Keywords
animals - alveolar bone loss - biomarkers - cadherins - epithelial–mesenchymal transition
- periodontitis - vimentin
Introduction
Periodontal disease is the sequalae of exaggerated immune response to the presence
of dysbiotic microbiota on the tooth surface and within the gingival sulcus.[1] The epithelium provides the first anatomical and immunological barrier blocking
the ingress of periodontal pathogens and their virulence factors to deeper tissues
and systemic circulation.[2] Keratinocytes are interconnected by cellular junctions, which consist of special
structural proteins.[3] Dissolution of those junctional proteins compromise the epithelial barrier, providing
a direct route for bacterial ingress into the periodontal connective tissue resulting
in tissue breakdown.[4] Epithelial–mesenchymal transition (EMT) is a process leading to phenotype shifting
from an epithelial to a mesenchymal phenotype. It is defined by key sequential events
including the dissolution of the epithelial junctions, loss of apical–basal polarity
with acquisition of front–rear polarity, reorganization of the cytoskeleton, downregulation
of epithelial genes and proteins, and upregulation of their counterparts, which define
the mesenchymal phenotype. Consequently, cells exhibit increased motility, and acquire
invasive characteristics that induce extracellular matrix degradation.[5] EMT has been extensively studied during embryonic development, malignancy, and tissue
fibrosis. Evidence from several in vitro studies now confirms that EMT is induced by periodontal pathogens, particularly gram-negative
species, such as Porphyromonas gingivalis, Fusobacterium nucleatum, Treponema denticola, and Aggregatibacter actinomycetemcomitans.[6]
[7]
E-cadherin-dependent adherens junctions are a prerequisite for the functional assembly
of other intercellular junctions including tight junctions and desmosomes.[8] Epithelial cells undergoing EMT are typically characterized by a shift in their
cadherins from E-cadherin to N-cadherin, and partial or complete replacement of keratin
with vimentin.[5] The EMT regulatory consortium comprises four different elements: transcriptional
factors (TFs), small noncoding RNAs, differential splicing, as well as translation
and posttranslational regulators. EMT key TFs include Snail and Twist, which are best
characterized within the context of tissue development, fibrosis, and cancer. Reported
results from in vitro and in vivo studies have documented their role in activating EMT programs in epithelial cells
in both developmental and pathological conditions.[5]
[9] Snail and Twist exhibit unique expression profiles induced by a variety of microbial,
inflammatory, and ecological stressors.[10]
[11] In EMT, Snail and Twist collaboratively or separately suppress E-cadherin expression
and simultaneously induce N-cadherin expression.[12] Additionally, the motility and invasive potential of the transitioned cells are
enhanced by increased vimentin expression induced by these TFs.[13] Furthermore, as part of a regulatory loop, Snail and Twist enhance the expression
of each other, and act synergistically on their targeted genes, including several
other EMT-TFs.[12]
[14]
To date, no studies have explored the expression of EMT key biomarkers in experimental
periodontitis (EP) models. Consequently, this study was conducted to investigate the
EMT phenotype expression within this context.
Materials and Methods
Animals and Ethical Approval
The experimental protocol for surgical procedures and animal treatment was approved
by the Ethics Committee, College of Dentistry, University of Baghdad (Project no.
65122 on September 13, 2022). All animals received human care according to the Animal
Research: Reporting of In Vivo Experiments (ARRIVE) guidelines.[15] A pilot study was conducted to identify the TLs analogous to different stages of
periodontitis. EP was induced in 24 animals for 21 days, 3 animals were scarified
at days 1, 3, 5, 7, 9, 14, and 21 ([Fig. 1]). According to the results of the pilot study, days 3, 7, 14, and 21 were selected
in this study representing the periodontitis terminal phases of stages 1, 2, 3, and
4, respectively.
Fig. 1 Design of ligature-induced periodontitis pilot study. (Top) For the pilot study,
experimental animals were allocated into two groups: ligation (L) and unligated control
(3 animals) . The animals in the ligation group (21 animal) were subdivided into 7
groups corresponding to experimental periodontitis' induction timelines 1, 3, 5, 7,
9, 14, and 21 days' timelines. (A) Representative clinical image of the ligated molar tooth, After sacrificing the
rats, the maxilla was resected to prepare bone and soft tissue for analysis. (B) The analyses included measurement of alveolar bone loss by morphometric analysis
from the cementoenamel junction to the alveolar bone crest (black lines), (C) measuring the thickness of the epithelium, and (D) counting the number of inflammatory cells. EP, experimental periodontitis.
Sample size was calculated using resource equation approach[16]:
n = DF/K + 1,
where DF stands for the degree of freedom, K for the number of groups (5 experimental timelines including control), and n for the number of animals per group. Since the acceptable range of DF is 10 as the minimum and 20 as the maximum, accordingly the minimum and maximum n were calculated as follows:
Minimum n = 10/K + 1.
Maximum n = 20/K + 1.
Accordingly, the maximum number required was 5 animals/timeline. Taking into account
10% as attrition rate, the total number was 6 animals/timeline. For this study, a
total of 30 Wistar albino male rats aged 2 months, disease free with an average body
weight of 250 to 300 g, were purchased from Al Dhya'a Advanced Pet Hospital, Baghdad,
Iraq. The rats were kept in the animal facility under standardized conditions: humidity
at 60%, temperature controlled at 22 ± 2°C, and a 12-hour light/dark cycle. The rats
acclimated for 1 week prior to the first procedure within plastic cages identified
by their number, group, and date. During the entire experimental period, animal monitoring
was undertaken daily.
Ligature-Induced Experimental Periodontitis
Simple randomization was used to assign rats into nonligated controls and EP timeline
groups. The animals were anesthetized using intramuscular (IM) injection, mixture
of 10% ketamine hydrochloride (80 mg/kg) and 2% xylazine hydrochloride (10 mg/kg),
with a dose of 0.05 ml/kg of body weight intraperitoneally. Full anesthesia set in
within 3 to 5 minutes and was confirmed by the absence of response when the rat toe
was pinched. To induce bone loss, a 5–0 silk ligature was tied around the right maxillary
second molar, as described by Pereira SSC et al. ([Fig. 1A]).[17] Ligatures were checked daily throughout the experimental period. All rats were euthanized
humanely at the end of treatment under general anesthesia overdose. To ensure the
animals were dead, a secondary method, cervical dislocation, was used. A pilot study
was initially conducted to confirm the induction of EP and to characterize TLs reflecting
the severity of bone loss corresponding to clinical stages of periodontitis in humans.
All experiments were repeated in triplicate both technically and biologically ([Fig. 1]).
Sample Preparation
The maxillae were resected and bisected from both sides, posterior to the incisor
teeth. The gingival tissue around ligated teeth was sampled from the resected maxillae
using a no. 11 scalpel blade by an initial horizontal shallow sulcular incision around
the maxillary molars, avoiding deep cutting as this point may cause the thin maxilla
to sever. The second horizontal incision was made parallel to the first near the palatal
midline. Both incisions were joined by two vertical incisions. The entire gingival
band was carefully removed by blunt dissection. The collected samples were cleaned
by tap water and immediately stored in plastic containers containing 10% neutral buffered
formalin solution for tissue fixation for 24 hours prior to further processing and
histological analysis ([Fig. 1]).
Bone Loss Morphometric Analysis
The hemimaxillae were fixed in formalin for 24 hours prior to mechanical defleshing.
To facilitate removal of the excess tissue, the hemimaxillae were immersed in 5% sodium
hypochlorite for 3 hours. Specimens were cleaned under running water and air dried
at room temperature. To enhance cementoenamel junction (CEJ) visualization, the specimens
were stained for 1 minute with 1% methylene blue.[18] For standardization, the specimens were fitted in a mold made of heavy body impression
putty within a hard paper frame with their palatal side facing upward and the occlusal
plane parallel to the surface below. The long axis of each specimen was aligned perpendicular
to the camera, and a millimeter-scale ruler fitted in the mold used as a reference,
prior to digital photographic documentation. Standardized digital photography was
used to capture images of the specimens stained with methylene blue using macro lens
(105 mm, Sigma, Japan), with a 1:1 ratio, mounted on a digital camera (Nikon D7200,
24.2 megapixels, Japan). A calibrated examiner performed the morphometric analysis
using ImageJ Software (NIH, United States). The distance from the CEJ to the alveolar
bone crest (ABC) on the palatal face of the maxillary second molar was measured. Three
points were measured for every specimen: one in the center of the mesial and distal
root and one in the furcation area, and the mean of these reading was then computed
([Fig. 1B]).
Epithelial Thickness
The gingival samples were longitudinally trimmed (mesiodistal direction) and routinely
prepared to obtain hematoxylin and eosin (H&E) tissue sections. Microscopic images
for all H&E slides of EP TLs were captured under ×10 magnification (Leica Qwin 500,
Wetzlar, Germany) using a digital video camera and an image analysis system (Motic,
ToupTek, ToupView, ×86, 3.7.4183, and 2014; [Fig. 1C]). The epithelial thickness of each EP TLs was determined by averaging the epithelial
thickness. The change in epithelial thickness was calculated as the difference between
the nonligated control and EP TLs.
Inflammatory Cell Count
For all H&E slides, quantitative assessment for inflammatory infiltrate in the gingival
connective tissue at the palatal side was performed under ×40 magnification by calibrated
examiners. Images were evaluated for inflammation intensity using multipoint tool
of ImageJ software. Five microscopic areas were randomly selected, and each field
was divided into 20 squares by an ImageJ grid tool ([Fig. 1D]). The mean count of inflammatory cells was determined and classified as follows:
score 0 (0–25 cells), score 1 (26–50 cells), score 2 (51–75 cells), and score 3 (>75
cells).[19]
Immunohistochemical Staining
Fixed samples were dehydrated in ascending grades of alcohol, cleared in xylol, embedded
in paraffin blocks. From each paraffin-embedded tissue block, serial sections, 4-μm
thickness, were dissected in the mesiodistal direction and mounted on positively charged
slides (Fisher Scientific and Escho superfrost plus, United States) for immunohistochemical
staining (IHS). The following antibodies were used: monoclonal anti-E-cadherin (1:200,
E-AB-70249), polyclonal anti-N-cadherin (1:300, E-AB-70061), polyclonal antivimentin
(1:200, E-AB-70081), polyclonal anti-Snail (1:100, E-AB-65888), and polyclonal anti-Twist
(1:200, E-AB-66494). The antigen–antibody binding signal was amplified by 2-step plus
Poly-HRP Anti Rabbit/Mouse IgG Detection System with DAB solution (E-IR-R217). All
products were purchased from Elabscience Biotechnology Inc., United States. Positive
controls were obtained from the tissues recommended by the manufacturer, while negative
controls for all biomarkers were prepared by adding phosphate buffered saline (PBS)
instead of the primary antibodies ([Supplementary Fig. 1], available in the online version only).
Immunohistochemical Scoring
Digital images of IHC-stained slides were evaluated by the main investigator, and
blindly calibrated with an oral pathologist until an agreement ≥90% for the histoscore
was achieved. A semiquantitative approach was used to grade the immunoreactivity for
membranous staining of E-cadherin, N-cadherin, Snail, and Twist. These were measured
using the following scale: –, negative expression; +, weak expression; ++, moderate
expression; and +++, strong expression. The percentage of cells stained at each intensity
level was graded as 0 (<5%), 1 (5 ± 25%), 2 (26 ± 50%), 3 (51 ± 75%), and 4 (>75%).
Both were multiplied to calculate the final histoscores as previously described.[20]
[21] For vimentin, the presence/absence of cytoplasmic expression was recorded as a dichotomous
score (1, 0).[22]
[23]
Statistical Analysis
GraphPad Prism Software (Version9.0; La Jolla, CA, United States) was used to analyze
the raw data. Both parametric and nonparametric distribution of the datasets was verified
using the Shapiro–Wilk test. Accordingly, inferential comparisons were performed by
the Kruskal–Wallis test. To determine the correlation between the dependent and independent
variables, Spearman's correlation test was used. The level of significant difference
was determined when p < 0.05.
Results
Morphometric Analysis of Alveolar Bone Loss
The alveolar bone loss (ABL) observed in all representative TLs was significantly
different from controls except for day 3 ([Fig. 2]). The significant difference of ABL was evident between successive TLs till day
14, which was not significantly different from day 21 ([Fig. 3], left panel).
Fig. 2 Left panel: Morphometric analysis showed that bone loss was 15% at day 3, 33% at
day 7, greater than 33% at day 14, and greater than 33% at day 21, associated with
tooth loss (arrow). Middle panel: Representative photomicrographs of the gingival samples. Control
and day 3 showing mild elongation and normal configuration of the epithelial ridges.
In contrast, days 7, 14, and 21 showed marked rete peg elongation and epithelial hyperplasia.
Note the tortuous appearance of the epithelial ridges. Hematoxylin and eosin (H&E)
stain, ×10 magnification. Right panel: Photomicrographs of the gingival samples showing
the gradual increase the number of inflammatory cell counts. Note the marked reduction
in inflammatory cell counts observed on day 21. H&E stain, ×40 magnification.
Fig. 3 Left: Significant alveolar bone loss was observed at day 7 in comparison with control
(C), which continued to days 14 and 21. Middle: Epithelium was significantly thicker
at days 7 to 21 in comparison with day 3 and control (C). Right: Inflammatory infiltrates
were significantly higher at all time points than control (C). Significant differences
were at p < 0.05 using the Kruskal–Wallis test. The superscript letter “a” represents the highest
value, while shared letters indicate nonsignificant differences. AB, alveolar bone;
CEJ, cementoenamel junction.
Epithelial Thickness
The gingival epithelial thickness showed a marked increased thickening in the EP model
([Fig. 2]). The gingival epithelium at days 14 and 21 displayed pronounced epithelial hyperplasia
and were significantly different from controls. Elongated epithelial ridges extending
deeply into the underlying lamina propria were observed during days 7, 14, and 21.
In contrast, epithelial ridges at day 3 resembled those in the control group ([Fig. 3], middle panel).
Inflammatory Cell Counts
All EP-TLs displayed higher inflammatory scores when compared with the nonligated
control ([Fig. 2]). The mean inflammatory cell count showed a gradual increase beginning from the
first day postligation and continued to peak at day 14, followed by a decline at the
end of the experimental period. Significantly, there were lower inflammatory cell
counts in controls compared with days 14 and 21 ([Fig. 3], left panel).
Immunohistochemical Expression of EMT Markers
E-Cadherin to N-Cadherin Switch
IHC analysis demonstrated that the E-cadherin histoscore was strongly expressed in
control and days 3 and 7, and declined at days 14 and 21. Inferential analysis showed
significantly higher E-cadherin histoscore in controls as compared with days 14 and
21. In addition, the E-cadherin expression was significantly lower at day 21 than
its expression at day 3. In contrast to N-cadherin, data showed constant upregulation
and the highest N-cadherin histoscore was observed at day 21, which was significantly
different from controls ([Fig. 4]).
Fig. 4 Cadherin switching in the experimental periodontitis model. (A) Immunohistochemical (IHC) staining showed that E-cadherin (E-cad) IHC scores were
decreased significantly at days 14 and 21 compared with control (C). (B) Concomitantly, N-cadherin (N-cad) IHC scores were significantly increased at days
7, 14, and 21 compared with day 3 and control (C). Significant differences were identified
at p < 0.05 using the Kruskal–Wallis test. The superscript letter “a” represents the highest
value, while shared letters indicate non-significant differences. Scale bar 80µm.
EMT-Transcriptional Factors (Snail and Twist)
IHC images showed that the positive expression of nuclear EMT-TFs, Snail1 and Twist,
gradually increased throughout the EP period ([Fig. 3]). Snail1 histoscores were significantly higher after 7, 14, and 21 days postligation
in comparison with controls. Twist histoscores showed similar patterns to Snail1 except
that significantly higher differences from control were not observed until days 14
and 21 ([Fig. 5]).
Fig. 5 Immunohistochemical (IHC) staining of epithelial–mesenchymal transition-transcriptional
factors (Snail1 and Twist1). Nuclear IHC expression of both transcriptional factors
was significantly increased toward the end of the experimental period. Significant
differences from control (C) appeared earlier with (A) Snail1, at day 3, (B) while Twist1 showed significant difference at day 7. Significant differences at
p < 0.05 using the Kruskal–Wallis test. The superscript letter “a” represents the highest
value, while shared letters indicate nonsignificant differences. Scale bar: 80 µm.
Vimentin
Analyses showed that both control and the day 3 group did not exhibit positive cytoplasmic
expression of vimentin ([Fig. 4]). Positive expression was initially evident at day 7 in few discrete cells (14.3%)
at the basal layer of the epithelium adjacent to the basement membrane, and a consistent
trend of expression was evident at days 14 and 21 ([Fig. 6]).
Fig. 6 Immunohistochemical (IHC) staining for vimentin. (A) Nonligated controls exhibited negative vimentin expression, (B) while positive cytoplasmic IHC expression of vimentin (green arrows) was observed at day 7 in 14.3% of cell populations, (C) which gradually increased to 33.3% at day 21. Scale bar: 100 µm. EMT, epithelial–mesenchymal
transition; EP, experimental periodontitis.
Correlation of EMT Key Biomarkers and Transcriptional Factors
The IHC histoscores of E-cadherin exhibited a significant negative correlation with
N-cadherin and Snail1. EMT-TFs, Snail1 and Twist, were significantly and positively
correlated. In addition, vimentin showed nonsignificant positive and negative correlations
with Twist and E-cadherin, respectively ([Table 1]).
Table 1
Correlation of epithelial–mesenchymal transition histoscores
|
|
E-cadherin
|
N-cadherin
|
Vimentin
|
Snail1
|
Twist
|
|
E-cadherin
|
r
|
|
–0.620
|
–0.13
|
–0.40
|
–0.105
|
|
p value[a]
|
|
<0.001
|
0.41
|
0.01
|
0.51
|
|
N-cadherin
|
r
|
–0.616
|
|
0.046
|
0.223
|
0.178
|
|
p value[a]
|
<0.001
|
|
0.77
|
0.16
|
0.27
|
|
Vimentin
|
r
|
–0.133
|
0.045
|
|
–0.070
|
0.258
|
|
p value[a]
|
0.41
|
0.77
|
|
0.66
|
0.10
|
|
Snail1
|
r
|
–0.394
|
0.223
|
–0.070
|
|
0.481
|
|
p value[a]
|
0.02
|
0.16
|
0.66
|
|
0.002
|
|
Twist
|
r
|
–0.105
|
0.178
|
0.258
|
0.481
|
|
|
p value[a]
|
0.51
|
0.27
|
0.10
|
0.002
|
|
Note: r: Spearman's correlation coefficient.
a Significant correlation at p < 0.05 indicated by bold font.
Discussion
The present study investigated the longitudinal IHC expression of EMT-key biomarkers
in an EP model. Data confirmed the EMT-associated cadherin switch together with a
positive expression of vimentin as well as upregulation of EMT-TFs, Snail1 and Twist.
These novel results supported the potential role of EMT as a pathogenic mechanism
involved in initiation and progression of periodontitis. The ligature EP is a common
technique for driving bone resorption in animal studies, which is based on inducing
dysbiosis of the microbiome mimicking that of periodontitis in a relatively short
period. The maxillary second molar was selected for ligation due to faster bone resorption
and more rapid onset of periodontitis as a result of the porosity of the maxilla.[24]
[25] Additionally, this site provides enhanced suture retention in the two interdental
areas.[17]
In this study, E-cadherin expression was decreased and associated with increased N-cadherin
expression, a hallmark of EMT.[26]
[27] These changes will compromise the integrity of the gingival epithelial barrier,
which is evident during periodontal disease initiation and progression.[11]
[22]
[28] This event is likely attributed to the microbial dysbiosis induced by the ligature
as the literature indicates the role of these bacteria and their virulence factors
in driving epithelial barrier function-related gene/protein expression via E-cadherin
dysregulation. It has been shown that arginine- and lysine-specific gingipains of
P. gingivalis, such as HRgpA, RgpB, and Kgp, have a hydrolytic effect on E-cadherin.[29] Challenging human gingival epithelial cells (HGEC) with P. gingivalis lipopolysaccharide (P. gingivalis-LPS) leads to disruption of the epithelial barrier, which accelerated the penetration
of P. gingivalis-LPS through the cell monolayer.[30] Whole live A. actinomycetemcomitans application to rat gingival sulcus and cultured HGEC decreased the level and expression
of E-cadherin, zonula occludens-1, and connexin. Additionally, treating HGEC with
recombinant Cdt of A. actinomycetemcomitans resulted in detachment of the keratinized outer layer, distention of spinous and
basal cells in the oral epithelium, and disruption of the rete pegs.[31] Furthermore, F. nucleatum–encoded adhesion protein FadA binds to E-cadherin, inducing its phosphorylation on
the membrane, internalization, and decreased expression.
EMT-TFs have key expression profiles, and their regulation of the EMT process targeted
particular cells and tissues to activate signaling pathways for EMT induction.[32] This study has identified a positive correlation between Snail1 and Twist supporting
their known role in repressing E-cadherin expression by directly and indirectly binding
to its promoter.[33] Indeed, Saliem et al previously confirmed positive Snail1 expression accompanied
by reduced E-cadherin levels in inflamed gingival tissues.[22] Moreover, Snail1 and Twist exhibit a dual role by repressing E-cadherin and inducing
N-cadherin, indicating that the cadherin switch is key to the transcriptional reprograming
process of EMT.[34] Notably, cadherin switching affects the epithelial barrier at multiple levels; N-cadherin-mediated
adherens junction are weaker than E-cadherin junction, and consequently influence
keratinocyte shifting from an adhesive epithelial state into a motile mesenchymal
invasive state. This phenomenon is mediated by several signaling pathways including
Wnt/β-catenin, PI3K/Akt, RhoA activation, and MAPK-Erk, along with expression of a
wide range of proteolytic metalloproteinases.[35] The cleaving and proteolytic action of the latter facilitates downregulation and
shedding of E-cadherin, which increase epithelial permeability, inducing basement
membrane microulceration, and enable invasion of the connective tissue by periodontal
pathogens.[36]
Periodontal pathogens and/or their virulence factors are considered potent Snail1
inducers.[30] Challenging rat oral keratinocytes with P. gingivalis and F. nucleatum induced expression of Snail1 together with other EMT regulators.[11] Indeed, the inflammatory response induced by the periodontal dysbiotic microenvironment
is enriched with a wide array of Snail1 inducers.[37] These are responsible for inducing complex downstream signaling networks affecting
Snail1 expression, including receptor tyrosine kinases signaling, which act through
the Ras-MAPK or PI3K-Akt, TGF-β, Notch, Wnt, NF-κB, TNF-α, and BMPs pathways Furthermore,
posttranslational modifications such as Gsk3β-mediated phosphorylation of Snail1 are
initiated by these aforementioned signaling pathways, thus controlling its cellular
localization, stability, degradation, and activity.[38]
Vimentin was previously identified in the human gingival crevicular fluid by quantitative
proteomic analysis and was more closely associated with periodontitis compared with
gingivitis and gingival health.[39] During EMT, cells shift their IF composition from being keratin dominant to vimentin
dominant. Although both are responsible for trafficking organelles and membrane-associated
proteins, they show different protein preferences. While keratins direct E-cadherin
to the cellular membrane enabling cellular adhesion, vimentin interacts with motor
proteins enabling cellular motility.[40] Our IHC findings revealed a gradual increase in the number of vimentin-positive
cells subjacent to the basement membrane. These data were consistent with in vitro evidence of epithelial cells expressing vimentin upon exposure to periodontal pathogens.[11] In addition, positive vimentin expression in inflamed human tissue samples of periodontitis
patient was also reported.[22]
[28] EMT is a multistage process and vimentin-positive cellular expression is associated
with the advanced state. Current findings support this notion since the positive expression
was observed at the advanced EP stage.
The latest guidelines recommend a multileveled investigational panel to confirm EMT,
including EMT molecular markers along with changes in the cellular and behavior properties
rather than relying on the sole influence of the molecular markers. In the current
study, EMT was investigated in the rat inflamed gingival tissue samples utilizing
only the immunohistochemical expression profile of EMT biomarker and TFs. Such limitation
arise since this study was self-funded. Further panels of analysis were not applicable
due to the limited laboratory and technical support as well as lack of institutional
financial grants.[5] Currently, and, to the best of our knowledge, the results of this preclinical study
are novel and demonstrate the positive expression of EMT biomarkers in EP. These findings
could pave the way for future experimental studies and for exploring unidentified
mechanistic links between periodontitis and EMT. Additionally, this model could provide
an opportunity for testing EMT inhibitors preclinically as a novel approach to treat
the disease.
Conclusion
The results of this EP model demonstrated induction of an EMT phenotype in diseased
periodontal tissue. This was confirmed by cadherin switching and positive vimentin
expression along with nuclear translocation of Snail1 and Twist. Further studies are
required to better characterize the induction of the EMT phenotype in the EP model
and these data could then provide a platform to better understand disease progression
and identify pharmaceuticals for disease management and treatment.
