analysis of the size of resorption
lacunae. (A) Flow chart of the experimental design for allocation of animals and schedule for drug
administration. (B) Intraoral pictures of the orthodontic appliance and retention device. (C) Reconstruc￾tion of a 3D model of the maxillary left first molar and separation of the distal buccal root by Mimics 21.0
software. (D) Calculation of the volumes of resorption lacunae on the compression side of distobuccal
roots. Considering the surface of the teeth was roughly a smooth convex curve initially, the pink areas in
the cross-section view represented the resorption lacunae drawn by Mimics.
Li et al e219
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
by the same technician (Fig 1, D). By calculating the dif￾ference value of the root volume with and without the
assumed surface, the size of the resorption lacunae can
be determined. For each sample, the total volume of
resorption lacunae was divided by its root length and ex￾pressed as resorption lacunae per millimeter of root
length.
After microCT scanning, the left half of the maxilla of
each animal was fixed and decalcified for paraffin
embedding. Five-mm serial sections in a mesiodistal di￾rection parallel to the long axis of the distal root of
the first molar were cut on a microtome (HM 355S; Mi￾crom International, Walldorf, Germany) and mounted on
glass slides. Selected sections were treated with hema￾toxylin and eosin (G1120; Solarbio, Beijing, China) and
tartrate-resistant acid phosphatase (TRAP) (Sigma,
St Louis, Mo) staining and examined under a light mi￾croscope (Eclipse 80i microscope; Nikon, Toyko, Japan).
The number of TRAP-positive cells on the compression
side of the periodontal area was counted and expressed
as cell numbers per millimeter of root length.
For immunohistochemical staining, tissue sections
were placed in 3% hydrogen peroxide for 30 minutes in
the dark. Subsequently, sections were blocked in a block￾ing solution containing 4% bovine serum albumin for
20 minutes to prevent unspecific background staining.29
Then sections were incubated with primary antibodies
diluted in blocking solution with different dilution rates:
EphB4 (20883-1-ap, 1:200; Proteintech, Wuhan, China);
ephrinB2 (ET1705-33, 1:200; Huabio, Hangzhou, China);
cementum attachment protein (CAP dilution 1:50, SC-
53947; Santa Cruz, Shanghai, China); ALP (ET1601-21,
1:400; Huabio); osteopontin (OPN, 0806-6, 1:200; Hua￾bio); Runx2 (ET1612-47, 1:400; Huabio); osteoprotegerin
(OPG, R1608-4, 1:250; Huabio); RANKL (ab169966,
1:200; Abcam, Shanghai, China) and sclerostin (SOST,
21933-1-AP, 1:200; Proteintech), at 4C overnight and
then 37 C for 1 hour. After a rinse, slides were incubated
with goat antirabbit or goat antimouse IgG secondary
antibody conjugated to horseradish peroxidase (SP-
9000; Zhongshan Bio-Tech, Beijing, China) for 30 min
at 37C. The immunoreaction was visualized by using a
3.3’-diaminobenzidine kit (ZLI-9017; Zhongshan Bio￾Tech, Beijing, China) according to the manufacturers’ in￾structions. The regions of interest were defined as the
cementum in an apical third area of the compression
side of the distal buccal root and the layer of cells (presume
cementoblasts) lining on it.30-32 Within the regions of
interest, areas with the proper stained color, which
reflected positive immunoreactivities, were first selected.
Integrated optical density of selected areas was then
calculated by Image-Pro Plus (version 6.0; Media Cyber￾netics, Bethesda, Md). Finally, the means of integrated
optical density in each group were calculated to reflect
the average intensity of immunohistochemical staining.
Statistical analysis
Statistical analyses were performed using the SPSS
(version 22; IBM Corporation, Armonk, NY). Data are
presented as mean 6 standard deviation from at least
3 independent experiments. Statistical comparison was
performed using a 1-way analysis of variance. P \0.05
was considered statistically significant.
RESULTS
The process of root regeneration was first examined
by hematoxylin and eosin staining (Fig 2, A). Compared
with the CON group, periodontal fibers of the PHY and
INH groups lined irregularly, and large resorption
lacunae were seen on the root surface on day 0. Along
with the root repair process, force compressed PDLs lined
more regularly, and the root surface became smoother.
At the end of the experiment, the root surface of the
PHY and INH groups appeared smoother.
To compare the root regeneration process between
groups quantitatively, we used 3D models of all the
samples that were reconstructed using Mimics 21.0
software (Fig 2, B). In general, total volumes of resorp￾tion lacunae decreased over time, and the roughness
of the reconstructed root surface also reduced. On day
0, no significant difference was found between PHY
and INH groups. After that, faster cementum repair
was found in the PHY group, with significantly smaller
resorption volume than that of the INH group from
day 7 to 28 (Fig 3, A).
TRAP-positive cells resembling osteoclast were
mainly observed on the compression side of the peri￾odontal membrane (Fig 4). Few positive cells were found
in the CON group throughout the experimental period
under physiological conditions. On day 0, right after
heavy force-induced root resorption, a considerable
number of TRAP-positive cells appeared in both PHY
and INH groups, which decreased along with the root
regeneration process. On days 7 and 14, the positive
cell number was significantly smaller in the PHY group
than in the INH group. No significant difference was
found between groups on day 28 (Fig 3, B).
The expression of EphB4 and ephrinB2 was examined
with immunohistochemistry on the apical third of the
distal buccal root of the mandibular first molars (Fig 5,
A and B). Semiquantitative analysis revealed that the
expression of EphB4 in the PHY group was upregulated
during the regeneration process compared with the CON
group. EphB4 in the PHY group reached its peak after
7 days of repair and was significantly greater than the
e220 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Fig 2. Dynamic changes in root surface during the process of root resorption repair. (A) Hematoxylin
and eosin staining on the compression side of distobuccal roots. White arrows showed the areas of root
resorption on the root surface. Periodontal regeneration was inhibited by daily injection of EphB4 inhib￾itor (day 28). Scale bar: 250 mm. (B) A 3D model of the distobuccal root of maxillary left first molar re￾constructed by Mimics 21.0 software. White arrows showed the resorption lacunae on the root surface.
Obvious resorption lacunae were initially found on the compression side on both the PHY and INH
groups and gradually ameliorated after 14 days of repair. Scale bar: 0.5 mm. C, cementum; P,
periodontal ligament; B, alveolar bone.
Li et al e221
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
INH group from day 7 to 14. Because of daily injections
of EphB4 inhibitor, EphB4 in the INH group was in￾hibited and remained low throughout the entire experi￾mental process (Fig 3, C). EphrinB2 expression peaked
on day 14. The expression level in the PHY and INH
groups were significantly greater than the CON group
after 0, 7, and 14 days of repair (Fig 3, D).
As shown in Figure 6, strong CAP positive staining
was observed in the PHY group after 7 and 14 days of
repair, whereas weakly positive immunolabeling was
shown on both the CON and INH groups. Semiquan￾titative analysis revealed that the CAP expression in
the INH group was affected by the EphB4 inhibitor
and was significantly smaller than that of the PHY
group from day 7 to 14 (Fig 9, A). Similarly, strong
ALP positive staining was observed in the PHY group
after 7 and 14 days of repair (Fig 7, A). The ALP
expression in the INH group was also significantly
smaller than that of the PHY group from day 7 to
14 but was still significantly greater than that of the
CON group (Fig 9, B). Significantly intense OPN and
Runx2 staining was shown in the PHY group
compared with CON and INH groups from day 7 to
14, whereas no significant difference was found be￾tween CON and INH groups from day 7 to 28 (Figs
7, B, and 8, A, C and D). On day 0, the SOST expres￾sion in PHY and INH groups was significantly greater
than in the CON group. Then, SOST expression
decreased more rapidly in the PHY group. As a result,
the SOST expression level in the INH group was signif￾icantly greater than in the CON and PHY groups from
day 7 to 14 (Figs 8, B and 11, A).
Fig 3. Quantitative analysis of the volume of resorption lacunae and the number of TRAP-positive cells
between groups and semiquantitative analysis of EphB4 and ephrinB2 expression on the compression
side of the distal buccal root of the maxillary first molar. (A) The total volume of resorption lacunae
decreased over time in the PHY and INH groups. The volume of resorption lacunae in the INH group
was significantly greater than the PHY group from day 7 and 28. (B) TRAP-positive cell number
decreased over time and was significantly greater in the INH group than in the PHY group on day 7
and 14. (C) The means of the integrated option density (IOD) of EphB4. The expression of EphB4
reached its peak after 7 days of repair and was significantly greater in the PHY group after 7 and
14 days compared with CON and INH groups. (D) The means of IOD of ephrinB2. The expression level
in PTH and INH groups was significantly greater than the CON group after 7 and 14 days of PTH injec￾tion. n 5 5; *P \0.05. Bars indicate standard deviations.
e222 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Generally, the expression level of RANKL increased
from day 0 to 7 and then decreased afterward (Fig 10,
A). RANKL expression was significantly greater in the
INH group and PHY group than in the CON group
from day 0 to 7 and day 0 to 14, respectively (Fig 11,
B). On day 7, the expression of OPG in the PHY and
INH groups was significantly greater than in the CON
group. No significant difference was found between
groups at all other time points (Figs 10, B and 11, C).
The ratio of RANKL to OPG was calculated to evaluate
osteoclastic activities. At the beginning of root regener￾ation on day 0, the ratio of RANKL to OPG in the PHY
and INH groups was significantly greater than the CON
group. This ratio declined remarkably on day 7 in the
PHY group and was significantly smaller than the CON
and INH groups. From day 14 to 28, no significant dif￾ference was found between groups (Fig 11, D).
DISCUSSION
Methodologically, microCT methods were used in our
study for the diagnosis of root resorption. These fast,
high-resolution and 3D inspection methods ensured
reliability in the number of resorption lacunae calculated
in our study and provided a clear external appearance of
the impaired root.33,34 Even under physiological condi￾tions, the root surface was not completely smooth. As
a result, a constant small quantity of resorption volume
was calculated in the CON group throughout the process
and was used as a negative control to evaluate the root
regeneration degree in the other 2 groups. The root
repair model was successfully established because signif￾icantly greater resorption volume was detected in PHY
and INH groups on day 0. At the same time, no
significant difference was found between the resorption
volume of PHY and INH groups, indicating a
well-controlled intergroup variance before the adminis￾tration of EphB4 inhibitor.
Histologic observation indicated an obvious trend of
root restoration in both the PHY and INH groups. Hema￾toxylin and eosin staining showed that the once com￾pressed PDLs lined more regularly, and the root surface
became smoother along with the root repair process,
which was consistent with the results of reconstructed
3D appearance of the root surface. To be specific, quan￾titative analysis of the volume of resorption lacunae
Fig 4. TRAP staining images of the third apical periodontium of the compression side of the distal
buccal root. More TRAP-positive cells appeared in the INH group at the early stage of root repair
than in the PHY group (day 7 and 14). C, cementum; P, periodontal ligament; B, alveolar bone. Scale
bar: 50 mm.
Li et al e223
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
Fig 5. Representative immunohistochemical images of EphB4 and ephrinB2 on the compression side
of distobuccal roots. (A) The staining of EphB4 was more obvious in the PHY group than in the CON
and INH groups after 7 and 14 days of repair. Scale bar: 50 mm. (B) Strong ephrinB2 staining was de￾tected in the PHY and INH groups from day 0 to 14. C, cementum; P, periodontal ligament; B, alveolar
bone. Scale bar: 50 mm.
e224 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
revealed that the resorption volume of both groups
declined from day 0 to day 28. However, the resorption
lacunae were larger in the roots surface of the INH group
than in the PHY group from day 7 to day 28, indicating
an inhibited reparative process compared with normal
regeneration. We speculated that inhibition of EphB4
might delay the formation of new cementum.
To verify our assumption concerning the regulatory
effect of EphB4/ephrinB2 signaling, we employed
immunohistochemical staining in our study to deter￾mine the expression of EphB4 and ephrinB2 during
root regeneration. In our study, increased EphB4 expres￾sion was found in animal samples of the PHY group from
day 7 to day 14 and increased ephrinB2 expression of the
PHY and INH groups from day 0 to day 14, indicating
active signaling activities during physiological root
regeneration. We also noticed that the EphB4 expression
was suppressed in the INH group, which confirmed the
inhibited function of EphB4/ephrinB2 signaling. The
intervention of EphB4/ephrinB2 signaling could influ￾ence the formation of new cementum after heavy
force-induced damage. In summary, we proved that
the EphB4/ephrinB2 signaling played an important
role in the normal root regeneration process. Consid￾ering the similarities between bone and cementum, our
results were consistent with previous studies that the
EphB4/ephrinB2 signaling was essential for bone ho￾meostasis.35,36 Inactivation of the EphB4/ephrinB2
pathway inhibited bone formation and bone resorption
in mice under pathologic conditions.37
Currently, mechanisms concerning EphB4/ephrinB2
signaling regulating root regeneration have not been
fully elucidated. In this study, both representative ce￾mentogenic- and osteoclastic-related factors were
tested, hoping to bring a better understanding of
EphB4/ephrinB2 signaling on cementum regeneration.
As a cementoblast-specific product, CAP has recently
been isolated and consequently regarded as a reliable
specific marker of cementoblasts in several related
studies.38,39 In general, factors like ALP, OPN, and
Runx2 are considered important osteogenic markers
that contribute to osteogenesis. However, some in vivo
studies recently reported that these osteoclastic-related
factors could also be secreted by cementoblasts,40-42
and osteogenic markers OPN and osteocalcin have
been used in one recent animal study to reflect
cementoblast activities during tooth development.43
Therefore, we believed that these factors could also serve
as proper indicators to reflect the functional activities of
cementoblasts.
Fig 6. Representative immunohistochemical images of CAP on the compression side of distobuccal
roots. The CAP staining was obvious after 7 and 14 days of root repair. More positive staining was de￾tected in the PHY group than the CON and INH groups. Representative positive staining cells were indi￾cated with white arrows. Scale bar: 50 mm.
Li et al e225
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
Fig 7. Representative immunohistochemical images of ALP and OPN on the compression side of dis￾tobuccal roots. (A) The ALP staining was obvious after 7 and 14 days of root repair, and more positive
staining was detected in the PHY group. Scale bar: 50 mm. (B) Strong positive OPN staining was de￾tected throughout the repair process. The PHY group has deeper positive staining (day 0 to day 14).
C, cementum; P, periodontal ligament; B, alveolar bone. Scale bar: 50 mm.
e226 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Fig 8. Representative immunohistochemical images of Runx2 and SOST on the compression side of
distobuccal roots. (A) The Runx2 staining was more obvious in the PHY group than in the CON and INH
groups from day 7 to day 28. Scale bar: 50 mm. (B) Strong positive SOST staining was detected on day
0 and more obvious in the INH group than in the PHY group from day 7 to 14. C, cementum; P,
periodontal ligament; B, alveolar bone. Scale bar: 50 mm.
Li et al e227
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
In our study, CAP was adopted as the key indicator
for cementogenesis and cementoblast regeneration.
Meanwhile, the expression of ALP, OPN, and Runx2
was also evaluated to help better understand the
anabolic effects of EphB4/ephrinB2 signaling on ce￾mentoblasts. Strong positive staining was observed in
the PHY group after 7 and 14 days of repair, representing
an active functional status of cementoblasts during root
regeneration. Because of inhibition of EphB4, the CAP
expression in the INH group failed to increase and was
significantly smaller than the PHY group. This staining
result of CAP showed that the normal functions of ce￾mentoblasts were affected in the INH group. Because
the expression of ALP, OPN, and Runx2 was consistent
with the expression of CAP and also decreased because
of inhibited EphB4/ephrinB2 signaling, we speculated
that it was very likely that the EphB4/ephrinB2 signaling
promoted cementum repair through regulating these
factors.
Secreted by osteocytes, SOST acts as an osteogenesis
inhibitor, which mainly exerts its biological effects
through inhibition of Wnt signaling.44 Similarly, recent
studies found that SOST could also affect cementogen￾esis by inhibiting cementoblast proliferation and differ￾entiation and promoting osteoclastogenesis.45-47 In our
study, we found that SOST expression in both PHY and
INH groups was significantly greater than
the CON group on day 0, indicating inhibited
cementogenesis right after the exertion of heavy force.
Then, SOST expression decreased in the PHY group,
whereas it remained strong in the INH group from day
7 to day 14. This result was in agreement with the CAP
staining. We speculated that the EphB4/ephrinB2
signaling was essential for regulating the functional
Fig 9. Semiquantitative analysis of CAP, ALP, OPN, and Runx2 expression on the compression side
of the distobuccal root of the maxillary first molar. (A) The means of integrated option density (IOD) of
CAP. The expression of CAP was significantly greater in the PHY group than CON and INH groups
after 7 and 14 days of repair. (B) The means of IOD of ALP. The expression of ALP was significantly
greater in the PHY group than in the INH group and significantly greater in the INH group than in the
CON group after 7 and 14 days of repair. (C) The means of IOD of OPN. The expression of OPN in
the PHY group was significantly greater than in the CON and INH groups on days 7 and 14. (D) The
means of IOD of Runx2. The expression of Runx2 reached its peak after 7 days of repair and was signif￾icantly greater in the PHY group than in CON and INH groups after 7 and 14 days.
e228 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Fig 10. Representative immunohistochemical images of RANKL and OPG on the compression side of
distobuccal roots. (A) The RANKL staining was more obvious in PHY and INH groups in the early
period of root repair. Scale bar: 50 mm. (B) Strong positive OPG staining was detected on day 7 and
more obvious in the PHY and INH groups than in the CON group. C, cementum; P, periodontal liga￾ment; B, alveolar bone. Scale bar: 50 mm.
Li et al e229
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
status of cementoblast. During the early period of root
regeneration, it could accelerate the downregulation of
SOST and recruit more cells into a functional state.
Afterward, the EphB4/ephrinB2 signaling was able to
maintain the secretion level of various cementogenic￾related factors and promote root regeneration. If the
signaling were inhibited, the normal cementogenesis
process would be affected, as shown in the histologic
staining.
An obvious downward trend was found in the num￾ber of TRAP-positive cells, representing osteoclasts, in
PHY and INH groups. As a result, few TRAP-positive cells
were found on day 28. No significant difference was
found between groups at this point. It might be due to
catabolic activities were almost complete at this point,
which covered up the between-group variance. Affected
by the inhibition of EphB4/ephrinB2 signaling, the oste￾oclast number was significantly greater in the INH group
than in the PHY group on day 7 and day 14. To deter￾mine the mechanisms of EphB4/ephrinB2 signaling
on the biological activities of osteoclasts, we determined
the expression levels of RANKL and OPG and calculated
the ratio of RANKL to OPG for analysis. It is known
that the OPG/RANK/RANKL system is a critical determi￾nant for bone homeostasis.48,49 The balance in the
counteraction between RANKL and OPG regulates the
Fig 11. Semiquantitative analysis of SOST, RANKL, and OPG expression on the compression side of
the distobuccal root of the maxillary first molar and calculation of the ratio of RANKL to OPG. (A) The
means of integrated option density (IOD) of SOST. The expression level in PHY and INH groups was
significantly greater than the CON group on day 0 and was significantly greater in the INH group after 7
and 14 days compared with CON and PHY groups. n 5 5; *P\0.05. Bars indicate standard deviations.
(B) The means of IOD of RANKL. The expression of RANKL was significantly greater in PHY and INH
groups compared with the CON group from day 0 to day 7 and significantly greater in the PHY group
compared with the CON group after 14 days of repair. (C) The means of IOD of OPG. The OPG expres￾sion of PHY and INH groups was significantly greater than the CON group on day 7. (D) Calculation of
the RANKL to OPG ratio. The ratio of RANKL to OPG was significantly greater in the PHY and INH
groups than in the CON group on day 0. Then, the RANKL to OPG ratio of the PHY group was signif￾icantly smaller than the CON and INH groups on day 7. n 5 5; *P \0.05. Bars indicate standard
deviations.
e230 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
development and activation of osteoclasts and bone
metabolism.50 When the ratio of RANKL to OPG de￾creases, osteoclast activity is inhibited; when the ratio
increases, osteoclast activity is promoted.51 In our study,
the ratio of RANKL to OPG was significantly greater in
the PHY and INH groups on day 0, representing active
osteoclast activities during root resorption. We found
that the ratio of RANKL to OPG in the INH group was
significantly greater than the PHY group on day 7, which
could partly explain the TRAP staining result. Previous
studies have demonstrated that the effects of EphB4/
ephrinB2 signaling on osteoclast activities were impor￾tant for maintaining bone homeostasis.37,52 One study
reported that the NFATc1 target gene Efnb2 (encoding
ephrinB2) was expressed by osteoclasts, and the reverse
signaling through ephrinB2 into osteoclast precursors
may suppress osteoclast differentiation by inhibiting
the osteoclastogenic c-Fos-NFATc1 cascade, which is
required for osteoclastogenesis and regulates many
genes important for osteoclast differentiation and func￾tion.16 Our results were in accordance with these find￾ings, and we supplemented that the EphB4/ephrinB2
signaling might be able to regulate osteoclastic activities
through the OPG/RANK/RANKL system.
Taken together, we demonstrated that inhibition of
EphB4 hindered the regeneration of root resorption after
heavy orthodontic force. The EphB4/ephrinB2 pathway
is essential for the normal regeneration of OIRR. It reg￾ulates the functional status of cementoblast through
suppression of the cementogenesis inhibitor SOST and
the promotion of cementogenic-related factors CAP,
ALP, OPN, and Runx2. Simultaneously, the EphB4/eph￾rinB2 signaling had an inhibitory effect on the osteoclast
in the early phase of root resorption repair through the
regulation of the OPG/RANK/RANKL system.
One limitation of the current study is that the exact
resorption volumes on the root surface were not exam￾ined directly during the root repair process. Though
negative controls (untreated control group) are set,
and a reasonable measurement method is adopted, small
variations among different groups might impair the reli￾ability of the final results. Taking microCT dynamically
for the teeth of the same group of animals and superim￾posing the corresponding 3D images to determine the
volume of resorption might be the best way. However,
the accuracy of our in vivo microCT equipment failed
to show the surface of the root clearly if the whole rat
was scanned; thus, we were unable to adopt this method
in our study. Using teeth on the contralateral side of the
same rats as negative controls might be another solu￾tion, provided that the highly consistent root shape of
both sides of the teeth are confirmed by preliminary
studies. With the improvement of measurement
techniques, future studies will shed more light on the
process of cementum regeneration during root repair.
CONCLUSIONS
Our study demonstrated that the EphB4/ephrinB2
pathway might be essential for the regeneration of
OIRR. This signaling might promote cellular functions
of cementoblasts through the regulation of various
cementogenesis-related factors, as well as inhibit osteo￾clast activities through the OPG/RANK/RANKL system.
EphB4/ephrinB2 signaling might be a promising thera￾peutic target for novel therapeutic approaches to accel￾erate OIRR regeneration.
REFERENCES
1. Brezniak N, Wasserstein A. Orthodontically induced inflammatory
root resorption. Part I: The basic science aspects. Angle Orthod
2002;72:175-9.
2. Krishnan V. Critical issues concerning root resorption: a contem￾porary review. World J Orthod 2005;6:30-40.
3. Bosshardt DD. Are cementoblasts a subpopulation of osteoblasts
or a unique phenotype? J Dent Res 2005;84:390-406.
4. Grzesik WJ, Cheng H, Oh JS, Kuznetsov SA, Mankani MH,
Uzawa K, et al. Cementum-forming cells are phenotypically
distinct from bone-forming cells. J Bone Miner Res 2000;15:
52-9.
5. Yamamoto T, Hasegawa T, Yamamoto T, Hongo H, Amizuka N.
Histology of human cementum: its structure, function, and devel￾opment. Jpn Dent Sci Rev 2016;52:63-74.
6. Nowrin SA, Jaafar S, Ab Rahman N, Basri R, Alam MK, Shahid F.
Association between genetic polymorphisms and external apical
root resorption: a systematic review and meta-analysis. Korean J
Orthod 2018;48:395-404.
7. Weltman B, Vig KW, Fields HW, Shanker S, Kaizar EE. Root resorption
associated with orthodontic tooth movement: a systematic review.
Am J Orthod Dentofacial Orthop 2010;137:462-76: discussion 12A.
8. Grzesik WJ, Narayanan AS. Cementum and periodontal wound
healing and regeneration. Crit Rev Oral Biol Med 2002;13:474-84.
9. Saygin NE, Giannobile WV, Somerman MJ. Molecular and cell
biology of cementum. Periodontol 2000 2000;24:73-98.
10. Brudvik P, Rygh P. Transition and determinants of orthodontic
root resorption-repair sequence. Eur J Orthod 1995;17:177-88.
11. Pasquale EB. Eph-ephrin bidirectional signaling in physiology and
disease. Cell 2008;133:38-52.
12. Pitulescu ME, Adams RH. Eph/ephrin molecules—a hub for
signaling and endocytosis. Genes Dev 2010;24:2480-92.
13. Taylor H, Campbell J, Nobes CD. Ephs and ephrins. Curr Biol 2017;
27:R90-5.
14. Genander M, Frisen J. Ephrins and Eph receptors in stem cells and
cancer. Curr Opin Cell Biol 2010;22:611-6.
15. Matsuo K, Otaki N. Bone cell interactions through Eph/ephrin:
bone modeling, remodeling and associated diseases. Cell Adh
Migr 2012;6:148-56.
16. Zhao C, Irie N, Takada Y, Shimoda K, Miyamoto T, Nishiwaki T,
et al. Bidirectional ephrinB2-EphB4 signaling controls bone ho￾meostasis. Cell Metab 2006;4:111-21.
17. Li C, Shi C, Kim J, Chen Y, Ni S, Jiang L, et al. Erythropoietin pro￾motes bone formation through EphrinB2/EphB4 signaling. J Dent
Res 2015;94:455-63.
Li et al e231
American Journal of Orthodontics and Dentofacial Orthopedics March 2021
Vol 159
Issue 3
18. Diercke K, Kohl A, Lux CJ, Erber R. Strain-dependent up￾regulation of ephrin-B2 protein in periodontal ligament fibro￾blasts contributes to osteogenesis during tooth movement. J
Biol Chem 2011;286:37651-64.
19. Zhu SY, Wang PL, Liao CS, Yang YQ, Yuan CY, Wang S, et al. Trans￾genic expression of ephrinB2 in periodontal ligament stem cells
(PDLSCs) modulates osteogenic differentiation via signaling cross￾talk between ephrinB2 and EphB4 in PDLSCs and between PDLSCs
and pre-osteoblasts within co-culture. J Periodontal Res 2017;52:
562-73.
20. Hou J, Chen Y, Meng X, Shi C, Li C, Chen Y, et al. Compressive force
regulates ephrinB2 and EphB4 in osteoblasts and osteoclasts
contributing to alveolar bone resorption during experimental
tooth movement. Korean J Orthod 2014;44:320-9.
21. Zhao N, Nociti FH Jr, Duan P, Prideaux M, Zhao H, Foster BL, et al.
Isolation and functional analysis of an immortalized murine ce￾mentocyte cell line, IDG-CM6. J Bone Miner Res 2016;31:430-42.
22. Zhao N, Foster BL, Bonewald LF. The Cementocyte-an osteocyte
relative? J Dent Res 2016;95:734-41.
23. Duan P, Bonewald LF. The role of the wnt/beta-catenin signaling
pathway in formation and maintenance of bone and teeth. Int J
Biochem Cell Biol 2016;77:23-9.
24. Arifin WN, Zahiruddin WM. Sample size calculation in animal
studies using resource equation approach. Malays J Med Sci
2017;24:101-5.
25. Amuk NG, Kurt G, Baran Y, Seyrantepe V, Yandim MK, Adan A,
et al. Effects of cell-mediated osteoprotegerin gene transfer and
mesenchymal stem cell applications on orthodontically induced
root resorption of rat teeth. Eur J Orthod 2017;39:235-42.
26. Wan Hassan WN, Stephenson PA, Waddington RJ, Sloan AJ. An
ex vivo culture model for orthodontically induced root resorption.
J Dent 2012;40:406-15.
27. He D, Kou X, Luo Q, Yang R, Liu D, Wang X, et al. Enhanced M1/M2
macrophage ratio promotes orthodontic root resorption. J Dent
Res 2015;94:129-39.
28. Zhao N, Liu Y, Kanzaki H, Liang W, Ni J, Lin J. Effects of local os￾teoprotegerin gene transfection on orthodontic root resorption
during retention: an in vivo micro-CT analysis. Orthod Craniofac
Res 2012;15:10-20.
29. Hu Z, Ma C, Rong X, Zou S, Liu X. Immunomodulatory ECM-like
microspheres for accelerated bone regeneration in diabetes melli￾tus. ACS Appl Mater Interfaces 2018;10:2377-90.
30. Bosshardt D, Schroeder HE. Evidence for rapid multipolar and
slow unipolar production of human cellular and acellular
cementum matrix with intrinsic fibers. J Clin Periodontol 1990;
17:663-8.
31. Furseth R. The fine structure of the cellular cementum of young
human teeth. Arch Oral Biol 1969;14:1147-58.
32. Bosshardt DD, Schroeder HE. Initial formation of cellular intrinsic
fiber cementum in developing human teeth. A light- and electron￾microscopic study. Cell Tissue Res 1992;267:321-35.
33. Chan E, Dalci O, Petocz P, Papadopoulou AK, Darendeliler MA.
Physical properties of root cementum: Part 26. Effects of micro￾osteoperforations on orthodontic root resorption: A microcom￾puted tomography study. Am J Orthod Dentofacial Orthop 2018;
153:204-13.
34. Huang TTY, Elekdag-Turk S, Dalci O, Almuzian M, Karadeniz EI,
Gonzales C, et al. The extent of root resorption and tooth
movement following the application of ascending and descending
magnetic forces: a prospective split mouth, microcomputed￾tomography study. Eur J Orthod 2017;39:547-53.
35. Zhang F, Zhang Z, Sun D, Dong S, Xu J, Dai F. EphB4 promotes os￾teogenesis of CTLA4-modified bone marrow-derived mesenchymal
stem cells Through cross talk with Wnt pathway in xenotransplan￾tation. Tissue Eng A 2015;21:2404-16.
36. Zhang F, Zhang Z, Sun D, Dong S, Xu J, Dai F. Periostin: a down￾stream mediator of EphB4-induced osteogenic differentiation of
human bone marrow-derived mesenchymal stem cells. Stem Cells
Int 2016;2016:7241829.
37. Wu M, Ai W, Chen L, Zhao S, Liu E. Bradykinin receptors and
EphB2/EphrinB2 pathway in response to high glucose-induced
osteoblast dysfunction and hyperglycemia-induced bone deterio￾ration in mice. Int J Mol Med 2016;37:565-74.
38. Alvarez-Perez MA, Narayanan S, Zeichner-David M, Rodrıguez
Carmona B, Arzate H. Molecular cloning, expression and immuno￾localization of a novel human cementum-derived protein (CP-23).
Bone 2006;38:409-19.
39. Nunez J, Sanz M, Hoz-Rodr ~ ıguez L, Zeichner-David M, Arzate H.
Human cementoblasts express enamel-associated molecules
in vitro and in vivo. J Periodont al Res 2010;45:809-14.
40. Lyu C, Zhang C, Li T, Huang L, Yin X, Zou S. CD81 T lymphocytes
enhance the anabolic effect of intermittent parathyroid hormone
on cementoblasts. Int Immunopharmacol 2019;77:105927.
41. Li Y, Hu Z, Zhou C, Xu Y, Huang L, Wang X, et al. Intermittent para￾thyroid hormone (PTH) promotes cementogenesis and alleviates NVP-BHG712
the catabolic effects of mechanical strain in cementoblasts. BMC
Cell Biol 2017;18:19.
42. Diercke K, Konig A, Kohl A, Lux CJ, Erber R. Human primary ce- €
mentoblasts respond to combined IL-1beta stimulation and
compression with an impaired BSP and CEMP-1 expression. Eur
J Cell Biol 2012;91:402-12.
43. Xu Y, Lv C, Zhang J, Li Y, Li T, Zhang C, et al. Intermittent parathy￾roid hormone promotes cementogenesis in a PKA- and ERK1/2-
dependent manner. J Periodontol 2019;90:1002-13.
44. Moester MJ, Papapoulos SE, L€owik CW, van Bezooijen RL. Sclero￾stin: current knowledge and future perspectives. Calcif Tissue Int
2010;87:99-107.
45. Bao X, Liu X, Zhang Y, Cui Y, Yao J, Hu M. Strontium promotes ce￾mentoblasts differentiation through inhibiting sclerostin expres￾sion in vitro. BioMed Res Int 2014;2014:487535.
46. Bao X, Liu Y, Han G, Zuo Z, Hu M. The effect on proliferation and
differentiation of cementoblast by using sclerostin as inhibitor. Int
J Mol Sci 2013;14:21140-52.
47. Lehnen SD, G€otz W, Baxmann M, J€ager A. Immunohistochemical
evidence for sclerostin during cementogenesis in mice. Ann Anat
2012;194:415-21.
48. Wittrant Y, Theoleyre S, Couillaud S, Dunstan C, Heymann D,
Redini F. Relevance of an in vitro osteoclastogenesis system to
study receptor activator of NF-kB ligand and osteoprotegerin bio￾logical activities. Exp Cell Res 2004;293:292-301.
49. de Moraes M, de Lucena HF, de Azevedo PR, Queiroz LM, Costa
Ade L. Comparative immunohistochemical expression of RANK,
RANKL and OPG in radicular and dentigerous cysts. Arch Oral
Biol 2011;56:1256-63.
50. Nonaka CF, Cavalcante RB, Nogueira RL, de Souza LB, Pinto LP.
Immunohistochemical analysis of bone resorption regulators
(RANKL and OPG), angiogenic index, and myofibroblasts in syn￾drome and non-syndrome odontogenic keratocysts. Arch Oral
Biol 2012;57:230-7.
51. Han G, Chen Y, Hou J, Liu C, Chen C, Zhuang J, et al. Effects of
simvastatin on relapse and remodeling of periodontal tissues after
tooth movement in rats. Am J Orthod Dentofacial Orthop 2010;
138:550.e1-7: discussion 550-1.
52. Baek JM, Cheon YH, Kwak SC, Jun HY, Yoon KH, Lee MS, et al.
Claudin 11 regulates bone homeostasis via bidirectional EphB4-
EphrinB2 signaling. Exp Mol Med 2018;50:50.
e232 Li et al
March 2021
Vol 159
Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics