PRAME Expression in Mammary and Extramammary Paget Disease

PRAME (PReferentially-expressed Antigen in MElanoma), also known as CT130 (cancer testis antigen 130), is a testis cancer antigen that has been extensively studied since some years, namely for the diagnosis of melanocytic tumors. Indeed, nuclear PRAME expression is much more often seen in melanomas compared with benign naevi and may thereby be of diagnostic help in the setting of ambiguous melanocytic tumors.1–5 In normal skin, PRAME is expressed by sebaceous glands, namely in the cytoplasm of peripheral, less differentiated sebocytes, and in apocrine sweat glands.6–8 Accordingly, PRAME is expressed in several sebaceous proliferations, with a decreasing frequency and intensity in tumors with increasing biologic aggressiveness.3,7,9–11 Data on PRAME expression by other adnexal skin tumors are much more limited. One study found that a subset (ca 19%) of benign sweat gland tumors, including namely papillary hidradenomas and hidrocystomas, expressed PRAME, whereas sweat gland carcinomas and benign hair follicle tumors were PRAME negative.10 PRAME expression has also been reported in a subset of Merkel cell carcinomas and, albeit with a lower intensity, in some basal cell and squamous cell carcinomas of the skin.3,5,7,8,10,12 During the routine use of PRAME for the diagnosis of melanocytic and adnexal proliferations, we noted that, apart from sebaceous glands (Fig. 1A) and apocrine sweat glands (Fig. 1B), PRAME was also expressed by a subset of eccrine secretory cells located toward the lumen of the gland (Fig. 1C), that is, where the dark, mucous-secreting cells, as seen by electron microscopy, are found (Fig. 1D). The previously reported expression of PRAME by a subset of benign sweat gland proliferations10 is, therefore, not totally unexpected. Extramammary Paget disease is regarded as an intraepidermal carcinoma with apocrine differentiation, which may be restricted to the skin (primary extramammary Paget disease) or associated with an underlying carcinoma, usually of the colon or the urogenital system (secondary extramammary Paget disease).13 Its mammary counterpart, mammary Paget disease, is an intraepidermal carcinoma affecting the nipple/areola skin and is almost invariably associated with an underlying breast carcinoma.14 Both mammary and extramammary Paget disease are characterized by the invasion of the epidermis by large, round atypical cells with large nuclei and a clear cytoplasm, the so-called Paget cells. So far, very limited data exist on the expression of cancer testis antigens by mammary and extramammary Paget disease. One study found that the cancer testis antigens New York esophageal squamous cell carcinoma 1 (NY-ESO-1) and melanoma-associated antigen A (MAGE-A) were expressed by 16% and 39% of extramammary Paget disease cases, respectively.15 Concerning PRAME, only 1 study on sebaceous carcinomas included (among other mimics) 2 cases of extramammary Paget disease, which were found to be PRAME negative.11 These facts prompted us to investigate the immunohistochemical expression of PRAME in mammary and extramammary Paget disease.FIGURE 1.: Normal human skin. A, PRAME is strongly expressed in the cytoplasm of peripherally located sebocytes. B, PRAME is strongly expressed in the cytoplasm of the secretory cells of an apocrine sweat gland. C, Eccrine secretory sweat gland (secretory coil). PRAME is expressed in the cytoplasm of a minority of cells located toward the glandular lumen. D, By transmission electron microscopy, these cells correspond to mucous-secreting (dark) cells (A–C: Immunoperoxidase revealed with fast-red, counterstained with Mayer hematoxylin). D, Transmission electron microscopy (ultrathin sections stained with uranyl acetate and lead citrate).Cases diagnosed as mammary and extramammary Paget disease for the past 5 years in the laboratories of Pathology of the Lyon-Sud Hospital (Hospices Civils de Lyon, France) and the AHEPA Hospital (Aristotle University of Thessaloniki, Greece) were retrieved and retrospectively studied. The diagnosis had been made by examination of hematoxylin-eosin-stained sections and appropriate immunostainings (namely for keratins 7 and 20, ER, PR, Her2/neu, GATA-3, EMA, TTF1, CDX-2), and after consideration of clinical data retrieved from the patients' medical files. A total of 35 cases were studied. There were 14 patients with mammary Paget disease localized on the nipple and/or the areola (all women, mean age 71 years, range 56–92) and 21 patients with in situ or invasive extramammary Paget disease (13 women/8 men, mean age 76.5 years, range 60–94). The 5-μm thick tissue sections cut from the paraffin blocks were immunostained for PRAME, using the ERP20330 rabbit IgG monoclonal antibody (Abcam, dilution 1/500) and an automated immunohistochemistry device (Ventana), with fast-red as chromogen and Mayer hematoxylin as counterstain. Sebaceous and/or apocrine glands present in each section served as positive internal controls. In each case, the percentage of positive Paget cells was evaluated and scored semiquantitatively (0: negative, 1: ≤20%, 2: 21%–60%, 3: ≥61%). The staining intensity was also evaluated semiquantitatively (0: absent, 1: mild, 2: moderate, 3: strong), considering as 3 the highest intensity observed in the most intensely labeled control skin structures present in each slide (sebaceous or apocrine sweat glands). An immunoscore was calculated for each case by multiplying the 2 scores (percentage of positive cells X staining intensity, ranging from 0 to 9). The detailed results are given in Tables 1 and 2. All but 1 of the 14 cases of mammary Paget disease (93%) expressed PRAME with variable immunoscores (mean value 5,5/9). PRAME expression was predominantly nuclear (8 cases), and cytoplasmic or both nuclear and cytoplasmic in the remaining 5 cases (Fig. 2A). In 5 of 13 cases, Paget cells found in the lowermost epidermal layers were more strongly stained compared with Paget cells located more superficially in the epidermis. Of the 21 cases of extramammary Paget disease, 19 (90.5%) expressed PRAME with variable immunoscores (mean value 5.2/9). PRAME expression was predominantly (13/19 cases) cytoplasmic and nucleocytoplasmic, or merely nuclear in the remaining 4 cases (Figs. 2B–D). As with mammary Paget disease, weaker expression of PRAME in the more superficially located Paget cells was noted in 7 of 19 cases (Fig. 2B; Table 2). TABLE 1. - PRAME Immunoreactivity in Mammary Paget Disease Case n° Sex/Age (yr) % Positive cells* Labeling intensity Immunoscore M1 F/77 3 2 6 M2 F/89 3 3 9 M3 F/80 3 3 9 M4 F/86 3 3 9 M5 F/56 2 1 2 M6 F/90 2 1 2 M7 F/43 2 2 4 M8 F/63 3 3 9 M9 F/92 1 1 1 M10 F/67 2 3 6 M11 F/53 2 3 6 M12 F/65 3 3 9 M13 F/69 1 1 1 M14 F/67 0 0 0 *1: ≤20% positive cells, 2: 21%–60% positive cells, 3: ≥61% positive cells.F, female. FIGURE 2.: A, Mammary Paget disease. Nucleocytoplasmic PRAME expression is seen in most Paget cells (case n° M2, immunoscore 9). B, Extramammary (vulvar) Paget disease. PRAME is expressed in the cytoplasm of most Paget cells, with a decreasing intensity toward the uppermost epidermal layers (case n° E2, immunoscore 9). C, Extramammary (inguinal) Paget disease. Variable nucleocytoplasmic PRAME expression by intraepidermal (nested) Paget cells is seen, which is slightly weaker than the underlying dermal tumor (case n° E9, immunoscore 4). D, Extramammary (perineal) Paget disease. Strong cytoplasmic PRAME expression is seen in most intraepidermal Paget cells (case n° E14, immunoscore 9) (all panels: Immunoperoxidase revealed with fast-red, counterstained with Mayer hematoxylin). TABLE 2. - PRAME Immunoreactivity in Extramammary Paget Disease Case n° Sex/Age (yr) Localization % Positive cells* Labeling intensity Immunoscore E1 F/81 Vulvar 3 3 9 E2 F/80 Vulvar 3 3 9 E3 F/83 Vulvar 3 3 9 E4 F/70 Vulvar 3 2 6 E5 F/73 Vulvar 2 2 4 E6 F/73 Vulvar 2 2 4 E7 F/71 Vulvar 3 2 6 E8 F/94 Vulvar 3 3 9 E9 M/93 Inguinal 2 2 4 E10 M/61 Inguinal 2 2 4 E11 M/90 Inguinal 2 1 2 E12 M/60 Inguinoscrotal 2 1 2 E13 M/86 Perineoscrotal 2 1 2 E14 F/84 Perineal 3 3 9 E15 F/75 Perineal 0 0 0 E16 M/83 Perianal 2 1 2 E17 F/76 Perianal 3 3 9 E18 M/84 Scrotal 3 2 6 E19 F/63 Vaginal 1 1 1 E20 F/63 Suprapubic 0 0 0 E21 M/64 Preauricular 1 1 1 *1: ≤20% positive cells, 2: 21%–60% positive cells, 3: ≥61% positive cells.F, female, M, male. Our findings show that most (91.4%) cases of mammary and extramammary Paget disease express PRAME to various degrees. They are at variance with the results of a previous study, which found no expression of PRAME in 2 cases of extramammary Paget disease.11 Whether this difference is because of a different antibody used cannot be assessed, because that previous study did not mention the antibody used.11 Our results expand the spectrum of cutaneous epithelial tumors, and more specifically glandular ones, that express PRAME. They are not totally surprising, considering that cutaneous apocrine sweat glands strongly express PRAME, and the fact that Paget cells show an apocrine differentiation, as suggested by the coexpression of markers of apocrine gland cells, such as keratin 7 and GCDFP-15.13,14 Histologically, both mammary and extramammary Paget disease should be differentiated from other epidermotropic neoplasms with clear cells, such as pagetoid Bowen disease, pagetoid melanoma, and sebaceous carcinoma, to name but a few.16 The usefulness of PRAME expression in the differential diagnosis of mammary and extramammary Paget disease from these mimics seems limited because some of the latter (namely sebaceous tumors and melanomas) also express PRAME (although in a nuclear pattern in the case of melanocytic tumors, contrasting with the predominantly cytoplasmic expression pattern in extramammary Paget disease). However, the very frequent expression of PRAME in cases of mammary and extramammary Paget disease should be known by dermatopathologists who use PRAME as a diagnostic biomarker. Future studies should address the usefulness of PRAME expression in the differential diagnosis of primary (purely cutaneous) from secondary (associated with underlying malignancies) cases of extramammary Paget disease. The prognostic value of PRAME expression in mammary Paget disease merits also further study, considering the fact that the percentage of PRAME-positive tumor cells in breast carcinomas has been shown to inversely correlate with disease-free survival.17