Share this article on:

Endoglin (CD105) and Claudin-5 Expression in Cutaneous Angiosarcoma

Hara, Hiroyuki MD

The American Journal of Dermatopathology: October 2012 - Volume 34 - Issue 7 - p 779–782
doi: 10.1097/DAD.0b013e318252fc32
Letters to the Editor

Division of Anatomical Science, Department of Functional Morphology, Nihon University School of Medicine, Tokyo, Japan

The authors declares no conflict of interest.

To the Editor:

Angiosarcoma is a malignant tumor derived from endothelium that occurs in a variety of anatomic sites including the skin. Angiosarcomas primarily affect elderly persons and are predominantly found in men. The most common site of involvement is the head and neck, particularly the area of the scalp. Angiosarcoma is an aggressive tumor that tends to recur locally and to metastasize, despite aggressive multimodal therapy. The overall prognosis is poor. Immunohistochemical analysis is an important adjunctive diagnostic approach for angiosarcoma. Antibodies directed against VIII-related antigen, Ulex europaeus agglutinin type 1 (UEA-1), CD34, and CD31 have been used for the study of angiosarcoma. Almost 80%–90% of angiosarcomas stain positively for CD31. Staining of CD31 is highly sensitive, and CD31 expression is relatively specific for angiosarcomas.1

Endoglin (CD105), a 180-kDa homodimeric transmembrane glycoprotein, is a receptor for transforming growth factor β-1. Endoglin is highly expressed on endothelial cells during tumor angiogenesis and inflammation but shows weak or negative expression on vascular endothelium of normal tissues.2 Endoglin is also strongly expressed in the vasculature of a wide range of solid tumors.3 Several studies have shown that endoglin is a more specific and sensitive angiogenesis marker than other commonly used panendothelial antibodies.3,4

Claudins are the main compartments of tight junction, which consist of at least 20 different claudin types.5 Claudins maintain cellular polarity and play a role in signal transduction. Claudin-5 has been shown to be expressed in vascular endothelia of the lung, kidney, intestine,6 and dermal vascular endothelial cells.7 Jakab et al8 reported that claudin-5 was expressed in the neoplastic endothelial cells of canine hemangiosarcoma. Little is known about the expression of endoglin and claudin-5 in human cutaneous angiosarcoma. The aim of this study was to determine whether the expression of either endoglin or claudin-5 is a marker of human cutaneous angiosarcoma.

An archival search for cutaneous angiosarcoma was performed at the Department of Dermatology, Nihon University School of Medicine. The study material comprised 17 cases of cutaneous angiosarcoma. Diagnoses were based on conventional staining with hematoxylin and eosin complemented with CD31 and CD34 immunostaining.

The histopathological appearance of cutaneous angiosarcoma was divided into 3 histologic types: sinusoidal dissecting type, solid epithelioid/spindled type, and mixed type.9 Sinusoidal dissecting type is characterized by well-formed anastomosing vascular channels dissecting the dermis and subcutaneous fat and was lined by a single of multiple layers of atypical endothelial cells. Solid epithelioid/spindled type consists of islands and diffuse sheet of spindled and epithelioid cells with little intervening stroma. Mixed type consists of a mixture of the dissecting and solid patterns.

The following primary antibodies used for immunostaining were designed for use in formalin-fixed paraffin-embedded tissues. They were monoclonal mouse antiendoglin antibody (diluted 1:500, clone 4G11, NCL-CD105; Novocastra Labs, Newcastle, United Kingdom), monoclonal mouse anti–claudin-5 antibody (diluted 1:100, clone 4C3C2; Zymed, Inc, San Francisco, CA), and monoclonal mouse anti-CD31 antibody (diluted 1:500, clone 1A10, NCL-CD31-1A10; Novocastra Labs, Newcastle, United Kingdom). The bound antibodies were revealed by streptoavidin–biotin peroxidase detection system. Diaminobenzidine served as a chromogen and substrate. Nuclei were counterstained by Meyer hematoxylin. The intensity of the staining was further graded in a semiquantitative manner (weak, moderate, or strong).

Patients with cutaneous angiosarcoma comprised 15 men and 2 women with an age range of 62–88 years. The primary lesion was located on the forehead in 2 cases and on the scalp in 15 cases. Ten of 17 cases showed this sinusoidal dissecting type. Only 1 of 17 cases showed the solid epithelioid/spindled type. Six of 17 cases showed mixed type.

Table 1 summarizes the results of the immunohistochemical stain for endoglin, claudin-5, and CD31 in cutaneous angiosarcoma. Because CD31 is known to be the most sensitive and specific marker of endothelial cells of cutaneous angiosarcoma,1 we compared the expression of endoglin and claudin-5 with that of CD31. CD31 gave granular cytoplasmic and membranous immunoreactivity. All 17 cases showed that CD31 reacted with endothelial cells of the vascular channels and solid components. CD31 were very selective for the blood vessel endothelium and reacted specifically with anastomosing vascular channels with equal intensity (Fig. 1A), whereas CD31 and endoglin were weakly expressed in the solid components (Fig. 1B).





The endoglin (CD105) antibody 4G11 labeled all 17 cases. Labeling for endoglin was predominantly cytoplasmic and membranous. Antiendoglin antibody stained the vascular endothelial cells in the vascular network or sinusoids with linear staining (Fig. 2A). In contrast to the vascular channel or sinusoids, which showed intense membranous positivity for endoglin, endoglin was positive in the solid components with varying intensity (Fig. 2B). Notably, endoglin was highly expressed on the endothelial cells of microvessels and anastomosing channels with equal intensity.



Immunoreactivity for claudin-5 expression was detected in all 17 cases. Labeling for claudin-5 was predominantly cytoplasmic and membranous. Immunoreactivity was noted to be of moderate to strong intensity in the vascular network or sinusoids stained (Fig. 3A), while being strong and diffuse in the solid components of the tumor in all cases which were classified as the solid pattern and mixed pattern (Fig. 3B).



A previous study of angiosarcoma demonstrated the diagnostic value of CD31, CD34, and UEA-1 immunostaining. Of these markers, CD31 is known to be the most sensitive and specific marker of angiosarcoma.10 We compared the expression of endoglin and claudin-5 with that of CD31 in cutaneous angiosarcomas using immunohistochemical staining. Both endoglin and claudin-5 were consistently present in all cases of cutaneous angiosarcoma. We have demonstrated that cutaneous angiosarcomas show strong and diffuse endoglin and claudin-5 positivity. Endoglin and claudin-5 as well as CD31 were very selective for the blood vessel endothelium and reacted specifically with anastomosing vascular channels with equal intensity. Endoglin antibody was more discriminatory in this staining of intratumoral microvessels. Strong intensity of staining for claudin-5 was observed in the solid components, whereas CD31 and endoglin were weakly expressed in the solid components.

Endoglin is a receptor for the transforming growth factor β-1 molecule and is upregulated in proliferating endothelial cells and is strongly expressed in the vasculature of a wide range of solid tumors.3 Antibodies against the panendothelial markers such as CD31, CD34, and UEA-1 generally react well with endothelial cells of larger vessels, but their expression is sometimes diminished or lost in tumor microvessels. The endoglin antibody reacts exclusively with endothelium of peritumoral and intratumoral microvessels, but endoglin expression is weak or negative in vascular endothelium of normal tissues, suggesting that endoglin is preferentially expressed in vessels undergoing neoangiogenesis.11 Endoglin has been shown to be a more useful marker as tumor neovascularization than panendothelial markers such as CD31.4 Furthermore, it has been shown that endoglin serves as a better prognostic marker of patient outcomes than CD31.4,12 Limited studies about the endoglin expression in neoplastic cells have been reported. Fosmire et al13 investigated endoglin expression in several solid malignancies of different histotypes. Consequently, endoglin weakly stained the cytoplasm in 1 of 2 angiosarcomas, 1 of 3 leiosarcomas, and 1 of 2 meningiomas. The present study is only the second report, to our knowledge, of the expression of endoglin in angiosarcoma. In the present study, endoglin was highly expressed in the intratumoral microvessels in addition to its expression in large vessels.

Claudin-5 is a transmembrane protein that is expressed at tight junctions and is also a component of the blood–brain barrier and the blood–testis barrier.14 Claudin-5 has been reported to be expressed in endothelial cells. The presence of claudin-5 in endothelial cells suggests that tumor cells that express claudin-5 could have some additional affinity for blood vessels.13 There have been a limited number of studies of claudin-5 staining of angiosarcomas. Soini et al15 studied the expression of claudins-1, -2, -3, -4, -5, and -7 in epithelial and nonepithelial tumors. In the nonepithelial tumors, claudin-5 was expressed only in vascular lesions (22 hemangiomas and 3 angiosarcomas). Jakab et al8 reported that claudin-5 was expressed in the neoplastic endothelial cells of canine hemangiosarcoma (12 spleen lesions, 4 right atrium lesions, and 6 skin lesions). It is interesting to note that in addition to membranous staining, cytoplasmic staining was also shown in angiosarcomas.17 It is suggested that this mislocation may be the result of abnormal pathway activation in cancer.18 The results of the present study were in agreement with these previous studies and clearly demonstrated that claudin-5 was expressed in endothelial atypical cells that lined the vascular channels and in the cytoplasm of solid epithelioid/spindled cells.

In conclusion, our data confirm the endothelial cell origin of cutaneous angiosarcoma and demonstrate that both endoglin and claudin-5 show promise as markers for cutaneous angiosarcoma. Further studies are needed to explore the expression of these proteins in vascular tumor pathology.

Back to Top | Article Outline


1. Yang XJ, Zheng JW, Zhou Q, et al.. Angiosarcomas of the head and neck: a clinico-immunohistochemial study of 8 consecutive patients. Int J Oral Maxillofac Surg. 2010;39:568–572.
2. Kumar P, Wang JM, Bernabeu C. CD105 and angiogenesis. J Pathol. 1996;178:363–366.
3. Marioni G, Gaio E, Giacomelli L, et al.. Endoglin (CD105) expression in head and neck basaloid squamous cell carcinoma. Acta Oto-Laryngologica. 2005;125:307–311.
4. Saad RS, Liu YL, Celebrezze J, et al.. Endoglin (CD105) and vascular endothelial growth factor as prognostic marker in colorectal cancer. Mod Pathol. 2004;17:197–203.
5. Tsukita S, Furuse M. Claudin-based barrier in simple and stratified cellular sheets. Curr Opin Cell Biol. 2002;14:531–536.
6. Morita K, Sasaki H, Furuse M, et al.. Endothelial claudin: claudin-5/TMVCF constitutes tight junction strands in endothelial cells. J Cell Biol. 1999;147:185–194.
7. Morita K, Sasaki H, Furuse K, et al.. Expression of claudin-5 in dermal vascular endothelia. Exp Dermatol. 2003;12:289–305.
8. Jakab C, Halász J, Kiss A, et al.. Claudin-5 protein is a new differential marker for histopathological differential diagnosis of canine hemangiosarcoma. Histol Histopathol. 2009;24:801–813.
9. Morgan MB, Swann M, Somach S, et al.. Cutaneous angiosarcoma: a case series with prognostic correlation. J Am Acad Dermatol. 2004;50:867–874.
10. Orchard GE, Zelger B, Jones EW, et al.. An immunocytochemical assessment of 19 cases of cutaneous angiosarcoma. Histopathology. 1996;28:235–240.
11. Brewer CA, Setterdahl JJ, Li MJ, et al.. Endoglin expression as a measure of microvessels density in cervical cancer. Obstet Gynecol. 2000;96:224–228.
12. El-Gohary YM, Silverman JF, Olson PR, et al.. Endoglin (CD105) and vascular endothelial growth factor as prognostic markers in prostatic adenocarcinoma. Am J Clin Pathol. 2007;127:572–579.
13. Fosmire SP, Dikerson EB, Scott AM, et al.. Canine malignant hemangiosarcoma as a model of primitive angiogenic endothelium. Lab Invest. 2004;84:562–572.
14. Nitta T, Hata M, Gotoh S, et al.. Size-selective loosening of the blood-brain barrier in claudin-5-deficient mice. J Cell Biol. 2003;161:653–660.
15. Soini Y, Tommola S, Helin H, et al.. Claudins 1, 3, 4 and 5 in gastric carcinoma, loss of claudin expression associates with the diffuse subtype. Virchow Arch. 2006;448:52–58.
16. Soini Y. Expression of claudins 1, 2, 3, 4, 5 and 7 in various types of tumors. Histopathology. 2005;46:551–560.
    17. Barthlow TL, Chandran UR, Becich MJ, et al.. Immunohistochmical profiles of claudin-3 in primary and metastatic prostatic adenocarcinoma. Diagn Pathol. 2011;6:12–17.
    18. Rangel LE, Agarwal R, D'Souza T, et al.. Tight junction proteins claudin-3 and claudin-4 are frequently overexpressed in ovarian cancer but not in ovarian cystadenomas. Clin Cancer Res. 2003;9:2567–2575.
    © 2012 Lippincott Williams & Wilkins, Inc.