Primary cutaneous perivascular epithelioid cell tumor: A clinicopathological and molecular reappraisal Yann Charli-Joseph, MD,a,b Andrea Saggini, MD,a,b Swapna Vemula, MS,a Jingly Weier, PhD,a Sonia Mirza, MBBS,a and Philip E. LeBoit, MDa,b,c San Francisco, California Background: Perivascular epithelioid cell tumor (PEComa) is a rare neoplasm of uncertain histogenesis with a mixed myomelanocytic immunophenotype, rarely arising in the skin (primary cutaneous PEComa [pcPEComa]). Objective: We analyzed the clinicopathological features of 8 pcPEComas, assayed for DNA copy number changes and for initiating mutations common in melanocytic neoplasms. Methods: pcPEComas were evaluated using immunohistochemistry, comparative genomic hybridization, and DNA sequencing. Results: pcPEComas were erythematous nodules, mostly in the lower extremities of women (5/8), composed of large pale-staining epithelioid cells. The patient’s age range was 26 to 67 (mean 46) years. The percentages of tumors staining positively were as follows: micro-ophthalmia-associated transcription factor, NKI/C3, bcl-1, E-cadherin, and cathepsin K (100%); HMB-45, 4E-binding protein 1, and CD68 (88%); smooth muscle actin and muscle-specific actin (40%); S100 (38%); calponin (20%); desmin (13%); and melan-A, SOX10, and keratin (0%). No chromosomal copy number changes or initiating mutations were identified. Limitations: Small sample size is a limitation. Conclusions: pcPEComas have a different molecular signature than extracutaneous tumors and are unrelated to tuberous sclerosis. However, the common expression of 4E-binding protein 1 points to a role of the mTOR pathway in their pathogenesis. Because pcPEComas are diagnostically challenging, we propose that micro-ophthalmia-associated transcription factor, NKIC3, smooth muscle actin, desmin, bcl-1, cathepsin K, and 4E-binding protein 1 can be used when evaluating a possible pcPEComa. ( J Am Acad Dermatol 2014;71:1127-36.) Key words: array-based comparative genomic hybridization; cutaneous clear cell myomelanocytic tumor; initiating mutations; mTOR pathway; perivascular epithelioid cell tumor.
any cutaneous mesenchymal tumors show, as can melanocytic neoplasms, clear cell morphology, including adipocytic, xanthomatous, myoepithelial, and fibrohistiocytic proliferations.1 Among these is an exceedingly rare dermal neoplasm, composed of a perivascular and diffuse proliferation of large and pale epithelioid cells, with a mixed myomelanocytic
M
immunophenotype. This neoplasm, originally designated ‘‘cutaneous clear cell myomelanocytic tumor,’’ was subsequently included in the family of perivascular epithelioid cell tumors (PEComas) that usually affect internal organs, and now are commonly termed ‘‘cutaneous perivascular epithelioid cell tumors’’ (primary cutaneous PEComas [pcPEComas]).2 The precise lineage of
From Pathology,a Dermatology,b and Helen A. Diller Family Comprehensive Cancer Center,c University of California San Francisco. Funding sources: None. Conflicts of interest: None declared. Accepted for publication August 10, 2014. Reprint requests: Philip E. LeBoit, MD, Dermatopathology Service, University of California San Francisco, 1701 Divisadero
St, Suite 280, San Francisco, CA 94115. E-mail: philip.leboit@ ucsf.edu. Published online September 26, 2014. 0190-9622/$36.00 Ó 2014 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2014.08.016
1127
J AM ACAD DERMATOL
1128 Charli-Joseph et al
DECEMBER 2014
this enigmatic tumor remains elusive because were purified using ExoSAP-IT (USB Corp, the proposed perivascular epithelioid cell of Cleveland, OH) and sequenced directly using a 3500 origin awaits proper identification and previous DNA Analyzer (Applied Biosystems, Foster City, CA). immunohistochemical staining results are inconThe sequences were analyzed using Mutation sistent. Furthermore, the relationship between Surveyor software (Softgenetics, State College, PA). extracutaneous PEComas and pcPEComas is The Committee on Human Research from the uncertain.3 In this study we describe a substantial University of California San Francisco granted series of these tumors, approval for this study to further characterize their originally titled ‘‘Histologic, CAPSULE SUMMARY clinicopathological and immunohistochemical, and immunohistochemical feagenetic study on cutaneous Cutaneous perivascular epithelioid cell tures. Finally, via arrayneoplasms’’ under the tumors are tumors with a mixed based comparative genomic study number 11-05569. myomelanocytic immunophenotype. hybridization (aCGH) and This covers archival paraffin Cutaneous perivascular epithelioid cell DNA sequencing, we blocks. tumors lack gross chromosomal assayed for initiating mutaaberrations and initiating mutations tions commonly observed in RESULTS common to melanocytic tumors whereas melanocytic neoplasms and Clinical features 4E-binding protein 1 expression DNA copy number changes Clinical findings are sumplausibly relates them to the mTOR as often seen in melanoma. marized in Table II. pathway. pcPEComas occurred in 5 METHODS An immunohistochemical panel women and 3 men. The Eight pcPEComas were including micro-ophthalmia-associated mean age of patients at the identified in the author’s (P. transcription factor, NKIC3, smooth time of presentation was E. L.) consultation files. muscle actin, desmin, bcl-1, cathepsin K, 46 years. pcPEComas Clinical information was oband 4E-binding protein 1 is predominated in the lower tained through contributing recommended for their evaluation. extremities (5/8). Most were clinicians. In each case, described as indurated formalin-fixed paraffinnodules (Fig 1). In no case embedded tissue sections had been stained was the correct diagnosis inferred clinically; the most with hematoxylin-eosin. We included only common impression being dermatofibroma (3/8). lesions centered in the dermis that fulfilled the None of the patients had history of tuberous sclerosis presently accepted criteria for PEComa.3 A broad complex (TSC). Follow-up data were available for 6 panel of immunohistochemical stains was perof 8 patients and ranged between 24 and 171 (mean formed on formalin-fixed paraffin-embedded 66.5) months. No local recurrences or metastases tissue. Antibodies, clones, dilutions, and sources developed in these cases. are outlined in Table I. Appropriate controls were used in parallel. Histopathologic and immunohistochemical features Array-based comparative genomic The histologic size of pcPEComas ranged from 4 hybridization to 20 mm. Neoplasms spared the epidermis and DNA was extracted from formalin-fixed paraffinoccupied the entire dermis except for a thin layer of embedded tissue as described previously.4,5 aCGH uninvolved papillary dermis (grenz zone) (5/8), or was carried out with 500 to 1000 ng of genomic DNA occurred exclusively in the reticular dermis (3/8). A on Agilent Human 4x180k oligomicroarrays diffuse, highly cellular and sheetlike growth pattern (G4449A, Agilent Technologies, Santa Clara, CA). was observed in 4 cases, the rest displayed less aCGH methodology details and data analysis were as cellularity and lesional cells arranged as thin described in a recent publication.6 aggregates, strands and singly in an interstitial array with perivascular and periadnexal accentuation (Fig 2). The interface between the tumors and the DNA sequencing surrounding tissues was well circumscribed. In all Polymerase chain reaction assays using primers cases, lateral borders were irregular and cells were specific for BRAF exon 15; NRAS exons 1 and 2; KIT interposed between thickened collagen bundles exons 11, 13, 17, and 18; GNAQ exon 5; and GNA11 (Fig 3). A similar configuration was observed in the exon 5 were used to amplify DNA as described deep margin in 2 cases where lesional cells did not previously.7 Polymerase chain reaction products d
d
d
J AM ACAD DERMATOL
Charli-Joseph et al 1129
VOLUME 71, NUMBER 6
Abbreviations used: aCGH: MiTF: pcPEComa: PEComa: SMA: TSC: 4EBP1:
array-based comparative genomic hybridization micro-ophthalmia-associated transcription factor primary cutaneous perivascular epithelioid cell tumor perivascular epithelioid cell tumor smooth muscle actin tuberous sclerosis complex 4E-binding protein 1
extend into the subcutaneous tissue. Six cases showed tumoral extension into the subcutis, and the deep border was smooth and rounded (Fig 2); 2 cases also showed a focal lacelike extension between adipocytes. Lesional cells were large (2-4 times the size of a spinous-layer keratinocyte) and oval to polygonal, except 1 case that displayed spindled morphology. Cells were optically clear with pale and coarsely vacuolated cytoplasm (slightly granular in 2 cases) containing a large central vesicular nucleus (1-3 times the size of a mid-layer epidermal keratinocyte nucleus) with, in most, a single central nucleolus (Fig 4). Cells intermediate between typical PEComa cells and adipocytes were absent. The degree of nuclear pleomorphism was negligible and a few had 2 nucleoli, some eosinophilic. Mitotic activity was low and ranged from 0 to 2 mitoses per 10 high-power fields (mean \1/10). No abnormal mitoses were identified. Infiltrating lymphocytes were present in 5 cases, in a perivascular distribution. Thickened collagen fibers were observed in close apposition to lesional cells in every case, and a prominent vasculature of thin-walled delicate capillaries. The rosetting of lesional cells in and around blood vessels in a radial pattern sometimes seen in visceral PEComas was not identified, nor were hyalinized vessel walls found. The overlying epidermis was unaltered in 5 cases, but central attenuation was present in 2 whereas 1 showed slight acanthosis. The immunohistochemical staining pattern of pcPEComas is summarized in Table III and illustrated in Figs 5 and 6. All tumors strongly and diffusely expressed micro-ophthalmia-associated transcription factor (MiTF), NKI/C3, bcl-1, and cathepsin K. E-cadherin immunoreactivity displayed a membranous pattern in all with lack of cytoplasmic labeling. Most pcPEComas (7/8) expressed 4E-binding protein 1 (4EBP1); CD68 stained neoplastic cells avidly in 4 of 7 and faintly in 3 of 7 cases. S100 reactivity was observed in 7 cases, but was fainter than in the appropriate internal controls
(melanocytes). However, after digestion with diastase only 3 of 5 cases retained cytoplasmic immunoreactivity, albeit with considerably lower intensity than before digestion. A blush of HMB-45 positivity appeared in very few of the lesional cells in most tumors (7/8) whereas antibodies to melan-A failed to stain lesional cells. Staining for smooth muscle actin (SMA) was limited to a small number of cells in immediate apposition to thin-walled capillaries (\5% of tumoral volume). One of the tumors was positive for muscle-specific actin, another for desmin, whereas only 1 was faintly labeled with calponin. In all, 20% to 30% of lesional cells in 3 of 8 cases demonstrated nuclear staining for c-myc; few cells were positive for p16 in 3 of 6. Tumors were negative for keratin cocktail, CD56, and SOX10. Proliferation rate, assessed by Ki67 reactivity, was low (\5%) except for 1 case that displayed a 30% proliferation rate. Genomic features None of the cases showed chromosomal imbalances (aCGH analysis) or initiating mutations because DNA sequencing demonstrated only wild-type alleles for BRAF, NRAS, KIT, GNAQ, and GNA11.
DISCUSSION The term ‘‘perivascular epithelioid cell tumor’’ was first used in 1996 when authors describing a pancreatic tumor noted striking similarities with visceral tumors previously reported under several terms (eg, ‘‘clear cell sugar tumor,’’ ‘‘angioleiomyoma,’’ ‘‘lymphangioleiomyoma’’).8 All these neoplasms were composed at least partly by large epithelioid to spindle cells with clear cytoplasm present in close apposition to thin vascular structures, and commonly labeled with both melanocytic markers and myoid immunoreactants.9 To date both benign and less commonly malignant tumors, considered to fall under the unifying PEComa designation, have been reported at multiple anatomic sites and interestingly a subset arise in the setting of TSC.10,11 pcPEComas represent 8% of reported cases within the family of PEComas, but may be somewhat more common as the dermatology and dermatopathology literature has seen only scant reports of pcPEComa, unlike the many reports in the pathology literature, given the predilection for noncutaneous sites.3 Including our cases, 34 legitimate pcPEComas have thus far been described.2,12-20 All but 1 behaved in an indolent fashion and most appear to display similar morphologic attributes on histology.2,12-20 The differential diagnosis posed by a pandermal proliferation of vacuolated cells, discussed in detail
J AM ACAD DERMATOL
1130 Charli-Joseph et al
DECEMBER 2014
Table I. Immunohistochemical reagents used in this study Antigen
S100 MiTF Melan-A HMB-45 SOX10 SMA MSA Desmin Calponin Keratin cocktail NKI/C3 CD68 CD56 BCL-1 p16 4EBP1 CatK E-cadherin c-Myc Ki67
Clone
Dilution
Source
Polyclonal D5 A103 HMB-45 N-20 1A4 HHF45 D33 CALP AE1/AE315D3 NKI/C3 PG-M1 IB6 EP12 ink4a Thr37/46 Polyclonal NCH-38 Y69 Polyclonal
1:20,000 1:100 1:200 1:200 1:100 1:100 1:200 1:25 1:50 Prediluted 1:4 1:200 Prediluted 1:200 1:10 1:1000 1:50
Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Santa Cruz Biotechnology, Dallas, TX Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA BioCare, Concord, CA BioGenex, Fremont, CA Dako, Carpenteria, CA Leica, Richmond, IL Dako, Carpenteria, CA Ventana, Tucson, AZ Cell Signaling, Danvers, MA abcam, Cambridge, MA Dako, Carpenteria, CA Epitomics, Burlingame, CA Dako, Carpenteria, CA
1:50 1:75
BCL-1, B-cell lymphoma-1; CatK, cathepsin K; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; NKI/C3, anti-melanoma-associated antigen; SMA, smooth muscle actin; SOX10, Sry-related HMB box-10; 4EBP1, 4Ebinding protein 1.
Table II. Clinical features of primary cutaneous perivascular epithelioid cell tumors Case
Gender/age, y
Location
Clinical description, impression
Treatment
Follow-up, duration [mo]
1 2 3
M/60 F/67 M/59
Thigh Thigh Lower aspect of leg
Complete excision Complete excision Complete excision
NED [64] NED [72] NED [171] DUE
4 5 6 7 8
F/37 F/32 M/45 F/26 F/47
Arm Back Abdomen Leg Lower aspect of leg
DF Indurated papule, R/O DF 1-cm Tender nodule, R/O neuroma vs glomus tumor 0.5-cm Pink-brown papule Lesion, details NA 1-cm Nodule DF DF
Marginal excision Complete excision Complete excision Complete excision Complete excision
NA NED [24] NED [41] NA NED [27]
DF, Dermatofibroma; DUE, died of an unrelated event; F, female; M, male; NA, not available; NED, no evidence of disease; R/O, rule-out.
in previous articles, includes both melanocytic and nonmelanocytic neoplasms.1 Melanocytic tumors most likely to be mistaken for pcPEComa include clear cell sarcoma, hypomelanotic blue nevus, balloon cell nevus, and melanoma. Nonmelanocytic neoplasms that invite confusion with pcPEComa include xanthomas, xanthogranulomas, clear cell angiofibroma, and dermatofibroma, among others.1 The reported immunophenotypes in visceral and pcPEComas vary considerably, both between these groups and between studies (Table IV). In our series, among melanocytic markers studied, MiTF (melanocytic marker also positive in some histiocytes and Schwannian lesions)24 and NKIC3 (initially described for melanoma but later found to stain a broad spectrum of neoplastic tissues such as neural
and histiocytic proliferations)25 showed strong and diffuse expression in all cases, coincident with previous reports. Likewise, HMB-45 stained most cases (88% vs 96% in previous reports) but positivity was faint and observed in a minority of lesional cells. This contrasts with visceral PEComas that are commonly positive in more of 50% of neoplastic cells.26,27 S100 positivity has been reported in up to 33% of visceral PEComas.27 Seven of 8 cases in our series showed cytoplasmic labeling with S100, though fainter than that of appropriate internal controls, and only 3 of 5 cases retained some positivity after digestion with diastase. The residual staining was at a lower intensity than before digestion. Positivity may thus stem, at least in part, from nonspecific binding of the immunoreactant to
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
Fig 1. Cutaneous perivascular epithelioid cell tumor. Nonspecific 1-cm erythematous nodule on the lower aspect of abdomen.
Fig 2. Cutaneous perivascular epithelioid cell tumor. The low-power magnification of this lesion shows a dome-shaped, diffuse, and highly cellular proliferation of large and pale epithelioid cells with a sheetlike growth pattern involving the entirety of the dermis. (Hematoxylineosin stain; original magnification: 320.) Inset: The base of the neoplasm shows a well-circumscribed pushing border with a dumb-shaped configuration impinging upon the subcutaneous tissue.
intracytoplasmic glycogen. This nonspecific staining with S100 has been observed in skeletal muscle (oral communication, Daniel Santa Cruz, MD, April 1993), which gave us the impetus to apply diastase digestion to our cases. Melan-A was consistently negative in our cases, and although this finding is lower than average in pcPEComas (approximately 50%), others have found percentages as low as 14%. Thus, we believe this stain should not be relied on for their identification.26,27 Our results reproduce the report of Llamas-Velasco et al19 on the lack of SOX10 expression in pcPEComas. Among muscle markers, SMA and desmin are reportedly the most sensitive to label pcPEComas (28% and 36% positive immunoreactivity, respectively). Eighty eight percent
Charli-Joseph et al 1131
Fig 3. Cutaneous perivascular epithelioid cell tumor. The peripheral aspects of the neoplasm show pale-staining epithelioid and spindle cells interposed between markedly thickened collagen bundles. (Hematoxylin-eosin stain; original magnification: 340.)
Fig 4. Cutaneous perivascular epithelioid cell tumor. This high-power magnification demonstrates that neoplastic cells are large, are oval to polygonal, have pale and coarsely vacuolated cytoplasm, have a negligible degree of nuclear pleomorphism, and are arrayed in close apposition to thin-walled capillaries. (Hematoxylin-eosin stain; original magnification: 3400.)
of our cases showed SMA positivity, limited to less than 5% of cells in immediate apposition to thinwalled capillaries. It is not clear whether these are admixed pericytes or true neoplastic cells. Desmin was positive in only 1 case whereas muscle-specific actin and calponin signals were variable and unreliable. CD68 immunopositivity was present among neoplastic cells in 88% of our cases, in contrast to 45% in previous reports.2,14,15,20 Histologic criteria for establishing malignancy have been proposed for visceral PEComas and suggested to evaluate pcPEComas.16,28 None of our cases fulfilled such criteria because they were small, had a virtually null mitotic count, and lacked necrosis or vascular invasion. In addition, staining with Ki67 (a cell proliferation marker) showed a very low labeling rate (\5%) in all but 1 case (that showed no increase in mitoses nor cytologically atypia).
J AM ACAD DERMATOL
1132 Charli-Joseph et al
DECEMBER 2014
Table III. Detailed immunohistochemical findings in the current study Case
S100 S100-D MiTF Melan-A HMB-45 SOX10 SMA MSA Desmin Calponin Keratin cocktail NKIC3 CD68 CD56 BCL-1 p16 4EBP1 CatK E-cadherin c-Myc Ki67
1
2
3
4
5
6
7
8
Faint 1 Focal 1 Focal 1 1 1 1 1 1 Focal 1 ;30%
1 Faint 1 Focal 1 Focal 1 1 1 Faint 1 Few cells 1 1 1 1 \5%
1 Faint 1 Focal 1 1 1 Focal 1 Few cells 1 1 1 1 \5%
Faint ND 1 Focal 1 Focal 1 ND ND 1 Faint ND 1 ND 1 1 1 Few cells 1 \5%
ND 1 Focal 1 Focal 1 ND ND ND 1 ND ND 1 ND 1 1 1 Few cells 1 \5%
Faint ND 1 Focal 1 Focal 1 ND ND ND 1 1 ND 1 ND 1 1 \5%
1 1 Focal 1 ND Focal 1 Faint 1 1 1 1 1 1 \5%
1 Faint 1 Focal 1 Focal 1 Focal 1 1 Faint 1 Few cells 1 1 1 1 Few cells 1 \5%
BCL-1, B-cell lymphoma-1; CatK, cathepsin K; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; ND, not done; NKI/C3, anti-melanoma-associated antigen; S100-D, S100 after digestion with diastase; SMA, smooth muscle actin; SOX10, Sry-related HMB box-10; 4EBP1, 4E-binding protein 1.
Fig 5. Cutaneous perivascular epithelioid cell tumor. A, Faint cytoplasmic reactivity. B, After digestion with diastase, the degree of cytoplasmic immunoreactivity is weaker than before digestion. C, Faint and focal cytoplasmic immunopositivity. D, Strong nuclear reactivity among all neoplastic cells. E, Strong reactivity among all neoplastic cells. F, Only few cells in immediate apposition to thin-walled capillaries show convincing cytoplasmic reactivity. (Immunohistochemistry with S100 [A and B], HMB-45 [C], micro-ophthalmia-associated transcription factor [D], NKIC3 [E], and smooth muscle actin [F]; original magnifications: A, B, D, and E, 3400; C and F, 3600.)
Bcl-1 (cyclin D1) is an important regulator of the G1 checkpoint of cell cycle and is overexpressed in mantle cell lymphomas.29 Bcl-1 avidly labels neoplastic cells in certain malignant extracutaneous
PEComas and was suggested to participate in the pathogenesis of malignant transformation.30,31 Our cases, which have thus far behaved in an indolent fashion, strongly labeled with bcl-1 suggesting that
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
Charli-Joseph et al 1133
Fig 6. Cutaneous perivascular epithelioid cell tumor. A, Strong nuclear reactivity is evident among all neoplastic cells. B, Avid nuclear reactivity among all neoplastic cells. C, Immunopositivity is shown in a membranous pattern with lack of cytoplasmic labeling. D, Strong cytoplasmic immunoreactivity among all lesional cells. (Immunohistochemistry with BCL-1 [A], 4E-binding protein 1 [B], E-cadherin [C], and cathepsin K [D]; original magnifications: A to D, 3400.)
although its overexpression may play a role in pcPEComas it is not a marker of malignant progression, and did not correspond to increased copy numbers of chromosome 11q, as it does in some malignancies.32 The pathogenesis of pcPEComas remains elusive. Among extracutaneous cases a subset arises predominantly in the setting of TSC (genetic disease associated with dysregulation of the mTOR pathway) or are associated with gene rearrangements in the transcription factor E3 (a member of MiTF family).10,11,33 Moreover, a frequent allelic alteration on chromosome 16p (where the TSC2 gene is located) occurs in both extracutaneous TSC-related and eunrelated PEComas.34 However, there is no report of a cutaneous case associated with TSC and a recent study failed to identify immunoreactivity or gene rearrangements for transcription factor E3 in cutaneous lesions.19 Similarly, our genomic analysis revealed that gross chromosomal aberrations were not responsible for the development of pcPEComas and that unlike bona fide melanocytic neoplasms, initiating mutations in BRAF, NRAS, KIT, GNAQ, and GNA11 did not play a role in their pathogenesis. Recently, kinase gene fusions
have been found in the majority of Spitz nevi, in the absence of such initiating mutations. These are sometimes accompanied by DNA copy number changes in the responsible kinase gene, or its fusion partner.35 However, the absence of copy number changes in our cases makes this mechanism unlikely. The molecular disparity between visceral and pcPEComas is intriguing but could relate to an incorrect lumping of these based on pathological features. The fact that cutaneous cases have not been associated with TSC and that they lack chromosomal aberrations and transcription factor E3 rearrangements, and the differences noted in immunophenotyping, support this hypothesis. Another possibility, however, is that the threshold used to analyze the comparative genomic hybridization data in extracutaneous PEComas was too narrow, leading to an overinterpretation of deviations in the comparative genomic hybridization curve. Independently of this possibility, the analytical software used differs and hence, an accurate conclusion for such discrepancies requires comparatively reviewing the raw data obtained in each study.
0/7 0/2 1/10 0/1 ND/NR 0/1 ND/NR 0/1 3/8 33 12
2005 2007 2008 2008 2010 2010 2013 2013 2014
S100
0/1 0/1
2003 2005 7/7 2/2 10/10 1/1 0/1 1/1 ND/NR 1/1 7/8 33 94
1/1 1/1
HMB-45
1/7 2/2 5/7 1/1 1/1 ND/NR ND/NR 0/1 0/7 26 35
ND/NR ND/NR
Melan-A
MiTF
7/7 ND/NR 5/5 ND/NR 1/1 1/1 ND/NR 1/1 8/8 23 100
ND/NR ND/NR
SOX10
ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR 0/5 ND/NR 0/7 12 0
ND/NR ND/NR
NKIC3
6/6 ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR 8/8 14 100
ND/NR ND/NR
SMA
1/7 1/2 1/10 1/1 1/1 0/1 ND/NR 0/1 7/8 33 42
1/1 1/1
MSA
0/6 ND/NR ND/NR ND/NR ND/NR 0/1 ND/NR ND/NR 2/5 12 17
ND/NR ND/NR 1/6 0/2 5/10 ND/NR 1/1 1/1 ND/NR 0/1 1/8 33 30
0/1 1/1
Desmin
2/4 ND/NR 0/10 ND/NR ND/NR ND/NR ND/NR 0/1 1/5 20 15
ND/NR ND/NR
Calponin
ND/NR ND/NR 0/10 ND/NR 1/1 ND/NR ND/NR 0/1 ND 13 8
ND/NR 0/1
Caldesmon
CK
0/7 0/2 0/10 0/1 0/1 0/1 ND/NR 0/1 0/6 31 0
0/1 0/1
CD68
5/7 2/2 2/10 ND/NR ND/NR ND/NR ND/NR 0/1 7/8 28 57
ND/NR ND/NR
EMA
ND/NR 0/2 0/10 0/1 0/1 ND/NR ND/NR ND/NR ND 14 0
ND/NR ND/NR
This literature review does not include the cases described by Folpe et al,21 Diment and Colecchia,22 or Harris et al23 because it was not specified whether or not these soft-tissue tumors involved cutaneous structures. CK, Keratin cocktail; EMA, epithelial membrane antigen; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; ND, not done; NKI/C3, anti-melanoma-associated antigen; NR, not reported; SOX10, Sry-related HMB box-10; SMA, smooth muscle actin.
Crowson et al12 de Saint Aubain Somerhausen et all3 Mentzel et al2 Tan et al14 Liegl et al15 Calder et al16 Ghazali et al17 Chaplin et al18 Llamas-Velasco et al19 Ieremia and Robson20 Current study Total cases assessed Positive cases, %
Table IV. Immunohistochemical staining pattern in primary cutaneous perivascular epithelioid cell tumors
1134 Charli-Joseph et al J AM ACAD DERMATOL DECEMBER 2014
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
4EBP1 is a downstream target in the mTOR pathway and its phosphorylation status (which controls messenger RNA translation and cellular proliferation) can be evaluated via immunohistochemistry.36 However, to our knowledge this antibody has not been previously evaluated in PEComas. In this study we found robust expression of 4EBP1 in all but 1 case, suggesting that activation of the mTOR pathway is indeed present in pcPEComas independently on the mutational status of TSC1/TSC2. This result is congruent to the finding that the mTOR pathway is activated via positive immunohistochemical staining in other PEComas for phospho-S6.37 E-cadherin (membrane-associated glycoprotein that mediates cellular adhesion) has been shown to display abnormal cytoplasmic immunolocalization in some extracutaneous PEComas, theoretically as a result of TSC2 gene loss.38,39 We found normal membranous immunoreactivity and absence of intracytoplasmic labeling in all cases, in keeping with the fact that none of our patients had evidence of TSC. Cathepsin K (transcriptional target of the MiTF family) is expressed in PEComas at various anatomic sites40 but has not yet been reported in cutaneous lesions. We detected strong positive reactions with cathepsin K antibody in all our cases, supporting it as a powerful marker for the identification of PEComas independently of location. This result reinforces the notion that PEComas represent a new member of the MiTF family of tumors. C-Myc is a gene involved in cell growth, apoptosis, and metabolism. Changes in its configuration or expression have been linked to a variety of human cancers because its dysregulation induces genomic instability.41 However, its expression profile in PEComas is unknown. We found that only 20% to 30% of the cells in 3 of 8 tumors expressed c-myc, a low percentage that plausibly correlates with the lack of genomic instability observed in our aCGH analysis. In summary, pcPEComas enter into the clinical differential diagnosis of nonspecific erythematous nodules and dermatofibroma, but are an exotically rare explanation for them. Our results provide additional evidence that pcPEComas have a different molecular signature from that of extracutaneous cases and are not associated with TSC. However, the cumulative immunohistochemical evidence points to at least a partial role of the mTOR pathway in the pathogenesis of these neoplasms, which are conceivably also related to the MiTF family of tumors. pcPEComas have a very consistent appearance in hematoxylin-eosinestained sections,
Charli-Joseph et al 1135
enabling their diagnosis but only for those familiar with the entity. We propose that a panel of immunoperoxidase stains, including MiTF, NKIC3, SMA, desmin, bcl-1, cathepsin K, and 4EBP1, should be used when evaluating a possible pcPEComa. The authors would like to thank Dr Carlo Cota for providing valuable information along with Kazuko Olson and Mary-Kate Fitzsimon for their help in various aspects of the research process. REFERENCES 1. Biswas A, Mahalingam M. Cutaneous clear cell neoplasms: a histopathological reappraisal. Am J Dermatopathol 2012;34: 237-54. 2. Mentzel T, Reisshauer S, R€ utten A, Hantschke M, Soares de Almeida LM, Kutzner H. Cutaneous clear cell myomelanocytic tumor: a new member of the growing family of perivascular epithelioid cell tumors (PEComas). Clinicopathological and immunohistochemical analysis of seven cases. Histopathology 2005;46:498-504. 3. Walsh SN, Sang€ ueza OP. PEComas: a review with emphasis on cutaneous lesions. Semin Diagn Pathol 2009;26:123-30. 4. Bastian BC, Olshen AB, LeBoit PE, Pinkel D. Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol 2003;16:1765-70. 5. Jalas JR, Vemula S, Bezrookove V, Leboit PE, Simko JP, Bastian BC. Metastatic melanoma with striking adenocarcinomatous differentiation illustrating phenotypic plasticity in melanoma. Am J Surg Pathol 2011;35:1413-8. 6. Charli-Joseph Y, Saggini A, Doyle L, Fletcher CD, Weier J, Mirza S, et al. DNA copy number changes in tumors within the spectrum of cellular, atypical and metastasizing fibrous histiocytoma. J Am Acad Dermatol 2014;71:256-63. 7. Langer R, Becker K, Feith M, Friess H, H€ ofler H, Keller G. Genetic aberrations in primary esophageal melanomas: molecular analysis of c-KIT, PDGFR, KRAS, NRAS and BRAF in a series of 10 cases. Mod Pathol 2011;24:495-501. 8. Zamboni G, Pea M, Martignoni G, Zancanaro C, Faccioli G, Gilioli E, et al. Clear cell ‘‘sugar’’ tumor of the pancreas: a novel member of the family of lesions characterized by the presence of perivascular epithelioid cells. Am J Surg Pathol 1996;20: 722-30. 9. Hornick JL, Fletcher CD. PEComa: what do we know so far? Histopathology 2006;48:75-82. 10. Moss J, Avila NA, Barnes PM, Litzenberger RA, Bechtle J, Brooks PG, et al. Prevalence and clinical characteristics of lymphangioleiomyomatosis (LAM) in patients with tuberous sclerosis complex. Am J Respir Crit Care Med 2001;164:669-71. 11. Pan CC, Jong YJ, Chai CY, Huang SH, Chen YJ. Comparative genomic hybridization study of perivascular epithelioid cell tumor: molecular genetic evidence of perivascular epithelioid cell tumor as a distinctive neoplasm. Hum Pathol 2006;37: 606-12. 12. Crowson AN, Taylor JR, Magro CM. Cutaneous clear cell myomelanocytic tumor-perivascular epithelioid cell tumor: first reported case. Mod Pathol 2003;16:400A. 13. de Saint Aubain Somerhausen N, Gomez Galdon M, Bouffioux B, Courtin C, Theunis A, Vogeleer MN, et al. Clear cell ‘‘sugar’’ tumor (PEComa) of the skin: a case report. J Cutan Pathol 2005;32:441-4. 14. Tan J, Peach AH, Merchant W. PEComas of the skin: more common in the lower limb? Two case reports. Histopathology 2007;51:135-6.
J AM ACAD DERMATOL
1136 Charli-Joseph et al
15. Liegl B, Hornick JL, Fletcher CD. Primary cutaneous PEComa: distinctive clear cell lesions of skin. Am J Surg Pathol 2008;32: 608-14. 16. Calder KB, Schlauder S, Morgan MB. Malignant perivascular epithelioid cell tumor (‘‘PEComa’’): a case report and literature review of cutaneous/subcutaneous presentations. J Cutan Pathol 2008;35:499-503. 17. Ghazali N, Cascarini L, Norris P, Barrett AW, Lavery KM. Perivascular epithelioid cell tumor (PEComa) of the cheek. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010;110: e26-31. 18. Chaplin A, Conrad DM, Tatlidil C, Jollimore J, Walsh N, Covert A, et al. Primary cutaneous PEComa. Am J Dermatopathol 2010;32:310-2. 19. Llamas-Velasco M, Mentzel T, Requena L, Palmedo G, Kasten R, Kutzner H. Cutaneous PEComa does not harbor TFE3 gene fusions: immunohistochemical and molecular study of 17 cases. Histopathology 2013;63:122-9. 20. Ieremia E, Robson A. Cutaneous PEComa: a rare entity to consider in an unusual site. Am J Dermatopathol. http://dx.doi.org/10. 1097/DAD.0000000000000041. Published online December 20, 2013. 21. Folpe AL, McKenney JK, Li Z, Smith SJ, Weiss SW. Clear cell myomelanocytic tumor of the thigh: report of a unique case. Am J Surg Pathol 2002;26:809-12. 22. Diment J, Colecchia M. Myomelanocytic tumor of the thigh. Am J Surg Pathol 2003;27:1288. 23. Harris GC, McCulloch TA, Perks G, Fisher C. Malignant perivascular epithelioid cell tumor (‘‘PEComa’’) of soft tissue: a unique case. Am J Surg Pathol 2004;28:1655-8. ~ez NG. Value of melanocytic-associated immuno24. Ord on histochemical markers in the diagnosis of malignant melanoma: a review and update. Hum Pathol 2014;45: 191-205. 25. Orchard GE. Comparison of immunohistochemical labeling of melanocyte differentiation antibodies melan-A, tyrosinase and HMB 45 with NKIC3 and S100 protein in the evaluation of benign nevi and malignant melanoma. Histochem J 2000;32: 475-81. 26. Schoolmeester JK, Howitt BE, Hirsch MS, Dal Cin P, Quade BJ, Nucci MR. Perivascular epithelioid cell neoplasm (PEComa) of the gynecologic tract: clinicopathologic and immunohistochemical characterization of 16 cases. Am J Surg Pathol 2014;38:176-88. 27. Folpe AL, Mentzel T, Lehr HA, Fisher C, Balzer BL, Weiss SW. Perivascular epithelioid cell neoplasms of soft tissue and
DECEMBER 2014
28.
29. 30.
31.
32.
33.
34.
35.
36. 37. 38.
39. 40.
41.
gynecologic origin: a clinicopathologic study of 26 cases and review of the literature. Am J Surg Pathol 2005;29:1558-75. Folpe AL. Neoplasms with perivascular epithelioid differentiation (PEComas). In: Fletcher CDM, Unni KK, Mertens F, editors. WHO classification of tumors: tumors of soft tissue and bone. Lyon (France): IARC Press; 2002. pp. 221-2. Pestell RG. New roles of cyclin D1. Am J Pathol 2013;183:3-9. Weinreb I, Howarth D, Latta E, Ghazarian D, Chetty R. Perivascular epithelioid cell neoplasms (PEComas): four malignant cases expanding the histopathological spectrum and a description of a unique finding. Virchows Arch 2007;450: 463-70. Yamamoto H, Oda Y, Yao T, Oiwa T, Kobayashi C, Tamiya S, et al. Malignant perivascular epithelioid cell tumor of the colon: report of a case with molecular analysis. Pathol Int 2006; 56:46-50. Hawryluk EB, Baran JL, Gerami P, Sepehr A. ‘‘Monster cell’’ melanoma with pulmonary metastasis and cyclin D1 amplification. J Cutan Pathol 2013;40:61-5. Argani P, Aulmann S, Illei PB, Netto GJ, Ro J, Cho HY, et al. A distinctive subset of PEComas harbors TFE3 gene fusions. Am J Surg Pathol 2010;34:1395-406. Pan CC, Chung MY, Ng KF, Liu CY, Wang JS, Chai CY, et al. Constant allelic alteration on chromosome 16p (TSC2 gene) in perivascular epithelioid cell tumor (PEComa): genetic evidence for the relationship of PEComa with angiomyolipoma. J Pathol 2008;214:387-93. Wiesner T, He J, Yelensky R, Esteve-Puig R, Botton T, Yeh I, et al. Kinase fusions are frequent in Spitz tumors and Spitzoid melanomas. Nat Commun 2014;5:3116. Guertin DA, Sabatini DM. Defining the role of mTOR in cancer. Cancer Cell 2007;12:9-22. Folpe AL, Kwiatkowski DJ. Perivascular epithelioid cell neoplasms: pathology and pathogenesis. Hum Pathol 2010;41:1-15. Barnes EA, Kenerson HL, Jiang X, Yeung RS. Tuberin regulates E-cadherin localization: implications in epithelial-mesenchymal transition. Am J Pathol 2010;177:1765-78. Wang Z, Gong Q, Fan Q. Expression of E-cadherin in angiomyolipoma. Hum Pathol 2012;43:2348-53. Rao Q, Cheng L, Xia QY, Liu B, Li L, Shi QL, et al. Cathepsin K expression in a wide spectrum of perivascular epithelioid cell neoplasms (PEComas): a clinicopathological study emphasizing extrarenal PEComas. Histopathology 2013;62:642-50. Bretones G, Delgado MD, Le on J. Myc and cell cycle control. Biochim Biophys Acta. http://dx.doi.org/10.1016/j.bbagrm.2014. 03.013. Published online April 1, 2014.
any cutaneous mesenchymal tumors show, as can melanocytic neoplasms, clear cell morphology, including adipocytic, xanthomatous, myoepithelial, and fibrohistiocytic proliferations.1 Among these is an exceedingly rare dermal neoplasm, composed of a perivascular and diffuse proliferation of large and pale epithelioid cells, with a mixed myomelanocytic
M
immunophenotype. This neoplasm, originally designated ‘‘cutaneous clear cell myomelanocytic tumor,’’ was subsequently included in the family of perivascular epithelioid cell tumors (PEComas) that usually affect internal organs, and now are commonly termed ‘‘cutaneous perivascular epithelioid cell tumors’’ (primary cutaneous PEComas [pcPEComas]).2 The precise lineage of
From Pathology,a Dermatology,b and Helen A. Diller Family Comprehensive Cancer Center,c University of California San Francisco. Funding sources: None. Conflicts of interest: None declared. Accepted for publication August 10, 2014. Reprint requests: Philip E. LeBoit, MD, Dermatopathology Service, University of California San Francisco, 1701 Divisadero
St, Suite 280, San Francisco, CA 94115. E-mail: philip.leboit@ ucsf.edu. Published online September 26, 2014. 0190-9622/$36.00 Ó 2014 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2014.08.016
1127
J AM ACAD DERMATOL
1128 Charli-Joseph et al
DECEMBER 2014
this enigmatic tumor remains elusive because were purified using ExoSAP-IT (USB Corp, the proposed perivascular epithelioid cell of Cleveland, OH) and sequenced directly using a 3500 origin awaits proper identification and previous DNA Analyzer (Applied Biosystems, Foster City, CA). immunohistochemical staining results are inconThe sequences were analyzed using Mutation sistent. Furthermore, the relationship between Surveyor software (Softgenetics, State College, PA). extracutaneous PEComas and pcPEComas is The Committee on Human Research from the uncertain.3 In this study we describe a substantial University of California San Francisco granted series of these tumors, approval for this study to further characterize their originally titled ‘‘Histologic, CAPSULE SUMMARY clinicopathological and immunohistochemical, and immunohistochemical feagenetic study on cutaneous Cutaneous perivascular epithelioid cell tures. Finally, via arrayneoplasms’’ under the tumors are tumors with a mixed based comparative genomic study number 11-05569. myomelanocytic immunophenotype. hybridization (aCGH) and This covers archival paraffin Cutaneous perivascular epithelioid cell DNA sequencing, we blocks. tumors lack gross chromosomal assayed for initiating mutaaberrations and initiating mutations tions commonly observed in RESULTS common to melanocytic tumors whereas melanocytic neoplasms and Clinical features 4E-binding protein 1 expression DNA copy number changes Clinical findings are sumplausibly relates them to the mTOR as often seen in melanoma. marized in Table II. pathway. pcPEComas occurred in 5 METHODS An immunohistochemical panel women and 3 men. The Eight pcPEComas were including micro-ophthalmia-associated mean age of patients at the identified in the author’s (P. transcription factor, NKIC3, smooth time of presentation was E. L.) consultation files. muscle actin, desmin, bcl-1, cathepsin K, 46 years. pcPEComas Clinical information was oband 4E-binding protein 1 is predominated in the lower tained through contributing recommended for their evaluation. extremities (5/8). Most were clinicians. In each case, described as indurated formalin-fixed paraffinnodules (Fig 1). In no case embedded tissue sections had been stained was the correct diagnosis inferred clinically; the most with hematoxylin-eosin. We included only common impression being dermatofibroma (3/8). lesions centered in the dermis that fulfilled the None of the patients had history of tuberous sclerosis presently accepted criteria for PEComa.3 A broad complex (TSC). Follow-up data were available for 6 panel of immunohistochemical stains was perof 8 patients and ranged between 24 and 171 (mean formed on formalin-fixed paraffin-embedded 66.5) months. No local recurrences or metastases tissue. Antibodies, clones, dilutions, and sources developed in these cases. are outlined in Table I. Appropriate controls were used in parallel. Histopathologic and immunohistochemical features Array-based comparative genomic The histologic size of pcPEComas ranged from 4 hybridization to 20 mm. Neoplasms spared the epidermis and DNA was extracted from formalin-fixed paraffinoccupied the entire dermis except for a thin layer of embedded tissue as described previously.4,5 aCGH uninvolved papillary dermis (grenz zone) (5/8), or was carried out with 500 to 1000 ng of genomic DNA occurred exclusively in the reticular dermis (3/8). A on Agilent Human 4x180k oligomicroarrays diffuse, highly cellular and sheetlike growth pattern (G4449A, Agilent Technologies, Santa Clara, CA). was observed in 4 cases, the rest displayed less aCGH methodology details and data analysis were as cellularity and lesional cells arranged as thin described in a recent publication.6 aggregates, strands and singly in an interstitial array with perivascular and periadnexal accentuation (Fig 2). The interface between the tumors and the DNA sequencing surrounding tissues was well circumscribed. In all Polymerase chain reaction assays using primers cases, lateral borders were irregular and cells were specific for BRAF exon 15; NRAS exons 1 and 2; KIT interposed between thickened collagen bundles exons 11, 13, 17, and 18; GNAQ exon 5; and GNA11 (Fig 3). A similar configuration was observed in the exon 5 were used to amplify DNA as described deep margin in 2 cases where lesional cells did not previously.7 Polymerase chain reaction products d
d
d
J AM ACAD DERMATOL
Charli-Joseph et al 1129
VOLUME 71, NUMBER 6
Abbreviations used: aCGH: MiTF: pcPEComa: PEComa: SMA: TSC: 4EBP1:
array-based comparative genomic hybridization micro-ophthalmia-associated transcription factor primary cutaneous perivascular epithelioid cell tumor perivascular epithelioid cell tumor smooth muscle actin tuberous sclerosis complex 4E-binding protein 1
extend into the subcutaneous tissue. Six cases showed tumoral extension into the subcutis, and the deep border was smooth and rounded (Fig 2); 2 cases also showed a focal lacelike extension between adipocytes. Lesional cells were large (2-4 times the size of a spinous-layer keratinocyte) and oval to polygonal, except 1 case that displayed spindled morphology. Cells were optically clear with pale and coarsely vacuolated cytoplasm (slightly granular in 2 cases) containing a large central vesicular nucleus (1-3 times the size of a mid-layer epidermal keratinocyte nucleus) with, in most, a single central nucleolus (Fig 4). Cells intermediate between typical PEComa cells and adipocytes were absent. The degree of nuclear pleomorphism was negligible and a few had 2 nucleoli, some eosinophilic. Mitotic activity was low and ranged from 0 to 2 mitoses per 10 high-power fields (mean \1/10). No abnormal mitoses were identified. Infiltrating lymphocytes were present in 5 cases, in a perivascular distribution. Thickened collagen fibers were observed in close apposition to lesional cells in every case, and a prominent vasculature of thin-walled delicate capillaries. The rosetting of lesional cells in and around blood vessels in a radial pattern sometimes seen in visceral PEComas was not identified, nor were hyalinized vessel walls found. The overlying epidermis was unaltered in 5 cases, but central attenuation was present in 2 whereas 1 showed slight acanthosis. The immunohistochemical staining pattern of pcPEComas is summarized in Table III and illustrated in Figs 5 and 6. All tumors strongly and diffusely expressed micro-ophthalmia-associated transcription factor (MiTF), NKI/C3, bcl-1, and cathepsin K. E-cadherin immunoreactivity displayed a membranous pattern in all with lack of cytoplasmic labeling. Most pcPEComas (7/8) expressed 4E-binding protein 1 (4EBP1); CD68 stained neoplastic cells avidly in 4 of 7 and faintly in 3 of 7 cases. S100 reactivity was observed in 7 cases, but was fainter than in the appropriate internal controls
(melanocytes). However, after digestion with diastase only 3 of 5 cases retained cytoplasmic immunoreactivity, albeit with considerably lower intensity than before digestion. A blush of HMB-45 positivity appeared in very few of the lesional cells in most tumors (7/8) whereas antibodies to melan-A failed to stain lesional cells. Staining for smooth muscle actin (SMA) was limited to a small number of cells in immediate apposition to thin-walled capillaries (\5% of tumoral volume). One of the tumors was positive for muscle-specific actin, another for desmin, whereas only 1 was faintly labeled with calponin. In all, 20% to 30% of lesional cells in 3 of 8 cases demonstrated nuclear staining for c-myc; few cells were positive for p16 in 3 of 6. Tumors were negative for keratin cocktail, CD56, and SOX10. Proliferation rate, assessed by Ki67 reactivity, was low (\5%) except for 1 case that displayed a 30% proliferation rate. Genomic features None of the cases showed chromosomal imbalances (aCGH analysis) or initiating mutations because DNA sequencing demonstrated only wild-type alleles for BRAF, NRAS, KIT, GNAQ, and GNA11.
DISCUSSION The term ‘‘perivascular epithelioid cell tumor’’ was first used in 1996 when authors describing a pancreatic tumor noted striking similarities with visceral tumors previously reported under several terms (eg, ‘‘clear cell sugar tumor,’’ ‘‘angioleiomyoma,’’ ‘‘lymphangioleiomyoma’’).8 All these neoplasms were composed at least partly by large epithelioid to spindle cells with clear cytoplasm present in close apposition to thin vascular structures, and commonly labeled with both melanocytic markers and myoid immunoreactants.9 To date both benign and less commonly malignant tumors, considered to fall under the unifying PEComa designation, have been reported at multiple anatomic sites and interestingly a subset arise in the setting of TSC.10,11 pcPEComas represent 8% of reported cases within the family of PEComas, but may be somewhat more common as the dermatology and dermatopathology literature has seen only scant reports of pcPEComa, unlike the many reports in the pathology literature, given the predilection for noncutaneous sites.3 Including our cases, 34 legitimate pcPEComas have thus far been described.2,12-20 All but 1 behaved in an indolent fashion and most appear to display similar morphologic attributes on histology.2,12-20 The differential diagnosis posed by a pandermal proliferation of vacuolated cells, discussed in detail
J AM ACAD DERMATOL
1130 Charli-Joseph et al
DECEMBER 2014
Table I. Immunohistochemical reagents used in this study Antigen
S100 MiTF Melan-A HMB-45 SOX10 SMA MSA Desmin Calponin Keratin cocktail NKI/C3 CD68 CD56 BCL-1 p16 4EBP1 CatK E-cadherin c-Myc Ki67
Clone
Dilution
Source
Polyclonal D5 A103 HMB-45 N-20 1A4 HHF45 D33 CALP AE1/AE315D3 NKI/C3 PG-M1 IB6 EP12 ink4a Thr37/46 Polyclonal NCH-38 Y69 Polyclonal
1:20,000 1:100 1:200 1:200 1:100 1:100 1:200 1:25 1:50 Prediluted 1:4 1:200 Prediluted 1:200 1:10 1:1000 1:50
Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Santa Cruz Biotechnology, Dallas, TX Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA Dako, Carpenteria, CA BioCare, Concord, CA BioGenex, Fremont, CA Dako, Carpenteria, CA Leica, Richmond, IL Dako, Carpenteria, CA Ventana, Tucson, AZ Cell Signaling, Danvers, MA abcam, Cambridge, MA Dako, Carpenteria, CA Epitomics, Burlingame, CA Dako, Carpenteria, CA
1:50 1:75
BCL-1, B-cell lymphoma-1; CatK, cathepsin K; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; NKI/C3, anti-melanoma-associated antigen; SMA, smooth muscle actin; SOX10, Sry-related HMB box-10; 4EBP1, 4Ebinding protein 1.
Table II. Clinical features of primary cutaneous perivascular epithelioid cell tumors Case
Gender/age, y
Location
Clinical description, impression
Treatment
Follow-up, duration [mo]
1 2 3
M/60 F/67 M/59
Thigh Thigh Lower aspect of leg
Complete excision Complete excision Complete excision
NED [64] NED [72] NED [171] DUE
4 5 6 7 8
F/37 F/32 M/45 F/26 F/47
Arm Back Abdomen Leg Lower aspect of leg
DF Indurated papule, R/O DF 1-cm Tender nodule, R/O neuroma vs glomus tumor 0.5-cm Pink-brown papule Lesion, details NA 1-cm Nodule DF DF
Marginal excision Complete excision Complete excision Complete excision Complete excision
NA NED [24] NED [41] NA NED [27]
DF, Dermatofibroma; DUE, died of an unrelated event; F, female; M, male; NA, not available; NED, no evidence of disease; R/O, rule-out.
in previous articles, includes both melanocytic and nonmelanocytic neoplasms.1 Melanocytic tumors most likely to be mistaken for pcPEComa include clear cell sarcoma, hypomelanotic blue nevus, balloon cell nevus, and melanoma. Nonmelanocytic neoplasms that invite confusion with pcPEComa include xanthomas, xanthogranulomas, clear cell angiofibroma, and dermatofibroma, among others.1 The reported immunophenotypes in visceral and pcPEComas vary considerably, both between these groups and between studies (Table IV). In our series, among melanocytic markers studied, MiTF (melanocytic marker also positive in some histiocytes and Schwannian lesions)24 and NKIC3 (initially described for melanoma but later found to stain a broad spectrum of neoplastic tissues such as neural
and histiocytic proliferations)25 showed strong and diffuse expression in all cases, coincident with previous reports. Likewise, HMB-45 stained most cases (88% vs 96% in previous reports) but positivity was faint and observed in a minority of lesional cells. This contrasts with visceral PEComas that are commonly positive in more of 50% of neoplastic cells.26,27 S100 positivity has been reported in up to 33% of visceral PEComas.27 Seven of 8 cases in our series showed cytoplasmic labeling with S100, though fainter than that of appropriate internal controls, and only 3 of 5 cases retained some positivity after digestion with diastase. The residual staining was at a lower intensity than before digestion. Positivity may thus stem, at least in part, from nonspecific binding of the immunoreactant to
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
Fig 1. Cutaneous perivascular epithelioid cell tumor. Nonspecific 1-cm erythematous nodule on the lower aspect of abdomen.
Fig 2. Cutaneous perivascular epithelioid cell tumor. The low-power magnification of this lesion shows a dome-shaped, diffuse, and highly cellular proliferation of large and pale epithelioid cells with a sheetlike growth pattern involving the entirety of the dermis. (Hematoxylineosin stain; original magnification: 320.) Inset: The base of the neoplasm shows a well-circumscribed pushing border with a dumb-shaped configuration impinging upon the subcutaneous tissue.
intracytoplasmic glycogen. This nonspecific staining with S100 has been observed in skeletal muscle (oral communication, Daniel Santa Cruz, MD, April 1993), which gave us the impetus to apply diastase digestion to our cases. Melan-A was consistently negative in our cases, and although this finding is lower than average in pcPEComas (approximately 50%), others have found percentages as low as 14%. Thus, we believe this stain should not be relied on for their identification.26,27 Our results reproduce the report of Llamas-Velasco et al19 on the lack of SOX10 expression in pcPEComas. Among muscle markers, SMA and desmin are reportedly the most sensitive to label pcPEComas (28% and 36% positive immunoreactivity, respectively). Eighty eight percent
Charli-Joseph et al 1131
Fig 3. Cutaneous perivascular epithelioid cell tumor. The peripheral aspects of the neoplasm show pale-staining epithelioid and spindle cells interposed between markedly thickened collagen bundles. (Hematoxylin-eosin stain; original magnification: 340.)
Fig 4. Cutaneous perivascular epithelioid cell tumor. This high-power magnification demonstrates that neoplastic cells are large, are oval to polygonal, have pale and coarsely vacuolated cytoplasm, have a negligible degree of nuclear pleomorphism, and are arrayed in close apposition to thin-walled capillaries. (Hematoxylin-eosin stain; original magnification: 3400.)
of our cases showed SMA positivity, limited to less than 5% of cells in immediate apposition to thinwalled capillaries. It is not clear whether these are admixed pericytes or true neoplastic cells. Desmin was positive in only 1 case whereas muscle-specific actin and calponin signals were variable and unreliable. CD68 immunopositivity was present among neoplastic cells in 88% of our cases, in contrast to 45% in previous reports.2,14,15,20 Histologic criteria for establishing malignancy have been proposed for visceral PEComas and suggested to evaluate pcPEComas.16,28 None of our cases fulfilled such criteria because they were small, had a virtually null mitotic count, and lacked necrosis or vascular invasion. In addition, staining with Ki67 (a cell proliferation marker) showed a very low labeling rate (\5%) in all but 1 case (that showed no increase in mitoses nor cytologically atypia).
J AM ACAD DERMATOL
1132 Charli-Joseph et al
DECEMBER 2014
Table III. Detailed immunohistochemical findings in the current study Case
S100 S100-D MiTF Melan-A HMB-45 SOX10 SMA MSA Desmin Calponin Keratin cocktail NKIC3 CD68 CD56 BCL-1 p16 4EBP1 CatK E-cadherin c-Myc Ki67
1
2
3
4
5
6
7
8
Faint 1 Focal 1 Focal 1 1 1 1 1 1 Focal 1 ;30%
1 Faint 1 Focal 1 Focal 1 1 1 Faint 1 Few cells 1 1 1 1 \5%
1 Faint 1 Focal 1 1 1 Focal 1 Few cells 1 1 1 1 \5%
Faint ND 1 Focal 1 Focal 1 ND ND 1 Faint ND 1 ND 1 1 1 Few cells 1 \5%
ND 1 Focal 1 Focal 1 ND ND ND 1 ND ND 1 ND 1 1 1 Few cells 1 \5%
Faint ND 1 Focal 1 Focal 1 ND ND ND 1 1 ND 1 ND 1 1 \5%
1 1 Focal 1 ND Focal 1 Faint 1 1 1 1 1 1 \5%
1 Faint 1 Focal 1 Focal 1 Focal 1 1 Faint 1 Few cells 1 1 1 1 Few cells 1 \5%
BCL-1, B-cell lymphoma-1; CatK, cathepsin K; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; ND, not done; NKI/C3, anti-melanoma-associated antigen; S100-D, S100 after digestion with diastase; SMA, smooth muscle actin; SOX10, Sry-related HMB box-10; 4EBP1, 4E-binding protein 1.
Fig 5. Cutaneous perivascular epithelioid cell tumor. A, Faint cytoplasmic reactivity. B, After digestion with diastase, the degree of cytoplasmic immunoreactivity is weaker than before digestion. C, Faint and focal cytoplasmic immunopositivity. D, Strong nuclear reactivity among all neoplastic cells. E, Strong reactivity among all neoplastic cells. F, Only few cells in immediate apposition to thin-walled capillaries show convincing cytoplasmic reactivity. (Immunohistochemistry with S100 [A and B], HMB-45 [C], micro-ophthalmia-associated transcription factor [D], NKIC3 [E], and smooth muscle actin [F]; original magnifications: A, B, D, and E, 3400; C and F, 3600.)
Bcl-1 (cyclin D1) is an important regulator of the G1 checkpoint of cell cycle and is overexpressed in mantle cell lymphomas.29 Bcl-1 avidly labels neoplastic cells in certain malignant extracutaneous
PEComas and was suggested to participate in the pathogenesis of malignant transformation.30,31 Our cases, which have thus far behaved in an indolent fashion, strongly labeled with bcl-1 suggesting that
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
Charli-Joseph et al 1133
Fig 6. Cutaneous perivascular epithelioid cell tumor. A, Strong nuclear reactivity is evident among all neoplastic cells. B, Avid nuclear reactivity among all neoplastic cells. C, Immunopositivity is shown in a membranous pattern with lack of cytoplasmic labeling. D, Strong cytoplasmic immunoreactivity among all lesional cells. (Immunohistochemistry with BCL-1 [A], 4E-binding protein 1 [B], E-cadherin [C], and cathepsin K [D]; original magnifications: A to D, 3400.)
although its overexpression may play a role in pcPEComas it is not a marker of malignant progression, and did not correspond to increased copy numbers of chromosome 11q, as it does in some malignancies.32 The pathogenesis of pcPEComas remains elusive. Among extracutaneous cases a subset arises predominantly in the setting of TSC (genetic disease associated with dysregulation of the mTOR pathway) or are associated with gene rearrangements in the transcription factor E3 (a member of MiTF family).10,11,33 Moreover, a frequent allelic alteration on chromosome 16p (where the TSC2 gene is located) occurs in both extracutaneous TSC-related and eunrelated PEComas.34 However, there is no report of a cutaneous case associated with TSC and a recent study failed to identify immunoreactivity or gene rearrangements for transcription factor E3 in cutaneous lesions.19 Similarly, our genomic analysis revealed that gross chromosomal aberrations were not responsible for the development of pcPEComas and that unlike bona fide melanocytic neoplasms, initiating mutations in BRAF, NRAS, KIT, GNAQ, and GNA11 did not play a role in their pathogenesis. Recently, kinase gene fusions
have been found in the majority of Spitz nevi, in the absence of such initiating mutations. These are sometimes accompanied by DNA copy number changes in the responsible kinase gene, or its fusion partner.35 However, the absence of copy number changes in our cases makes this mechanism unlikely. The molecular disparity between visceral and pcPEComas is intriguing but could relate to an incorrect lumping of these based on pathological features. The fact that cutaneous cases have not been associated with TSC and that they lack chromosomal aberrations and transcription factor E3 rearrangements, and the differences noted in immunophenotyping, support this hypothesis. Another possibility, however, is that the threshold used to analyze the comparative genomic hybridization data in extracutaneous PEComas was too narrow, leading to an overinterpretation of deviations in the comparative genomic hybridization curve. Independently of this possibility, the analytical software used differs and hence, an accurate conclusion for such discrepancies requires comparatively reviewing the raw data obtained in each study.
0/7 0/2 1/10 0/1 ND/NR 0/1 ND/NR 0/1 3/8 33 12
2005 2007 2008 2008 2010 2010 2013 2013 2014
S100
0/1 0/1
2003 2005 7/7 2/2 10/10 1/1 0/1 1/1 ND/NR 1/1 7/8 33 94
1/1 1/1
HMB-45
1/7 2/2 5/7 1/1 1/1 ND/NR ND/NR 0/1 0/7 26 35
ND/NR ND/NR
Melan-A
MiTF
7/7 ND/NR 5/5 ND/NR 1/1 1/1 ND/NR 1/1 8/8 23 100
ND/NR ND/NR
SOX10
ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR 0/5 ND/NR 0/7 12 0
ND/NR ND/NR
NKIC3
6/6 ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR ND/NR 8/8 14 100
ND/NR ND/NR
SMA
1/7 1/2 1/10 1/1 1/1 0/1 ND/NR 0/1 7/8 33 42
1/1 1/1
MSA
0/6 ND/NR ND/NR ND/NR ND/NR 0/1 ND/NR ND/NR 2/5 12 17
ND/NR ND/NR 1/6 0/2 5/10 ND/NR 1/1 1/1 ND/NR 0/1 1/8 33 30
0/1 1/1
Desmin
2/4 ND/NR 0/10 ND/NR ND/NR ND/NR ND/NR 0/1 1/5 20 15
ND/NR ND/NR
Calponin
ND/NR ND/NR 0/10 ND/NR 1/1 ND/NR ND/NR 0/1 ND 13 8
ND/NR 0/1
Caldesmon
CK
0/7 0/2 0/10 0/1 0/1 0/1 ND/NR 0/1 0/6 31 0
0/1 0/1
CD68
5/7 2/2 2/10 ND/NR ND/NR ND/NR ND/NR 0/1 7/8 28 57
ND/NR ND/NR
EMA
ND/NR 0/2 0/10 0/1 0/1 ND/NR ND/NR ND/NR ND 14 0
ND/NR ND/NR
This literature review does not include the cases described by Folpe et al,21 Diment and Colecchia,22 or Harris et al23 because it was not specified whether or not these soft-tissue tumors involved cutaneous structures. CK, Keratin cocktail; EMA, epithelial membrane antigen; HMB-45, human melanoma black 45; MiTF, micro-ophthalmia-associated transcription factor; MSA, muscle-specific actin; ND, not done; NKI/C3, anti-melanoma-associated antigen; NR, not reported; SOX10, Sry-related HMB box-10; SMA, smooth muscle actin.
Crowson et al12 de Saint Aubain Somerhausen et all3 Mentzel et al2 Tan et al14 Liegl et al15 Calder et al16 Ghazali et al17 Chaplin et al18 Llamas-Velasco et al19 Ieremia and Robson20 Current study Total cases assessed Positive cases, %
Table IV. Immunohistochemical staining pattern in primary cutaneous perivascular epithelioid cell tumors
1134 Charli-Joseph et al J AM ACAD DERMATOL DECEMBER 2014
J AM ACAD DERMATOL VOLUME 71, NUMBER 6
4EBP1 is a downstream target in the mTOR pathway and its phosphorylation status (which controls messenger RNA translation and cellular proliferation) can be evaluated via immunohistochemistry.36 However, to our knowledge this antibody has not been previously evaluated in PEComas. In this study we found robust expression of 4EBP1 in all but 1 case, suggesting that activation of the mTOR pathway is indeed present in pcPEComas independently on the mutational status of TSC1/TSC2. This result is congruent to the finding that the mTOR pathway is activated via positive immunohistochemical staining in other PEComas for phospho-S6.37 E-cadherin (membrane-associated glycoprotein that mediates cellular adhesion) has been shown to display abnormal cytoplasmic immunolocalization in some extracutaneous PEComas, theoretically as a result of TSC2 gene loss.38,39 We found normal membranous immunoreactivity and absence of intracytoplasmic labeling in all cases, in keeping with the fact that none of our patients had evidence of TSC. Cathepsin K (transcriptional target of the MiTF family) is expressed in PEComas at various anatomic sites40 but has not yet been reported in cutaneous lesions. We detected strong positive reactions with cathepsin K antibody in all our cases, supporting it as a powerful marker for the identification of PEComas independently of location. This result reinforces the notion that PEComas represent a new member of the MiTF family of tumors. C-Myc is a gene involved in cell growth, apoptosis, and metabolism. Changes in its configuration or expression have been linked to a variety of human cancers because its dysregulation induces genomic instability.41 However, its expression profile in PEComas is unknown. We found that only 20% to 30% of the cells in 3 of 8 tumors expressed c-myc, a low percentage that plausibly correlates with the lack of genomic instability observed in our aCGH analysis. In summary, pcPEComas enter into the clinical differential diagnosis of nonspecific erythematous nodules and dermatofibroma, but are an exotically rare explanation for them. Our results provide additional evidence that pcPEComas have a different molecular signature from that of extracutaneous cases and are not associated with TSC. However, the cumulative immunohistochemical evidence points to at least a partial role of the mTOR pathway in the pathogenesis of these neoplasms, which are conceivably also related to the MiTF family of tumors. pcPEComas have a very consistent appearance in hematoxylin-eosinestained sections,
Charli-Joseph et al 1135
enabling their diagnosis but only for those familiar with the entity. We propose that a panel of immunoperoxidase stains, including MiTF, NKIC3, SMA, desmin, bcl-1, cathepsin K, and 4EBP1, should be used when evaluating a possible pcPEComa. The authors would like to thank Dr Carlo Cota for providing valuable information along with Kazuko Olson and Mary-Kate Fitzsimon for their help in various aspects of the research process. REFERENCES 1. Biswas A, Mahalingam M. Cutaneous clear cell neoplasms: a histopathological reappraisal. Am J Dermatopathol 2012;34: 237-54. 2. Mentzel T, Reisshauer S, R€ utten A, Hantschke M, Soares de Almeida LM, Kutzner H. Cutaneous clear cell myomelanocytic tumor: a new member of the growing family of perivascular epithelioid cell tumors (PEComas). Clinicopathological and immunohistochemical analysis of seven cases. Histopathology 2005;46:498-504. 3. Walsh SN, Sang€ ueza OP. PEComas: a review with emphasis on cutaneous lesions. Semin Diagn Pathol 2009;26:123-30. 4. Bastian BC, Olshen AB, LeBoit PE, Pinkel D. Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol 2003;16:1765-70. 5. Jalas JR, Vemula S, Bezrookove V, Leboit PE, Simko JP, Bastian BC. Metastatic melanoma with striking adenocarcinomatous differentiation illustrating phenotypic plasticity in melanoma. Am J Surg Pathol 2011;35:1413-8. 6. Charli-Joseph Y, Saggini A, Doyle L, Fletcher CD, Weier J, Mirza S, et al. DNA copy number changes in tumors within the spectrum of cellular, atypical and metastasizing fibrous histiocytoma. J Am Acad Dermatol 2014;71:256-63. 7. Langer R, Becker K, Feith M, Friess H, H€ ofler H, Keller G. Genetic aberrations in primary esophageal melanomas: molecular analysis of c-KIT, PDGFR, KRAS, NRAS and BRAF in a series of 10 cases. Mod Pathol 2011;24:495-501. 8. Zamboni G, Pea M, Martignoni G, Zancanaro C, Faccioli G, Gilioli E, et al. Clear cell ‘‘sugar’’ tumor of the pancreas: a novel member of the family of lesions characterized by the presence of perivascular epithelioid cells. Am J Surg Pathol 1996;20: 722-30. 9. Hornick JL, Fletcher CD. PEComa: what do we know so far? Histopathology 2006;48:75-82. 10. Moss J, Avila NA, Barnes PM, Litzenberger RA, Bechtle J, Brooks PG, et al. Prevalence and clinical characteristics of lymphangioleiomyomatosis (LAM) in patients with tuberous sclerosis complex. Am J Respir Crit Care Med 2001;164:669-71. 11. Pan CC, Jong YJ, Chai CY, Huang SH, Chen YJ. Comparative genomic hybridization study of perivascular epithelioid cell tumor: molecular genetic evidence of perivascular epithelioid cell tumor as a distinctive neoplasm. Hum Pathol 2006;37: 606-12. 12. Crowson AN, Taylor JR, Magro CM. Cutaneous clear cell myomelanocytic tumor-perivascular epithelioid cell tumor: first reported case. Mod Pathol 2003;16:400A. 13. de Saint Aubain Somerhausen N, Gomez Galdon M, Bouffioux B, Courtin C, Theunis A, Vogeleer MN, et al. Clear cell ‘‘sugar’’ tumor (PEComa) of the skin: a case report. J Cutan Pathol 2005;32:441-4. 14. Tan J, Peach AH, Merchant W. PEComas of the skin: more common in the lower limb? Two case reports. Histopathology 2007;51:135-6.
J AM ACAD DERMATOL
1136 Charli-Joseph et al
15. Liegl B, Hornick JL, Fletcher CD. Primary cutaneous PEComa: distinctive clear cell lesions of skin. Am J Surg Pathol 2008;32: 608-14. 16. Calder KB, Schlauder S, Morgan MB. Malignant perivascular epithelioid cell tumor (‘‘PEComa’’): a case report and literature review of cutaneous/subcutaneous presentations. J Cutan Pathol 2008;35:499-503. 17. Ghazali N, Cascarini L, Norris P, Barrett AW, Lavery KM. Perivascular epithelioid cell tumor (PEComa) of the cheek. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010;110: e26-31. 18. Chaplin A, Conrad DM, Tatlidil C, Jollimore J, Walsh N, Covert A, et al. Primary cutaneous PEComa. Am J Dermatopathol 2010;32:310-2. 19. Llamas-Velasco M, Mentzel T, Requena L, Palmedo G, Kasten R, Kutzner H. Cutaneous PEComa does not harbor TFE3 gene fusions: immunohistochemical and molecular study of 17 cases. Histopathology 2013;63:122-9. 20. Ieremia E, Robson A. Cutaneous PEComa: a rare entity to consider in an unusual site. Am J Dermatopathol. http://dx.doi.org/10. 1097/DAD.0000000000000041. Published online December 20, 2013. 21. Folpe AL, McKenney JK, Li Z, Smith SJ, Weiss SW. Clear cell myomelanocytic tumor of the thigh: report of a unique case. Am J Surg Pathol 2002;26:809-12. 22. Diment J, Colecchia M. Myomelanocytic tumor of the thigh. Am J Surg Pathol 2003;27:1288. 23. Harris GC, McCulloch TA, Perks G, Fisher C. Malignant perivascular epithelioid cell tumor (‘‘PEComa’’) of soft tissue: a unique case. Am J Surg Pathol 2004;28:1655-8. ~ez NG. Value of melanocytic-associated immuno24. Ord on histochemical markers in the diagnosis of malignant melanoma: a review and update. Hum Pathol 2014;45: 191-205. 25. Orchard GE. Comparison of immunohistochemical labeling of melanocyte differentiation antibodies melan-A, tyrosinase and HMB 45 with NKIC3 and S100 protein in the evaluation of benign nevi and malignant melanoma. Histochem J 2000;32: 475-81. 26. Schoolmeester JK, Howitt BE, Hirsch MS, Dal Cin P, Quade BJ, Nucci MR. Perivascular epithelioid cell neoplasm (PEComa) of the gynecologic tract: clinicopathologic and immunohistochemical characterization of 16 cases. Am J Surg Pathol 2014;38:176-88. 27. Folpe AL, Mentzel T, Lehr HA, Fisher C, Balzer BL, Weiss SW. Perivascular epithelioid cell neoplasms of soft tissue and
DECEMBER 2014
28.
29. 30.
31.
32.
33.
34.
35.
36. 37. 38.
39. 40.
41.
gynecologic origin: a clinicopathologic study of 26 cases and review of the literature. Am J Surg Pathol 2005;29:1558-75. Folpe AL. Neoplasms with perivascular epithelioid differentiation (PEComas). In: Fletcher CDM, Unni KK, Mertens F, editors. WHO classification of tumors: tumors of soft tissue and bone. Lyon (France): IARC Press; 2002. pp. 221-2. Pestell RG. New roles of cyclin D1. Am J Pathol 2013;183:3-9. Weinreb I, Howarth D, Latta E, Ghazarian D, Chetty R. Perivascular epithelioid cell neoplasms (PEComas): four malignant cases expanding the histopathological spectrum and a description of a unique finding. Virchows Arch 2007;450: 463-70. Yamamoto H, Oda Y, Yao T, Oiwa T, Kobayashi C, Tamiya S, et al. Malignant perivascular epithelioid cell tumor of the colon: report of a case with molecular analysis. Pathol Int 2006; 56:46-50. Hawryluk EB, Baran JL, Gerami P, Sepehr A. ‘‘Monster cell’’ melanoma with pulmonary metastasis and cyclin D1 amplification. J Cutan Pathol 2013;40:61-5. Argani P, Aulmann S, Illei PB, Netto GJ, Ro J, Cho HY, et al. A distinctive subset of PEComas harbors TFE3 gene fusions. Am J Surg Pathol 2010;34:1395-406. Pan CC, Chung MY, Ng KF, Liu CY, Wang JS, Chai CY, et al. Constant allelic alteration on chromosome 16p (TSC2 gene) in perivascular epithelioid cell tumor (PEComa): genetic evidence for the relationship of PEComa with angiomyolipoma. J Pathol 2008;214:387-93. Wiesner T, He J, Yelensky R, Esteve-Puig R, Botton T, Yeh I, et al. Kinase fusions are frequent in Spitz tumors and Spitzoid melanomas. Nat Commun 2014;5:3116. Guertin DA, Sabatini DM. Defining the role of mTOR in cancer. Cancer Cell 2007;12:9-22. Folpe AL, Kwiatkowski DJ. Perivascular epithelioid cell neoplasms: pathology and pathogenesis. Hum Pathol 2010;41:1-15. Barnes EA, Kenerson HL, Jiang X, Yeung RS. Tuberin regulates E-cadherin localization: implications in epithelial-mesenchymal transition. Am J Pathol 2010;177:1765-78. Wang Z, Gong Q, Fan Q. Expression of E-cadherin in angiomyolipoma. Hum Pathol 2012;43:2348-53. Rao Q, Cheng L, Xia QY, Liu B, Li L, Shi QL, et al. Cathepsin K expression in a wide spectrum of perivascular epithelioid cell neoplasms (PEComas): a clinicopathological study emphasizing extrarenal PEComas. Histopathology 2013;62:642-50. Bretones G, Delgado MD, Le on J. Myc and cell cycle control. Biochim Biophys Acta. http://dx.doi.org/10.1016/j.bbagrm.2014. 03.013. Published online April 1, 2014.
