Virchows Archiv

, Volume 451, Issue 5, pp 959–966

Altered expression of desmocollin 3, desmoglein 3, and β-catenin in oral squamous cell carcinoma: correlation with lymph node metastasis and cell proliferation

Authors

    • Department of Bioengineering, School of sciencesGraduate School of Northeastern University
    • Department of Diagnostic Oral Pathology, Oral Restitution,Oral health SciencesGraduate School of Tokyo Medical and Dental University
  • Tingjiao Liu
    • Sectiont of Oral Pathology, College of StomatologyDalian Medical University
  • Yao Wang
    • Department of Bioengineering, School of sciencesGraduate School of Northeastern University
  • Lei Cao
    • Department of Oral Pathology, Oral Restitution, Oral health SciencesGraduate School of Tokyo Medical and Dental University
  • Mai Nishioka
    • Department of Diagnostic Oral Pathology, Oral Restitution,Oral health SciencesGraduate School of Tokyo Medical and Dental University
  • Rodelio L. Aguirre
    • Department of Diagnostic Oral Pathology, Oral Restitution,Oral health SciencesGraduate School of Tokyo Medical and Dental University
  • Ayataka Ishikawa
    • Department of Pathology, School of MedicineTokyo Women’s Medical University
  • Li Geng
    • Department of Bioengineering, School of sciencesGraduate School of Northeastern University
  • Norihiko Okada
    • Department of Diagnostic Oral Pathology, Oral Restitution,Oral health SciencesGraduate School of Tokyo Medical and Dental University
Original Article

DOI: 10.1007/s00428-007-0485-5

Cite this article as:
Wang, L., Liu, T., Wang, Y. et al. Virchows Arch (2007) 451: 959. doi:10.1007/s00428-007-0485-5

Abstract

Desmocollin 3 (Dsc3) and desmoglein 3 (Dsg3) are both transmembrane glycoproteins that belong to the cadherin family of calcium-dependent cell adhesion molecules. β-Catenin is a member of the cadherin–catenin complex that mediates homotypic cell–cell adhesion and is also an important molecule in the wnt signaling pathway. In this study, we examined the simultaneous expression level of Dsc3, Dsg3, and β-catenin in oral squamous cell carcinomas (OSCCs) and normal oral epithelia using immunohistochemistry. There was a significant correlation (p < 0.05) among the following variables in OSCCs: reduced or loss of expression of Dsc3, Dsg3, and β-catenin compared to normal oral epithelium, reduced or loss of expression of Dsc3 and histological grade (moderately or poorly differentiated), and reduced or loss of expression of β-catenin and lymph node metastasis. Furthermore, a positive correlation was found between reduced or loss of β-catenin staining and reduced or loss of Dsc3 staining in lymph node metastatic cancer tissue (r = 0.734, p < 0.05). These results suggest an abnormal expression of Dsc3, Dsg3, and β-catenin induced in the progression of oral carcinomas and that the Dsc3 expression level might be related to the regulation of β-catenin in lymph node metastasis and cell proliferation in OSCCs.

Keywords

Oral squamous cell carcinomaDesmosomal cadherinsβ-CateninLymph node metastasisCell proliferation

Introduction

Desmosomes are intercellular junctional apparatuses that connect intracellular intermediate filaments to the cell surface and mediate strong cell–cell adhesion [5]. Malignant transformation is often characterized by major changes in the organization of the cytoskeleton, decreased adhesive strength, and aberrant adhesion-mediated signaling [5, 6]. Disruption of normal cell–cell adhesion in transformed cells may contribute to enhancement of tumor cell migration and proliferation, leading to invasion and metastasis [6].

The main components of desmosomes consist of the products of three gene superfamilies: The desmosomal cadherins, the armadillo family of proteins, and the plakins family of cytolinkers [11, 12, 38].

Desmosomal cadherins are divided into two subfamilies: desmogleins (Dsg) and desmocollins (Dsc). There are four main types of epidermis-specific Dsg proteins, Dsg1, Dsg2, Dsg3, and Dsg4, and three major Dsc proteins, Dsc1, Dsc2, and Dsc3, each cadherin subtype being encoded by a unique gene [12]. They belong to the cadherin family of calcium-dependent cell adhesion molecules. Plakoglobin belongs to the armadillo family and binds directly to the cytoplasmic tail of both Dsg and Dsc [27, 26] and is closely related to the adherens junction molecule β-catenin [39].

β-Catenin interacts with the cytoplasmic tail of the calcium-dependent cell–cell adhesion molecule and through the association with plakoglobin creates a link between cadherin and the actin cytoskeleton [39]. More specifically, β-catenin is involved in the Wingless/Wnt signaling pathway [2] and interacts with epidermal growth factor receptor [16], the adenomatous polyposis coli protein product [35], and proto-oncogene c-ErbB2 [25]. β-Catenin also interacts with the two nuclear transcription factors, lymphoid-enhancing binding factor-1/T cell factor-4 [24].

Many studies have demonstrated a reduction in or loss of cell surface desmosomal cadherin expression, β-catenin membranous expression, and β-catenin mutation, which correlated with esophageal carcinoma [23], oral carcinoma [37, 18], and other carcinomas in systemic organs [29, 19, 1]. It was also shown that downregulation of desmosomal protein and abnormal β-catenin expression was associated with invasion and cell proliferation in pharyngeal [9], oral [9, 34], hepatocellular [7], and pancreatic [31] carcinomas.

In this study, we examined the simultaneous expression levels of Dsc3, Dsg3, and β-catenin in both oral squamous cell carcinomas (OSCCs) and normal oral epithelia using immunohistochemistry. We analyzed the relationship between clinicopathological findings, especially metastasis or cell proliferation and Dsc3, Dsg3, and β-catenin expression levels. We also examined the intracellular localization of Dsc3, Dsg3, and β-catenin expression.

Materials and methods

Human samples and clinicopathologic data

Between January 2001 to December 2003, we collected 48 formalin-fixed, paraffin-embedded human primary OSCC biopsy specimens (without previously treated), which included 24 cases without lymph node metastases and 24 cases with lymph node metastases, and 26 normal oral epithelia from the patients admitted to Tokyo Medical and Dental University Hospital, Japan. The patients with OSCC consisted of 20 women and 28 men with an age range of 28–82 years and an average age at the diagnosis of 59 years. Tumors were located in the tongue in 20 cases (41.7%), in the gingiva in 12 cases (25%), in the floor of the mouth in nine cases (18.8%), and in the buccal mucosa in seven cases (14.6%). Clinical details of each patient included age, gender, tumor size, clinical stage, lymph node status, and histological grade of tumor differentiation. Histological grading was done according to the World Health Organization (WHO) classification [8, 28].

Serial sections of 4 μm thick were cut from the paraffin-embedded tissue blocks and mounted on silanized slides. One section was stained with hematoxylin–eosin and examined to confirm the original diagnosis and tumor grade, and the other sections were submitted to immunohistochemical staining.

Immunohistochemical analysis

Anti-β-catenin (1:100, E-5, mouse monoclonal, Santa Cruz Biotechnology, Santa Cruz, CA) antibody, Dsc3 (U-114, mouse monoclonal, ready to use, Progen Biotechnik, Germany) antibody, and Dsg3 (1:25, 5G11, mouse monoclonal, Zymed, Berlin, Germany) antibody were used for all specimens. After deparaffinization and rehydration, endogenous peroxidase activity was blocked by immersing the sections into methanol containing 0.3% hydrogen peroxide at room temperature for 20 min. For antigen retrieval, histological sections were heated twice in 10 Mm citrate buffer (pH 6.0) for 10 min each time in a 500-W microwave oven, and sections were subsequently incubated with 10% normal rabbit serum to block nonspecific immunological reactions for 30 min. Each section was incubated with primary antibody overnight in a moist chamber, followed by incubation with biotinylated anti-mouse IgA+G+M secondary antibody and streptavidin–peroxidase complex (Nichirei, Tokyo, Japan). Finally, 3,3'-diaminobenzidine was used as a chromogen to visualize labeling. Nuclei were counterstained with methyl green. Negative controls were prepared by substituting the primary antibody with phosphate-buffered saline.

Scoring of immunostaining

The adjacent normal oral epithelium was used as an immunopositive control for Dsc3, Dsg3, and β-catenin staining that is normally seen at cell–cell junctions. A grading system was devised to obtain a semiquantitative evaluation of the distribution of each protein in the tissue slides. Scores were awarded according to Alazawi et al. [1] as follows: 0, negative staining throughout the epithelium, I, minimal membrane staining, II, patchy nonpericellular membrane staining, III, pericellular staining in one third of the epithelium, IV, pericellular staining in two thirds of the epithelium, and V, pericellular staining in all three thirds of the epithelium.

Statistical analysis

The statistical data was stored and analyzed by SPSS 13.0 for Windows software (SPSS). The Chi-squared test was used to analyze the association between clinical parameters (age, gender, T status, clinical stage, regional lymph node status, and histopathological grade) and immunohistochemical staining results. The relationship between the expression of β-catenin and Dsc3/Dsg3 were analyzed using Pearson’s coefficient of correlation method. A p value less than 0.05 was accepted as statistically significant.

Results

Immunohistochemical examination showed Dsc3, Dsg3, and β-catenin expressions in all samples as follows.

Expression of Dsc 3

As shown in Table 1, Dsc3 membranous immunoreactivity was confined to the basal and spinous cells of normal oral epithelium (15 cases scored V points, ten cases scored IV points, and only one case scored III points, Fig. 1a), and intracytoplasmic expression was not observed. On the other hand, the reduced or absent membranous expression of Dsc3 was found in the OSCC cases (30 of 48 cases showing 0–II scores) with only 18 cases showing III–IV scores (Fig. 1b,c). The scores of Dsc3 expression were significantly lower in OSCCs than in normal oral epithelium (p < 0.01). In addition, Dsc3 exhibited diffuse intracytoplasmic immunostaining, and nuclear staining was not observed in OSCC tumor cells (Fig. 1d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-007-0485-5/MediaObjects/428_2007_485_Fig1_HTML.jpg
Fig. 1

a Dsc3 immunohistochemical staining in normal oral epithelium showing strong linear membranous expression (150×). b Dsc3 immunohistochemical staining showing partially reduced membranous and cytoplasmic expression in moderately differentiated OSCC (150×). c Dsc3 staining is completely absent in poorly differentiated OSCC (150×). d Abnormal Dsc3 immunohistochemical localization in the cytoplasm of tumor cells (750×)

Table 1

Comparison of immunohistochemical staining score of β-catenin, Dsc3, and Dsg 3 in normal oral mucosa and OSCCs

 

Dsc3

Dsg3

β-Catenin

0

I

II

III

IV

V

Total

0

I

II

III

IV

V

Total

0

I

II

III

IV

V

Total

NOMa

0

0

0

1

10

15

26

0

0

0

3

8

15

26

0

0

0

4

11

11

26

OSCCb

6

10

14

10

8

0

48

4

13

15

6

8

2

48

4

10

18

12

4

0

48

P

<0.01

<0.01

<0.01

aNOM Normal oral mucosa

bOSCC Oral squamous cell carcinoma

Expression of Dsg 3

The staining results of Dsg3 are summarized in Table 1. Dsg3 membranous staining was located in the basal and spinous cells of normal oral epithelium (15 cases scored V points, eight cases scored IV points, and only three cases scored III points, Fig. 2a), and cytoplasmic expression was not observed, whereas reduced or absent membranous staining of Dsg3 was found in OSCCs (32 of 48 showing 0–II scores, 16 of 48 showing III–V scores, Fig. 2b). The staining score of Dsg3 was significantly lower in OSCCs than in normal oral epithelium (p < 0.01). In addition, intracytoplasmic accumulation of Dsg3 was observed in OSCC tumor cells (Fig. 2c).
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-007-0485-5/MediaObjects/428_2007_485_Fig2_HTML.jpg
Fig. 2

a In normal oral epithelium, immunohistochemical staining of Dsg3 shows strong linear membranous expression (150×). b Dsg3 immunohistochemical staining showing partially reduced membranous and cytoplasmic expression in moderately differentiated OSCC (150×). c Abnormal Dsg3 localization in the cytoplasm of tumor cells (750×)

Expression of β-catenin

The expression of β-catenin in normal oral epithelia and OSCCs is presented in Table 1. Among 26 specimens of normal oral epithelium, β-catenin expression was showed as a strong, linear membranous pattern (11 cases scored V points, 11 cases scored IV points, and four cases scored III points, Fig. 3a). Faint staining was detected in the cytoplasm, but no nuclear staining was found, whereas reduced or absent membranous expression of β-catenin was recognized in OSCCs (32 of 48 showing 0–II scores) with only 16 cases with score III or IV (Fig. 3b). There was a statistically significant difference in low expression level of β-catenin between OSCCs and normal oral epithelium (p < 0.01). In addition, β-catenin exhibited cytoplasmic and nuclear immunoreactivity in OSCC tumor cells (Fig. 3c).
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-007-0485-5/MediaObjects/428_2007_485_Fig3_HTML.jpg
Fig. 3

a In normal oral epithelium, β-catenin shows normal membranous expression (150×). b Peripheral loss of membranous of β-catenin in well differentiated OSCC (150×). c Nuclear β-catenin expression in tumor cells of OSCCs (750×)

Relationship between clinicopathological features and expression of Dsc3 Dsg3 and β-catenin in OSCCs

The variety of Dsc3, Dsg3, and β-catenin expression in relation to clinicopathological features are presented in Table 2. No correlation was found between the reduced or loss of membranous staining of Dsc3, Dsg3, β-catenin and clinical parameters such as age, sex, T status, and clinical stage (p > 0.05, respectively). Whereas, the statistical analyses showed a positive correlation between reduced or loss of membranous staining of β-catenin and lymph node metastasis (p < 0.05) and also between reduced or loss of membranous staining of Dsc3 and histological grade (moderately or poorly differentiated; p < 0.05).
Table 2

Correlation between β-catenin, Dsc3, and Dsg 3 immunohistochemical staining score and clinicopathological parameters in OSCCs

 

Dsc3

Dsg3

β-Catenin

0~IIa

III~Vb

Total

0~IIa

III~Vb

Total

0~IIa

III~Vb

Total

Age

≥60

16

10

26

17

9

26

18

8

26

<60

14

8

22

15

7

22

14

8

22

p value

>0.05

>0.05

>0.05

Gender

Male

20

8

28

19

9

28

20

8

28

Female

10

10

20

13

7

20

12

8

20

p value

>0.05

>0.05

>0.05

T status

T1 + T2

18

8

26

16

10

26

19

7

26

T3 + T4

12

10

22

16

6

22

13

9

22

p value

>0.05

>0.05

>0.05

Clinical staging

Stage 1–2

16

6

22

13

9

22

15

7

22

Stage 3–4

14

12

26

19

7

26

17

9

26

p value

>0.05

>0.05

>0.05

Histological differentiation

Grade 1

6

16

22

13

9

22

17

5

22

Grade 2

18

1

19

18

1

19

12

7

19

Grade 3

6

1

7

3

4

7

3

4

7

p value

<0.05

>0.05

>0.05

N status

N−c

14

10

24

14

10

24

10

14

24

N+d

16

8

24

18

6

24

22

2

24

p value

>0.05

>0.05

<0.05

a0~II Score 0, I, and II

bIII~V score III, IV, and V

cN Negative lymph node metastasis

dN+ Positive lymph node metastasis

β-Catenin expression levels correlate with decrease in Dsc3 and Dsg3 expression levels in lymph node metastasis in OSCC

The reduced or absent expression of β-catenin had a positive correlation with reduced or absent expression of Dsc3 showing scores 0, I, and II in 24 cases with lymph node metastasis (p < 0.05, Table 3). There was no significant correlation between the expression of Dsc3/Dsg3 and β-catenin/Dsg3 revealing scores 0, I, and II in 24 cases with lymph node metastasis (p > 0.05, respectively, Table 3). However, there were no correlations among the decrease in β-catenin/Dsc3, β-catenin/Dsg3, and Dsc3/Dsg3 expression levels exhibiting scores III, IV, and V in 24 cases with lymph node metastasis (p > 0.05, respectively, Table 3). It is also showed in Table 3 that no correlation among the decrease in β-catenin/Dsc3, β-catenin/Dsg3, and Dsc3/Dsg3 expression levels exhibiting scores 0, I, and II and scores III, IV, and V in 24 cases without lymph node metastasis (p > 0.05, respectively).
Table 3

Correlation between expression levels of β-catenin and Dsc3/Dsg3 in lymph node metastasis

correlation

LNa positive for metastasis

LN negative for metastasis

Pearson’s correlation (coefficient, r)

p value

Pearson’s correlation (coefficient, r)

p value

0–IIb

β-Catenin/Dsc3

0.734

<0.05

0.219

>0.05

β-Catenin/Dsg3

0.428

>0.05

0.317

>0.05

Dsc3/Dsg3

0.348

>0.05

0.289

>0.05

III–Vc

β-Catenin/Dsc3

0.216

>0.05

0.209

>0.05

β-Catenin/Dsg3

0.308

>0.05

0.231

>0.05

Dsc3/Dsg3

0.231

>0.05

0.199

>0.05

aLymph node

bScore 0, I, and II

cScore III, IV, and V

Discussion

Desmosomes are intercellular junctional apparatuses that provide membranous anchors for intermediate filament cytoskeleton. A reduced number of desmosomes would represent reduced intercellular adhesiveness, which, in turn, would be expected to promote invasive and metastatic ability [34]. In this study, we used immunohistochemistry to examine the cellular expression of Dsc3, Dsg3, and β-catenin in OSCCs and normal oral epithelium. We analyzed the relationship between clinicopathological features, especially metastasis or cell proliferation and Dsc3, Dsg3, and β-catenin expression levels. We also assessed intracellular localization of Dsc3, Dsg3, and β-catenin.

Desmosomal proteins are well established markers of epithelial differentiation [10]. Our results showed that Dsc3 and Dsg3 are strongly expressed in the basal and lower spinous layers in normal oral epithelium. Moreover, our results also demonstrated that a reduction in or absence of expression of Dsc3 and Dsg3 was significantly higher (30 of 48 and 32 of 48 cases showing 0–II scores, respectively) in OSCCs than in normal oral epithelium. The results are generally in agreement with previous studies [1, 34, 15, 22].

Hynes [17] reported that adhesion molecules transduce signals across the cell membrane both inward and outward and that such signals can modulate cell behavior, proliferation, and gene expression. A relationship between suprabasal misexpression of Dsc3 and cell proliferation/early terminal differentiation is suggested by Hardman et al. [14]. Dsc and Dsg misexpression has been examined and founded in actinic keratoses, Bowen’s disease, and squamous cell carcinoma (SCC) of the skin and oral cavity formerly [22, 34], this study furthers our understanding by showing that there was a definite relation between the reduction in or absence of expression of Dsc3 and histological differentiation (moderately or poorly differentiation) in OSCCs. The results of the present study suggest that desmosomes may have a possible role in suppression of cell proliferation in OSCCs.

It was reported that normal oral epithelium showed only membranous staining for β-catenin and membranous staining was lost or reduced with an increase in cytoplasmic staining in oral carcinoma [18]. In basal cell carcinomas of the skin, oral SCC and breast cancer, nuclear and diffuse intracytoplasmic staining of β-catenin was observed without membrane signal reinforcement [18, 19, 32]. Thus, membranous staining may be lost or markedly reduced according to the cytoplasmic internalization/nuclear staining of β-catenin. Our results showed that a reduction in or absence of membranous β-catenin staining was significantly different in OSCCs (0–II scores in 32 cases) from normal oral epithelium (III–V scores in 26 cases). These results are supported by other studies showing a similar correlation between β-catenin and the progression of oral cancer [36] and skin tumor [3].

Hiraki et al. [15] and Kurzen et al. [22] had recently reported an association between the abnormal expression of β-catenin at the time of invasion/metastasis and progression in oral cancer and skin tumor. The current study promotes our understanding by showing that β-catenin is also significantly associated with lymph node metastasis in OSCCs, but no correlation was found between the expression or absence of β-catenin and clinicopathological parameters such as age, gender, T status, clinical stage, as well as histological differentiation.

Dsc also plays an important role in epidermal integrity and differentiation. Dsc3 expression might be more compatible with keratinocytic proliferation and dynamic cell–cell adhesion. Plakophilin 2 binds β-catenin in vitro, and experimental upregulation enhances β-catenin signaling; at the same time, plakophilin 2 is also expressed in the basal cell layer of epidermis, where it should interact with Dsc3 [4]. Aberrant expression of Dsc3 in the epidermal suprabasal layers regulated epidermal differentiation and affected β-catenin expression [14, 13]. Hardman et al. [14] have reported a correlation between Dsc3 and β-catenin stability, providing further evidence for the integration of desmosomal cadherins and cell-signaling pathways fundamental to epidermal proliferation and differentiation programs. A study regarding the relationship between β-catenin, Dsc3, and Dsg3 expression in 24 cases with lymph node metastases showed that while there was a significant positive correlation between the reduction in or loss of β-catenin and the reduction in or loss of Dsc3 (Pearson’s correlation coefficient r = 0.734, p < 0.05), no correlation was found between β-catenin and Dsg3 and between Dsc3 and Dsg3 expression showing scores 0, I, and II. In other words, our results showed that the decrease in Dsc3 expression levels positively regulates β-catenin expression levels in lymph node metastasis in OSCCs. The research on relation between desmosomal components and β-catenin signaling have been carried out, where it was shown that Dsc3 regulated β-catenin in suprabasal keratinocytes [14] and human odontoblasts [33], by inducing β-catenin stabilization and altering signaling.

Additionally, our results also showed that a diffuse perinuclear cytoplasmic internalization of Dsc3 and Dsg3 and a cytoplasmic and nuclear staining pattern of β-catenin were found in OSCC tumor cells. This change of distribution was accompanied by a reduction in membranous expression in most tumors and significantly correlated with disease progression in accordance with other studies [18, 20, 21, 32, 30]. Thus, a reduction in or absence of membranous pattern of cadherin proteins and β-catenin, together with internalization to the cytoplasm, may be associated with tumor progression.

Conclusion

We have shown that there was a significant correlation between loss of or reduction in β-catenin staining and loss of or reduction in Dsc3 staining in lymph node metastasis in OSCCs. Our observations suggest that the abnormal expression of Dsc3, Dsg3, and β-catenin were correlated with the progression of oral carcinomas. Dsc3 expression levels might be related to the regulation of β-catenin in metastatic activities and cell proliferation in OSCCs, and internalization of Dsc3, Dsg3, and β-catenin may also be associated with progression of OSCCs.

Acknowledgments

The authors thank members of Oral Pathology, Graduate School of Tokyo Medical and Dental University for their excellent technical assistance.

Copyright information

© Springer-Verlag 2007