J Clin Pathol 1999;52:829–832 829

Papillary : immunohistochemical J Clin Pathol: first published as 10.1136/jcp.52.11.829 on 1 November 1999. Downloaded from analysis of steroid receptor profile with a focus on apocrine diVerentiation

Annamaria OYdani, Anna Campanati

Abstract sometimes been shown to express oestrogen Aim—To make a quantitative evaluation receptors and progesterone receptors.1–3 How- by image analysis of oestrogen receptors, ever, there have been few reports on the steroid progesterone receptors, and androgen re- receptor profile of papillary hidradenoma, ceptors in papillary hidradenomas and which has unique clinicopathological features anogenital sweat glands. as it develops only in postpuberal and post- Methods—20 papillary hidradenomas and menopausal women and is localised almost the anogenital sweat glands detected in exclusively to vulvar, perineal, and perianal surgical specimens selected from 10 vul- skin.45 Moreover in recent years several vectomies for squamous , eight histological studies of the female anogenital haemorrhoidectomies, and one anal region have carefully examined a new variant of polypectomy, all from female patients, cutaneous glands, the so called “anogenital were investigated by the avidin– sweat glands,” which are considered the most streptavidin peroxidase testing system. likely source of papillary hidradenomas.6 Our Results—90% of papillary hidradenomas objective in this immunohistochemical study and almost all the anogenital sweat glands was to make a quantitative evaluation of showed immunoreactivity for oestrogen oestrogen receptors, progesterone receptors, receptor and, more weakly, for progester- and androgen receptors. By focusing attention one receptor, with immunolabelled nu- on the apocrine diVerentiation of papillary epi- clear area ranging from 10% to 90%. thelium, we aimed to provide further immuno- Conversely conventional sweat glands did histochemical evidence for the supposed his- not show any nuclear staining. Overex- togenetic linkage of papillary hidradenoma pression of androgen receptors occurred with anogenital sweat glands, as well its steroid in 20% of papillary hidradenomas, with hormone dependence. nuclear staining strictly bordering papil- lary with apocrine diVerentia-

tion. There was no immunoreactivity for Methods http://jcp.bmj.com/ androgen receptors in anogenital sweat The patient data and specimens used in this glands. study were retrieved by AC from the files of the Conclusions—Oestrogen and progesterone department of pathology and the dermatology receptors seem to represent reliable mark- clinic of the University of Ancona from 1986 to ers for diVerentiating between anogenital 1997. The investigation was performed on 19 sweat glands and conventional sweat excisional biopsies, including 20 papillary hid- glands, and a further link to explain why radenomas. In addition, anogenital sweat papillary hidradenomas occur almost ex- glands were analysed in the specimens from 10 on September 28, 2021 by guest. Protected copyright. clusively in the female anogenital region. vulvectomies performed for squamous carci- Positivity for oestrogen/progesterone re- noma, eight hemorrhoidectomies, and one anal ceptors suggests that epithelia either of polypectomy, all from female patients. Sections anogenital sweat glands or of papillary hid- from archival formalin fixed, paraYn embed- radenomas are controlled by ovarian ster- ded tissues were stained using conventional oid hormones. Androgen receptor nuclear procedures, and by the alcian blue (pH 2.5) staining of the epithelium with apocrine and periodic acid–SchiV (PAS) methods for diVerentiation in vulvar papillary hidrad- mucins. Serial sections were also incubated enoma strengthens its homology with with a panel of primary antibodies, all known breast duct . to react with formaldehyde fixed, paraYn (J Clin Pathol 1999;52:829–832) embedded tissue, against the following: low molecular weight keratins (CAM 5.2; Becton Keywords: steroid receptors; anogenital sweat glands; Dickinson), high molecular weight keratins Dermatology Clinic, papillary hidradenomas; apocrine diVerentiation University of Ancona, (AE 1–3; Diagnostic Products Corporation), Ospedale Umberto 1, carcinoembrionic antigen (anti-CEA; Ylem), Largo Cappelli I, 60123 Oestrogen and progesterone receptors have human milk fat globule protein (anti-HMFG 1 Ancona, Italy AOYdani recently been carefully investigated in both and 2; Oxoid), S 100 (anti-S protein; Dako- A Campanati benign and malignant eccrine sweat gland patts), actin (HHF 35; Dakopatts), vimentin tumours, primarily to identify a marker for dif- (anti-vimentin; Dakopatts), lysozyme (anti- Correspondence to: ferentiating breast metastases from lysozyme; Dakopatts), EMA (anti-EMA; Da- Professor OYdani eccrine . , chondroid sy- kopatts), and receptors for oestrogen (oestro- Accepted for publication ringoma, eccrine hidradenoma, eccrine carci- gen receptor ID5; BioGenex), progesterone 6 July 1999 noma, and microcystic adnexal carcinoma have (PGR-1A6; BioGenex), and androgen 830 OYdani, Campanati

(F39.4.1; BioGenex). The avidin-streptavidin- tin, human milk fat globules (HMFG), and J Clin Pathol: first published as 10.1136/jcp.52.11.829 on 1 November 1999. Downloaded from peroxidase testing system was employed. All lysozyme but not for those against carcinoem- preparations of papillary hidradenomas were brionic antigen (CEA). The cells with apocrine briefly counterstained with ethyl green and diVerentiation were distinctly immunoreactive those of anogenital sweat glands with haema- to epithelial membrane antigen, keratin, and toxylin. Non-immune mouse and rabbit sera lysozyme, but not to carcinoembrionic antigen. were substituted as negative controls. Appro- priate positive controls were run concurrently RECEPTORS for all the antibodies tested. Tissue samples Anogenital sweat glands reacted strongly with from breast and prostate , known to antibodies to oestrogen receptor and moder- bear oestrogen/progesterone receptors and ately to progesterone receptor, whereas there androgen receptors, acted as external controls. was no immunoreactivity to androgen recep- Stromal vimentin staining (fibroblasts, en- tor. Interestingly, in glands with an intermedi- dothelium) was used as an internal control of ate morphology encompassing anogenital antigenicity as its epitope shows a vulnerability sweat glands with “a small segment of typical to fixation or processing comparable to that eccrine epithelium”,6 only nuclei of anogenital 7 shown by steroid receptor antigens. The epithelium stained with oestrogen receptor and oestrogen and progesterone receptors were progesterone receptor antibodies. quantified on slides with the Cell Analysis Sys- Eighteen papillary hidradenomas (90%) 8 tem 200 instrument (CAS 200). Information showed immunoreactivity for oestrogen/pro- concerning clinical presentation, treatment, gesterone receptor which was scored according and outcome was sought for all patients. Com- to per cent positive nuclear surface11: 0, negative plete excision of hidradenomas was easily (two cases); 1, less than 1–10% positive (no accomplished and postoperative course was cases); 2, 1–10% positive (no cases); 3, 10–33% uneventful. positive (five cases); 4, 34–66% positive (seven cases); 5, 67–100% positive (sic cases). We chose Results a score of 10% as the cut oV to include border- CLINICAL FEATURES line positive cases in the positive category. The All patients were in postpuberal and postmeno- staining intensity was calculated on the basis of pausal age groups: the youngest was 26 years 1, weak; 2, moderate; and 3, intense nuclear old and the oldest 92 (mean 51 years). One staining. The oestrogen receptor score staining patient had two hidradenomas. Nine of the 19 was 3 in all positive cases, whereas progesterone patients were asymptomatic; in three, the receptor staining score was 2 in 13 cases and 1 in tumour was a fortuitous finding in specimens five cases. There was no relation between stain- from haemorrhoidectomies. Bleeding and ul- ing degree and the age of the patients. No cerated lesions had occurred in three patients. immunostaining was obtained for oestrogen and Four patients reported fluctuation in size of the progesterone receptors in epithelium with apo- tumor with the menstrual cycle. The clinical crine diVerentiation; in contrast, it showed over- diagnosis of the lesion was correct in five cases; expression of androgen receptors as almost all http://jcp.bmj.com/ misdiagnoses included in nine, haemor- the nuclei were immunolabelled. Two papillary rhoids in three, and and carcinoma hidradenomas did not show any staining for in one case each. Tumour sizes ranged from 0.3 either steroid receptors or vimentin filaments. cm to 4 cm (mean: 1.2 cm)

HISTOPATHOLOGY Discussion All 20 tumours fulfilled the diagnostic criteria Although an unusual lesion, papillary hidrad- of papillary hidradenoma, as reported in enoma represents the most common benign on September 28, 2021 by guest. Protected copyright. standard textbooks of dermatopathology.910In adnexal tumour of vulvar skin.12 The hypoth- five cases we detected areas of apocrine diVer- esis that this tumour originates directly from entiation of papillary epithelium where lining apocrine sweat glands conflicts with its main cells appeared columnar, cuboidal, or flat- clinicopathological features: it has not been tened, depending entirely on the location of the reported in men, it does not develop in the lesion on papillary folds or in tension cysts. axillary regions, and it does not appear to occur These cells were equipped with dome shaped, in black women.4 The common factor in these abundant, granular, and strongly eosinophilic exceptions is the presence of high concentra- cytoplasm and pale, globoid nuclei with one or tions of apocrine sweat glands. two nucleoli. Anogenital sweat glands were Several recent studies of the microscopic selected according to the morphological de- anatomy of the female anogenital region have scription provided by van der Putte.6 examined in detail a new variant of cutaneous glands, the so called anogenital sweat glands, IMMUNOHISTOCHEMISTRY which are considered to be the most likely The epithelium of both papillary hidradeno- source of papillary hidradenomas6 (fig 1). In mas and anogenital sweat glands rested on a this immunohistochemical investigation, we outer layer of myoepithelial cells that gave a evaluated oestrogen receptors, progesterone strong reaction with antibodies directed against receptors, and androgen receptors in a large smooth muscle actin (HHF35), S100 protein, series of anogenital sweat glands in which the and, more weakly, CAM 5.2. The inner epithelial cells were distinctly reactive to luminal layer of columnar cells, mainly with oestrogen receptors (fig 2) and to a lesser apical “snouts,” expressed distinct reactivity for degree to progesterone receptor, but not to antibodies against low molecular weight kera- androgen receptor. Steroid receptor profile in papillary hidradenoma 831

patients. Typical eccrine and apocrine sweat J Clin Pathol: first published as 10.1136/jcp.52.11.829 on 1 November 1999. Downloaded from glands included in specimens selected were devoid of either oestrogen receptors or proges- terone receptors, while immunoreactivity for androgen receptor was detected in eccrine secretory coils and apocrine luminal cells. These results agree with those previously reported.13–15 Thus oestrogen and progesterone receptors seem reliable markers for diVerenti- ating female anogenital glands from conven- tional sweat glands. This report, based on the quantitative evalu- ation by image analysis of oestrogen, progester- one, and androgen receptors in 20 papillary hidradenomas, is one of the most complete, as other reports have included fewer cases,16 or have involved testing for oestrogen receptor alone.2 Ninety per cent of papillary hidradeno- mas showed immunoreactivity for oestrogen receptor and more weakly for progesterone receptor (fig 3), and labelling ranged from 90% to 10% of nuclear area. We think that the nega- tivity in two cases was caused by loss of antigenicity during fixation and processing rather than because of a true lack of steroid receptors. Figure 1 Anogenital sweat gland with characteristic Thus the presence of oestrogen and proges- localisation in vulvar sulcus interlabialis: coiled duct is lined terone receptors tends to support the assump- by columnar cells with intraluminal snouts and apocrine tion that papillary hidradenoma is derived from type secretion. Haematoxylin and eosin stain. anogenital epithelium expressing these recep- tors rather than from conventional apocrine glands. In 20% of our papillary hidradenomas, immunoreactivity for androgen receptor was present but was strictly limited to luminal cells bearing the phenomenon of so called “apocrine diVerentiation” (fig 4), in the same way as in breast proliferative disorders, where it has been investigated in detail.17 18 http://jcp.bmj.com/ This finding, in combination with the morphological features, underscores the ho- mology between papillary hidradenoma and breast duct papilloma and supports the view that papillary hidradenoma originates from ectopic breast tissue, which may occur any-

where along the embryonic milk line. The on September 28, 2021 by guest. Protected copyright. lies in this line and is rich in apocrine glands, modification of which may give rise to accessory well developed breast tissue. Nevertheless this occurrence is exceedingly rare, since only 35 cases had been documented in published reports up to 1988.19 We did not find any well formed breast tissue in our speci- mens, and papillary hidradenomas were mainly found in the neighbourhood of anogenital sweat glands. Moreover they were also located Figure 2 Anogenital sweat gland: immunohistochemical in perianal skin, which is considered to be out detection of oestrogen receptor on formalin fixed, paraYn of the milk line and, on the average, they were embedded tissue. Nuclear area was strongly stained up to 95% with anti-oestrogen receptor antibody (ER-ID5). of smaller size than ectopic breast tissue. In Immunoperoxidase staining counterstained with addition, no patients in our series described haematoxylin. vulvar symptoms during pregnancy and lacta- tion, which might be expected as ectopic breast Interestingly, in anogenital sweat glands of tissue is capable of behaving in a similar way to intermediate type equipped with both “ano- the normal breast and responds to the hormo- genital” and typical eccrine epithelium, only nal influences.20 The presence of oestrogen the former was immunoreactive to oestrogen receptors, and even more so of progesterone and progesterone receptors. There was no cor- receptors, which are a phenotypical markers of relation between number of nuclei expressing hormonal action,21 22 suggests that the luminal oestrogen/progesterone receptor and age of the cells of anogenital sweat glands and papillary 832 OYdani, Campanati

papillary epithelium tends to support the J Clin Pathol: first published as 10.1136/jcp.52.11.829 on 1 November 1999. Downloaded from assumption that papillary hidradenomas are derived from these glands, rather than directly from conventional apocrine sweat glands, the latter being devoid of oestrogen and progester- one receptors. Finally the coexpression of ster- oid receptors has enabled us to classify the papillary hidradenoma in the group of hor- mone related tumours.

1 Wallace ML, Longacre TA, Smoller BR. Estrogen and pro- gesterone receptors and anti-gross cystic disease fluid pro- tein 15 (BRST-2) fail to distinguish metastatic breast carci- noma from eccrine neoplasms. Mod Pathol 1995;8:897– 901. 2 Swanson PE, Mazoujian G, Mills SE, et al. Immunoreactiv- ity for estrogen receptor protein in sweat gland tumors. Am J Surg Pathol 1991;15:835–41. 3 Wallace ML, Smoller BR. Progesterone receptor positivity supports hormonal control of syringoma. J Cutan Pathol 1995;22:442–5. 4 Woodworth H, Dockerty MB, Wilson RB, et al. Papillary hidradenoma of the vulva: a clinicopathologic study of 69 cases. Am J Obstet Gynecol 1971;110:501–8. Figure 3 Papillary hidradenoma: immunohistochemical detection of progesterone receptor 5 Meeker JH, Neubecker RD, Helwig EB. Hidradenoma on formalin fixed, paraYn embedded tissue. Almost all nuclei lining papillary folds were papilliferum. Am J Clin Pathol 1962;37:182–95. 6 Van der Putte SCJ. Anogenital “sweat” glands. Histology stained with anti-progesterone receptor antibody (PGR-1A6). Conversely the cells with and pathology of a gland that may mimic mammary glands. apocrine diVerentiation did not show any reactivity. Immunoperoxidase staining weakly Am J Dermatopathol 1991;13:557–67. counterstained with ethyl green. 7 Battifora H. Assessment of antigen damage in immunohis- tochemistry. The vimentin internal control. Am J Clin Pathol 1991;96:669–71. 8 Esteban JM, Kandalaft PL, Mehta P, et al. Improvements on the quantification of estrogen and progesterone receptors on paraYn-embedded tumors by image analysis. Am J Clin Pathol 1993;99:32–8. 9 McKee PH. Hidradenoma papilliferum. Pathology of the skin with clinical correlations, 2nd ed. London: Mosby-Wolfe, 1996:15.75–6. 10 Elder D, Elenitsas R, Ragsdaler BD. Hidradenoma papil- liferum. In: Elder D, ed. Lever’s histopathology of the skin, 8th ed. Philadelphia: Lippincot-Raven, 1997:770–1. 11 Elias JM, Cartun RA, England DM, et al. Interlaboratory comparison of estrogen receptor analysis in paraYn sections by a monoclonal antibody to estrophilin (H222). J Histotechnol 1993;16:57–63. 12 Michael H, Roth LW. Papillary hidradenomas. In: Wilkin- son EJ ed. Pathology of the vulva and vagina. New York: Churchill Livingstone, 1987:31–2. 13 Lloveras B, Googe PB, Goldberg DE, et al. Estrogen recep- tors in skin appendage tumors and extramammary Paget’s disease. Mod Pathol 1991;4:487–90. http://jcp.bmj.com/ 14 Shikata N, Kurokawa I, Andachi H, et al. Expression of androgen receptors in skin appendage tumors. An immuno- histochemical study. J Cutan Pathol 1995; 22:149–53. 15 Choundhry R, Hodgin MB, Van der Kwast TH, et al. Localisation of androgen receptors in human skin by immunohistochemistry: implication for the hormonal regulation of hair growth, sebaceous glands and sweat glands. J Endocrinol 1992;133:467–75. 16 IoVe OB, Alonsozana E, Drachenberg CB, et al. Hormone Figure 4 Papillary hidradenoma: immunohistochemical detection of androgen receptor on receptors in anogenital hidradenoma papilliferum (ahp): formalin fixed, paraYn embedded tissue. Only cells with apocrine diVerentiation stained for correlation with morphology and comparison with benign on September 28, 2021 by guest. Protected copyright. anti-AR antibody (F 39.4.1) while epithelium lining folds was negative. breast proliferations [abstract]. Mod Pathol 1998;11:50A. Immunoperoxidase staining weakly counterstained with ethyl green. 17 Eusebi V, Damiani S, Losi L, et al. Apocrine diVerentiation in breast epithelium. Adv Anat Pathol 1997;4:139–55. hidradenomas are controlled by ovarian steroid 18 Galatica Z. Immunohistochemical analysis of apocrine breast lesions. Consistent over-expression of androgen hormones. receptor accompanied by the loss of estrogen and On the basis of this immunohistochemical progesterone receptors in apocrine metaplasia and apo- crine carcinoma in situ. Pathol Res Pract 1997; 193:753–8. analysis, we believe that oestrogen receptors 19 Simon KE, Dutcher JP, Runowicz CD, et al. and progesterone receptors can be useful addi- arising in vulvar breast tissue. Cancer 1988;62:2234–8. 20 Donati P, Amantea A. of anogenital mammary- tional markers for distinguishing between ano- like glands. Am J Dermatopathol 1996;18:73–6. genital sweat glands and typical sweat glands, 21 Nenci I, Beccati MD, Arslan Pagnini C. Estrogen receptors and post-receptor markers in human breast cancer. Tumori which is never easy on morphological grounds 1978;64:161–74. alone. In addition the presence of nuclear 22 Horwitz KB, McGuire W. Estrogen control of progesterone staining for oestrogen receptor/progesterone receptor in human breast cancer. Correlation with nuclear processing of estrogen receptors. J Biol Chem 1978;253: receptors in both anogenital sweat glands and 2223–8.