Journal of the Egyptian National Cancer Institute (2012) 24, 175–184

Cairo University

Journal of the Egyptian National Cancer Institute

www.nci.cu.adu.egwww.sciencedirect.com

Original article

Potential diagnostic utility of CD56 and claudin-1

in papillary thyroid carcinoma and solitary follicular

thyroid nodules

Rasha M. Abd El Atti *, Lobna S. Shash

Pathology Department, Faculty of Medicine, Ain Shams University, Cairo, Egypt

Received 30 August 2012; accepted 11 October 2012Available online 6 November 2012

A

va

of

un*

E-

Pe

C

11

ht

KEYWORDS

Papillary carcinoma;

Follicular nodules;

CD56;

Claudin-1

bbreviations: PTC, papillary

riant of papillary carcinoma;

unknown malignant poten

known malignant potential.

Corresponding author. Tel.:

mail address: rashashaban90

er review under responsibil

airo University.

Production an

10-0362 ª 2012 National Ca

tp://dx.doi.org/10.1016/j.jnci.

thyroid

WDT-U

tial; FT

+20 100

00@yaho

ity of Th

d hostin

ncer Inst

2012.10.0

Abstract Background and Aim: The pathological diagnosis of papillary thyroid carcinoma (PTC)

is usually easily achieved. However distinguishing the follicular variant of papillary carcinoma

(FVPC) from other follicular thyroid lesions is an area of controversy. In this study we investigated

the role of CD56 and claudin-1 in the discriminating the FVPCs from other solitary follicular pat-

terned nodules. We also evaluated the application of these two markers in reclassifying the contro-

versial cases of the well differentiated tumors of unknown malignant potential (WDTs-UMP).

Materials and methods: The immunohistochemical expression of CD56 and claudin-1 was evalu-

ated in 86 samples of thyroid lesions together with 10 samples of normal thyroid tissue. Thyroid

lesions included: 29 PTCs [classic papillary carcinoma (n= 13) and FVPC (n= 16)], 47 solitary fol-

licular patterned nodules [follicular adenomas (n= 12), hyperplastic nodules (n= 32) and follicu-

lar tumor of unknown malignant potential (n= 3)] and 10 WDTs-UMP.

Results: The statistical analysis showed significantly different expressions of each of CD56 and

claudin-1 in the FVPCs versus other solitary follicular patterned nodules. Claudin-1 sensitivity

(100%) was higher than CD56 sensitivity (81.3%). However claudin-1 specificity (80.9%) was <

CD56 specificity (89.4%). The combined use of CD56 and claudin-1(claudin-1+/CD56�) showed

carcinoma; FVPC, follicular

MP, well differentiated tumor

-UMP, follicular tumor of

5221271.

o.com (R.M. Abd El Atti).

e National Cancer Institute,

g by Elsevier

itute, Cairo University. Production and hosting by Elsevier B.V.

02

Open access under CC BY-NC-ND license.

176 R.M. Abd El Atti, L.S. Shash

specificity (100%), positive predictive value (100%) and sensitivity (81.3%) in the differentiation

between the FVPCs and other follicular nodules. In the light of this statistical outcome, 5/10 cases

of WDTs-UMP expressing the (claudin-1+/CD56�) panel could be rediagnosed as PTC.

Conclusion: Combined utility of CD56 and claudin-1 is helpful in diagnosing the FVPC and its dif-

ferentiation from other follicular patterned nodules. Application of these two markers may greatly

aid in the reevaluation of the WDTs-UMP and interpretation of their expected behavior.

ª 2012 National Cancer Institute, Cairo University. Production and hosting by Elsevier B.V.

Open access under CC BY-NC-ND license.

Introduction

Papillary thyroid carcinoma (PTC) is the most common typeof thyroid cancer comprising approximately 80% of thyroidepithelial malignancies [1,2].

The ‘‘gold standard’’ in diagnosis of thyroid nodules is path-ologic evaluation using routine hematoxylin and eosin (H&E)staining. However, morphologic overlap between follicular le-

sions especially the follicular variant of papillary carcinoma(FVPC) is common. In such cases an objective consistent diag-nosis based merely on morphologic assessment is sometimesimpossible [3].

Consequently, immunohistochemical and molecular meth-ods were investigated to aid in the diagnosis of these problem-atic cases [4–7]. Several immunohistochemical markers such as

galectin-3, cytokeratin-19 and HBME-1 have been recom-mended to help in the discrimination between these controver-sial thyroid nodules [8]. Nonetheless up till now there is no

agreed consensus about an immunohistochemical panel thatwould reliably overcome such diagnostic obstacles.

Claudin-1, a member of claudin family, is identified as a

tight junction component [9]. Tight junctions are transmem-brane and cytoplasmic proteins which mainly prevent the freediffusion of solutes and maintain cell polarity [10,11]. Increasedclaudin-1 expression has been reported in PTC [12,13]. On the

other hand, CD56, a neural cell adhesion molecule (NCAM)[14], has been reported to be expressed in normal thyroid follic-ular cells with frequent low expression in malignant thyroid tu-

mors especially PTC [15–18]. CD56 regulates cell motility andhemophilic binding between neurons, hence its expressionmay affect the migratory capacity of tumor cells [19].

Claudin-1 and CD56 have been candidates for investiga-tions in several studies for the diagnosis of PTC. Howevernone of these studies, to the best of our knowledge, appliedthe combination of both claudin-1 and CD56 in an immuno-

histochemical panel to test the probability of an additive statis-tical significant outcome.

Therefore, the present study investigated the expression of

claudin-1 and CD56 separately and in combination in PTC, sol-itary thyroid nodules with follicular pattern, as well as normalthyroid tissue. Our aim was to identify the possible diagnostic

role of these two markers in the follicular morphologicalmimics.

Materials and methods

The material of this retrospective study included 86 specimens

of surgically removed, formalin-fixed and paraffin embeddedthyroid lesions that were received at the pathology departmentof Ain Shams University Hospital during the period fromJanuary 2009 to January 2011. All specimens with residual

adequate tissue that would allow proper immunohistochemicalassessment of the markers under research were included.

Furthermore, another 10 samples of randomly chosen nor-mal thyroid tissue obtained from radical laryngectomies forlaryngeal squamous cell carcinomas were included to assess

basic expression of the two markers under investigation innon lesional thyroid tissue.

Our 86 selected thyroid lesions included:

(a) 29 PTC cases.(b) 47 cases of solitary follicular patterned nodules.(c) 10 cases of well differentiated tumors of unknown malig-

nant potential (WDTs-UMP).

The 29 PTC cases were further classified into 13 cases of

classic PTC and 16 cases of FVPCs.The 47 cases of solitary follicular patterned nodules in-

cluded 32 cases of hyperplastic nodules, 12 cases of follicularadenomas, and 3 cases of follicular tumors of unknown malig-

nant potential (FTs-UMP).Confirmation of the original diagnosis was achieved through

separate revision of H&E stained slides by the two authors inde-

pendently using the well established diagnostic criteria for eachlesion. For the diagnosis of PTC we followed the histologicalcriteria proposed by Chan [20] which are divided into major

and minor features. The major features include: (1) nuclei areovoid rather than rounded; (2) nuclei are crowded, often man-ifesting as a lack of polarization in the cells that line the follicle;

(3) nuclei show a clear or pale chromatin pattern; (4) psam-moma bodies are found. If one of the four features was lacking,four ormore of the following featuresmay occur: (1) presence ofabortive papillae; (2) predominantly elongated or irregular

shaped follicles; (3) dark staining colloid; (4) presence of rarenuclear pseudoinclusions; or (5) multinucleated histiocytes infollicle lumen. The term FT-UMP was proposed for encapsu-

lated tumors composed of well differentiated follicular cellswith questionable or incomplete capsular invasion in absenceof PTC type nuclear changes. This term has been used by some

authors to designate follicular carcinoma exhibiting only mini-mal capsular invasion (irregular capsular profile, with initialpenetration by the tumor which never reached full thickness)

[21]. Follicular adenomas were defined as completely encapsu-lated follicular tumors with homogenous morphology, lackingthe nuclear features of papillary carcinoma and without capsu-lar or vascular invasion. While the term WDT-UMP was

proposed for an encapsulated tumor composed of well differen-tiated follicular cells with questionable PTC type nuclearchanges and shows absent or questionable capsular invasion

[21]. In our study, the WDT-UMP cases showed follicular cellswith some degree of PTC nuclear changes in the form of occa-sional optically clear nuclei, nuclear grooves and pseudoinclu-

Potential diagnostic utility of CD56 and claudin-1 in papillary thyroid carcinoma 177

sions. No evidence of papillary structures, capsular invasion or

vascular invasion. There was no controversy regarding the re-vised diagnoses among both reviewers.

Immunohistochemistry

All 96 samples (86 thyroid lesions and 10 normal thyroid tis-sues) were subjected to immunohistochemical staining with

claudin-1 and CD56 antibodies. The paraffin embedded tissuesections were deparaffinized in xylene and rehydrated throughabsolute alcohol. Antigen retrieval in citrate buffer (pH9 Lab

vision cat#AP9003) was used after the sections were treatedin a microwave at 8 w for 5–6 min, then at 3 w for 10 min,the sections were then left to cool for 20 min. Peroxidase and

protein block were done. After that the slides were incubatedovernight with each of the primary antibodies at room temper-ature using CD56 antibody (mouse monoclonal antibody 7 ml,Cat# MS-204-R7 Lab vision corporation, CA, USA) and clau-

din-1 rabbit polyclonal antibody 7 ml, Cat# 359A-18. Cellmarque, CA, USA), followed by rinsing in PBS (pH 7.6). Thiswas followed by the secondary biotin conjugated antibody for

1 h and finally the peroxidase conjugated streptavidin for an-other hour. Diaminbenzidine tetrachloride (DAB) (freshly pre-pared) was added for 25 min, then counterstained in Harris

Hematoxylin, followed by dehydration, clearing and mount-ing. Positive control for CD56 antibody was neuroblastomawhile positive control for claudin-1 was normal skin. Negativecontrols were done by excluding the primary antibody and its

replacement with a non-immune antibody.

Interpretation of immnohistochemical staining of CD56

The results of immnohistochemical staining were assessed byboth authors and a consensus regarding controversial cases

was reached at a multiheaded microscope.According to Park et al. [22], strong and complete membra-

nous expression with or without cytoplasmic staining of the

cells qualified the case as positive for CD56.Results were expressed in a semiquantitative manner with

respect to the percentage of positive tumor cells: 0, staining in<10% of the cells; 1, staining in 10–25% of the cells; 2, staining

in 26–50% of the cells; 3, staining in >50% of the cells.

Interpretation of immnohistochemical staining of claudin-1

According to Nemeth et al. [23], cases exhibiting membranousclaudin-1 staining in >5% of the cells were considered posi-

tive. Claudin-1 expression was scored as follows: 0, stainingin <5% of the cells; 1, staining in 5–25% of the cells; 2, stain-ing in 26–50% of the cells; 3, staining in >50% of the cells.

Statistical analysis

Data were analyzed using the SPSS program Version 15.Qualitative data were presented using the frequency and re-

lated percentage. Comparison of qualitative variables betweengroups was done using the Chi-square or the Exact test. Sensi-tivity, specificity, predictive value positive and predictive value

negative were calculated for the examined markers.Percentage of agreement was done using the Kappa statistic

(K).

A ‘‘P’’ value of 0.05 was chosen as the level of significance.

The K value can be interpreted as follows (Altman, 1991):

Value of K Strength of agreement

<0.20 Poor

0.21–0.40 Fair

0.41–0.60 Moderate

0.61–0.80 Good

0.81–1.00 Very good

Literature

Altman (1991) practical statistics for medical research. Lon-don: Chapman and Hall.

Results

CD56 and claudin-1 expressions in normal thyroid tissue

CD56 positive expression was found in the 10 samples ofnormal thyroid tissue 10/10 (100%), which showed strongmembranous CD56 expression in >50% of the cells (score3).

Meanwhile negative claudin-1 expression was observed in the10 normal thyroid tissue samples (10/10) (100%).

CD56 in the studied thyroid lesions

Among the solitary follicular patterned thyroid nodules; posi-tive CD56 expression was observed in 42 out of 47 cases

(89.4%) (Table 1), which included 11 out of 12 cases of follic-ular adenomas (91.7%) (Fig. 1a), 28 out of 32 cases of hyper-plastic nodules (87.5%) (Fig. 2a) and 3 out of 3 cases of FTs-

UMP (100%) (Fig. 3). All of the positive cases displayed strongCD56 expression in >50% of the cells (score 3), while the 5negative cases showed CD56 expression in <10% of the cells.

On the other hand, assessment of CD56 staining in the 29PTC cases showed negative CD56 expression in 24 out of the29 cases (82.8%). These cases included 11 out of 13 cases ofclassic PTC (84.6%) and 13 out of 16 cases of FVPC

(81.3%) (Fig. 4d). Positive CD56 expression was observed inonly 5 cases of PTC (17.2%); two of which were classic PTCshowing focal expression in 26–50% of the tumor cells (score

2), the other 3 cases were FVPCs and revealed focal expressionin 10–25% of the tumor cells (score 1). No statistical signifi-cant difference was found between classic PTCs and FVPCs

as regards CD56 expression (P = 1).CD56 distinguished FVPCs from other follicular patterned

nodules (follicular adenomas, hyperplastic nodules and

FTs-UMP) with a high statistically significant difference(P< 0.001) (Table 1).

Claudin-1 in the studied thyroid lesions

In contrast to CD56, the solitary follicular patterned nodulesshowed negative claudin-1 expression (staining in <5% of

the cells) in 38 out of 47 cases (80.9%) (Table 2), whichincluded 9 out of 12 cases of follicular adenomas (75%)(Fig. 1b), 27 out of 32 cases of hyperplastic nodules (84.4%)

(Fig. 2b) and 2 out of 3 cases of FTs-UMP (66.7%). Positiveclaudin-1 expression was observed in only 9 cases (19.1%).Five out of these 9 cases showed claudin-1 positivity in

Table 1 Comparison between follicular variant of papillary carcinomas (FVPCs) and other solitary follicular patterned nodules as

regards CD56 expression.

Follicular patterned

thyroid nodules

FVPC Total P value

CD56

Negative

Count 5 13 18 P < 0.001*

K= 0.442% within diagnosis 10.6% 81.3% 28.6%

Positive

Count 42 3 45

% within diagnosis 89.4% 18.7% 71.4%

Total 47 16 63

K: Kappa agreement = 0.442 (moderate strength of agreement).* Highly significant.

Figure 1 (a) Follicular adenoma showing diffuse positive CD56

immunostaining (CD56 · 400). (b) Follicular adenoma showing

negative claudin-1 immunostaining (claudin-1 · 200).

Figure 2 (a) Hyperplastic nodule showing positive CD56

expression (CD56 · 400). (b) Hyperplastic nodule showing nega-

tive claudin-1 expression (claudin-1 · 200).

178 R.M. Abd El Atti, L.S. Shash

26–50% of the cells (score 2); two of which were hyperplasticnodules and 3 were follicular adenomas. Whereas 4 of the clau-

din-1 positive cases showed focal expression in 5–25% of thecells (score 1) and these comprised 3 cases of hyperplastic nod-ules and 1 case of FT-UMP.

Meanwhile regarding the 29 PTC cases, strong and dif-fuse claudin-1 expression in 80–90% of the tumor cells(score 3) was observed in all cases, including 13/13

cases of classic PTC (Fig. 4a) and 16/16 cases of FVPCs(Fig. 4c).

Figure 3 Positive CD56 expression in follicular tumor of

unknown malignant potential (CD56 · 100).

Potential diagnostic utility of CD56 and claudin-1 in papillary thyroid carcinoma 179

There was a high statistically significant difference between

FVPCs and other follicular patterned nodules as regards clau-din-1 expression (P < 0.001) (Table 2).

Figure 4 (a) Classic papillary thyroid carcinoma showing strong dif

Follicular variant of papillary carcinoma (HE · 400). (c) Follicular vari

immunostaining (claudin-1 · 200). (d) Follicular variant of papillary c

Results of immunohistochemical expression of (claudin-1+/CD56�) panel in the studied thyroid lesions

(Claudin-1+ve/CD56�) expression panel was expressed only

in 13 out of 16 cases (81.3%) of FVPCs. On the other hand,it was absent in all 47 cases of follicular patterned nodules(100%). There was a statistically significant difference between

FVPC cases and other solitary follicular patterned thyroid nod-ules as regards the expression of (claudin-1+/CD56�) panel(P< 0.001) (Table 3).

Diagnostic validity of CD56, claudin-1 and (claudin-1+/CD56�) panel

CD56 was 81.3% sensitive and 89.4% specific in distinguishingFVPCs from other solitary follicular nodules, in comparison toclaudin-1 which was 100% sensitive and 80.9% specific.

Although not significant, the sensitivity of claudin-1 was higherthan that of CD56, whereas its specificity was lower than that ofCD56. The immunohistochemical panel of (claudin-1+/

CD56�) distinguished FVPCs with diagnostic accuracy of95.2% and specificity of 100%which were the highest comparedto those yielded by CD56 and claudin-1 individual expression(Table 4).

fuse membranous claudin-1 immunostaining (claudin-1 · 400). (b)

ant of papillary carcinoma showing diffuse membranous claudin-1

arcinoma showing negative CD56 immunostaining (CD56 · 400).

Table 2 Comparison between FVPCs and solitary follicular patterned nodules as regards claudin-1expression.

Follicular patterned

nodules

FVPC Total P value

Claudin-1

Negative

Count 38 0 38 P < 0.001*

K= 0.682% within diagnosis 80.9% 0% 60.3%

Positive

Count 9 16 25

% within diagnosis 19.1% 100% 39.7%

Total 47 16 63

K: Kappa agreement = 0.682 (good strength of agreement).* Highly significant.

Table 3 Comparison between FVPCs and other solitary follicular patterned nodules as regards (claudin-1+/CD56�) expression.Follicular patterned nodules FVPC Total P value

Antibody

Absent

Count 47 3 50 P < 0.001*

K= 0.866% within diagnosis 100% 18.7% 79.4%

Claudin1+/CD56

Present

Count 0 13 13

% within diagnosis 0 81.3% 20.6%

Total 47 16 63

K: Kappa agreement = 0.866 (very good strength of agreement).* Highly significant.

Table 4 The sensitivity, specificity, positive predictive value, negative predictive values and accuracy for distinguishing FVPCs from

the other solitary follicular patterned nodules.*

Antibody Sensitivity Specificity PPV NPV Accuracy

CD56�ve 81.3%

(57–93.4)

89.4%

(77.4–95.4)

72.2%

(49–87.5)

93.3%

(82–98)

(13 + 42)/63 = 87.3%

(77–93)

Claudin-1+ve 100%

(80.6–100)

80.9%

(67.4–89.6)

64%

(45–80)

100%

(91–100)

(38 + 16)/63 = 85.7%

(75–92)

Claudin-1+/CD56� 81.3%

(57–93.4)

100%

(92.4–100)

100%

(77–100)

94%

(84–98)

(47 + 13)/63 = 95.2%

(87–98)

PPV: positive predictive value, NPV: negative predictive value.* Estimated measurements 95% confidence interval.

180 R.M. Abd El Atti, L.S. Shash

The results of CD56 and claudin-1 expressions in well differen-tiated tumors of unknown malignant potential (WDTs-UMP)

In the light of the previous results, ten cases of WDTs-UMP were then subjected to immunostaining using CD56and claudin-1, where 3 out of 10 cases (30%) showedstrong positive CD56 immunoreactivity in 26–50% of the

cells (score 2) and the remaining 7 cases (70%) wereCD56 negative.

Eight cases of WDTs-UMP (80%) showed strong positive

claudin-1 expression in 80–90% of the cells (score 3) while 2cases (20%) showed negative claudin-1 expression.

The results of combined expression of CD56 and claudin-1in the 10 cases of WDTs-UMP showed: 2 cases (20%) negativefor both claudin-1 and CD56 (claudin-1�/CD56�), 3 cases(30%) positive for both claudin-1 and CD56 (claudin-

1+/CD56+) and 5 cases (50%) expressing the (claudin-1+/CD56�) panel (Fig. 5b and c).

Discussion

The diagnosis of PTC is, usually but not always, easily

achieved with almost minimal interobserver variability.However, in the absence of papillary architecture, distin-

Figure 5 (a) Well differentiated tumor of unknown malignant potential showing some optically clear nuclei, nuclear grooves and nuclear

pseudoinclusions, (HE · 400). (b) The same previous case showing positive claudin-1 expression (claudin-1 · 400). (c) The same case

showing negative CD56 expression (CD56 · 400).

Potential diagnostic utility of CD56 and claudin-1 in papillary thyroid carcinoma 181

guishing the FVPCs from cellular adenomatous nodules maybe challenging [24]. Therefore this study investigated the

possible role of CD56 and claudin-1 in resolving suchproblem.

CD56 has been reported to be an antigen related to the dif-ferentiation of the follicular epithelium [25] and many previous

studies reported high CD56 expression in normal thyroidtissue and benign thyroid follicular lesions as follicular adeno-mas and nodular hyperplasias [3,18,22,26–28]

In accordance with those studies, we currently reported ahigh positive CD56 expression in normal thyroid tissue com-pared to PTC cases. The present study also confirmed the

strong and diffuse positive CD56 expression in 89.4% of thesolitary follicular patterned thyroid nodules (Follicular adeno-mas, FTs-UMP and hyperplastic nodules).

On the other hand, the current study showed negativeCD56 expression in 82.8% of all PTC cases. Similarly, previ-ous studies reported negative CD56 expression in all or mostof their studied PTC cases [3,16–18,22,27,28].

The underlying molecular context of CD56 expression inthyroid cancer remains to be elucidated. However it is spec-ulated that CD56 expression might be involved in the activa-

tion of epithelial mesenchymal transition (EMT) (whichleads to more migratory and invasive cancers), and modula-tion of genes regulating metastasis as the vascular endothe-

lial growth factor (VEGF) [29–32]. This could explain themaintained elevated CD56 expression in some PTC cases

which may acquire later a more aggressive and metastaticphenotype.

Based on the previously mentioned results and in the lightof our finding that there was no statistically significant differ-

ence between CD56 expression in FVPCs and its expression inclassic PTCs (P > 0.05), we investigated the applicability ofusing CD56 as a marker to differentiate between FVPCs and

other solitary follicular nodules. As a result a statistically sig-nificant difference between these two groups as regards CD56expression was found (P < 0.001). Therefore we were able to

emphasize that lack of CD56 expression in the FVPCs wasvery helpful in their discrimination from other follicular nod-ules. The sensitivity and specificity of CD56 as a negative mar-

ker of FVPCs was 81.3% and 89.4% respectively and itsdiagnostic accuracy was 87.3%. On the other hand, Etem elal. [33] found no statistically significant difference betweenhis studied group of FVPCs and the other group of follicular

tumors (FTs-UMP, follicular adenomas and follicular carcino-mas) as regards CD56 expression.

In addition to CD56, claudin-1 immunohistochemistry was

previously evaluated in various thyroid lesions [34].The role of caludin-1 as a tight junction protein in cancer

initiation and progression has been intensively investigated.

182 R.M. Abd El Atti, L.S. Shash

Reduced expression, elevated levels or subcellular relocaliza-

tion of tight junction proteins have been reported in varioushuman malignancies and are variably associated with tumordifferentiation and survival [35–41]. The overexpression ofthese proteins in cancers (which typically lose their tight

junctions) is unexpected but may be related to roles that areunrelated to tight junction formation [42].

The upstream signaling pathways influencing claudin-1

expression in thyroid tumors remain elusive. Tzelepi et al.[34] demonstrated that papillary carcinoma being a welldifferentiated thyroid carcinoma shows high claudin-1 expres-

sion, and that dedifferentiation of thyroid tumors involvestight junction impairment via claudin-1 down regulation. Inaddition, loss of tight junction integrity leads to an increased

influx of growth factors, nutrients and other tumor promot-ing molecules, therefore providing an advantage for tumordevelopment and progression. On the other hand, Hucz etal. [12] and Nemeth et al. [23] associated claudin-1 expression

with the invasive and metastatic phenotype of PTC due topreserved claudin-1 strong expression in the lymph nodemetastasis.

In the current study all cases of PTCs (100%) showedstrong and diffuse claudin-1 expression .While all normal thy-roid tissue samples (100%) and 80.9% of the solitary follicular

patterned nodules showed negative claudin-1 expression. Sim-ilar results were elucidated by Hucz et al. [12] and Nemeth etal. [23] who reported high claudin-1 expression in PTC casescompared to negative expression in normal thyroid tissue

and in follicular adenomas. This differential expression in thenormal versus neoplastic tissues may provide new opportuni-ties for targeted cancer therapy [42]. In addition, our results

showed a high statistically significant difference betweenFVPCs and other solitary follicular patterned nodules as re-gards claudin-1 expression (P < 0.001). Thus claudin-1 immu-

nohistochemistry is proved to be very useful in differentiatingFVPCs from other follicular nodules with 100% sensitivityand 80.9% specificity, 100% NPV and 85.7% diagnostic accu-

racy. Although the specificity of claudin-1 is lowered comparedto that of CD56, claudin-1 negative expression can rule out thepossibility of papillary carcinoma in any suspicious follicularnodule.

Our recent observations encouraged us to assess the possi-ble value of the (claudin-1+/CD56�) panel in the differentialdiagnosis of the studied thyroid nodules with a better statisti-

cal significant outcome. As a result the sensitivity and NPV of(claudin-1+/CD56�) combination were lowered compared tothose of claudin-1 alone. However this panel was able to dis-

criminate FVPCs among other follicular patterned noduleswith diagnostic accuracy (95.2%), specificity (100%), andPPV (100%), which were the highest compared to those of

CD56� and claudin-1+. Therefore adding CD56 to claudin-1 especially in claudin-1 positive cases could greatly aid inreaching the final diagnosis.

Based on the previous results of CD56 and claudin-1 immu-

noreactivities in our studied cases, this work proposed the fol-lowing protocol in the evaluation of the follicular thyroidnodules:

(a) Diagnosis of papillary carcinoma (FVPC) in any contro-versial case should be definitely excluded if it shows neg-

ative claudin-1 expression (claudin-1 sensitivity is 100%and its NPV is 100%).

(b) Diagnosis of papillary carcinoma (FVPC) is definite in

any controversial case if it shows (claudin-1+/CD56�)immunoreactivity (this panel has 100% specificity and100% PPV).

(c) The diagnosis of PTC (FVPC) is excluded by 94% in anycontroversial case if it shows (claudin-1+/CD56+)panel [NPV of (claudin-1+/CD56�) is 94%].

Williams et al. [21] have proposed the term WDT-UMP foran encapsulated tumor composed of well differentiated follicu-lar cells with questionable PTC type nuclear changes, whether

capsular invasion is absent or questionable and in absence ofvascular invasion. The problem with this term is that it doesn’treally solve the interobserver variability between pathologists

regarding the diagnosis of FVPC. Instead it allows safe expres-sion of differences in opinion and help in avoiding excessivetreatment in controversial cases that cannot be readily placed

in a definite benign or a definite malignant category.Therefore the application of the above mentioned protocol

on the atypical category of WDTs-UMP might be helpful inthe reclassification and interpretation of the biologic behavior

of these tumors.Accordingly, by evaluating the claudin-1 and CD56 immu-

noreactivities in the 10 studied cases of WDTs-UMP, the

diagnosis of papillary carcinoma was likely to be excludedin the two cases which showed negative claudin-1 expression.Moreover, the diagnosis of papillary carcinoma would prob-

ably be excluded by 94% in the 3 cases which showed (clau-din-1+/CD56+) panel. Hence we could suggest that thefocal and incompletely developed PTC nuclear changes ob-served in these cases may be due to a mere artifact of fixa-

tion. On the other hand, the diagnosis of papillarycarcinoma could be suggested in the 5 cases which showed(claudin-1+/CD56�) panel.

The more apparent nuclear pseudoinclusions and groovesseen in these 5 cases may prove to be the expression of a path-ological abnormality rather than a technical error. In this sit-

uation the nuclear features argue in favor of rediagnosis ofthe (claudin-1+/CD56+) cases as PTC or at least suggest alink between WDT-UMP and PTC as the former may possibly

be a precursor of the latter.In conclusion, although the differential diagnoses of thy-

roid follicular nodules are based on histologic and cytomor-phologic criteria, combined utility CD56 and caludin-

1(claudin1+/CD56�) might be useful in the diagnosis ofPTC. However a question is raised ‘‘Does (claudin-1+/CD56+) expression panel has a practical diagnostic role in

the sense of absolute pushing of the WDT-UMP cases intothe benign category?’’ Our study demonstrated that this couldbe achieved by 94%. However the satisfactory answer to this

question will require long term follow up in order to seewhether the biologic effects of malignancy will influence thesecases or not. Therefore the current study recommends the eval-

uation of the role of claudin-1 and CD56 panel in more exten-sive studies including other variants of follicular cell derivednodules especially the borderline category of WDT-UMP toevaluate their expected behavior.

Competing interests

The authors declare that they have no competing interests.

Potential diagnostic utility of CD56 and claudin-1 in papillary thyroid carcinoma 183

Authors’ contributions

RMA conceived, designed and coordinated the study, re-

viewed the histologic diagnosis, evaluated immunohistochem-istry, carried out photographing and drafted the manuscript.LSS performed data collection, reviewed the histological diag-nosis, evaluated immunohistochemistry, helped in the study

design, helped to draft the manuscript and critically reviewedthe manuscript. Both authors read and approved the finalmanuscript.

References

[1] Pacini F, Schlumberger M, Dralle H, Elisei R, Smit JW,

Wiersinga W. European consensus for the management of

patients with differentiated thyroid carcinoma of the follicular

epithelium. Eur J Endocrinol 2006;154:787–803.

[2] Mc Nicol A. Pathology of thyroid tumors. Surgery

2007;25:458–62.

[3] El Demellawy D, Nasr AL, Babay S, Alowami S. Diagonistic

utility of CD56 immunohistochemistry in papillary carcinoma of

the thyroid. Pathol Res Pract 2009;205:303–9.

[4] Raphael SJ, Mckeown-Eyssen G, Asa SL. High molecular

weight cytokeratin and cytokeratin-19 in the diagnosis of

thyroid tumors. Mod Pathal 1994;7:295–300.

[5] Prasad ML, Pellegata NS, Huang Y, Nagraga HN, De la

chapelle A, Kloos RT. Galectin-3, Fibronectin-1, CI TED-1,

HBME1 and cytokeratin-19 immunohistochemistry is useful for

the differential diagnosis of thyroid tumors. Mod Pathol

2000;18:48–57.

[6] Huang Y, Prasad M, Lemon WJ, Hampel H, Wright FA,

Kornacher K, et al. Gene expression in papillary thyroid

carcinoma reveals highly consistent profiles. Proc Natl Acad

Sci USA 2001;98:15044–9.

[7] Beesley MF, Mclaren KM. Cytokeratin-19, galectin-3

immunohistochemistry in the differential diagnosis of solitary

thyroid nodule. Histopathology 2002;41:236–43.

[8] Liu YY, Morreau H, Kievit J, Romijn JA, Carrasco N, Smit

JW. Combined immunostaining with galectin-3, fibronectin-1,

CITED-1, HBME1, cytokeratin-19, peroxisome proliferator –

activated receptor – gamma and sodium/iodide symporter

antibodies for the differential diagnosis of non medullary

thyroid carcinoma. Eur J Endocrinol 2008;158:375–84.

[9] Furuse M, Fujita K, Hiiragi T, Fujimoto K, Tsukita S. Claudin-

1 and-2: novel integral membrane proteins localizing at tight

junctions with no sequence similarity to occludin. J Cell Biol

1998;141:1539–50.

[10] Sawada N, Murata M, Kikuchi K, Osanai M, Tobioka H,

Kojima T, et al. Tight junctions and human diseases. Med

Electron Microsc 2003;36:147–56.

[11] Gonzalez Mariscal L, Bentazos A, Nava P, Jaramillo BE. Tight

junction proteins. Prog Bio Mol Biol 2003;18:1–44.

[12] Hucz J, Kowalska M, Jarzab M, Wiench M. Gene expression of

metalloproteinase 11, claudin-1 and selected adhesion related

genes in papillary thyroid cancer. Endokrynol Pol 2006;57:18–25.

[13] Fluge O, Bruland O, Akslen LA, Lillehaug JR, Varhaug JE.

Gene expression in poorly differentiated thyroid carcinomas.

Thyroid 2006;16:161–75.

[14] Lanier LL, Testi R, Bindl J, Phillips JH. Identity of leu-19

(CD56) leukocyte differentiation and neural cell adhesion

molecule. J Exp Med 1989;169:2233–8.

[15] Zeromski J, Dworacki G, Jenek J, Niemir Z, Jezewska E, Jenek

R, et al. Protein and mRNA expression of CD56/N-CAM on

follicular epithelial cells of the human thyroid. Int J

Immunopathol Pharmacol 1999;12:23–30.

[16] Satoh F, Umemura S, Yasuda M, Osamura RY.

Neuroendocrine marker expression in thyroid epithelial

tumors. Endocr Pathol 2001;12:291–9.

[17] Scarpino S, Di Napoli A, Melotti F, Talerico C, Cancrini A,

Ruco L. Papillary carcinoma of the thyroid: low expression of

NCAM (CD56) is associated with down regulation of VEGF-D

production by tumor cells. J Pathol 2008;212:411–9.

[18] El Demellawy D, Nasr A, Alowami S. Application of CD56, P63

and CK19 immunohistochemistry in the diagnosis of papillary

carcinoma of the thyroid. Diagn Pathol 2008;3:5.

[19] Huerta S, Srivatsan ES, Venkatesan N, Peters J, Moatamed F,

Renner S, et al. Alternative mRNA splicing in colon cancer

causes loss of expression of neural cell adhesion molecules.

Surgery 2001;130:834–43.

[20] Chan JK. Strict criteria should be applied in the diagnosis of

encapsulated follicular variant of papillary thyroid carcinoma.

Am J Clin Pathol 2002;117:16–8.

[21] Williams ED. Two proposals regarding the terminology of

thyroid tumors. Int J Surg Pathol 2000;8:181–3.

[22] Park WY, Jeong SM, Lee JH, Kang HJ, Sin DH, Choi KU,

et al. Diagnostic value of decreased expression of CD56 protein

in papillary carcinoma of the thyroid gland. Basic Appl Pathol

2009;2:63–8.

[23] Nemeth J, Nemeth Z, Tatrai P, Peter I, Somoracz A, Szasz AM,

et al. High expression of claudin-1 protein in papillary thyroid

tumor and its regional lymph node metastasis. Pathol Oncol

Res. 2010;16:19–27.

[24] Nasr MR, Mukhopadhyay S, Zhang S, Katzenstein AL.

Immunohistochemical markers in diagnosis of papillary

thyroid carcinoma: utility of HBME 1 combined with CK-19

immunostaining. Mod Pathol 2006;19:1631–7.

[25] Migita K, Eguchi K, Kawakami A, Ida H, Fukuda T, Kurata A,

et al. Detection of leu-19 (CD56) antigen on human thyroid

epithelial cells by an immunohistochemical method.

Immunology 1991;72:246–9.

[26] Zeromski J, Biczysko M, Stajgis P, Lawniczak M, Biczysko W.

CD56 (NCAM) antigen in glandular epithelium of human

thyroid: light microscopic and ultrastructural study. Folia

Histochem Cytobiol 1999;37:11–7.

[27] Yang AH, Chen JY, Lee CH, Chen JY. Expression of NCAM

and OCIAD 1 in well differentiated thyroid carcinoma:

correlation with the risk of distant metastasis. J Clin Pathol

2012;65:206–12.

[28] Shin MK, Kim JW, Ju Y. CD56 and high molecular weight

keratin as dignostic markers of papillary thyroid carcinoma.

Korean J Pathol 2011;45:477–84.

[29] Vasko VV, Saji M. Molecular mechanisms involved in

differentiated thyroid cancer invasion and metastasis. Curr

Opin Oncol 2007;19:11–7.

[30] Ringel MD. Molecular markers of aggressiveness of thyroid

cancer. Curr Opin Endocrinol Diabetes Obes 2009;16:361–6.

[31] Lehember F, Yilmaz M, Wicki A, Schomber T, Strittmatter K,

Ziegler D, et al. NCAM induced focal adhesion assembly: a

functional switch upon loss of E-cadherin. EMBO J

2008;27:2603–15.

[32] Vasko V, Espinosa AV, Scouten W, He H, Auer H,

Liyanarachchi S, et al. Gene expression and functional

evidence of epithelial-to-mesenchymal transition in papillary

thyroid carcinoma invasion. Proct Natl Acad Sci USA

2007;104:2803–8.

[33] Etem H, Ozekinci S, Mizrak B, Senturk S. The role of CD56,

HBME-1, and p63 in follicular neoplasms of the thyroid. J

Pathol 2010;26:238–42.

[34] Tzelepi V, Tsamandas AC, Vlotinou HD, Vagianos CE, Ssopa

CD. Tight junctions in thyroid carcinogenesis: diverse

expression of claudin-1, claudin-4, claudin-7 and occludin in

thyroid neoplasms. Mod Pathol 2008;21:22–30.

184 R.M. Abd El Atti, L.S. Shash

[35] Resnick MB, Konkin T, Routhier J, Sabo E, Pricolo VE.

Claudin-1 is a strong prognostic indicator in stage II colonic

cancer: a tissue microarray study. Mod Pathol 2005;18:511–8.

[36] Kominsky SL, Argani P, Korz D, Evron E, Raman V,

Garrett E, et al. Loss of tight junction protein claudin-7

correlates with histologic grade in both ductal carcinoma

in situ and invasive ductal carcinoma of the breast. Oncogene

2003;22:2021–33.

[37] Usami Y, Chiba H, Nakayama F, Ueda J, Matsuda Y, Sawada

N, et al. Reduced expression of claudin-7 correlates with

invasion and metastasis in squamous cell carcinoma of the

oesophagus. Hum Pathol 2006;37:569–77.

[38] Michl P, Barth C, Buchholz M, Lerch M, Rolke M, Holzmann

K, et al. Claudin-4 expression decreases invasiveness and

metastatic potential of pancreatic cancer. Cancer Res

2003;63:6265–71.

[39] Rangel LB, Agarwal R, D’ Souza T, Pizer E, Alo’ PL, Lancaster

WD. Tight junction proteins claudin-3 and claudin-4 are

frequently over expressed in ovarian cancer but not in ovarian

cystadenomas. Clin Cancer Res 2003;9:2567–75.

[40] Lee JW, Lee SJ, Seo J, Song SY, Ahn G, Park CS. Increased

expression of claudin-1 and claudin-7 during the progression of

cervical neoplasia. Gynecol Oncol 2005;97:53–9.

[41] Hewitt KJ, Agarwal R, Morin PJ. The claudin gene family:

expression in normal and neoplastic tissue. BMC Cancer

2006;6:186.

[42] Morin PJ. Claudin proteins in human cancer: promising new

targets for diagnosis and therapy. Cancer Res 2005;65:9603–6.