OUP user menu

p63 Immunohistochemical Staining Is Limited in Soft Tissue Tumors

Vickie Y. Jo MD, Christopher D.M. Fletcher MD, FRCPath
DOI: http://dx.doi.org/10.1309/AJCPXNUC7JZSKWEU 762-766 First published online: 1 November 2011


p63 is a p53 homolog that is expressed in various normal epithelial tissues and epithelial malignancies. Its expression in mesenchymal lesions has not been examined in depth; therefore, we studied p63 expression by immunohistochemical analysis in 650 soft tissue tumors. We found that p63 expression is limited in soft tissue tumors. The majority of tumors studied were p63–, including all cases of angiosarcoma, lipomatous neoplasms, dermatofibrosarcoma protuberans, solitary fibrous tumor, schwannoma, neurofibroma, gastrointestinal stromal tumor, and leiomyosarcoma. Nuclear p63 reactivity was found in a subset of soft tissue myoepithelioma and myoepithelial carcinoma of soft tissue, cellular neurothekeoma, soft tissue perineurioma, Ewing sarcoma/peripheral neuroectodermal tumor, diffuse-type giant cell tumor, and giant cell tumor of soft parts. Infrequent, weak, or focal p63-staining patterns were observed in low-grade fibromyxoid sarcoma, malignant peripheral nerve sheath tumor, extraskeletal myxoid chondrosarcoma, myxofibrosarcoma, proximal-type epithelioid sarcoma, synovial sarcoma, embryonal rhabdomyosarcoma, desmoplastic small round cell tumor, atypical fibroxanthoma, and spindle cell melanoma. Absent p63 expression is typical for most soft tissue tumors, including most (but not all) that would be in the differential diagnosis of spindle cell squamous carcinoma.

Key Words:
  • p63
  • Sarcoma
  • Soft tissue
  • Tumor

p63, which is encoded on chromosome 3q27–28, is a p53 homolog required for epithelial development and maintenance of epithelial stem cell populations.1,2 Expression of p63 is present in various normal epithelial tissues, such as squamous epithelium, urothelium, basal cells of skin appendages, thymus, and myoepithelial/basal cells of salivary, breast, and prostatic glandular epithelium.3 In stratified epithelia, p63 is strongly expressed in the basal layers, and expression decreases as cells migrate more superficially and become more terminally differentiated. Expression of p63 has been well documented in various epithelial malignancies, including squamous cell carcinoma, basal cell carcinoma, thymic tumors, urothelial carcinoma, and the myoepithelial components of salivary gland neoplasms.

The immunohistochemical detection of p63 has several diagnostic uses, particularly in the evaluation of epithelial neoplasms. Identification of myoepithelial and basal cell layers in breast and prostate lesions helps distinguish between in situ and invasive components of glandular neoplasia. Immunohistochemical studies for p63 are widely regarded as being useful in confirming squamous differentiation, particularly in the evaluation of poorly differentiated or spindle cell lesions and in the distinction from atypical fibroxanthoma, spindle cell malignant melanoma, and other mesenchymal neoplasms, but this diagnostic use has not been carefully validated.

To date, the distribution of p63 expression in soft tissue tumors has not been well-characterized. Studies using tissue microarrays of soft tissue tumors have reported rare to absent p63 immunostaining3,4; however, no systematic or comprehensive study has been performed to our knowledge. Expression of p63 is present in a certain proportion of myoepithelial tumors of soft tissue (7%–23%),5,6 which also show S-100 protein, glial fibrillary acidic protein (GFAP), and keratin immunoreactivity, analogous to their salivary gland counterparts. There have been several reports of p63 expression in giant cell tumor of bone (69%–100%),4,7 with staining limited to the mononuclear component. We performed the present study to examine p63 expression in a large series of soft tissue tumors.

Materials and Methods

Cases were retrieved from the consultation files of one of us (C.D.M.F.). All diagnoses were based on standard and widely accepted criteria810 and supported by appropriate immunohistochemical and cytogenetic studies.

Immunohistochemical studies for p63 were performed on whole sections of formalin-fixed, paraffin-embedded tissue using a mouse monoclonal antibody raised against human p63 protein (clone 4A4, NeoMarkers, Fremont, CA) at a 1:600 dilution with antigen retrieval by citrate buffer and microwave. Appropriate positive and negative control samples were used. Tumors were evaluated for nuclear staining, and results were recorded as positive or negative. The presence of any reactivity in tumor cell nuclei was considered positive, particularly because, in routine practice, it seems that even very limited positivity is often taken as supporting a diagnosis of spindle cell squamous carcinoma. Positive cases were further classified as follows: multifocal if staining was present in greater than 10% of nuclei and distributed through the tissue section and focal if staining was present in fewer than 10% of nuclei.


A total of 650 soft tissue tumors were examined for p63 expression by immunohistochemical staining. The frequency of p63 expression in all entities is shown in Table 1. No expression of p63 was seen in the following: angiosarcoma (0/20), nodular Kaposi sarcoma (0/10), atypical lipomatous tumor/well-differentiated liposarcoma (0/20), dedifferentiated liposarcoma (0/20), myxoid liposarcoma (0/20), spindle cell lipoma (0/20), cellular benign fibrous histiocytoma (0/20), dermatofibrosarcoma protuberans (0/20) Image 1A, desmoid fibromatosis (0/20), nodular fasciitis (0/10), solitary fibrous tumor (0/20), leiomyosarcoma (0/20), alveolar rhabdomyosarcoma (0/10), anaplastic lymphoma kinase–positive inflammatory myofibroblastic tumor (0/10), neurofibroma (0/20), schwannoma (0/20), so-called angiomatoid malignant fibrous histiocytoma (0/10), gastrointestinal stromal tumor (0/20), PEComa (or perivascular epithelioid cell neoplasm; 0/10), ossifying fibromyxoid tumor (0/10), alveolar soft part sarcoma (0/10), clear cell sarcoma (0/10), and follicular dendritic cell sarcoma (0/10).

View this table:
Table 1

As would have been predicted, strong nuclear reactivity for p63 was observed in 9 soft tissue myoepitheliomas (9/20 [45%]) and 8 cases of myoepithelial carcinoma of soft parts (8/20 [40%]). Soft tissue myoepithelioma staining was multifocal in 6 cases and focal in 3 cases, and weak staining was observed in 4 cases. Of the positive myoepithelial carcinomas of soft tissue, 4 cases showed focal staining and 4 showed multifocal staining Image 1B. All 40 cases studied showed convincing myoepithelial differentiation by using immunohistochemical studies for keratins (pankeratin, AE1/AE3, CAM 5.2), GFAP, and S-100.

Image 1

p63 Immunohistochemical nuclear staining in selected soft tissue tumors. No staining was seen in most entities, such as dermatofibrosarcoma protuberans (A, ×20). Multifocal p63 staining was present in subsets of myoepithelial carcinoma of soft tissue (B, ×40), cellular neurothekeoma (C, ×40), soft tissue perineurioma (D, ×40), and Ewing sarcoma/peripheral neuroectodermal tumor (E, ×40). p63 staining was seen in the mononuclear components of diffuse-type giant cell tumor (F, ×40). Multifocal p63 staining was seen in rare cases of spindle cell melanoma (G, ×20), malignant peripheral nerve sheath tumor (H, ×40), and atypical fibroxanthoma (I, ×40).

Expression of p63 was present in 12 cases of cellular neurothekeoma (12/20 [60%]). In 10 cases, there was multifocal staining in tumor cells, and 2 cases showed focal staining. Most cases of p63+ cellular neurothekeoma showed weak staining Image 1C, appearing fainter than the adjacent basal squamous epithelium. All cases diagnosed as cellular neurothekeoma showed typical morphologic features and immunohistochemical profiles (reactivity for smooth muscle actin, NKI-C3, and neuron-specific enolase).

Of 20 cases of soft tissue perineurioma, 4 were p63+ (10%) Image 1D, of which 1 case showed only focal staining and 2 showed just rare positive cells. All examined cases had typical morphologic features of perineurioma and showed perineurial differentiation by epithelial membrane antigen immunoreactivity.

Of 20 cases of Ewing sarcoma/peripheral neuroectodermal tumor (EWS/PNET), 7 (35%) were p63+ Image 1E, 2 of which were only focal. All cases evaluated had typical morphologic features and diffuse membranous CD99 positivity and EWSR1 translocation detected by fluorescence in situ hybridization on paraffin-embedded, formalin-fixed tissue in our laboratory.

In 5 (25%) of 20 cases of diffuse-type giant cell tumor, there was weak, scattered nuclear p63 positivity; 3 cases had focal expression Image 1F. Nuclear reactivity was present predominantly in the mononuclear cells. Weak and focal p63 staining was seen in 3 cases of giant cell tumor of soft tissue (3/10 [30%]), with staining restricted to the mononuclear component.

Some tumor types showed infrequent weak or focal positivity. Single cases demonstrating multifocal p63 were seen in embryonal rhabdomyosarcoma (1/10), spindle cell melanoma (1/20) Image 1G, extraskeletal myxoid chondrosarcoma (1/10), malignant peripheral nerve sheath tumor (MPNST; 1/20) Image 1H, and low-grade fibromyxoid sarcoma (1/10). One case of synovial sarcoma showed focal p63 reactivity that was confined to the glandular component. Weak staining for p63 was observed in 1 case of atypical fibroxanthoma (1/20) Image 1I, with fainter nuclear reactivity in tumor cells than in the adjacent basal squamous epithelium and adnexal structures. All diagnoses of atypical fibroxanthoma followed strict criteria, specifically, origin in sun-damaged skin, lack of expression of markers for squamous or melanocytic differentiation, absence of necrosis, and confinement to the dermis. Weak focal p63 was observed in 1 case of desmoplastic small round cell tumor (1/10) and 1 case of myxofibrosarcoma (1/20). In 2 cases of proximal-type epithelioid sarcoma, there was weak staining, 1 focal and 1 multifocal (2/20).


We found that the extent of p63 immunoreactivity in soft tissue tumors is limited. Of 650 total soft tissue tumors, only 59 tumors (9.1%) were p63+, the majority of which showed focal and/or weak staining. As expected, subsets of myoepithelioma of soft tissue and myoepithelial carcinoma of soft tissue showed more prominent p63 reactivity. In addition, we found that some cases of cellular neurothekeoma, soft tissue perineurioma, EWS/PNET, diffuse-type giant cell tumor, and giant cell tumor of soft tissue were also p63+. Our findings are concordant with previous reports that p63 expression is infrequent in mesenchymal lesions. Di Como et al3 reported infrequent p63 staining in rhabdomyosarcoma (2/31) and MPNST (4/18) using a tissue microarray, and Lee at al4 found p63 positivity in 1 of 60 undifferentiated pleomorphic sarcomas, 1 of 19 synovial sarcomas, 3 of 42 leiomyosarcomas, 3 of 15 chondroblastomas, and 1 of 25 aneurysmal bone cysts, also using tissue microarrays. Our large series using whole sections of tumor confirms that p63 staining is largely absent in soft tissue tumors.

In our series, 45% of myoepitheliomas of soft tissue and 40% of myoepithelial carcinomas of soft tissue were p63+, about half of which showed focal staining. This finding is in keeping with prior studies5,6 that p63 is present in a subset of myoepitheliomas and myoepithelial carcinomas of soft tissue. Similar to its salivary gland counterpart, myoepithelial carcinoma of soft tissue is a heterogeneous neoplasm showing varying myoepithelial differentiation in terms of morphologic features and immunostaining patterns with keratins, epithelial membrane antigen, S-100, and GFAP. A subset of myoepithelial carcinomas of soft tissue harbors EWSR1 translocations with fusion partners ZNF44, POU5F1, PBX1, and FUS.11 The differential diagnosis of myoepithelial carcinoma of soft tissue includes ossifying fibromyxoid tumor and extraskeletal myxoid chondrosarcoma. We found p63 staining to be absent in ossifying fibromyxoid tumor and multifocal in 1 case of extraskeletal myxoid chondrosarcoma. However, p63 immunostaining is useful only in this context as an adjunct to the more frequently positive immunohistochemical markers in myoepithelial carcinoma and with careful attention to the morphologic features.

The majority of mesenchymal lesions in our series were negative for p63. Single cases of low-grade fibromyxoid sarcoma, MPNST, extraskeletal myxoid chondrosarcoma, myxofibrosarcoma, synovial sarcoma, spindle cell melanoma, desmoplastic small round cell tumor, embryonal rhabdomyosarcoma, and atypical fibroxanthoma and 2 cases of epithelioid sarcoma showed p63 reactivity. Most of these p63+ cases showed focal or weak staining. The finding of sporadic positive cases suggests that p63 immunoreactivity in these mesenchymal lesions, when present, is nonspecific, or at least difficult to explain in biologic terms.

It is interesting that p63 expression was found relatively more frequently in several diagnostic entities. More than half of the cellular neurothekeomas examined showed at least weak p63 expression, and a subset of soft tissue perineuriomas and diffuse-type giant cell tumors also showed reactivity. Seven cases of EWS/PNET were p63+, but the staining was mostly weak. The significance of these findings is unknown and requires study of additional cases and molecular profiling to determine whether our observation has biologic implications and/or diagnostic usefulness.

Giant cell tumor of soft tissue showed weak p63 staining in a minority of cases; similar to giant cell tumor of bone, the staining was restricted to the mononuclear component. In comparison, the frequency of p63 reactivity in giant cell tumor of bone has been reported to be 69%7 or 100%.4 In the study by Lee et al,4 gene expression profiling demonstrated high expression of TP73L (which encodes the p63 protein) in giant cell tumor of bone relative to other bone and soft tissue tumors. Subsequent immunohistochemical studies using tissue microarrays of mesenchymal lesions showed positivity in 69% of giant cell tumors of bone; it is interesting that no cases of diffuse-type giant cell tumor were p63+. Dickson et al,7 using immunohistochemical studies on whole sections, reported 100% p63 positivity in giant cell tumor of bone (14 cases) and no staining in 22 cases of diffuse-type giant cell tumor. The conflicting results of p63 staining in these 2 studies may be related to sampling differences between tissue microarray and whole sections and the use of different p63 antibody sources and antigen retrieval conditions (NeoMarkers in our series, DAKO,4 and LabVision,7 all clone 4A4; see “Materials and Methods” sections for conditions). Because the biologic significance of p63 expression remains uncertain, resolution of the conflicting results requires further study.

In our series, 1 case of atypical fibroxanthoma and 1 case of spindle cell melanoma were p63+. When considering cytologically malignant cutaneous spindle cell lesions arising in sun-damaged skin, the use of immunohistochemical markers is often paramount in the differential diagnosis, most commonly including spindle cell squamous cell carcinoma, melanoma, and atypical fibroxanthoma. In this context, p63 is useful for distinguishing spindle cell squamous cell carcinoma, which is characterized by strong and diffuse nuclear staining. Our finding of weak p63 staining in single cases of melanoma and atypical fibroxanthoma likely represents nonspecific expression and does not challenge the usefulness of p63 in confirming squamous differentiation. In most circumstances, additional markers are also useful, such as pankeratin and cytokeratin 5 for squamous differentiation and melanocytic markers (in particular, S-100 protein).

In mesenchymally derived normal tissues, such as smooth muscle, fibroblasts, and endothelium, p63 is not present.3 p63 is expressed primarily in epithelial tissues and has distinct developmental regulatory roles, notably in limb and craniofacial development and maintenance of epithelial structures. Although p53 is a well-characterized tumor suppressor gene, the role and function of p63 in neoplasia remains unclear, and no specific mutation or other consistent aberration of the gene is known in human malignancies. p63 encodes multiple isoforms that have diverse biologic functions and complicated kinetics. While the ongoing body of literature studying the regulation of (and various roles of) p63 splice variants is beyond the scope of this discussion, some of the recognized functions of p63 isoforms include repression of p53, transcriptional activation of p53, induction of cell cycle arrest, promotion of apoptosis, and p21 induction.1,12,13 Neither a specific tumorigenic role nor the mechanism for overexpression of p63 has been identified in epithelial malignancies, although these may be downstream effects mediated through the p53 signaling pathway.

Our findings show that p63 immunohistochemical staining is limited in soft tissue tumors. Among the subset of p63+ mesenchymal neoplasms, the staining pattern usually varies from weak to focal, and p63 expression does not seem to be specific for any single soft tissue neoplasm with the exception of soft tissue myoepithelioma and myoepithelial carcinoma of soft tissue. No cases showed strong and diffuse p63 positivity. The significance of p63 staining in a subset of cellular neurothekeoma, soft tissue perineurioma, diffuse-type giant cell tumor, EWS/PNET, and giant cell tumor of soft tissue is unknown. We do not consider p63 to be a useful marker to distinguish between soft tissue neoplasms. In the evaluation of superficial spindle cell lesions, p63 remains a useful marker to identify squamous differentiation, so long as staining is strong and at least multifocal. We would not regard only focal or weak p63 positivity, in the absence of keratin positivity, as convincing evidence to support a diagnosis of spindle cell squamous cell carcinoma.


Upon completion of this activity you will be able to:

  • discuss the utility of p63 immunohistochemistry in the diagnosis of cutaneous spindle cell lesions.

  • cite the soft tissue entities in which p63 immunopositivity may be expected.

  • describe the expected p63 staining patterns in mesenchymal lesions.

The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. Physicians should claim only the credit commensurate with the extent of their participation in the activity. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.

The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Questions appear on p 819. Exam is located at www.ascp.org/ajcpcme.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
View Abstract