GLI activated epithelioid cell tumour: report of a case and proposed new terminology

In 2018, Antonescu et al. reported a series of six cases of an unusual tumour characterised by round to ovoid cells with a nested, cord-like and cribriform architecture, infiltrative growth, strong S100 expression, and GLI1 gene rearrangements, which usually arose in soft tissues and often showed clinically aggressive behaviour.1Antonescu C.R. Agaram N.P. Sung Y.S. et al.A distinct malignant epithelioid neoplasm with GLI1 gene rearrangements, frequent S100 protein expression, and metastatic potential: expanding the spectrum of pathologic entities with ACTB/MALAT1/PTCH1-GLI1 fusions.Am J Surg Pathol. 2018; 42: 553-560Crossref PubMed Scopus (53) Google Scholar They noted that the tumours appeared distinct compared with previously described actin positive perivascular myoid tumours displaying ACTB-GLI1 gene fusions known as ‘pericytoma with t(7;12) translocation’ and two recently reported gastric tumours, plexiform fibromyxoma and gastroblastoma, that both harboured MALAT1-GLI1 fusions. In view of the distinctive clinical, morphological, and immunohistochemical profile of their cases, Antonescu provisionally proposed the descriptive terminology ‘malignant epithelioid tumour with GLI1 rearrangement’. Recently, a case series was published describing a related group of tumours that showed morphological overlap with the earlier examples, but lacked the defining GLI1 translocation. Rather, these harboured GLI1 amplification as an alternative pathway of GLI1 activation.2Agaram N.P. Zhang L. Sung Y. et al.GLI1-amplifications expand the spectrum of soft tissue neoplasms defined by GLI1 gene fusions.Mod Pathol. 2019; 32: 1617-1626Crossref PubMed Scopus (22) Google Scholar Here, we report a case of GLI1 amplified tumour, expanding the clinicopathological spectrum of described features, and propose the term ‘GLI activated epithelioid cell tumour’ to more aptly describe it. We also highlight the utility of FISH for DDIT3, which lies in close proximity to GLI1 on chromosome 7, as a surrogate marker for GLI1 amplification. A 9-year-old female presented with a 5 year history of a slowly growing lesion on the left great toe. It came to medical attention as a painless mass causing functional restriction, and local excision was performed. The overlying skin had a nodular outline with focal ulceration (Fig. 1A,B). The cut surface had a uniform tan appearance. Microscopy revealed a dermally-based tumour with a multinodular growth pattern (Fig. 1C,D). The tumour nodules had an organoid arrangement with tight nests separated by thin fibrovascular septa reminiscent of paraganglioma (Fig. 1E). Some cribriform gland-like structures were present (Fig. 1F). The cells had uniform round to ovoid nuclei with granular chromatin and inconspicuous nucleoli (Fig. 1G). The cytoplasm was scant eosinophilic to clear. Necrosis was absent, but mitoses were readily identified (up to 8/mm2). There was prominent perivascular tumour growth, and vascular invasion with multiple intravascular tumour deposits (Fig. 1C,D) was noted. Immunohistochemically (Fig. 2A–E), the lesional cells showed strong diffuse positivity for BCL2, Glut A, and CD56. S100 was positive in isolated cells, probably representing associated dendritic cells. There was patchy staining for EMA and focal p63. SMA showed staining of perivascular cells at the periphery of the nests of tumour cells. The tumour was negative for desmin, calponin, CEA, CD21, Melan A, synaptophysin, chromogranin, myogenin, and a broad range of cytokeratins. As Ewing sarcoma was considered in the differential diagnosis, FISH testing for EWSR1 rearrangement was performed, which was negative. The tumour was found to be negative for GLI1 rearrangement on FISH testing but there was amplification of the GLI1 locus. Since GLI1 is lies in close proximity to DDIT3 on chromosome 12, FISH for DDIT3 was also performed (Fig. 2F). As expected, co-amplification of this gene was observed. Further investigations revealed no sites of metastatic disease. The patient was well at last follow-up, 10 months after her surgery.Fig. 2Immunohistochemistry and FISH findings. (A) The tumour cells are negative for cytokeratin AE1/AE3. (B) SMA highlights a subset of cells at the periphery of the tumour nodules, while the central cells are negative. (C) Patchy weak staining for EMA. (D) Strong diffuse staining for CD56. (E) S100 stains the occasional dendritic cells only. (F) Break-apart FISH for DDIT3 shows amplification, and is a useful surrogate marker for GLI1 amplification due to the proximity of the two loci.View Large Image Figure ViewerDownload Hi-res image Download (PPT) GLI1 is a transcription factor that functions as the terminal effector of the Sonic Hedgehog (Shh) signalling pathway. This pathway is critical for normal tissue growth, differentiation, and polarity, but is also implicated in the pathogenesis of a number of human cancers. Indeed, GLI1 was first identified in glioma,3Kinzler K. Bigner S. Bigner D. et al.Identification of an amplified highly expressed gene in human glioma.Science. 1987; 236: 70-73Crossref PubMed Scopus (542) Google Scholar where a 50-fold amplification of the gene was reported. It was subsequently found to be amplified and/or overexpressed in a number of other established tumour types.4Stein U. Eder C. Karsten U. et al.GLI gene expression in bone and soft tissue sarcomas of adult patients correlates with tumor grade.Cancer Res. 1999; 59: 1890-1895PubMed Google Scholar However, there has been inconsistency regarding the classification of mesenchymal neoplasms harbouring GLI1 associated chromosomal abnormalities reported in the literature to date (Table 1). GLI1 rearranged soft tissue tumours were first described in a series of five cases that showed the t(7;12) translocation by cytogenetic analysis.5Dahlen A. Fletcher C.D. Mertens F. et al.Activation of the GLI oncogene through fusion with the beta-actin gene (ACTB) in a group of distinctive pericytic neoplasms: pericytoma with t(7;12).Am J Pathol. 2004; 164: 1645-1653Abstract Full Text Full Text PDF PubMed Google Scholar This abnormality resulted in ACTB-GLI1 gene fusion in all cases, leading the authors to conclude that the ACTB promoter caused overexpression of GLI1 transcripts with subsequent activation of downstream genes. The cases reported in this initial series showed spindle cell morphology with a perivascular growth pattern and variable expression of SMA, and were therefore considered to have a pericytic phenotype. Further case reports of tumours with similar histological features and an identical molecular abnormality were then added to the literature, occurring in bone,6Bridge J.A. Sanders K. Huang D. et al.Pericytoma with t(7;12) and ACTB-GLI1 fusion arising in bone.Hum Pathol. 2012; 43: 1524-1529Crossref PubMed Scopus (27) Google Scholar stomach,7Castro E. Cortes-Santiago N. Ferguson L.M. et al.Translocation t(7;12) as the sole chromosomal abnormality resulting in ACTB-GLI1 fusion in pediatric gastric pericytoma.Hum Pathol. 2016; 53: 137-141Crossref PubMed Scopus (20) Google Scholar and ovary.8Koh N.W.C. Seow W.Y. Lee Y.T. et al.Pericytoma with t(7;12): the first ovarian case reported and a review of the literature.Int J Gyn Pathol. 2018; 38: 479-484Crossref Scopus (10) Google Scholar Interestingly, the last of these cases was negative for SMA and positive for S100, despite its designation as pericytoma with t(7;12). Subsequently, GLI1 translocations were shown to occur in two distinct rare entities in the stomach, namely gastroblastoma9Graham R.P. Nair A.A. Davila J.I. et al.Gastroblastoma harbors a recurrent somatic MALAT1-GLI1 fusion gene.Mod Pathol. 2017; 30: 1443-1452Crossref PubMed Scopus (46) Google Scholar and plexiform fibromyxoma.10Spans L. Fletcher C.D. Antonescu C.R. et al.Recurrent MALAT1-GLI1 oncogenic fusion and GLI1 up-regulation define a subset of plexiform fibromyxoma.J Pathol. 2016; 239: 335-343Crossref PubMed Scopus (50) Google Scholar Up to this point, all of the reported cases had pursued a benign clinical course. The spectrum of tumours with GLI1 gene abnormalities was further expanded in 2018 when Antonescu et al. published their aforementioned series of six tumours with GLI1 fusions, epithelioid morphology, and negativity for SMA.1Antonescu C.R. Agaram N.P. Sung Y.S. et al.A distinct malignant epithelioid neoplasm with GLI1 gene rearrangements, frequent S100 protein expression, and metastatic potential: expanding the spectrum of pathologic entities with ACTB/MALAT1/PTCH1-GLI1 fusions.Am J Surg Pathol. 2018; 42: 553-560Crossref PubMed Scopus (53) Google Scholar Unlike the t(7:12) pericytomas, these tumours preferentially occurred in the soft tissues and bone, and three of the cases metastasised. Recently, a further three cases of ACTB-GLI1 fusion associated tumour were reported,11Kerr D.A. Pinto A. Subhawong T.K. et al.Pericytoma with t(7;12) and ACTB-GLI1 Fusion. Reevaluation of an unusual entity and its relationship to the spectrum of GLI1 fusion-related neoplasms.Am J Surg Pathol. 2019; 43: 1682-1692PubMed Google Scholar and served as a possible link between the t(7;12) pericytomas and Antonescu's cases.1Antonescu C.R. Agaram N.P. Sung Y.S. et al.A distinct malignant epithelioid neoplasm with GLI1 gene rearrangements, frequent S100 protein expression, and metastatic potential: expanding the spectrum of pathologic entities with ACTB/MALAT1/PTCH1-GLI1 fusions.Am J Surg Pathol. 2018; 42: 553-560Crossref PubMed Scopus (53) Google Scholar Similar to our case, these tumours showed morphological overlap with the latter cohort, with packeted growth of round to oval cells and a prominent vascular component. In contrast, however, there was immunohistochemical evidence of at least partial myogenic differentiation. Follow-up revealed malignant behaviour in two of the tumours. The authors speculated that tumours with ACTB-GLI1 fusions may represent related entities that lie on a morphological spectrum and display variable pericytic phenotype and biologic potential.Table 1Clinicopathological features of mesenchymal neoplasms with chromosomal abnormalities involving the GLI1 gene reported to dateStudyAge/sexSiteMitoses (per 10 HPF)GLI1 abnormalityArchitectureCytologyS100/SMA/CKClinical behaviourFU (months)Dahlen et al.5Dahlen A. Fletcher C.D. Mertens F. et al.Activation of the GLI oncogene through fusion with the beta-actin gene (ACTB) in a group of distinctive pericytic neoplasms: pericytoma with t(7;12).Am J Pathol. 2004; 164: 1645-1653Abstract Full Text Full Text PDF PubMed Google Scholar61/MCalf<1ACTB-GLI1 fusionNested, arranged around thin-walled vesselsSpindle to ovoid, clear cytoplasm–/+/–NED6027/FTongue<1ACTB-GLI1 fusion–/+/–NED6011/MTongue<1ACTB-GLI1 fusion–/F/–NED2265/FStomach<1ACTB-GLI1 fusion–/F/–NED2412/FTongue<1ACTB-GLI1 fusion–/F/–NED120Bridge et al.6Bridge J.A. Sanders K. Huang D. et al.Pericytoma with t(7;12) and ACTB-GLI1 fusion arising in bone.Hum Pathol. 2012; 43: 1524-1529Crossref PubMed Scopus (27) Google Scholar67/MTalusLowACTB-GLI1 fusionSolid with prominent thin-walled vesselsSpindle to ovoid–/F/–NED168Castro et al.7Castro E. Cortes-Santiago N. Ferguson L.M. et al.Translocation t(7;12) as the sole chromosomal abnormality resulting in ACTB-GLI1 fusion in pediatric gastric pericytoma.Hum Pathol. 2016; 53: 137-141Crossref PubMed Scopus (20) Google Scholar9/FStomachACTB-GLI1 fusionSolid and cystic, plexiform vasculatureOvoid, eosinophilic to clear cytoplasm–/–/–NED6Antonescu et al.1Antonescu C.R. Agaram N.P. Sung Y.S. et al.A distinct malignant epithelioid neoplasm with GLI1 gene rearrangements, frequent S100 protein expression, and metastatic potential: expanding the spectrum of pathologic entities with ACTB/MALAT1/PTCH1-GLI1 fusions.Am J Surg Pathol. 2018; 42: 553-560Crossref PubMed Scopus (53) Google Scholar20/MThigh<5ACTB-GLI1 fusionSolid-cystic, large malformed vesselsRound to oval to epithelioid nuclei, cytoplasm varies from scant to abundant and clear to eosinophilic+/–/–No FU data–16/MC2 spine<5MALAT-GLI1 fusionNests and cords, rich capillary network+/–/–No FU data–30/FFoot<5ACTB-GLI1 fusionNested, rich capillary network+/–/–Inguinal LN met2134/FNeck<5PTCH1-GLI1 fusionNests and cords, rich capillary network, myxoid stroma+/–/–LN and lung mets8079/FRP<5ACTB-GLI1 fusionCribriform/sieve-like, myxoid stroma–/–/–LN mets38/FChest wall<5ACTB-GLI1 fusionSolid sheets with scattered tubular structures–/–/FNo FU data–Koh et al.8Koh N.W.C. Seow W.Y. Lee Y.T. et al.Pericytoma with t(7;12): the first ovarian case reported and a review of the literature.Int J Gyn Pathol. 2018; 38: 479-484Crossref Scopus (10) Google Scholar11/FOvaryLowACTB-GLI1 fusionFascicles, sheets, nests; hypo- and hypercellular areas; arborizing capillaries and larger thick vesselsRound to spindled nuclei, eosinophilic cytoplasm+/–/–No FU data–Agaram et al.2Agaram N.P. Zhang L. Sung Y. et al.GLI1-amplifications expand the spectrum of soft tissue neoplasms defined by GLI1 gene fusions.Mod Pathol. 2019; 32: 1617-1626Crossref PubMed Scopus (22) Google Scholar4/FShoulder5AmplificationCord-like growth, rich capillary networkMonomorphic ovoid cellsF/–/–No FU data–10/MFinger15Amplification, fusionNested, rich capillary network, dilated vascular spacesRound and uniform, clear cytoplasm–/–/–No FU data–17/MThigh5AmplificationNested, rich capillary networkRound and uniform, eosinophilic to clear cytoplasm–/–/–No FU data–23/FThigh2AmplificationCribriform/rosette-like growth, trabecular areasRound and uniform, eosinophilic to clear cytoplasm–/–/+NED3626/FLung4AmplificationTrabecular and sinusoidal patternsRound and uniform, eosinophilic to clear cytoplasmF/–/–No FU data–39/MNeck>25AmplificationNests alternating with short fasciclesBiphasic, epithelioid and spindle cell component–/+/–Lung met2651FBack>25AmplificationSingle files to sheets in collagenous stromaEpithelioid, focally increased atypiaF/–/–Local recurrence1654FElbow15AmplificationMultinodular, nested, rich capillary networkRound and uniform, eosinophilic to clear cytoplasmF/F/–No FU data–60/MForearm10AmplificationNested, rich capillary networkRound and uniform, eosinophilic to clear cytoplasm–/–/–No FU data–65/MTongue5AmplificationCompact nested growth in fibrotic stromaEpithelioid uniform, clear cytoplasm–/–/–No FU data–Kerr et al.11Kerr D.A. Pinto A. Subhawong T.K. et al.Pericytoma with t(7;12) and ACTB-GLI1 Fusion. Reevaluation of an unusual entity and its relationship to the spectrum of GLI1 fusion-related neoplasms.Am J Surg Pathol. 2019; 43: 1682-1692PubMed Google Scholar57/FTibia5ACTB-GLI1 fusionMultinodular and trabecular growth, HPC-like and fine vesselsRound to oval, clear to eosinophilic cytoplasm–/F/–Met to 8th rib2762/MScapula<1ACTB-GLI1 fusionTight nests; delicate vasculature; areas of myxoid changeRound to oval, clear to eosinophilic cytoplasm–/F/–Lung, thigh met16841/FOvary<1ACTB-GLI1 fusionHypo and hypercellular areas; sheet-like growth; delicate vasculatureOvalF/F/+NED14Xu et al.12Xu B. Chang K. Folpe A.L. et al.Head and neck mesenchymal neoplasms with GLI1 gene alterations.Am J Surg Pathol. 2020; 44: 729-737Crossref PubMed Scopus (13) Google Scholar46/FTongue0AmplificationMultinodular or plexiform growth, with nested architecture and rich arborizing vascular network. Frequent protrusion of tumour into vascular spaces.Monomorphic round to oval nuclei, clear cytoplasmF/+/–NED360/MTongue11AmplificationF/–/NDNo FU data–38/MTongue1PTCH1-GLI1 fusion+/–/–NED237/MNeck0ACTB-GLI1 fusion–/F/–NED301/MTongue4ACTB-GLI1 fusionND/–/NDNED228/MTongue0ACTB-GLI1 fusion+/F/–No FU data–14/MTongue0ACTB-GLI1 fusion–/ND/–No FU data–56/FTongue8MALAT-GLI1 fusion–/–/FNo FU data–Aivazian et al. (our case)9/FGreat toe20AmplificationNested, rich capillary networkRound to epithelioid–/F/–NED10–, negative; +, positive; CK, cytokeratin; F, focally positive; FU, follow-up; met, metastasis; mo, months; ND, not done; NED, no evidence of disease; SMA, smooth muscle actin. Open table in a new tab –, negative; +, positive; CK, cytokeratin; F, focally positive; FU, follow-up; met, metastasis; mo, months; ND, not done; NED, no evidence of disease; SMA, smooth muscle actin. In contrast to the cases characterised by chromosomal translocations involving GLI1, amplification of the GLI1 gene as the defining molecular event has been previously described in only two case series, in a total of 12 patients.2Agaram N.P. Zhang L. Sung Y. et al.GLI1-amplifications expand the spectrum of soft tissue neoplasms defined by GLI1 gene fusions.Mod Pathol. 2019; 32: 1617-1626Crossref PubMed Scopus (22) Google Scholar,12Xu B. Chang K. Folpe A.L. et al.Head and neck mesenchymal neoplasms with GLI1 gene alterations.Am J Surg Pathol. 2020; 44: 729-737Crossref PubMed Scopus (13) Google Scholar The first included 10 cases that showed histological similarities to our case and the malignant epithelioid tumours previously reported by Antonescu and colleagues, but lacked GLI1 rearrangements.2Agaram N.P. Zhang L. Sung Y. et al.GLI1-amplifications expand the spectrum of soft tissue neoplasms defined by GLI1 gene fusions.Mod Pathol. 2019; 32: 1617-1626Crossref PubMed Scopus (22) Google Scholar These cases did not show any consistent morphological features or immunophenotypes, occurred at a variety of anatomical sites, and showed variable clinical behaviour. The morphological spectrum was wider than previously reported. Whilst all cases showed the characteristic nested epithelioid growth, some tumours also included biphasic epithelioid and spindled areas, cells growing in single file in a fibrous stroma, or significant cytological atypia. Intriguingly, one of the cases was from an acral location in a child and showed striking morphological similarities to our case. In addition to GLI1 gene amplification, this tumour had an unbalanced translocation involving the GLI1 locus. It is possible that the chromosomal instability inherent to unbalanced translocations resulted in amplification, the latter being responsible for some of the morphological features seen. The second series,12Xu B. Chang K. Folpe A.L. et al.Head and neck mesenchymal neoplasms with GLI1 gene alterations.Am J Surg Pathol. 2020; 44: 729-737Crossref PubMed Scopus (13) Google Scholar from the same group, contributed eight new cases (and three previously reported cases) from the head and neck region. Two of the new cases showed GLI1 amplification, the rest being characterised by GLI1 fusions. Again, a consistent immunophenotype failed to emerge: positivity for S100, SMA, and cytokeratin was seen in 6/10 (60%), 4/10 (40%), and 2/9 (22%) cases, respectively. Of the two patients with GLI1 amplified tumours, one had follow-up and was free of disease at 3 months. A number of differential diagnoses should be considered during work up of this neoplasm. Morphologically, the pattern most resembles a neuroendocrine neoplasm such as carcinoid or paraganglioma; however, the lack of synaptophysin or chromogranin excludes these possibilities. Myoepithelial tumours and epithelioid schwannoma may arise as possibilities in SMA positive and S100 positive cases, respectively. However, GLI1 activated tumours lack other markers of myoepithelial differentiation such as calponin and p63, while its infiltrative growth pattern is incompatible with a diagnosis of schwannoma. Given the acral location in our case, we also considered digital papillary adenocarcinoma and mucin producing sweat gland carcinoma. However, the morphological features and immunoprofile were not consistent with these alternatives. Our case expands the growing body of knowledge about the novel group of tumours associated with GLI1 abnormalities. Whilst a disease spectrum is emerging, the precise pathogenetic relationship between the reported cases is difficult to elucidate. Nevertheless, it is clear that a subset of tumours with either GLI1 gene fusions or amplifications can show similar epithelioid cytomorphology that can aid their recognition and prompt consideration of GLI1 molecular studies. For these reasons, we propose the term ‘GLI activated epithelioid cell tumour’ to aptly and succinctly describe them. Awareness and recognition of this entity may aid in a refined understanding of its pathogenesis. We thank Dr Diane Payton for the opportunity to review the case and Dr Cristina Antonescu and Dr Andrew Folpe for their input and performing GLI1 FISH studies. Support from colleagues at the Royal Prince Alfred Hospital, New South Wales Health Pathology, and Melanoma Institute Australia is also gratefully acknowledged. Karina Aivazian is supported by a Deborah and John McMurtrie Melanoma Pathology Fellowship through Melanoma Institute Australia . Richard A. Scolyer is supported by an Australian National Health and Medical Research Council Fellowship. The authors state that there are no conflicts of interest to disclose.