cancers
Article
N-Cadherin Distinguishes Intrahepatic Cholangiocarcinoma
from Liver Metastases of Ductal Adenocarcinoma of
the Pancreas
Tiemo S. Gerber 1 , Benjamin Goeppert 2, Anne Hausen 1, Hagen R. Witzel 1, Fabian Bartsch 3 ,
Mario Schindeldecker 1,4, Lisa-Katharina Gröger 3 , Dirk A. Ridder 1, Oscar Cahyadi 5, Irene Esposito 6 ,
Matthias M. Gaida 1, Peter Schirmacher 2, Peter R. Galle 7 , Hauke Lang 3, Wilfried Roth 1 and Beate K. Straub 1,*
1 Institute of Pathology, University Medical Center of the Johannes Gutenberg-University Mainz,
55131 Mainz, Germany; tiemo.gerber@unimedizin-mainz.de (T.S.G.);
anne.hausen@unimedizin-mainz.de (A.H.); hagen.witzel@unimedizin-mainz.de (H.R.W.);
mario.schindeldecker@unimedizin-mainz.de (M.S.); dirk.ridder@unimedizin-mainz.de (D.A.R.);
matthias.gaida@unimedizin-mainz.de (M.M.G.); wilfried.roth@unimedizin-mainz.de (W.R.)
2 Institute of Pathology and Neuropathology, RKH Klinikum Ludwigsburg, 71640 Ludwigsburg, Germany;
benjamin.goeppert@rkh-gesundheit.de (B.G.); peter.schirmacher@med.uni-heidelberg.de (P.S.)
3 Department of General, Visceral and Transplant Surgery, University Medical Center of the Johannes
Gutenberg-University Mainz, 55131 Mainz, Germany; fabian.bartsch@unimedizin-mainz.de (F.B.);
lisa-katharina.heuft@unimedizin-mainz.de (L.-K.G.); hauke.lang@unimedizin-mainz.de (H.L.)
4 Tissue Biobank, University Medical Center of the Johannes Gutenberg-University Mainz,
55131 Mainz, Germany
5 Institute of Pathology, University of Heidelberg, 69120 Heidelberg, Germany; o.cahyadi@klinikum-bochum.de
6 Institute of Pathology, University Clinic Düsseldorf, 40225 Düsseldorf, Germany;
irene.esposito@med.uni-duesseldorf.de
7 Department of Medicine I, University Medical Center of the Johannes Gutenberg-University Mainz,
55131 Mainz, Germany; peter.galle@unimedizin-mainz.de
Citation: Gerber, T.S.; Goeppert, B.; * Correspondence: beate.straub@unimedizin-mainz.de
Hausen, A.; Witzel, H.R.; Bartsch, F.;
Schindeldecker, M.; Gröger, L.-K.; Simple Summary: The differential diagnosis of primary liver cancer versus liver metastases of an
Ridder, D.A.; Cahyadi, O.; Esposito, extrahepatic primary tumor may be challenging, especially in carcinoma of the pancreatobiliary
I.; et al. N-Cadherin Distinguishes phenotype, as in intrahepatic cholangiocarcinoma or metastases of ductal adenocarcinoma of the
Intrahepatic Cholangiocarcinoma pancreas. Nevertheless, the distinction is of fundamental importance for therapy planning. In
from Liver Metastases of Ductal the frame of this study, we focus on the differential expression and localization of the cell–cell
Adenocarcinoma of the Pancreas. contact-proteins E-cadherin and N-cadherin, markers for cell differentiation in normal epithelia of
Cancers 2022, 14, 3091. https:// the pancreaticobiliary tract as well as in derived tumors, and demonstrate that N-cadherin positivity
doi.org/10.3390/cancers14133091 distinguishes intrahepatic cholangiocarcinoma from liver metastases of ductal adenocarcinoma of the
Received: 20 May 2022 pancreas. We propose that N-cadherin-positivity together with other biliary markers may be used
Accepted: 21 June 2022 for this important histopathological differential diagnosis and may thus improve the accuracy of
Published: 23 June 2022 cholangiocarcinoma diagnosis.
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in Abstract: Carcinomas of the pancreatobiliary system confer an especially unfavorable prognosis.
published maps and institutional affil- The differential diagnosis of intrahepatic cholangiocarcinoma (iCCA) and its subtypes versus liver
iations. metastasis of ductal adenocarcinoma of the pancreas (PDAC) is clinically important to allow the
best possible therapy. We could previously show that E-cadherin and N-cadherin, transmembrane
glycoproteins of adherens junctions, are characteristic features of hepatocytes and cholangiocytes. We
therefore analyzed E-cadherin and N-cadherin in the embryonally related epithelia of the bile duct
Copyright: © 2022 by the authors. and pancreas, as well as in 312 iCCAs, 513 carcinomas of the extrahepatic bile ducts, 228 gallbladder
Licensee MDPI, Basel, Switzerland.
carcinomas, 131 PDACs, and precursor lesions, with immunohistochemistry combined with image
This article is an open access article
analysis, fluorescence microscopy, and immunoblots. In the physiological liver, N-cadherin colocalizes
distributed under the terms and
with E-cadherin in small intrahepatic bile ducts, whereas larger bile ducts and pancreatic ducts are
conditions of the Creative Commons
Attribution (CC BY) license (https:// positive for E-cadherin but contain decreasing amounts of N-cadherin. N-cadherin was highly
creativecommons.org/licenses/by/ expressed in most iCCAs, whereas in PDACs, N-cadherin was negative or only faintly expressed.
4.0/). E- and N-cadherin expression in tumors of the pancreaticobiliary tract recapitulate their expression
Cancers 2022, 14, 3091. https://doi.org/10.3390/cancers14133091 https://www.mdpi.com/journal/cancers
Cancers 2022, 14, 3091 2 of 18
in their normal tissue counterparts. N-cadherin is a helpful marker for the differential diagnosis
between iCCA and PDAC, with a specificity of 96% and a sensitivity of 67% for small duct iCCAs
and 50% for large duct iCCAs.
Keywords: liver cancer; cell–cell contacts; adherens junctions; biliary tract cancer; differential diagnosis;
tumor biology
1. Introduction
Adenocarcinomas of the pancreatobiliary tract constitute a group of underestimated
carcinomas with a devastating prognosis: although tumors of the liver and pancreas account
only for the sixth and twelfth most frequent tumors, they present the third and seventh
most deadly tumors worldwide [1,2]. Patients with pancreatic ductal adenocarcinoma
(PDAC), the most frequent tumor of the pancreas, have a poor one-year survival rate
of only 18% for all stages of the disease [3]. Intrahepatic cholangiocarcinoma (iCCA) is
the second most common primary hepatic malignancy after hepatocellular carcinoma
(HCC), with increasing incidence and yet unfavorable prognosis between that of HCC and
PDAC [4–6]. In patients without metastatic spread, PDAC and iCCA are both treated with
locally curative resection. Metastatic disease is usually treated with chemotherapy [3–5].
However, PDAC and iCCA need different chemotherapy regimens. For example, while
5-FU is not the first choice for therapy of PDAC, it is a more common choice for iCCA.
As already stated by Lok et al., the differential diagnosis of PDAC metastases and iCCA
based on histomorphological criteria alone is challenging [7]. However, there is still no
sufficiently specific immunohistochemical marker for this difficult diagnostic dilemma.
Therefore, it is pivotal to improve our knowledge of underlying circumstances leading to
iCCA and PDAC, unravel potential subgroups, and thereby help improve therapy.
According to the current 2019 WHO classification [8], iCCA are classified into small
duct and large duct iCCAs. Depending on their location, extrahepatic cholangiocarcino-
mas are subdivided into perihilar (pCCA) and distal (dCCA) cholangiocarcinomas [9,10].
Due to their common embryonal origin, in conventional histology, the differential diag-
nosis of biliary tract carcinoma (BTC) and metastatic PDAC involving the liver may be
problematic [11,12].
Cell–cell contacts of adherens junctions are mechanical tethers of multicellular organ-
isms that are required for proper cellular function, embryogenesis, cell differentiation, and
maintenance of tissue integrity [13]. Due to their importance in regulating the invasion
and metastasis of malignant tumors, they have been extensively studied in tumorigenesis,
especially carcinogenesis [14]. The transmembrane domain of adherens junctions is formed
by proteins of the cadherin superfamily, Ca2+-dependent membrane glycoproteins com-
prising more than 100 members in humans [15]. The extracellular domains of cadherins
connect one cell to the extracellular domain of another cadherin, usually in a homotypic
fashion, whereas their intracellular domains are linked via (α-, β-, γ-) catenins to the actin
cytoskeleton [16,17]. Cadherin–catenin interaction, especially via the nuclear translocation
of β-catenin, influences cell signaling, such as the Wnt-signaling cascade, and plays an im-
portant role in the maintenance and homeostasis of polarized epithelial monolayers [18–20].
The best-studied cadherin in epithelial cells is E-cadherin (formerly known as uvomorulin
and L-CAM [21]), whereas N-cadherin (formerly known as A-CAM) has been predomi-
nantly described in mesenchymal and neuroectodermal cells [22–25]. Loss of E-cadherin in
the mouse embryo leads to early death due to defective preimplantation development [26].
Loss of N-cadherin shows defects in the formation of the neural tube and impaired heart
development, leading to death on day 10 of mouse development [27]. Due to their cell-
and tissue-specific expression patterns, cadherins may be used as immunohistochemical
markers to determine the cell of origin, e.g., in malignant tumors. For example, E- and
N-cadherin are routinely used in the differential diagnosis of malignant mesothelioma and
Cancers 2022, 14, 3091 3 of 18
adenocarcinoma of the lung [28,29]. In addition, loss of E-cadherin is a hallmark of lobular
invasive carcinoma of the breast [30] as well as signet cell carcinoma of the stomach [31].
We could previously unravel that hepatocytes harbor equal amounts of E- and N-
cadherin and are characterized by a certain type of adherens junction, forming cis-E:N-
cadherin heterodimers [32] at the basolateral membrane that are also retained in the respec-
tive hepatocellular tumors, such as hepatocellular adenoma and carcinoma. In preliminary
analyses, we also detected the colocalization of E- and N-cadherin in small and large bile
ducts, so we concluded that the coexpression of E- and N-cadherin may be a characteristic
of an endodermal cell lineage, differentiating the liver from other organs, where N-cadherin
has only been demonstrated in the process of epithelial-to-mesenchymal transition [32]. In
small cohorts of patients, studies by other authors have already demonstrated the presence
of N-cadherin, in addition to E-cadherin, in cholangiocarcinomas as well [33–35]. To test the
utility of E- and N-cadherin in the differential diagnosis of tumors of the pancreatobiliary
system, we have analyzed the expression pattern in the physiological biliary tract as well
as in a large cohort of patients with iCCA, pCCA, GBC, and PDAC, along with clinical
parameters. In addition, we propose a strategy in the workup of these entities that utilizes
both commonly available antibodies and N-cadherin.
2. Materials and Methods
2.1. Tissue Collection
Paraffin-embedded formalin-fixed (FFPE) tissue samples from 1184 patients with
iCCA, pCCA, dCCA, GBC, and PDAC, diagnosed at the Institutes of Pathology, University
Medical Center Mainz, between 2006 and 2020 and the University of Heidelberg, were
collected. Tissue samples were processed in accordance with the regulations of the Tissue
Biobank of the University Medical Center Mainz after approval by the local ethics com-
mittee (Ethics Committee of the State Medical Council of Rhineland-Palatinate and the
University Heidelberg). Tissue microarrays (TMAs) with cores of 2 mm were constructed.
For each patient, the primary tumor, corresponding normal tissue, and, if available, pre-
cursor lesions, as well as lymph nodes or foreign metastases, were represented. Tumor
samples were taken in duplicate, one sample each from central and peripheral tumor areas,
to mitigate the impact of intratumoral heterogeneity. Overall, 11,396 TMA cores were
evaluated. In addition, twelve cryopreserved tumor samples of iCCA, pCCA, GBC, and
PDAC were used for immunoblot analysis and immunofluorescence microscopy. Moreover,
cryopreserved normal human, mouse, rat, and bovine liver, gallbladder, and pancreas
tissues were used as previously published [32]. Cultured cells of human PDACs were from
the lines Capan-1 (ATCC, HTB-79) and Panc-1 (ATCC CRL-1469).
2.2. Antibodies and Reagents
Mouse monoclonal antibodies were against E-cadherin (Clone 36, Bioscientia Health-
care GmbH, Ingelheim, Germany), N-cadherin (Clone 32, Bioscientia Healthcare GmbH),
and actin (Clone C4, Merck Chemicals GmbH, Darmstadt, Germany). Mouse monoclonal
antibodies purchased from Dako Deutschland GmbH, Hamburg, Germany, were against
epithelial membrane antigen (EMA; Clone E29), CDX2 (Clone DAK-CDX2), CK7 (Clone OV-
TL,), CK20 (Clone K2 20.8), and Ca19-9 (Clone 1116-NS-19-9), as well as Ki-67 (Clone MIB-1).
In addition, a rabbit monoclonal antibody was used against E-cadherin (Clone EP700Y,
Epitomics, Burlingame, CA, USA), with rabbit polyclonal antibodies against N-cadherin
(Calbiochem, Darmstadt; QED Bioscience, San Diego, CA, USA).
2.3. Immunofluorescence Microscopy
Cryosections of normal liver and tumor tissues were cut at a thickness of 5 µm, air-
dried for 1 h, and fixed with acetone at −20 ◦C for 10 min. After a permeabilization
step for 4 min in 0.1% Triton X-100 and two washing steps in phosphate-buffered saline
(PBS), the primary antibodies were applied for 30 min to 1 h, followed by two washing
steps for 5 min in PBS and 30 min incubation with the respective secondary antibodies
Cancers 2022, 14, 3091 4 of 18
in a humid chamber (cy 3, rabbit, Dianova; Alexa 488 anti-mouse, MoBiTec). After two
subsequent washing steps for 5 min in PBS as well as a short washing step in distilled
water, the slides were dehydrogenated with 100% ethanol for 5 min and mounted with a
DAPI embedding medium. For confocal immunofluorescence microscopy, a laser scanning
microscope (LSM 510 Meta; Carl Zeiss AG, Oberkochen, Germany) equipped with Plan
Apochromat 63×/1.40 NA oil and Plan-Neofluar 40×/1.30 NA oil objectives was used.
AxioVision Release 4.6.3.0 and LSM Image Browser 3.2.0.115 software (Carl Zeiss AG,
Oberkochen, Germany) were used for image processing.
2.4. Immunoblot Analysis
Cryopreserved human tissue samples were cut into 100–150 5 µm thick sections with
a cryostat (Leica CM3050 S) and transferred into tissue homogenizing kit CKMix tubes
(Precellys). Afterwards, the tissue sections were supplemented with 500 µl RIPA buffer
(50 mM TRIS-HCl, 15 mM EGTA, 100 mM NaCl, 1% (v/v) Triton X-100, pH = 8.0) and
1:100 Halt™ protease and phosphatase inhibitor (Thermo ScientificTM). The lysis was
performed with a Precellys 24 homogenizer at 6500 rpm for 20 s. The homogenate was
centrifuged at 12,000× g for 5 min at 4 ◦C, and the pellet was discarded. The protein
concentration of the supernatant was determined with the Bradford protein assay (Bio-
Rad). The proteins of the total cell lysates were then separated by 8% SDS-PAGE and
transferred with a Trans-Blot® TurboTM transfer system (Bio-Rad) onto a nitrocellulose
membrane (AmershamTM ProteanTM 0.45 µm NC, GE Healthcare Life science). The
membrane was blocked overnight at 4 ◦C with 1× Tris-buffered saline (A5001, PanReac
AppliChem ITW Reagents) supplemented with 0.05% (v/v) Tween-20 and 5% (w/v) nonfat
dried milk powder (A0830, PanReac AppliChem ITW Reagents). Primary and secondary
antibodies were diluted in a blocking solution (1:2000 dilution for E- and N-cadherins and
1:10000 for actin). Gimp software 2.10.28 was used for pixel density analysis of the protein
bands. We calculated the E-cadherin/actin and N-cadherin/actin ratios for each band and
displayed them relative to the maximum.
2.5. Immunohistochemistry
Immunohistochemistry (IHC) was performed on 4 µm thick histological sections of
FFPE specimens, according to the manufacturer’s recommendations. Subsequently, slides
were digitalized by a whole slide scanner at 40×, with a pixel size of 0.2278 × 0.2278 µm
(Nanozoomer, Hamamatsu Photonics, Hamamatsu, Japan).
Specific membranous immunohistochemical stainings for E- and N-cadherin were
manually evaluated by scoring the intensity of staining from 0 (absence of staining), 1 (faint
membranous staining), and 2 (moderate membranous staining) to 3 (strong membranous
staining), as described before [36]. Immunohistochemical staining was performed on tissue
specimens present in duplicate, and the mean value was calculated. A staining intensity
that had a mean value greater than or equal to 1.5 was assigned to the high expression
group, while values less than 1.5 were assigned to the low expression group. The extent of
cytoplasmic staining for CK7, CK20, EMA, and CA19-9 and the nuclear staining for CDX2
were classified as negative, weak, moderate, and strong. As a measure of proliferation
activity indicative of aggressive tumor behavior, the Ki-67-staining was correlated with E-
and N-cadherin expression.
A quantitative assessment of E- and N-cadherin expression levels was undertaken us-
ing QuPath version 0.2.3, an open-source software [37]. TMA cores were de-arrayed. Stain
vectors and background vectors were individually set in each slide and further analyzed
using the “cell detection” algorithm in QuPath. Cellular chromogen 3,3’-diaminobenzidine-
tetrahydrochloride-dihydrate mean levels were classified into four categories, using empir-
ical threshold scores (0–0.21 for 0, >0.21 for 1+, >0.45 for 2+, and >0.7 for 3+). These levels
provided a relatively low threshold for the discrimination of negative stains (score 0) and
faint stains (1+) and a relatively high threshold for the highest achievable score. Ki-67 index
estimation was performed after de-arraying and color decomposition using the positive cell
Cancers 2022, 14, 3091 5 of 18
detection algorithm in QuPath. The tumor was annotated using a detection classifier. Due
to intertumoral heterogeneity, we used custom-tailored classifiers on a case-to-case basis to
ensure the proper separation of tumor and non-tumor tissue. This entailed the application
of the random trees classifier to train QuPath interactively to distinguish tumor cells from
stromal and inflammatory cells. The H-score for E- and N-cadherins was calculated from
the extent and intensity of staining, giving scores ranging from 0 to 300 [38]. Ki-67 index
estimation was conducted with only one threshold, resulting in a mean percentage of
positive tumor cells. TMA cores that could not be evaluated, e.g., due to loss of tissue
during staining, were excluded from the study.
2.6. Statistics
Statistical analyses were performed using IBM SPSS Statistics Version 27 (IBM Corp.,
Armonk, NY, USA). To assess the data for normal distribution, the Shapiro–Wilk test was
applied for sample sizes below 30. According to the central limit theorem, the distribution
above a sample size of 30 is considered normally distributed [39]. An unpaired t-test was
used to compare mean H-scores between each tumor entity as well as respective lymph
node metastases and normal tissues. Because of different sample sizes, we conducted
Hedges’ g-test to determine effect size. For non-normally distributed data, we used the
Mann–Whitney U-test and calculated the effect size r, according to Fritz et al. [40]. To
determine the relationship between two variables, we calculated the Spearman rank-order
correlation coefficient for non-metric data and the Pearson correlation coefficient for metric
and normally distributed data. We assessed the strength of correlation coefficient values
according to Chan [41]. Overall survival and recurrence-free survival were identified as
the time measured from initial diagnosis to death or cancer recurrence, respectively, and
censored at the last clinical follow-up. Using the E-/N-cadherin H-scores from our cohort
and the CDH1 and CDH2 mRNA data from the TCGA cohort [42], both optimal cut-off
values were calculated using the Charité Cutoff Finder [43]. Survival analyses were plotted
using the Kaplan–Meier model and compared by log-rank test. For the survival analysis
of our cohort, we only used primary iCCA, and we excluded cases with unresectable
tumors. To evaluate the discrimination of the dichotomous variables (high vs. low N-
cadherin expression), we used a variation of the chi-squared test, McNemar’s test [44].
A p-value ≤ 0.05 was considered statistically significant.
3. Results
3.1. Colocalization of E- and N-Cadherin Characterizes Epithelia of the Physiological Biliary Tract
After we discovered a novel type of adherens junction involving cis-E:N-cadherin
heterodimers in the hepatocytes of embryonic and adult liver that also stably characterizes
hepatocellular adenoma and carcinoma [32], we were interested in investigating E- and
N-cadherin in the embryonically related cells of the biliary tract and pancreas as well as in
derived tumors of the pancreatobiliary system.
Using cryopreserved mouse, rat, bovine, and human liver tissues, complete colocal-
ization of E- and N-cadherin at the basolateral membrane of hepatocytes was detected, as
previously published. Moreover, the epithelial cells of small and large intrahepatic bile
ducts also showed complete colocalization at the basolateral membrane, whereas stromal
cells of the portal tracts and endothelia of larger vessels were positive for N-cadherin and
negative for E-cadherin, as expected. In the gallbladder epithelium, however, variability
was observed among different species as rats did not harbor a gallbladder, mouse and
human gallbladder epitheliums were positive only for E-cadherin, and bovine gallbladder
was positive for both E- and N-cadherin. The extrahepatic bile ducts showed positivity
for E-cadherin; however, minor amounts of N-cadherin were observed when compared
to intrahepatic bile ducts. In the pancreas, acinar epithelial cells were only positive for
E-cadherin, islet cells were positive in part for N-cadherin and in part for E-cadherin, and
pancreatic ducts were only positive for E-cadherin, except for bovine pancreas, where
Cancers 2022, 14,  6  of  18 
 
whereas stromal cells of the portal tracts and endothelia of larger vessels were positive for 
N‐cadherin  and  negative  for  E‐cadherin,  as  expected.  In  the  gallbladder  epithelium, 
however,  variability was  observed  among  different  species  as  rats  did  not  harbor  a 
gallbladder,  mouse  and  human  gallbladder  epitheliums  were  positive  only  for  E‐
cadherin,  and  bovine  gallbladder  was  positive  for  both  E‐  and  N‐cadherin.  The 
extrahepatic bile ducts showed positivity for E‐cadherin; however, minor amounts of N‐
cadherin were observed when compared to intrahepatic bile ducts. In the pancreas, acinar 
Cancers 2022, 14, 3091 epithelial cells were only positive for E‐cadherin, islet cells were positive in part fo6r oNf 1‐8
cadherin  and  in  part  for  E‐cadherin,  and  pancreatic  ducts were  only  positive  for  E‐
cadherin, except for bovine pancreas, where pancreatic ducts demonstrated positivity for 
pbaonthc rEe‐a taincdd uNc‐tcsaddhemeroinn s(Ftriagtuerde p1)o.s Sitoivmitey laforgrebro hthumEa-na npdanNcr-ecaatdich edruincts(F digisuprleay1e)d.  Sfaoimnte 
laamrgoeurnhtus mofa Nn ‐pcaandchreerainti casd uwcetlsl.d  isplayed faint amounts of N-cadherin as well.
 
Figure 1.1C.  oClooclaolciazlaitziaotnioonf Eo-fa nEd‐ Na-cnadd hNer‐icnaidshceorninse  rivse dcoinnsneorrvmedal  eipni thneoliramoafl theeppitahneclriae atofb iltihaery 
tpraanctc.rDeaotoubbilleia-lrayb etlrlaacste. r sDcoaunbnlien‐glaibmelm  ulansoeflr usocraenscneinngce  mimicmrousncoofpluyosrhesocwenscceo mmpilcertoesccoolpoyc alsihzaotwiosn 
(croemsupltleinteg cionloyceallliozwaticoonl o(rreisnultthinegm ine rygeelloimwa cgoelso)r oinf tEh-ec amdehregrein im(Ea-gceasd) ,oAf Ele‐xcaad4h8e8r,ing r(Eee‐cna)d, ,w AitlhexNa -
c4a8d8h, egrrienen(N),- cwaidth, c Ny3‐c, areddh)eirnins m(Na‐llcabdil,e cdyu3c, trseidn) ninor smmaal lhl ubmilea ndulicvtesr i(nA n),oirnmaalla hrguembailne ldivuecrt i(nAn),o irnm aa l
large  bile duct  in normal bovine  liver  (B),  as well  as  in bovine  gallbladder  epithelium  (C)  and 
bpoavnicnreealtiivce rdu(Bc)t ,eapsiwtheellliuams i ninb novorinmeagl ablolbvlianded pearnecprietahse l(iDum). N(Co)tea ntdhep caonlcorceaaltiizcadtiuonct oefp  iEt‐h ealniudm Ni‐n
ncaodrmhearlinbso vinin neopramnaclr ehaesp(aDto).cyNtoeste  (tAh,eBc),o  ltohcea pliozasittiiovnityo foEf- vaensdseNls- cfaodr hoenrliyn sNin‐candohrmerainl  hbeupta ntoocty Ete‐s
(cAad,Bh)e,rtihne  p(Bo,sCit)i,v  iEty‐coadf hveersisne‐lpsofosirtiovnitlyy Nw-ictahd  hNe‐rciandbhuetrinno‐nteEg-actaidvhiteyr inin( Bp,Can),crEe-acsa dahceinriin  -(pDo)s, itaivnidty 
wdiftfherNen-ctaiadl hNe‐r iann-dn eEg‐actaidvhiteyrin epxapnrcersesaisona ciinn pi (aDnc),reaantdic disiflefet rceenlltsia (lDN).- Baanrd leEn-gcathd:h 2e0r iμnme.x pression in
pancreatic islet cells (D). Bar length: 20 µm.
The  respective  findings  were  reproduced  immunohistochemically  in  a  large 
collecTthioenr eosfp ehcutimveanfi nFdFiPnEgs‐mwaetreeriraelp  orof dsumcaedll iamnmd ulanroghei sbtoilceh  edmucictsa,l lpyeirnihaillaarr gbeilceo ldleucctitos,n 
of human FFPE-material of small and large bile ducts, perihilar bile ducts, gallbladder
epithelium, and pancreatic ducts, as evaluated manually as well as with quantitative
  analysis (Figure 2). In normal pancreatic ducts, using immunohistochemistry, apical
secretions are frequently strongly stained by the used N-cadherin antibody that is not
observed in immunofluorescence microscopy, resulting in a false-positive N-cadherin
H-score, as determined by QuPath (Figure S1).
Cancers 2022, 14,  7  of  18 
 
gallbladder  epithelium,  and  pancreatic  ducts,  as  evaluated manually  as well  as with 
quantitative analysis (Figure 2). In normal pancreatic ducts, using immunohistochemistry, 
apical secretions are frequently strongly stained by the used N‐cadherin antibody that is 
not observed in immunofluorescence microscopy, resulting in a false‐positive N‐cadherin 
H‐score, as determined by QuPath (Figure S1). 
In summary, during physiological conditions, such as hepatocytes, the epithelia of the 
intrahepatic  bile  ducts  harbor  high  amounts  of  both  E‐  and  N‐cadherin.  N‐cadherin 
Cancers 2022, 14, 3091 7 of 18
expression markedly declines  along  the biliary  tree  from  intrahepatic bile ducts with  a 
strong expression to extrahepatic bile ducts and pancreatic ducts with faint or no expression. 
 
Figure 2. E‐ and N‐cadherin expression  in physiological epithelia of  the human pancreatobiliary 
Figtruarcet. 2I.mEm- aunndohNis-tcoacdhheemriincael xapnraelsyssioisn  oinf pEh‐ yasniodl oNgi‐ccaaldehpeirtihnesl iadeomf tohnesthrautmesa nstproanngc reEa‐ctoabdihleiarriny  tract.
Imemxpurneosshiiosnto  icnh esmaiclla  l(Aan) aalnyds islaorgfeE -bailne ddNuc-tcsa  (dBh)e  irnin hsudmeamno  lnivsterra, tiens  as trpoenrighiEla-rc abdilhee  rdiuncet x(pCr)e, sisni on in
small (A) and large bile ducts (B) in human liver, in a perihilar bile duct (C), in gallbladder epithe-
lium (D) and in pancreas ducts and acini (E), whereas N-cadherin expression declines from strong
  expression in intrahepatic and perihilar bile ducts to faint expression in gall bladder epithelium, with
negative expression in pancreatic ducts. QuPath analysis is shown on the bottom right (A’–E’,A”–E”).
(A’): manual score 3.0, E-cadherin H-score 268.75; (A”): manual score 3.0, N-cadherin H-score 218.75;
(B’): manual score 3.0, E-cadherin H-score 268.56; (B”): manual score 1.5, N-cadherin H-score 67.73;
(C’): manual score 3.0, E-cadherin H-score 212.39; (C”): manual score 2.5, N-cadherin H-score 85.43;
(D’): manual score 3.0, E-cadherin H-score 290.83; (D”): manual score 2.0, N-cadherin H-score 58.47;
(E’): manual score 2.5, E-cadherin H-score 105; (E”): manual score 0.5, N-cadherin H-score 8.33.
Cells with colored border: white—stromal cells, blue—epithelial cell 0, yellow—epithelial cell 1+,
orange—epithelial cell 2+, red—epithelial cell 3+. Black bar: 50 µm (A–E,A’–E’,A”–E”) Black bar:
20 µm (A–E,A’–E’,A”–E”).
Cancers 2022, 14,  8  of  18 
 
gallbladder epithelium  (D) and  in pancreas ducts and acini  (E), whereas N‐cadherin expression 
declines from strong expression in intrahepatic and perihilar bile ducts to faint expression in gall 
bladder epithelium, with negative expression in pancreatic ducts. QuPath analysis is shown on the 
bottom right (A’–E’,A’’–E’’). (A’): manual score 3.0, E‐cadherin H‐score 268.75; (A’’): manual score 
Cancers 2022, 14, 3091 3.0, N‐cadherin H‐score 218.75; (B’): manual score 3.0, E‐cadherin H‐score 268.56; (B’’): manual s8coofr1e8 
1.5, N‐cadherin H‐score 67.73; (C’): manual score 3.0, E‐cadherin H‐score 212.39; (C’’): manual score 
2.5, N‐cadherin H‐score 85.43; (D’): manual score 3.0, E‐cadherin H‐score 290.83; (D’’): manual score 
2.0, NIn‐casduhmermina Hry‐s,cdourer 5in8.g47p; h(Ey’)s:i molaonguiacla slccooren 2d.5it, iEo‐ncasd, hsuercihn Has‐shcoerpea 1t0o5c; y(Ete’’s):, mthaenueapl istchoerlei a0.5o,f 
tNhe‐caindthrearhine pHa‐tsiccorbei l8e.3d3u. cCtesllhs awrbitohr choliogrheda mboorudenrt:s wohfibteo—thstrEo-maanl dceNlls-,c abdluhee—rienp.itNhe-lciaald cheell 0, yellow—epithelial cell 1+, orange—epithelial cell 2+, red—epithelial cell 3+. Black bar: 50 μm (Ari–n
eEx,Apr’–eEs’s,iAo’n’–Em’’a) rBklaecdkl ybadr:e 2c0l iμnmes (Aal–oEn,Ag’–thE’e,Ab’i’–liEa’r’)y.  tree from intrahepatic bile ducts with a
strong expression to extrahepatic bile ducts and pancreatic ducts with faint or no expression.
3.2. N‐Cadherin Distinguishes Intrahepatic Cholangiocarcinoma from PDAC 
3.2. N-Cadherin Distinguishes Intrahepatic Cholangiocarcinoma from PDAC
To analyze whether biliary  tumors and precursor  lesions recapitulate  the same E‐ 
and TNo‐caandahlyze whether biliary tuand N-cadheerriinn eexxpprreessssiioonn ppaatttteerrn
mors
n aass pph
aynsid precurhysioollooggiiccaalll
syo pr rleesseionnt sinr eincatrpaihtueplaatteict hbeiles admucetsE, -
we analyzed biliary intraepithelial neoplasia (BilIN) alsy wpresent in intrahepatic bile duwe analyzed biliary intraepithelial neoplasia (BilIN) aselwl aesl liCaCsAiC. ICnA d.oIunbldeo‐luabbelel -llaas
cts,
beerl 
lsacsaenrnscinagn niimngmiumnmofulunoorfleusocerensccee nmceicmroisccroopscyo, pbyi,libairlyia riyntirnaterpaeitphiethliealli anlenoepolpalsaiasi aanandd  iCiCCCAA 
ddeemmoonnssttrraatteeddp  apratriatilatlo  tcoo mcpomletpelecotel occaolliozcaatiloiznatoifonE - aonf dEN‐ -caanddh  eNri‐ncaadththeericne llatm  ethmeb  rcaenlle 
(mFiegmurbera3n).e (Figure 3). 
 
FFiigguurree 33.. EE-‐ aannd N-‐ccadheriin are rettained in human BilIN‐-lesions as wellll as in iCCA.. Doubllee‐-llaabbeell 
llaasseerr ssccaannnniinngg iimmuunnooflfluuoorreesscceennccee miiccrroossccooppyy sshhoowss ppaarrttiiaall ccoollooccaalliizzaattiioonn ooff EE‐-ccaaddhheerrinin ((EE‐c-caadd, ,
AAlleexxaa 448888,, ggrreeeenn)) aanndd NN-‐ccaaddhheerriinn  ((NN‐-ccaadd,, ccyy33,, rreedd))  iinn BBiillIINN‐-lleessiioonnss  ((AA––AA’”’)) aass wweellll aass  iinn ppoooorrlyly 
ddiiffffeerreennttiiaatteeddi iCCCCAA ((BB––BB”’’)).. BBeessiiddeess ccoommpplleettee ccoollooccaalliizzaattiioonn,, aass ddeemmoonnssttrraatteedd bbyy tthhee mmeerrggeedd yyeellloloww 
color,  the  cell membrane also  stretches with more N‐cadherin or more E‐cadherin,  as  seen. Bar 
cloelnogr,thth: e20c eμllmm. embrane also stretches with more N-cadherin or more E-cadherin, as seen. Bar length:
20 µm.
TToo ffuurrtthheerr aannaallyyzzee EE-‐ aanndd NN-‐ccaaddhheerriinn eexxpprreessssiioonn iinn aa llaarrggee ccoolllleeccttiioonn ooff ttuummoorrss ooff 
tthhee ppaannccrreeaattoobbiilliiaarryy ttrraacctt,, ttooggeetthheerr wwiitthh mmeettaassttaasseess pprreeccuurrssoorr lleessiioonnss,, iinn ccoommppaarriissoonn ttoo 
tthhee rreessppeeccttiivvee nnoorrmmaall ttiissssuueess,, wwee uusseedd iimmmmuunnoohhiissttoocchheemmiissttrryy ffoorr EE‐- aanndd NN‐-ccaaddhheerriinn iinn 
tissue microarrays comprising 1184 patient samples (Table 1). Nearly every biliary tract
cancer and PDAC, as well as their associated lymph node metastases, precursor lesions,
  and associated non-neoplastic tissue, showed a high expression of E-cadherin. However,
N-cadherin expression was strongest and most stable in small duct iCCA, and expression
declined in the following order: large duct iCCA, BilIN, pCCA, GBC, dCCA, with the
lowest to no expression in PDAC (Figure 4).
Cancers 2022, 14, 3091 9 of 18
Table 1. Immunohistochemical evaluation of E- andf N-cadherin in iCCA primary carcinoma, premalignant lesions, lymph node metastasis, and normal/
non-neoplastic adult tissue.
E-Cadherin N-Cadherin
Category Subcategory Cases Score H-Score High [%] † Score H-Score High [%] † Ki-67
Lymph node metastases iCCA small duct type 18 2.88 ± 0.28 232.59 ± 39.18 100 1.93 ± 0.98 64.06 ± 63.07 64.29 23.02 ± 14.96
iCCA large duct type 15 2.98 ± 0.06 206.04 ± 33.28 100 1.12 ± 0.85 41.78 ± 45.50 33.33 23.01 ± 11.51
pCCA 45 2.90 ± 0.32 226.83 ± 54.26 100 0.79 ± 0.63 36.78 ± 47.16 25.64 21.80 ± 17.08
GBC 34 2.86 ± 0.49 241.85 ± 39.49 96.97 0.47 ± 0.66 13.07 ± 37.30 6.45 26.80 ± 22.63
dCCA 13 3.00 ± 0 259.14 ± 15.95 100 0.73 ± 0.59 34.51 ± 47.32 8.33 24.79 ± 17.76
PDAC 67 2.75 ± 0.53 178.56 ± 62.63 97.01 0.20 ± 0.44 2.09 ± 7.54 3.13 n.a.
Primary carcinoma iCCA small duct type 196 2.90 ± 0.29 162.35± 76.25 98.96 1.76 ± 0.92 38.97 ± 53.32 67.01 16.06 ± 16.06
iCCA large duct type 116 2.85 ± 0.32 149.46 ± 72.36 99.13 1.30 ± 0.93 22.64 ± 38.24 50.34 19.73 ± 15.91
pCCA 328 2.94 ± 0.27 173.06 ± 62.25 99.05 0.63 ± 0.73 17.47 ± 33.88 17.57 17.81 ± 16.26
GBC 228 2.78 ± 0.47 137.94 ± 76.6 97.29 0.45 ± 0.67 7.17 ± 20.40 12.67 27.91 ± 18.18
dCCA 185 2.93 ± 0.19 155.78 ± 71.91 100 0.35 ± 0.59 5.67 ± 17.14 8.74 21.30 ± 17.75
PDAC 131 2.89 ± 0.24 177.36 ± 50.45 100 0.34 ± 0.44 3.02 ± 5.60 3.85 n.a.
Premalignant lesion BilIN 167 2.99 ± 0.06 240.74 ± 34.73 100 0.88 ± 0.65 34.22 ± 42.22 23.87 22.33 ± 17.07
PanIn 107 2.98 ± 0.24 218.49 ± 42.16 100 0.51 ± 0.79 17.58 ± 38.88 14.56 n.a.
Non-neoplastic tissue Small bile ducts 59 2.80 ± 0.34 165.88 ± 71.12 100 2.29 ± 0.59 88.32 ± 56.56 97.96 3.00 ± 5.48
Large bile ducts 42 2.97 ± 0.16 239.47 ± 36.2 100 1.03 ± 0.72 24.02 ± 29.17 34.15 5.49 ± 7.89
Perihilar bile ducts 65 3.00 ± 0 218.40 ± 50.39 100 0.98 ± 0.79 38.70 ± 41.70 33.33 5.82 ± 6.92
Gallbladder epithelium 148 3.00 ± 0.05 179.03 ± 54.39 100 0.47 ± 0.66 19.72 ± 40.23 12.84 8.64 ± 7.73
Pancreatic ducts 92 2.85 ± 0.34 188.45 ± 42.85 100 0.59 ± 0.70 50.58 ± 70.39 8.14 n.a.
† defined as an E-/N-cadherin score of ≥1.5. n.a.: not analyzed/data not available.
Cancers 2022, 14,  9  of  18 
 
tissue microarrays comprising 1184 patient samples (Table 1). Nearly every biliary tract 
cancer and PDAC, as well as their associated lymph node metastases, precursor lesions, 
and associated non‐neoplastic tissue, showed a high expression of E‐cadherin. However, 
N‐cadherin expression was strongest and most stable in small duct iCCA, and expression 
Cancers 2022, 14, 3091 declined  in  the  following order:  large duct  iCCA, BilIN, pCCA, GBC, dCCA, with10  tohfe1 8
lowest to no expression in PDAC (Figure 4). 
 
FFiigguurree 44. .S  tSatbalbelEe -cEa‐dchaderhienriann dadnidff edreifnfteiraelnNti-acla  dNh‐ecraidnheexrpinr esesxiopnreisnsicoanr ciinno mcaarcoifntohmeap aonfc rethateo -
pancreatobiliary  tract. With  immunohistochemistry,  high N‐cadherin  expression  is  detected  in 
biCilCiaAry (Atr,aBc)t .anWdi tphCiCmAm (uCn)o, hwishteorcehaes monisltyr yf,aihnitg hN‐Nca-cdahdehrienr ienxpexrepsrseisosnio ins disetdeectteecdt eind GinBiCC C(DA) (aAn,dB )
adnCdCpAC (CE)A; n(Co )N, w‐chaderheearsino nisl ydefateinctteNd -icna PdDheAriCn (eFx)p. Irnes csoionntriasstd, eEt‐eccatdedheirninG iBs Chi(gDh)lya nedxpdrCesCseAd (iEn) ;anllo 
Nca-rccaindohmerains oisf dtheete pctaendcrienaPtoDbAiliCar(yF )t.raInctc. oQnutrPaastth,  Ea-ncaaldyhsiesr ionni sthhei gbholtytoemxp rriegshste (dAi’n–Fa’l,lAc’a’–rcFi’n’)o. m(Aa’s’)o: f
tmhaenpuaanl csrceoarteo b3i, lNia‐rcyatdrhaectr.inQ HuP‐sactohrea n1a3l1y.2s4is; (oBn’’t)h: embaontutoaml scroigreh t2(.5A,’ N–F‐’c,aAd”h–eFr”i)n. (HA‐”s)c:omrea 8n0u.a4l0;s c(Cor’’e):3  ,
Nm-acnaudahle srcionrHe 2-s.5c,o Nre‐c1a3d1h.2e4r;in(B H”)‐:scmoaren u13a2l .s6c5o; r(eD2’’.5):, mNa-cnaudahl escrionreH 2-.s0c, oNre‐c8a0d.h40e;ri(nC H”)‐:smcoarneu 6a4l.1s8c;o (rEe’’2).:5  ,
Nm-acnaudahle srcinorHe -1s.c5o, rNe‐1c3a2d.h65e;ri(nD H”)‐:smcoarneu 4a6l.1sc7o; r(eF’2’).:0 ,mNa-ncuaadlh secroinreH 1-.0sc, oNre‐c6a4d.h18e;ri(nE ”H):‐smcoarneu 1a.l6s7c.o Creel1ls.5  ,
Nw-itcha dchoelorirnedH b-socrodreer4: 6w.1h7i;t(eF—”)s:tmroamnaula cleslclos,r eb1lu.0e,—Nt-ucmadohre 0ri,n yHel-lsocwor—e t1u.6m7o. rC e1l+l,s owriatnhgceo—lotruemd obro r2d+e, r:
wrehdi—tet—umstoror m3+a.l Bclealclsk, bbalur:e 5—0 tμumm o(Ar –0F, y,Ae’l–loFw’,A—’’t–uFm’’)o. r 1+, orange—tumor 2+, red—tumor 3+. Black
bar: 50 µm (A–F,A’–F’,A”–F”).
Consequently, N-cadherin immunohistochemistry distinguished small and large duct
  iCCAs from PDACs with a specificity of 96.30% and a sensitivity of 67.01% for small duct
iCCAs and a sensitivity of 50.43% for large duct iCCAs (each p < 0.01).
To test the practical use of N-cadherin for the differential diagnosis of iCCA to liver
metastases of PDAC in liver biopsies taken before therapy planning, we retrospectively
Cancers 2022, 14, x FOR PEER REVIEW  11  of  18 
 
Consequently,  N‐cadherin  immunohistochemistry  distinguished  small  and  large 
duct iCCAs from PDACs with a specificity of 96.30% and a sensitivity of 67.01% for small 
duct iCCAs and a sensitivity of 50.43% for large duct iCCAs (each p < 0.01). 
To test the practical use of N‐cadherin for the differential diagnosis of iCCA to liver 
Cancers 2022, 14, 3091 metastases of PDAC in liver biopsies taken before therapy planning, we1 r1eotfr1o8spectively 
analyzed  each  ten‐punch  biopsy  of  clinicopathologically  definite  iCCA  and  liver 
metastases  of  PDAC  and  long  clinical  follow‐up. Additionally,  in  punch  biopsies, N‐
acnaadlhyzeerdine apcrhotvened-p uton cbheb aio vpsayluoafbclien imcoapraktheor lotog idcailflfyerdenfitniiateteiC iCACAan adnlidv ePrDmAetCas.t aNsesvertheless, 
othf eP DaAnaClyansids lwonagsc clionmicapllfioclalotewd- ubpy.  Athded iftaiocnt atlhlya, ti nthpeu nocrhigbiinopasl iebsi,leN d-cuadcthse,r ians pwroevlel das residual 
thoebpeaatovcayluteasb leenmclaorskeedr  tiond tihffee rteunmtiaotre, isCtaCiAneadn dstPrDonAgCly. N peovseirttihveel efsosr, tNhe‐caandahlyesriisnw. as
complIincated by the fact that the original bile ducts, as well as residual hepatocytes enclosedin the tu maodrd, sittaioinne,d wstero  nagnlaylypzoseidti vtehfeo rkNn-ocawdnh epriann. creatobiliary  tract markers  CK7,  CK20, 
CA1I9n‐9ad, dEitMioAn,,w  aenadn aClyDzeXd2t hienk ncoowntnropal n(cTreaabtloebsi liSa2ry–Str6a)c.t mMaorskte rBs TCCK7s, CanKd20  ,PCDAA19C-s  showed 
9s,taEiMniAng, a rnedacCtiDoXn2s iangcaointsrto Cl (KT7ab alensdS E2–MS6A)., wMiotsht nBeTgCastiavnidtyP aDgAaCinssst hCoKw2e0d asntadin CinDgX2. CA19‐
r9e awctaiosn ws aegaakilnys teCxKp7reasnsdedE MinA s,mwaitlhl ndeugcatt iivCitCyAag aindst sChKo2w0eadnd sCtrDonXg2.erC Ast1a9in-9inwga sin PDACs. 
wHeoawkleyveexrp,r tehssee edxipnrsemssailol dnu wctaisC CofAteann dvasrhioawbleed. sNtrongee orfs ttahien iensgtainblPiDshAeCds .mHaorwkeervse ra, lone were 
tshuepeexrpiroers stioo nNw‐caasdohfteerninv airnia tbhlee. dNioffneereonf tthiaeteiostna bolfis ihCedCAm aarnkedr sPaDloAneCw.  ere superior to
N-cadThherei neixnptrheesdsiioffner epnatitatetironn ooff iCEC‐ Aanadnd NP‐DcAadCh. erin was  reproduced  in  the  immunoblot 
The expression pattern of E- and N-cadherin was reproduced in the immunoblot
aannaallyyssisiso of wf whohleolteis stiusesulyes altyessaotef ss ixofc rsyixo pcrreysoerpvreedseiCrvCeAds i(CthCreAese a(tchhroefet heeacshm aolfl athned small and 
llaarrggeed duuctctty tpyeps,easn, dantwd otweaoc heaocfhp CoCf ApC, GCBAC,, GanBdCP,D aAndC )P(FDigAuCre) 5()F.igure 5). 
 
Fiigguurree5 .5S. eSmemiquiqaunatintattiitvaetiivmem imunmobulnotoabnloalty asnisaslhyoswiss sehqouwalsl eevqeulsaol fleEv- ealnsd oNf -Eca‐ dahnedri nNi‐ncaiCdChAerin in iCCA 
anddp pCCCAA, w, hwehreearsePaDs APCDdAisCp ldayisspplraeydso mprinedanotmE-icnaadnhte rEin‐ceaxdphreesrsiinon e.x(pAr)e: Ismsimonun. (oAbl)o: tIamnamlyusnisoobf lot analysis 
wofh owlehtoislseu teislyssuaete lsyosfaitCeCs Aof( simCaCllAan (dsmlaraglel danucdt ilCaCrgAe (dbauncdt si1C–C3Aan d(b4a–n6)d),sp 1C–C3A a(nbdan 4d–s67))a, npdC8C),A (bands 7 
GanBdC 8(b),a GndBsC9 (abnadn1d0s) ,9a annddP D10A),C a(nbdan PdDsA11Ca n(bda1n2d)sw 1a1s apnrodb 1ed2)w witahs apnrtoibboeddie ws aitgha iannsttiEb-oadnides against E‐ 
Nan-cda dNh‐ecraind.hOefrninot.e O, wf hnolteet,i swsuheollyes atitsessuwee lryesNa-tceasd wheerrine- pNo‐scitaidvehienrsintr‐opmoaslictievlles ain dstvreosmsealsl. cEeqlulsa land vessels. 
aEmqouuanl tasmofopurnottesi nowf perreoltoeaidne dwaenred eloqaudiliebdra atenddw eiqthuailcitbinr.aMteodl ewcuitlahr amcatsins .m Marokleercsualraerd mepaicstse dmarkers are 
odnepthiectreigdh  tosnid  eth. Teh  reiugnhctr ospidpeed.  WTheset eurnnbcrlootpspareeds hWowenstienrnFi gbulroetSs 4.aGrer asphhoicwalnr epinre  sFeingtuatrieo nSo4f.  Graphical 
trheeppreixseelndteantisoitny oanf atlhyes ips ioxfetlh deerensspietyct iavneablyansdiss o(Bf )thaen dretshpeerecstipveec tbivaentdusm (oBr)( aCn).d the respective tumor (C). 
 
Cancers 2022, 14, 3091 12 of 18
In whole tissue lysates of iCCA, pCCA, and GBC, E- and N-cadherin were detected
in almost equal amounts at about 120 and 130 kDa, respectively. In accordance with our
immunohistochemical data, in PDAC, minor amounts of N-cadherin were still detected,
possibly due to the N-cadherin in the stromal cells and vessels included in the whole
tissue lysates.
To further substantiate the expression pattern of N-cadherin in PDAC, we used cul-
tured PDAC cells of the lines Capan-1 and Panc-1. Again, large amounts of E-cadherin
but only small amounts of N-cadherin were detected by immunoblot. N-cadherin staining
was observed on membranes of only individual cells by immunofluorescence microscopy,
whereas the vast majority of Capan-1 and Panc-1 cells were negative.
In summary, the expression pattern of E- and N-cadherin in pancreatobiliary tumors re-
capitulates their expression along the normal bile duct tree and pancreatic ducts. Therefore,
N-cadherin-positivity may be able to distinguish iCCA from PDAC.
3.3. Expression of E- and N-Cadherin in BTC Is Retained in Respective Metastases
Since E- and N-cadherin have generally been implicated in the process of epithelial-
to-mesenchymal transition (EMT) during the invasion and metastases of carcinomas, we
compared the respective primary carcinomas with their lymph node and foreign metas-
tases and evaluated the intensity of staining by means of a semiquantitative, automated
analysis (QuPath). The mean E-cadherin H-scores of primary BTC were significantly higher
compared to their lymph node metastases, while in PDAC, we found no respective differ-
ence. GBC showed the highest effect (p < 0.01; g = 1.426), decreasing from dCCA (p < 0.01,
r = 0.33) to iCCA of the small (p < 0.01, r = 0.25) and large duct type (p < 0.01, r = 0.23).
The mean E-cadherin H-score in the normal tissue was higher than in their respective
primary carcinomas, arguing for the downregulation of E-cadherin during carcinogenesis.
We found significant effects in GBC (p < 0.01; g = 0.585), dCCA (p < 0.01; g = 1.245), iCCA
of the large duct type (p < 0.01; g = 1.376), and pCCA (p < 0.01; g = 0.748). No significant
differences were found for iCCA of the small duct type and PDAC and their respective
normal tissues. The mean N-cadherin H-scores of dCCA (p < 0.01, r = 0.29) and pCCA
(p < 0.01, g = 0.542) were higher in lymph node metastases when compared to their respec-
tive primary carcinomas. iCCAs of the large and small duct types, GBC, and PDAC, as
well as their associated lymph node metastases, did not show a significant difference. The
mean N-cadherin H-score was higher in normal tissue than in associated carcinomas, again
arguing for the downregulation of N-cadherin during carcinogenesis. The effect size was
highest in PDAC (p < 0.01; g = 1.071), and it decreased from dCCA (p < 0.01; g = 0.924)
to iCCA of the small duct type (p < 0.01; g = 0.912), pCCA (p < 0.01; g = 0.602), and GBC
(p < 0.01; g = 0.439). There was no significant difference for iCCA of the large duct type.
Both E- and N-cadherin expression was thereby downregulated during the late stages of
carcinogenesis, such as invasion and metastasis formation.
3.4. Significance of E- and N-Cadherin for the Prognosis of Patients with Carcinomas of the
Pancreatobiliary System
To investigate the prognostic impact of E- and N-cadherin in intrahepatic cholangio-
carcinoma, we used the TCGA cohort as an independent cohort (n = 36). In the TCGA
cohort, low mRNA expression of both CDH1 (E-cadherin) and CDH2 (N-cadherin) was
significantly associated with unfavorable outcomes. In addition, an increase in nodal
status, T- and UICC-stage, and R-status, as well as vascular invasion, was associated with
worse survival.
In all tumors examined in our patient cohort, there was no clear trend between
the grading or the Ki-67 proliferation index and the E-cadherin expression (Figure S2),
while higher N-cadherin H-scores were weakly correlated with a higher proliferation rate
(r = 0.207, p < 0.01). Concerning cadherin expression in BTC, patients with lower E-cadherin
H-scores showed a higher T-stage (ρ = − 0.143, p < 0.01). Nevertheless, no correlation
was detected between E-cadherin and N-cadherin H-scores and lymphangioinvasion,
Cancers 2022, 14, 3091 13 of 18
vascular or perineural invasion, N-stage, grading, and tumor size; there was also no
correlation between E-cadherin H-score and proliferation (Ki-67-score) and N-cadherin
H-score and T-stage.
In iCCA, the E-cadherin H-score was not associated with a significant difference
in recurrence-free survival. In contrast, it was shown that a low N-cadherin score was
associated with a poorer prognosis (Figure S3; concerning patient details, see Table S1).
4. Discussion
In this study, we demonstrate that E- and N-cadherin are constitutively expressed in
cholangiocytes under physiological conditions and retained in BilIN and iCCA in a large
clinicopathologic cohort of patients. iCCA and the respective normal bile ducts have already
been shown to be positive for N-cadherin, especially in small duct iCCAs [34,35,45,46].
Nevertheless, N-cadherin is only present in minor amounts in physiological extrahepatic
bile ducts, as well as in dCCA and pCCA, and is negative or only faintly expressed in
physiological pancreatic ducts as well as in PDAC. Thereby, N-cadherin expression in
the physiological biliary tract and associated BTC decreases sequentially on a continuous
spectrum from small bile ducts and small duct iCCAs to larger and extrahepatic bile
ducts and their respective carcinomas. E- and N-cadherin expression in pancreatobiliary
carcinomas recapitulates their expression in the respective physiological epithelia of the
pancreatobiliary tree, as already illustrated for normal hepatocytes as well as hepatocellular
adenoma and carcinoma [32]. Coexpression of E- and N-cadherin may thereby constitute a
characteristic of epithelia and derived tumors of the hepatobiliary system. N-cadherin is
not restricted to mesenchymal and neuroectodermal cells, as postulated by [23–25], but is
also constitutively expressed in special endoderm-derived epithelia.
The characteristic expression pattern of N-cadherin in iCCA and its absence in PDAC
allow for an immunohistochemical differentiation of these entities. Only one study has
previously investigated the usefulness of N-cadherin to discriminate between metastatic
PDAC and iCCA—an indicative study with only 31 cholangiocarcinomas [33]. In our
study, we obtained comprehensive results in a large and well-characterized clinicopatho-
logic collection of 312 iCCAs as well as 131 PDACs, supporting the use of N-cadherin
in this differential diagnosis. Coexpression of E- and N-cadherin may help differentiate
the primary liver carcinomas HCC and iCCA from liver metastases originating from an
extrahepatic primary tumor. So far, N-cadherin has not been demonstrated in the vast
majority of extrahepatic malignancies apart from EMT. Own preliminary results in this
respect demonstrated the absence of N-cadherin in most gastrointestinal, tuboovarial,
mammary, and lung adenocarcinomas, as well as urothelial carcinomas, so N-cadherin
may even help in the differential diagnosis of primary liver cancer from liver metastases
of adenocarcinomas of unknown primary tumors. Further studies are underway in this
respect. In addition to the N-cadherin analyzed in the frame of this study, albumin RNA
in situ hybridization (albumin ISH) may be used in the differential diagnosis between
iCCA and metastatic PDAC. However, much like N-cadherin, only in the case of positive
expression can albumin ISH guarantee the diagnosis of iCCA [47], so both methods may be
used complementarily. Especially in the case of large duct iCCAs, this may hold true as
they are positive for albumin RNA ISH in only 18% of cases [48], whereas N-cadherin has a
positive rate of 50%.
Of note, loss or downregulation of E-cadherin together with de novo expression or
upregulation of N-cadherin has been implicated in the process of EMT in different types of
carcinoma and has been correlated to an unfavorable prognosis (the so-called “cadherin
switch”) [49,50]. However, the roles of E- and N-cadherin are much more diverse and
versatile, which is evident by their controversial roles in different tumors. For example, low
expression of N-cadherin is correlated with metastatic dissemination in neuroblastoma [51],
and aberrantly high expression of E-cadherin is a hallmark of ovarian carcinoma [52].
Remarkably, the inverse phenomenon of a mesenchymal–epithelial transition (so called
MET) has also been described during dedifferentiation in an example of a hematopoietic
Cancers 2022, 14, 3091 14 of 18
neoplasm [53]. Many more contradictory and context-dependent roles of different cad-
herins have been described [54]. Concerning primary liver carcinoma, besides iCCA, HCC
also seems to be an exception to the general assumption of EMT as hepatocytes and bile
duct epithelial cells contain high amounts of N-cadherin physiologically. Therefore, it is
not astonishing to find that in our study, both N- and E-cadherin expression in BTC and
PDAC was lower than in the associated normal tissues. Only in associated lymph node
metastases was a trend to higher expression of N-cadherin noted; yet, so far, this is of
unclear significance. Downregulation of N-cadherin, but not E-cadherin, was significantly
associated with poorer outcomes. In our analyses of the independent TCGA cohort, low
E-and N-cadherin mRNA levels were both associated with poor survival. Mechanistically,
alterations in E-cadherin have been shown to lead to the activation of Wnt signaling and
the Hippo signaling pathway. In addition, E-cadherin may interact with epidermal growth
factors, the dysregulation resulting in dysmorphogenesis and altered gene expression [55].
In a study by Techasen et al., invasion and migration of cholangiocarcinoma cell lines
were increased by siRNA-mediated reduction of E-cadherin levels with the upregulation of
vimentin. Furthermore, low immunohistochemical E-cadherin expression was significantly
associated with lymph node metastasis. There was no significant association with patient
age, sex, histology, and iCCA subtype. [56]. In another study, immunohistochemical ex-
amination of iCCA showed a significantly lower expression of E-cadherin than in adjacent
tissue and normal bile duct tissue, confirming our results [57]. Reduction of immunohis-
tochemical E-cadherin expression in iCCA compared to normal tissue has been reported
in many studies, often correlating with a higher tumor grade [58–62]. We, therefore, sug-
gest that the downregulation of E- and N-cadherins in BTC is a general phenomenon of
dedifferentiation, as it is also commonly observed for various other immunohistochemical
marker proteins.
In the frame of this study, we could successfully utilize digital image analysis of E-
and N-cadherin as membrane-bound markers as it may also prove helpful in the scoring
of other immunohistochemical stains. When compared, eyeballing estimation and digital
analyses were highly concordant when the digital analysis was adequately trained. Our
approach uses manual analysis to assess the expression level and digital analysis to assess
the quantitative staining reaction. This should provide a comprehensive assessment and
the future use of a manual evaluation in daily routine. Other frequently used membrane
biomarkers with relevance for pathology include Her2/neu in breast and stomach cancer,
as well as the various cluster of differentiation (CD) antigens in hematopathology. A com-
prehensive study on punch biopsies for Her2/neu has already shown that manual analysis
and digital analysis achieve high concordance rates [63].
To summarize, the expression of N-cadherin in BTC is a characteristic of its biliary
tract differentiation and may not reflect EMT.
5. Conclusions
In conclusion, in BTC, E- and N-cadherin are expressed in a tissue-specific manner.
N-cadherin is a highly specific marker to differentiate iCCA from liver metastases of PDAC.
High expression of N-cadherin in adenocarcinomas of pancreatobiliary morphology thereby
suggests the diagnosis of iCCA and not PDAC. Moreover, the differential expression of
CA19-9 and CK7 versus CDX-2 and CK20 may be used to confirm progeny from the
gastrointestinal tract and help diagnose iCCA (Figure 6). We recommend N-cadherin
immunohistochemical staining to distinguish biliary origin rather than pancreatic origin.
Cancers 2022, 14, x FOR PEER REVIEW  15  of  18 
  Cancers 2022, 14, 3091 15 of 18
 
FiguFrieg u6r.e S6c. hSecmheamtiact icimimmmuunnoohhiissttochemiissttrryyw worokrflkofwlowfo rfothre  tdhieff edriefnfetiraelndtiiaagln doisaisgonfomsiest aosft amticetastatic 
anda pnrdimpraimrya raydaedneoncoacracricninoommaass of the lliviveer.r.N N-c‐acdahdehrienrimn amy abye ubsee dustoed itffoe rdeinftfieartenthtieatoer igthine oofrigin of 
CK7C+K/C7+K/2C0K− 2a0n−da nCdDCXD2X‐2/C-/AC1A91‐9- 9−−/+/ +ppaannccrreeaattobiliarryya daedneoncoacrcairncoimnoams. aBsC. : BbrCea: sbt craeracsitn ocmarac. inoma. 
CRCC:R cCo:locorelocrteaclt aclacracricnionommaa. .EESSCC:: esopphhaaggeaelaslq suqaumaomuosucasr cianrocminao.mGCa.A G: gCaAst:r igcacsatrrciicn ocamrac.inHoCmCa: . HCC: 
hepahteopcaetollcuellalurl acracracrincionmomaa. .MMCCCC:: Merkeell ceclellcl acracirncoinmoam. Na.E NC:EnCe:u  rnoeeunrdooecnridnoeccrairncein ocmarac.inNoEmTa: . NET: 
neurnoeeunrodeoncdroincrei nteutmumoro. r.NNSSCCLLCC:: non-‐ssmaallllc eclellllu lnugncga nccaenr.cePrC. AP:CpAro: sptartoesctantece cr.aRnCceCr:. rRenCaCl c: erlelnal cell 
carccinarocminoam. SaC. SLCCL:C s:msmalall lcecelll llluung ccaancceer.r.U UCCAA: u: ruotrhoetlhiaellicaalr ccianrocminao. ma. 
Supplementary  Materials:  The  following  supporting  information  can  be  downloaded  at: 
Supplementary Materials: The following supporting information can be downloaded at: https:
ww/w/.wmwdwp.im.cdopmi.c/xoxmx//asr1t,i cFleig/u10r.e3 3S910:/ Nca‐nccaedrsh1e4r1i3n3‐0p9o1/sist1i,vFei gaupriecaSl1 :sNec-rceatdihoenrsin o-pf opsaitnivcereaaptiicca ldsueccrtes-; Figure 
S2: Eti‐o nasnodf Npa‐nccardehateicridnu ectxsp; FreigsusiroenS 2i:nE c-aarncdinNo-mcaadsh eorfi ntheex ppraesnsciorenaitnocbairlciianroym sayssotefmth;e Fpiagnucrreea Sto3b: iSl-urvival 
analiyarsyiss yosft eEm‐ ;aFnidgu Nre‐Sca3:dShuerrviinv aplraontaeliyns iesxopfrEe-sasniodnN i-nc aidChCerAins porfo oteuinr ecxophroesrst ioinn icnomiCpCaArsisoofno utor  CDH1 
andc 2o hmorRtNinAco emxparreissosinotno iCnD tHhe1 iannde2pmenRdNeAnte xTpCreGssAio cnoihnothrte; iFnidgeupreen dSe4n: tthTeC GuAnccroohpoprte;dF iWguerestSe4r:n blots; 
Tablteh eSu1n: cproaptipeendt Wcheastrearcntebrloistst;icTsa bolfe iSC1:CpAat ifeonrt cthaer ascutrevriisvtiacls aonf iaClyCsAisf o(rowthen scuorvhiovratl)a, nsaelpyasirsa(toewdn by high 
andc lohwor Nt),‐sceapdahraetreidn bHy‐hscigohreasn; dTlaobwleN S-2c:a dSuhemrimn Har-ysc ofr eesv; aTalubaletiSo2n:  Souf mCmKa7r ystoafineivnaglu; aTtiaobnleo fSC3:K o7f CA19‐
9 stasitnaiinign;g T; Taabblele SS43:: off ECMA1A9- 9stsatianininingg; ;TTablle S45:: ooffE CMDAXs2t asitnaiinngi;nTga;b alenSd5 T: oafbCleD SX62: sotfa iCnKin2g0;  asntadining in 
the cToabhloerSt6. : of CK20 staining in the cohort.
AutAhuotrh CoroCnotrnitbriubtuiotinons:s :DDaatata ccuurraattiioonn,, TT.S.S.G.G.,.B, .BG.G., .A, .AH..H, H.,. RH.W.R..,WF.B.,. ,FM.B..S, .,ML..S-K.,. GL..,‐KM..GM.,.G M. a.Mnd.G., and 
B.K.BS..K; .FS.o;rFmoraml aalnaanlaylyssisis,, TT..SS..GG..a annddB .BK..SK.;.SF.u; nFduinngdaincqgu aiscitqiounis, iBt.iKo.nS,. ;BIn.Kve.Sst.i;g Iantivoens, tTi.gSa.Gti.o, On,.C T. .aSn.Gd ., O.C., 
P.S.; Methodology, T.S.G. and B.K.S.; Project administration, B.K.S.; Software, T.S.G.; Supervision, I.E.,
andP  .RP..GS..,; WM.Re.,thHo.Ld.oalnodgBy.,K  .TS..;SV.Gal.i daatniodn , TB.S.K.G.S.,.D;  .AP.rRo.jeacntd  Ba.dKm.S.i;nViissturaaltiizoanti,o nB, T.K.S..SG..;, HS.oRf.tWw.,aMre.,S . T.S.G.; 
SupaenrvdiBsi.Kon.S,. ;I.WEr.,i tPin.Rg—.Go.r, iWgin.Ral.,d Hra.fLt,.,T a.Sn.Gd. Ba.nKd.SB..;K V.Sa.;liWdraittiinogn—, Tre.Sv.iGew., Dan.dAe.Rdi.t,i anng,dT B.S..KG..Sa.n; dVBis.Kua.Sl.ization, 
T.S.GAl.l, aHut.Rho.Wrs .h, aMve.Sre.,a danadnd Ba.gKr.eSe.d; Wto rthiteinpgu—blioshriegdinvearls  idornaoftf, thTe.Sm.Gan. uasncdri pBt..K.S.; Writing—review and 
editing,  T.S.G.  and  B.K.S.  All  authors  have  read  and  agreed  to  the  published  version  of  the 
manFuusncdriinpgt.: T.S.G. and D.A.R. are supported by the Clinician Scientist Fellowship “Else Kröner ResearchCollege: 2018_Kolleg.05”. In addition, D.A.R. was supported by the Level I Program of University
FunMdiendgic:a  lTC.Se.nGte. raMnadin Dz. .A.R.  are  supported  by  the  Clinician  Scientist  Fellowship  “Else  Kröner 
Research College: 2018_Kolleg.05”. In addition, D.A.R. was supported by the Level I Program of 
University Medical Center Mainz. 
Institutional Review Board  Statement: All  procedures  performed  in  studies  involving  human 
tissue  were  in  accordance  with  the  ethical  standards  of  the  institutional  and  regional  ethics 
committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical 
standards.  The  study was  approved  by  the  ethics  committee  of  the  State Medical  Council  of 
Rhineland‐Palatinate  (837.280.17  (11114)  and  2021‐15819)  and  by  the  ethics  committee  of  the 
University of Heidelberg (S‐206/2005, S‐207/2005, and S‐519/2019). 
 
Cancers 2022, 14, 3091 16 of 18
Institutional Review Board Statement: All procedures performed in studies involving human tissue
were in accordance with the ethical standards of the institutional and regional ethics committee and
with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The
study was approved by the ethics committee of the State Medical Council of Rhineland-Palatinate
(837.280.17 (11114) and 2021-15819) and by the ethics committee of the University of Heidelberg
(S-206/2005, S-207/2005, and S-519/2019).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: The data are not publicly available due to the data containing informa-
tion that could compromise the privacy of the research participants. All data included in this study
are available upon request from the corresponding author.
Acknowledgments: First: we want to thank Werner W. Franke for his help in the initiation of the
study and for fruitful discussions. Tissue samples were provided by the tissue bank of University
Medical Center Mainz as well as BioMaterialBank Heidelberg (BMBH) in accordance with the
regulations of the respective tissue banks. We thank Bonny Adami and Elisabeth Specht for their
excellent technical assistance. We especially thank the Else-Kröner-Fresenius-Research College “From
chronic inflammation to carcinogenesis: importance of the liver micromilieu for development and
progression of liver tumors” Mainz for their general support and Ivonne Dietzel for her help in
administrative organization. In 2021, the study was attributed a poster prize of Virtual Pathology
Days (ViPa) of the German Society of Pathology (DGP) to T.S.G. and B.K.S.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Sung, H.; Ferlay, J.; Siegel, R.L.; Laversanne, M.; Soerjomataram, I.; Jemal, A.; Bray, F. Global Cancer Statistics 2020: GLOBOCAN
Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J. Clin. 2021, 71, 209–249. [CrossRef]
[PubMed]
2. Global Burden of Disease Cancer Collaboration. Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost,
Years Lived with Disability, and Disability-Adjusted Life-Years for 29 Cancer Groups, 1990 to 2017: A Systematic Analysis for the
Global Burden of Disease Study. JAMA Oncol. 2019, 5, 1749–1768. [CrossRef]
3. Hidalgo, M.; Cascinu, S.; Kleeff, J.; Labianca, R.; Löhr, J.M.; Neoptolemos, J.; Real, F.X.; Van Laethem, J.L.; Heinemann,
V. Addressing the challenges of pancreatic cancer: Future directions for improving outcomes. Pancreatology 2015, 15, 8–18.
[CrossRef] [PubMed]
4. Rizvi, S.; Khan, S.A.; Hallemeier, C.L.; Kelley, R.K.; Gores, G.J. Cholangiocarcinoma—Evolving concepts and therapeutic strategies.
Nat. Rev. Clin. Oncol. 2018, 15, 95–111. [CrossRef] [PubMed]
5. Kelley, R.K.; Bridgewater, J.; Gores, G.J.; Zhu, A.X. Systemic therapies for intrahepatic cholangiocarcinoma. J. Hepatol. 2020,
72, 353–363. [CrossRef] [PubMed]
6. Schmuck, R.B.; de Carvalho-Fischer, C.V.; Neumann, C.; Pratschke, J.; Bahra, M. Distal bile duct carcinomas and pancreatic ductal
adenocarcinomas: Postulating a common tumor entity. Cancer Med. 2016, 5, 88–99. [CrossRef]
7. Lok, T.; Chen, L.; Lin, F.; Wang, H. Immunohistochemical distinction between intrahepatic cholangiocarcinoma and pancreatic
ductal adenocarcinoma. Hum. Pathol. 2014, 45, 394–400. [CrossRef]
8. WHO Classification of Tumours Editorial Board. WHO Classification of Tumours. Digestive System Tumours: WHO Classification of
Tumours; World Health Organization: Geneva, Switzerland, 2019; Volume 1.
9. Nagtegaal, I.D.; Odze, R.D.; Klimstra, D.; Paradis, V.; Rugge, M.; Schirmacher, P.; Washington, K.M.; Carneiro, F.; Cree, I.A. The
2019 WHO classification of tumours of the digestive system. Histopathology 2020, 76, 182–188. [CrossRef]
10. Banales, J.M.; Cardinale, V.; Carpino, G.; Marzioni, M.; Andersen, J.B.; Invernizzi, P.; Lind, G.E.; Folseraas, T.; Forbes, S.J.;
Fouassier, L.; et al. Cholangiocarcinoma: Current knowledge and future perspectives consensus statement from the European
Network for the Study of Cholangiocarcinoma (ENS-CCA). Nat. Rev. Gastroenterol. Hepatol. 2016, 13, 261–280. [CrossRef]
11. Zaccari, P.; Cardinale, V.; Severi, C.; Pedica, F.; Carpino, G.; Gaudio, E.; Doglioni, C.; Petrone, M.C.; Alvaro, D.;
Arcidiacono, P.G.; et al. Common features between neoplastic and preneoplastic lesions of the biliary tract and the pan-
creas. World J. Gastroenterol. 2019, 25, 4343–4359. [CrossRef]
12. Chu, P.G.; Schwarz, R.E.; Lau, S.K.; Yen, Y.; Weiss, L.M. Immunohistochemical Staining in the Diagnosis of Pancreatobiliary
and Ampulla of Vater Adenocarcinoma: Application of CDX2, CK17, MUC1, and MUC2. Am. J. Surg. Pathol. 2005, 29, 359–367.
[CrossRef] [PubMed]
13. Rudini, N.; Dejana, E. Adherens junctions. Curr. Biol. 2008, 18, R1080–R1082. [CrossRef] [PubMed]
14. Vasioukhin, V. Adherens junctions and cancer. Subcell. Biochem. 2012, 60, 379–414. [CrossRef] [PubMed]
15. Brasch, J.; Harrison, O.J.; Honig, B.; Shapiro, L. Thinking outside the cell: How cadherins drive adhesion. Trends Cell Biol. 2012,
22, 299–310. [CrossRef]
Cancers 2022, 14, 3091 17 of 18
16. Nelson, W.J. Regulation of cell-cell adhesion by the cadherin-catenin complex. Biochem. Soc. Trans. 2008, 36, 149–155. [CrossRef]
17. Hiroki, O. Evolution of the cadherin-catenin complex. Subcell. Biochem. 2012, 60, 9–35. [CrossRef]
18. Stemmler, M.P. Cadherins in development and cancer. Mol. Biosyst. 2008, 4, 835–850. [CrossRef]
19. Kourtidis, A.; Lu, R.; Pence, L.J.; Anastasiadis, P.Z. A central role for cadherin signaling in cancer. Exp. Cell Res. 2017, 358, 78–85.
[CrossRef]
20. Loh, C.Y.; Chai, J.Y.; Tang, T.F.; Wong, W.F.; Sethi, G.; Shanmugam, M.K.; Chong, P.P.; Looi, C.Y. The E-Cadherin and N-Cadherin
Switch in Epithelial-to-Mesenchymal Transition: Signaling, Therapeutic Implications, and Challenges. Cells 2019, 8, 1118.
[CrossRef]
21. Pagano, R.E.; Takeichi, M. Adhesion of phospholipid vesicles to Chinese hamster fibroblasts. Role of cell surface proteins. J. Cell
Biol. 1977, 74, 531–546. [CrossRef]
22. Volk, T.; Geiger, B. A 135-kd membrane protein of intercellular adherens junctions. EMBO J. 1984, 3, 2249–2260. [CrossRef]
[PubMed]
23. Halbleib, J.M.; Nelson, W.J. Cadherins in development: Cell adhesion, sorting, and tissue morphogenesis. Genes Dev. 2006,
20, 3199–3214. [CrossRef] [PubMed]
24. Gumbiner, B.M. Regulation of cadherin-mediated adhesion in morphogenesis. Nat. Rev. Mol. Cell Biol. 2005, 6, 622–634.
[CrossRef] [PubMed]
25. Takeichi, M. The cadherins: Cell-cell adhesion molecules controlling animal morphogenesis. Development 1988, 102, 639–655.
[CrossRef]
26. Riethmacher, D.; Brinkmann, V.; Birchmeier, C. A targeted mutation in the mouse E-cadherin gene results in defective preimplan-
tation development. Proc. Natl. Acad. Sci. USA 1995, 92, 855–859. [CrossRef]
27. Radice, G.L.; Rayburn, H.; Matsunami, H.; Knudsen, K.A.; Takeichi, M.; Hynes, R.O. Developmental defects in mouse embryos
lacking N-cadherin. Dev. Biol. 1997, 181, 64–78. [CrossRef]
28. Peralta Soler, A.; Knudsen, K.A.; Jaurand, M.C.; Johnson, K.R.; Wheelock, M.J.; Klein-Szanto, A.J.; Salazar, H. The differential
expression of N-cadherin and E-cadherin distinguishes pleural mesotheliomas from lung adenocarcinomas. Hum. Pathol. 1995,
26, 1363–1369. [CrossRef]
29. Ordóñez, N.G. Value of E-cadherin and N-cadherin immunostaining in the diagnosis of mesothelioma. Hum. Pathol. 2003,
34, 749–755. [CrossRef]
30. Moll, R.; Mitze, M.; Frixen, U.H.; Birchmeier, W. Differential loss of E-cadherin expression in infiltrating ductal and lobular breast
carcinomas. Am. J. Pathol. 1993, 143, 1731–1742.
31. Oda, T.; Kanai, Y.; Oyama, T.; Yoshiura, K.; Shimoyama, Y.; Birchmeier, W.; Sugimura, T.; Hirohashi, S. E-cadherin gene mutations
in human gastric carcinoma cell lines. Proc. Natl. Acad. Sci. USA 1994, 91, 1858–1862. [CrossRef]
32. Straub, B.K.; Rickelt, S.; Zimbelmann, R.; Grund, C.; Kuhn, C.; Iken, M.; Ott, M.; Schirmacher, P.; Franke, W.W. E-N-cadherin
heterodimers define novel adherens junctions connecting endoderm-derived cells. J. Cell Biol. 2011, 195, 873–887. [CrossRef]
[PubMed]
33. Hooper, J.E.; Morgan, T.K.; Grompe, M.; Sheppard, B.C.; Troxell, M.L.; Corless, C.L.; Streeter, P.R. The novel monoclonal antibody
HPC2 and N-cadherin distinguish pancreatic ductal adenocarcinoma from cholangiocarcinoma. Hum. Pathol. 2012, 43, 1583–1589.
[CrossRef] [PubMed]
34. Akita, M.; Sawada, R.; Komatsu, M.; Suleman, N.; Itoh, T.; Ajiki, T.; Heaton, N.; Fukumoto, T.; Zen, Y. An immunostaining panel
of C-reactive protein, N-cadherin, and S100 calcium binding protein P is useful for intrahepatic cholangiocarcinoma subtyping.
Hum. Pathol. 2021, 109, 45–52. [CrossRef] [PubMed]
35. Mosnier, J.F.; Kandel, C.; Cazals-Hatem, D.; Bou-Hanna, C.; Gournay, J.; Jarry, A.; Laboisse, C.L. N-cadherin serves as diagnostic
biomarker in intrahepatic and perihilar cholangiocarcinomas. Mod. Pathol. 2009, 22, 182–190. [CrossRef] [PubMed]
36. Gerber, T.S.; Bartsch, F.; Wagner, D.C.; Schindeldecker, M.; Heuft, L.-K.; Roth, W.; Lang, H.; Straub, B.K. Clinicopathological
Significance of Syndecan-1 in Cholangiocarcinoma: A Study Based on Immunohistochemistry and Public Sequencing Data. J.
Clin. Med. 2021, 10, 2745. [CrossRef]
37. Bankhead, P.; Loughrey, M.B.; Fernández, J.A.; Dombrowski, Y.; McArt, D.G.; Dunne, P.D.; McQuaid, S.; Gray, R.T.; Murray, L.J.;
Coleman, H.G.; et al. QuPath: Open source software for digital pathology image analysis. Sci. Rep. 2017, 7, 16878. [CrossRef]
38. McCarty, K.S., Jr.; Szabo, E.; Flowers, J.L.; Cox, E.B.; Leight, G.S.; Miller, L.; Konrath, J.; Soper, J.T.; Budwit, D.A.;
Creasman, W.T.; et al. Use of a monoclonal anti-estrogen receptor antibody in the immunohistochemical evaluation of
human tumors. Cancer Res. 1986, 46, 4244s–4248s.
39. Kwak, S.G.; Kim, J.H. Central limit theorem: The cornerstone of modern statistics. Korean J. Anesthesiol. 2017, 70, 144–156.
[CrossRef]
40. Fritz, C.O.; Morris, P.E.; Richler, J.J. Effect size estimates: Current use, calculations, and interpretation. J. Exp. Psychol. Gen. 2012,
141, 2–18. [CrossRef]
41. Chan, Y.H. Biostatistics 104: Correlational analysis. Singap. Med. J. 2003, 44, 614–619.
42. Farshidfar, F.; Zheng, S.; Gingras, M.C.; Newton, Y.; Shih, J.; Robertson, A.G.; Hinoue, T.; Hoadley, K.A.; Gibb, E.A.; Roszik, J.; et al.
Integrative Genomic Analysis of Cholangiocarcinoma Identifies Distinct IDH-Mutant Molecular Profiles. Cell Rep. 2017,
18, 2780–2794. [CrossRef] [PubMed]
Cancers 2022, 14, 3091 18 of 18
43. Budczies, J.; Klauschen, F.; Sinn, B.V.; Győrffy, B.; Schmitt, W.D.; Darb-Esfahani, S.; Denkert, C. Cutoff Finder: A comprehensive
and straightforward Web application enabling rapid biomarker cutoff optimization. PLoS ONE 2012, 7, e51862. [CrossRef]
[PubMed]
44. Kim, S.; Lee, W. Does McNemar’s test compare the sensitivities and specificities of two diagnostic tests? Stat. Methods Med. Res.
2017, 26, 142–154. [CrossRef] [PubMed]
45. Yu, T.H.; Yuan, R.H.; Chen, Y.L.; Yang, W.C.; Hsu, H.C.; Jeng, Y.M. Viral hepatitis is associated with intrahepatic cholangiocarci-
noma with cholangiolar differentiation and N-cadherin expression. Mod. Pathol. 2011, 24, 810–819. [CrossRef] [PubMed]
46. Yeh, Y.C.; Lei, H.J.; Chen, M.H.; Ho, H.L.; Chiu, L.Y.; Li, C.P.; Wang, Y.C. C-Reactive Protein (CRP) is a Promising Diagnostic
Immunohistochemical Marker for Intrahepatic Cholangiocarcinoma and Is Associated with Better Prognosis. Am. J. Surg. Pathol.
2017, 41, 1630–1641. [CrossRef]
47. Brackett, D.G.; Neyaz, A.; Arora, K.; Masia, R.; Mattia, A.; Zukerberg, L.; Misdraji, J.; Goyal, L.; Zhu, A.X.; Ferrone, C.R.; et al.
Cholangiolar pattern and albumin in situ hybridisation enable a diagnosis of intrahepatic cholangiocarcinoma. J. Clin. Pathol.
2020, 73, 23–29. [CrossRef]
48. Sigel, C.S.; Drill, E.; Zhou, Y.; Basturk, O.; Askan, G.; Pak, L.M.; Vakiani, E.; Wang, T.; Boerner, T.; Do, R.K.G.; et al. Intrahepatic
Cholangiocarcinomas Have Histologically and Immunophenotypically Distinct Small and Large Duct Patterns. Am. J. Surg.
Pathol. 2018, 42, 1334–1345. [CrossRef]
49. Luo, Y.; Yu, T.; Zhang, Q.; Fu, Q.; Hu, Y.; Xiang, M.; Peng, H.; Zheng, T.; Lu, L.; Shi, H. Upregulated N-cadherin expression
is associated with poor prognosis in epithelial-derived solid tumours: A meta-analysis. Eur. J. Clin. Investig. 2018, 48, e12903.
[CrossRef]
50. Kalluri, R.; Weinberg, R.A. The basics of epithelial-mesenchymal transition. J. Clin. Investig. 2009, 119, 1420–1428. [CrossRef]
51. Lammens, T.; Swerts, K.; Derycke, L.; De Craemer, A.; De Brouwer, S.; De Preter, K.; Van Roy, N.; Vandesompele, J.; Speleman,
F.; Philippé, J.; et al. N-cadherin in neuroblastoma disease: Expression and clinical significance. PLoS ONE 2012, 7, e31206.
[CrossRef]
52. Rodriguez, F.J.; Lewis-Tuffin, L.J.; Anastasiadis, P.Z. E-cadherin’s dark side: Possible role in tumor progression. Biochim. Biophys.
Acta 2012, 1826, 23–31. [CrossRef] [PubMed]
53. Franke, W.W.; Rickelt, S. Mesenchymal–epithelial transitions: Spontaneous and cumulative syntheses of epithelial marker
molecules and their assemblies to novel cell junctions connecting human hematopoietic tumor cells to carcinomatoid tissue
structures. Int. J. Cancer 2011, 129, 2588–2599. [CrossRef] [PubMed]
54. van Roy, F. Beyond E-cadherin: Roles of other cadherin superfamily members in cancer. Nat. Rev. Cancer 2014, 14, 121–134.
[CrossRef] [PubMed]
55. Mendonsa, A.M.; Na, T.Y.; Gumbiner, B.M. E-cadherin in contact inhibition and cancer. Oncogene 2018, 37, 4769–4780. [CrossRef]
[PubMed]
56. Techasen, A.; Loilome, W.; Namwat, N.; Khuntikeo, N.; Puapairoj, A.; Jearanaikoon, P.; Saya, H.; Yongvanit, P. Loss of E-cadherin
promotes migration and invasion of cholangiocarcinoma cells and serves as a potential marker of metastasis. Tumour Biol. 2014,
35, 8645–8652. [CrossRef]
57. Jiang, F.; Fang, D.B.; Lin, J.; Chen, Q.; Zhu, L.X.; Yu, H.Z. Correlation of LARP1 and E-cadherin expression with prognosis of
intrahepatic cholangiocarcinoma. Int. J. Clin. Exp. Pathol. 2018, 11, 3559–3566.
58. Mao, X.; Chen, D.; Wu, J.; Li, J.; Zhou, H.; Wu, Y.; Duan, X. Differential expression of fascin, E-cadherin and vimentin: Proteins
associated with survival of cholangiocarcinoma patients. Am. J. Med. Sci. 2013, 346, 261–268. [CrossRef]
59. Li, Q.; Wang, J.M.; Liu, C.; Xiao, B.L.; Lu, J.X.; Zou, S.Q. Correlation of aPKC-iota and E-cadherin expression with invasion and
prognosis of cholangiocarcinoma. Hepatobiliary Pancreat. Dis. Int. 2008, 7, 70–75.
60. Ashida, K.; Terada, T.; Kitamura, Y.; Kaibara, N. Expression of E-cadherin, alpha-catenin, beta-catenin, and CD44 (standard and
variant isoforms) in human cholangiocarcinoma: An immunohistochemical study. Hepatology 1998, 27, 974–982. [CrossRef]
61. Németh, Z.; Szász, A.M.; Somorácz, A.; Tátrai, P.; Németh, J.; Gyorffy, H.; Szíjártó, A.; Kupcsulik, P.; Kiss, A.; Schaff, Z. Zonula
occludens-1, occludin, and E-cadherin protein expression in biliary tract cancers. Pathol. Oncol. Res. 2009, 15, 533–539. [CrossRef]
62. Tian, X.; Cao, Z.; Ding, Q.; Li, Z.; Zhang, C. Prognostic value of multiple epithelial mesenchymal transition-associated proteins in
intrahepatic cholangiocarcinoma. Oncol. Lett. 2019, 18, 2059–2065. [CrossRef] [PubMed]
63. Jeong, Y.S.; Kang, J.; Lee, J.; Yoo, T.K.; Kim, S.H.; Lee, A. Analysis of the molecular subtypes of preoperative core needle biopsy
and surgical specimens in invasive breast cancer. J. Pathol. Transl. Med. 2020, 54, 87–94. [CrossRef] [PubMed]