skin (Fig. 2g). In contrast, blood vessels appeared similar in both

groups (Fig. 2g,h). Morphometric analysis using IP-LAB software

(Scanalytics, Fairfax, VA, USA) revealed that the average skin area

covered by lymphatic vessels was larger in the ears of claudin-5+/�

mice than WT mice on day 2 after 40 mJ/cm2 UVB exposure

(P < 0.05) (Fig. 2i). Moreover, the average size of lymphatic ves-

sels was significantly increased in the ear skin of claudin-5+/� mice

as compared to WT mice (P < 0.01) (Fig. 2j), whereas the numbers

of lymphatic vessels in the two groups were not significantly different

(Fig. 2k). In contrast, no significant differences were found in the

average size or density of blood vessels in the two groups (Fig. 2l–n).ConclusionAlthough the expression of claudin-5 was greatly downregulated in

claudin-5+/� mice, lymphatic vessels did not exhibit morphological

and functional abnormalities in a physiological condition. As tight

junctions were not only composed of claudin-5 but also other clau-

dins such as claudin-7 (11) and ZO-1 (6), claudin-5+/� mice might

maintain normal morphology of lymphatic vessels. However, claudin-

5+/� mice revealed pronounced higher sensitivity to UVB. Claudin-

5+/� mice enhanced oedema formation together with hyperpermeable

lymphatic vessels after low-dose UVB irradiation, whereas a low dose

of UVB had no effect on WT mice. In contrast, no abnormalities were

detected in blood vessels. In inflamed, formation of focal intercellular

gaps leads to plasma leakage from blood vessels. These gaps are differ-

ent from discontinuous button-like junctions of initial lymphatics

(8). Our results suggested that the lymphatic tight junction molecule,

claudin-5, plays an important role in UVB-induced impairment of

lymphatic vessel function, which contributes to the dermal oedema

and skin inflammation.

AcknowledgementsWe thank Fumika Miyohashi for her technical assistance. Y.M. and K.K

designed performed the research and wrote the paper, and M.F contributed

the claudin-5+/�mice and analysed the data.

Conflict of interestsThe authors state no conflict of interests.

Ethical approvalThis study was approved by the ethics committee of Shiseido Research

Center in accordance with the guideline of the National Institute of

Health.

References1 Kligman A M. Drugs 1989: 38: 1–8.2 Kripke M L. Cancer Res 1994: 54: 6102–6105.3 Kajiya K, Hirakawa S, Detmar M. Am J Pathol

2006: 169: 1496–1503.4 Kajiya K, Sawane M, Huggenberger R et al. J

Invest Dermatol 2009: 129: 1292–1298.5 Navid F, Kolbe L, Stab F et al. Exp Dermatol

2012: 21: 147–153.6 Kajiya K, Kidoya H, Sawane M et al. Am J Pathol

2012: 180: 1273–1282.

7 Tammela T, Alitalo K. Cell 2010: 140: 460–476.8 Baluk P, Fuxe J, Hashizume H et al. J Exp Med

2007: 204: 2349–2362.9 Tsukita S, Furuse M, Itoh M. Nat Rev Mol Cell

Biol 2001: 2: 285–293.10 Nitta T, Hata M, Gotoh S et al. J Cell Biol 2003:

161: 653–660.11 Nelson G M, Padera T P, Garkavtsev I, Shioda T,

Jain R K. Neoplasia 2007: 9: 1038–1045.

Supporting InformationAdditional Supporting Information may be found inthe online version of this article:Data S1. Materials and methods.

Please note: Wiley-Blackwell are not responsible forthe content or functionality of any supporting materialssupplied by the authors. Any queries (other than miss-ing material) should be directed to the correspondingauthor for the article.

DOI: 10.1111/j.1600-0625.2012.01530.x

www.blackwellpublishing.com/EXDLetter to the Editor

Intratumoral mast cell number is negatively correlated with tumorsize and mitosis in dermatofibrosarcoma protuberans

Minji Kim1, Kwang Hyun Cho1, Jong Hee Lee2, Mee Soo Chang3,4 and Soyun Cho1,2

1Department of Dermatology, Seoul National University College of Medicine, Seoul, Korea; 2Department of Dermatology, Seoul National

University Boramae Hospital, Seoul, Korea; 3Department of Pathology, Seoul National University College of Medicine, Seoul, Korea; 4Department

of Pathology, Seoul National University Boramae Hospital, Seoul, Korea

Correspondence: Soyun Cho, Department of Dermatology, Seoul National University Boramae Hospital, 41, Boramae Road, Dongjak-Gu, Seoul

156-707, Korea, Tel.: 82-2-870-2385, Fax: 82-2-870-3866, e-mail: sycho@snu.ac.kr

Abstract: Dermatofibrosarcoma protuberans (DFSP) is a rare

cutaneous tumor with intermediate malignancy. The purpose of

this study was to evaluate the utility of p53, smooth muscle actin

(SMA) and c-kit as immunohistochemical markers and toluidine

blue staining for mast cell that can correlate with the clinical

outcome and clarify role of mast cells in pathogenesis of tumor.

We analysed data for 32 lesions from 31 patients.

Fibrosarcomatous type DFSP showed high immunoreactivity for

SMA compared with other subtypes (P = 0.026). No differences in

p53 immunoreactivity were observed between subtypes. None of

tumor cells were immunoreactive with c-kit. The mast cell counts

showed a negative correlation with mitosis and tumor size

(P < 0.05), implying that mast cells do not have a causative

primary role in tumorigenesis but rather play a secondary role.

Key words: c-kit – dermatofibrosarcoma protuberans – mast cell – p-53

– smooth muscle actin.

Accepted for publication 7 May 2012

ª 2012 John Wiley & Sons A/SExperimental Dermatology, 2012, 21, 535–561 559

Letter to the Editor

BackgroundDermatofibrosarcoma protuberans (DFSP) is a locally infiltrative

dermal tumor. Most cases have indolent disease course, but some

cases undergo frequent local recurrences and severe tissue destruc-

tion (1). Variable histopathological subtypes are reported, and dif-

ferent subtypes could affect clinical prognosis or outcome. To

date, surgical margin (2) and fibrosarcomatous change (3) are

considered valuable histological prognostic factors of DFSP. How-

ever, there is no immunohistochemical indicator to help predict

the clinical prognosis. Mast cells that perform diverse functions in

various conditions (4) can infiltrate around tumor cells in various

cancers (5–8), and it is suggested that complex and multiple

functions of mast cells are important in tumor growth and

progression.

Questions addressedThe purpose of this study is to evaluate the utility of p53, SMA

and c-kit as prognostic immunohistochemical markers, which

could correlate with the clinical outcome. Moreover, we investi-

gated role of mast cells in pathogenesis of the tumor.

Experiment designFor details of methods, see the Data S1.

Results and discussionsClinical dataWe analysed data for 31 patients, and Table S1 summarizes the

clinical characteristics of the 31 patients who had DFSP.

Histological dataA total of 32 lesions from 31 patients were enrolled. On slide

review of the 31 patients, most of the cases showed conventional

histopathological features with dense and uniform proliferation

of spindle-shaped tumor cells in a storiform arrangement. The

next most common forms included myoid, myxoid, pigmented

and fibrosarcomatous transformation (DFSP-FS) (Table S2, Fig.

S1a–d). There were two cases of DFSP-FS that had areas of cellu-

lar fascicles with a herring-bone pattern and cells with pleomor-

phic nuclei and numerous mitoses. Among them, one case of

DFSP-FS coexisted with myoid differentiation. DFSP-FS tended

to invade more deeply (P = 0.042) and be of larger size

(P = 0.026) than other subtypes. All cases of DFSP-FS were sized

over 4 cm, but only 11.5% of other subtypes exceeded that size.

Except for invasiveness and size of tumor, differences in the clini-

cal manifestation including age, site, local recurrence rate and

metastasis according to histological subtypes were not observed.

Most authors concluded that DFSP-FS has a higher tendency to

have a definite risk of recurrence and metastasis than classic

DFSP (3,9). We could not find a direct correlation between

DFSP-FS and a poorer clinical outcome, but as tumor size and

invasiveness was relatively associated with frequent local recur-

rence in previous research (9), the results of the present series

are in line with previous observations.

In immunohistochemical review, five cases showed immunore-

activity for SMA (Fig. 1a) and 18 cases for p53 (Fig. 1b). DFSP-

FS showed a high immunoreactivity for SMA compared with

other subtypes. All of the DFSP-FS cases showed positive SMA

staining, although only 11.5% of other subtypes were positive

(P = 0.026, OR = 8.677, (95% CI = 2.990–25.123). The character-

istic spindle cell proliferation seen in DFSP may suggest an origin

in myoblast (10) or mesenchymal stem cells (11). This finding

supports previous observations by Calonje et al. (10), who sug-

gested myoid/myoblast differentiation during fibrosarcomatous

transformation; however, as our results were obtained from a

small number of DFSP-FS, further study is needed.

No differences in p53 immunoreactivity among the subtypes

were observed. p53 overexpression has been correlated with higher

grade and aggressiveness in soft tissue sarcomas (12–14). However,

there were some discordant results about p53 expression in DFSP.

In our study, only half of the DFSP-FS cases showed p53 positiv-

ity, and 69.2% of DFSP of other subtypes had p53 immunoreac-

tivity. Moreover, there was no correlation between p53 positivity

and clinicopathological factors. In this context, our results demon-

strate that sole evaluation of p53 immunoreactivity without evalu-

ation of other parameters for proliferative activity such as MIB-1

labelling index and DNA analysis cannot reflect malignancy and

aggressiveness of DFSP.

The mean number of mast cells was 10.5/mm2 by toluidine

blue stain. There was no statistically significant difference in the

number of positively staining cells by toluidine blue (Fig. 1c) and

c-kit staining (Fig. 1d, Fig. S2 P = 0.679). The number of mast

cells was negatively correlated with mitotic activity (r = �0.487,

P = 0.021) and size of tumor (r = �0.393, P = 0.043) (Fig. 2a,b).

However, there was no correlation between mitosis and tumor

size. The involvement of mast cells in tumor growth and progres-

sion is noteworthy but debatable. Several studies have shown that

in many carcinomas such as colorectal cancer (15), pancreatic

cancer (8) and uterine cervix cancer (7), mast cells were associated

with new vessel formation which appeared to be their primary

contribution in tumor development. However, studies on the role

that mast cells play in tumors that originate from breast cancer

(5,6) and soft tissue show discrepant results (16,17). Donhuijsen

et al. (16) described that increased number of mast cells was cor-

related with low grade of malignancy in various soft tissue sarco-

(a) (b)

(c) (d)

Figure 1. Positive immunoreactivity for smooth muscle actin (SMA) (a, cytoplasmicpositivity) and p53 (b, nuclear positivity) in Dermatofibrosarcoma protuberans(DFSP), (original magnification 9200), scale bar: 40 lm; (c) toluidine blue stainingshows mast cells with red-purple granules (black arrows). (original magnification9400), scale bar: 20 lm; (d) none of the tumor cells stained for c-kit, but mastcells were found infiltrating the tumor cells. (original magnification 9400), scalebar: 20 lm.

560ª 2012 John Wiley & Sons A/S

Experimental Dermatology, 2012, 21, 535–561

Letter to the Editor

mas and concluded that it reflected the inhibitory role of mast

cells in tumor progression. Using toluidine blue stain and c-kit

immunostain, we conducted a mast cell count for evaluation of

host reaction to DFSP. Through c-kit immunoreactivity, we

checked for both mast cell count and tumor cell’s immunoreactiv-

ity for CD117; we confirmed mast cell counts with toluidine blue

staining. None of the tumor cells stained for c-kit, but various

ranges of mast cells were found infiltrating the tumor. Our results

showed an inverse correlation between the number of mast cells

and tumor size and mitosis, which reflect malignancy grade

of DFSP. Therefore, we suggest that mast cells are recruited sec-

ondarily as a host-tumor reaction analogous to other soft tissue

sarcomas. Accordingly, we cautiously speculate that unlike in car-

cinomas where mast cells play a primary role in tumorigenesis, in

DFSP mast cells infiltrate to help with tumor regression and inhi-

bit tumor progression.

ConclusionIn summary, immunoreactivity with p53, SMA and c-kit are not

useful markers to predict the clinical outcome of DFSP. However,

a negative correlation was found between the number of mast cells

and malignancy grade, implying a secondary, not primary, role of

mast cells in the pathogenesis of DFSP. Mistaking DFSP for a

keloid and administering antihistamines might be harmful to the

patient.

AcknowledgementsThis study was supported in part by a clinical research grant from the La

Roche-Posay Asia-Pacific Foundation.

Author contributionMinji Kim performed the research and wrote the paper. Soyun Cho

designed, supervised the project and wrote the paper. Kwang Hyun Cho

and Jong Hee Lee provided materials and intellectual input, Mee Soo

Chang coordinated experiments and provided intellectual input.

Ethics approvalThe institutional review board of the Boramae Hospital of Seoul National

University approved the study. Here, we enclose the certificate of approval.

Conflict of interestsThe authors have declared no conflicting interests.

References1 Cooper J Z, Brown M D. Malignant fibrous

tumors of the dermis. In: Wolff K G L, Katz S I,Gilchrest B A, Paller A S, Leffell D J, eds. Fitzpa-trick’s Dermatology in General Medicine. NewYork: McGraw-Hill; 2008: 1159–1161.

2 Monnier D, Vidal C, Martin L et al. J Eur AcadDermatol Venereol 2006: 20: 1237–1242.

3 Mentzel T, Beham A, Katenkamp D, Dei Tos A P,Fletcher C D. Am J Surg Pathol 1998: 22: 576–587.

4 Kneilling M, Rocken M. Exp Dermatol 2009: 18:488–496.

5 Della Rovere F, Granata A, Monaco M, Basile G.Anticancer Res 2009: 29: 3157–3161.

6 Della Rovere F, Granata A, Familiari D, D’ArrigoG, Mondello B, Basile G. Anticancer Res 2007:27: 2465–2471.

7 Utrera-Barillas D, Castro-Manrreza M, Castell-anos E et al. Exp Mol Pathol 2010: 89: 190–196.

8 Chang D Z, Ma Y, Ji B, Wang H, Deng D, Liu Y,Abbruzzese J L, Liu Y J, Logsdon C D, Hwu P.Clin Cancer Res 2011: 17: 7015–7023.

9 Bowne W B, Antonescu C R, Leung D H et al.Cancer 2000: 88: 2711–2720.

10 Calonje E, Fletcher C D. J Cutan Pathol 1996:23: 30–36.

11 Sellheyer K, Nelson P, Krahl D. Br J Dermatol2009: 161: 1317–1322.

12 Das P, Kotilingam D, Korchin B et al. Cancer2007: 109: 2323–2333.

13 Szadowska A, Olborski B, Harezga-Bal B,Debiec-Rychter M. Pol J Pathol 1999: 50: 9–16.

14 Takahira T, Oda Y, Tamiya S et al. Hum Pathol2004: 35: 240–245.

15 Gulubova M, Vlaykova T. J Gastroenterol Hepa-tol 2009: 24: 1265–1275.

16 Donhuijsen K, Sastry M, Volker B, Leder L D.Pathol Res Pract 1992: 188: 61–66.

17 Tomita Y, Aozasa K, Myoui A et al. Int J Cancer1993: 54: 194–199.

Supporting InformationAdditional Supporting Information may be found inthe online version of this article:Figure S1. Histologic features of DFSP.

Figure S2. Comparison of mast cell count by tolui-dine blue stain and c-kit immunostain: there was nostatistically significant difference in the number of posi-tively staining cells by toluidine blue and c-kit staining(10.5 vs 11.7/mm2, P = 0.679).Table S1. Clinical characteristics of patients with

DFSP.Table S2. Immunohistochemical staining data.

Data S1. Material and methods.

Please note: Wiley-Blackwell are not responsible forthe content or functionality of any supporting materialssupplied by the authors. Any queries (other than miss-ing material) should be directed to the correspondingauthor for the article.

(a)

(b)

Figure 2. The number of mast cells was negatively correlated with mitotic activity(/10HPF) (a) r = �0.487, P = 0.021) and size of tumor (mm2) (b) r = �0.393,P = 0.043). However, there was no correlation between mitosis and size.

ª 2012 John Wiley & Sons A/SExperimental Dermatology, 2012, 21, 535–561 561

Letter to the Editor