Article Text
Abstract
FOS and FOSB proto-oncogens are involved in a wide variety of tumourigenic processes. FOS and FOSB gene rearrangements are observed in epithelioid haemangioma, pseudomyogenic haemangioendothelioma, osteoid osteoma/osteoblastoma/cementoblastoma and proliferative myositis/fasciitis. In this review, we provide an overview of FOS and FOSB, including their functions and the differences between lesions with known FOS/FOSB gene rearrangements. Additionally, we discuss the use of FOS/FOSB immunohistochemistry as a diagnostic tool for these lesions.
- soft tissue neoplasms
- bone and bones
- molecular biology
- immunohistochemistry
Statistics from Altmetric.com
Gene and protein structure
FOS (Fos proto-oncogene, activator protein 1 (AP-1) transcription factor subunit), a protein coding gene, was first detected as a cellular homologue of two viral v-FOS oncogenes, which induce osteosarcoma in rats and mice and is therefore also known as FBJ Murine Osteosarcoma Viral Oncogene Homolog, or cellular variant of FOS (c-FOS).1 The FOS gene belongs to the Fos gene family along with FOSB, FOSL1 and FOSL2.2 The FOS gene is located on chromosome 14q24.323 while the FOSB gene maps to 19q13.32.2 These genes encode nuclear phosphoproteins, including c-Fos, FosB, ΔFosB and their smaller splice variants Fos-related antigens (Fra-1) and Fra-2.4–6 The Fos family is part of the AP-1 transcription factor family.1 7 These proteins contain a conserved basic region Leucine Zipper (bZIP) domain, which is essential for their dimerisation.4
Fos proteins are expressed in various cell types and tissue, and their DNA binding necessitates heterodimerisation. By forming complexes with Jun family proteins, they form the AP-1 transcription factor complex, which binds to AP-1 regulatory elements in the promoter and enhancer regions of a wild range of genes.4 5 The AP-1 regulatory elements can be diverse, enabling the Fos-Jun family dimers to have different functions, even within different cell types. This versatility is further due to specific (extracellular) signals (eg, growth factors, cytokines, tumour-promoting agents and DNA damage)8 9 or interactions with non-bZIP DNA-binding proteins (eg, NFAT, Ets, Smad and bHLH), which influence the expression of Fos-Jun dimers. Additionally, binding specificities can alter through cross-family dimerisation of Fos-Jun proteins with other bZIP proteins, including different members of ATF, C/EBP and NF-E2 proteins.4
Fos protein levels are tightly regulated by both transcript and protein degradation. There are two translation-dependent mechanisms that contribute to rapid mRNA degradation.10 First, a length-dependent interaction between the poly-A tail and an exon 3 domain (known as the major coding region determinant of instability). The length of the poly-A tail affects the stability of the mRNA, with shorter tails being associated with increased degradation.11 Another mechanism involves the presence of AU-rich elements in the 3’ untranslated region of the mRNA. These AU-rich elements contribute to mRNA instability and degradation, independent of the poly-A tail length.12 Both mechanisms are likely to be disrupted by gene rearrangements with FOS. FOS gene translocations (figure 1) identified in epithelioid haemangioma (EH) and osteoblastoma (OB) result in premature stop codons, leading to loss of the c-terminal end of the protein. This renders the protein resistant to degradation, resulting in high expression levels within tumour cells.9 10 The FOSB fusions (figure 1) found in atypical EH and pseudomyogenic haemangioendothelioma (PHE) occur at the N-terminal region of the protein with preservation of the basic DNA binding motif, leucine zipper domain and the transactivating domain and are most likely induced by promoter swap events causing upregulation of FosB.13 14 However, further research is needed to determine the specific impact of FOS gene rearrangements on the described mechanisms of mRNA degradation.10 Furthermore, breakdown of the AP1-transcription factor complex is regulated by controlled proteasome degradation. Gomard et al demonstrated that the degradation of c-Fos and Fra-1 can occur independently of prior ubiquitination. This degradation mechanism may be specific to the Fos family and could also apply to Fra-2 and FosB, as they share conserved structural domains not found in other proteins.6 Additionally, van IJzendoorn et al found that the four C-terminal amino acids (LLAL) of Fos make it intrinsically susceptible to ubiquitin-independent degradation by the 20S proteasome. This ubiquitin-independent degradation is an essential mechanism that is bypassed by tumour Fos proteins.9
Function
The AP-1 transcription factor binds to TPA-responsive elements (5′-TGAC/GTCA-3′) of the promoter and enhancer regions of target genes, thereby regulating a wide variety of physiological (eg, osteoblast maturation, adipocyte differentiation) and tumourigenic processes, including cell proliferation, oncogenic transformation, tumour invasion, distant metastasis and angiogenesis, as well as control of stress responses, organogenesis, immune responses and control of cognitive functions.4–7 15
Stimulation of c-Fos has been observed to induce angiogenesis by triggering overexpression of vascular endothelial growth factor-D. This angiogenic effect is considered an important mechanism contributing to cancer progression.16 c-Fos has also been associated with the failure of peritoneal dialysis, due to its involvement in fibrotic and inflammatory processes.16 17 Furthermore, c-Fos serves as a crucial transcription factor in the initiation of adipocyte differentiation. The presence of a mutation that triggers a repression mechanism at the c-fos promoter raises the hypothesis that reduced c-Fos expression could potentially contribute to congenital generalised lipodystrophy by interfering with the normal development of adipocytes.18 c-Fos and ΔFosB have also been studied in the context of neuronal function and plasticity, however, their specific contribution to Alzheimer’s disease onset and progression is still being investigated.19 20 Fra plays a role in carcinoma progression.5 Downregulation of Fos is linked to pathological conditions with immunological, skeletal and neurological defects, to oncogenic transformation and tumour progression.7
Pathological lesions containing FOS/FOSB rearrangements
Epithelioid haemangioma
EH is a well-demarcated benign vascular neoplasm with a predilection for the skin and subcutis of head and neck but can also occur in deep soft tissue, bone, lymph nodes, lung, eye, heart, spleen and penis.21 The term ‘EH’ was introduced in the WHO classification of tumours in 1993. Prior to that, this tumour was known as ‘epithelioid endothelial cell tumour’. The updated terminology ‘EH’ was used to better reflect the distinctive histological appearance and vascular nature of the tumour.22 EH is composed of vascular channels arranged in a lobular growth pattern and lined by a monolayer of epithelioid endothelial cells and a pericytic layer. The endothelial cells can show intracytoplasmic vacuoles (blister cells). There is only mild cytological atypia and mitotic activity is low (figure 2A). EH has a wide morphology with intravascular growth, abundant inflammatory infiltrates and/or cellular/solid growth. The inflammatory infiltrate typically consists of mixed inflammatory cells and can show abundant lymphocytes and plasma cells, often with the formation of prominent lymphoid follicles, as well as eosinophils.9 23 The endothelial cells are positive for CD31, ERG and may also show positivity for keratins and epithelial membrane antigen (EMA). The pericytic cells stain for SMA.24 EH can be subclassified into three morphological subtypes: conventional, cellular (solid growth in >50%) and angiolymphoid with eosinophilia (ALHE).21 ALHE is defined as an EH showing centrifugal growth from a muscular vessel and a variable degree of inflammation.21 25 Described gene fusions in EH are ZFP36::FOSB, WWTR1::FOSB, FOS::LMNA, VIM::FOS and FOS::MBNL1.13 21 22 FOS gene rearrangements have a higher prevalence in cellular and intraosseous EH-variant, variant with mild inflammatory reaction or an elevated mitotic activity (2–5 mitoses/10 HPF).2 21 FOS gene rearrangements are not observed in the ALHE cases, which could suggest another subtype with different pathogenesis (and may suggest a reactive process) from classic and cellular EH.21 Also, FOSB gene rearrangements are more commonly found in EH of the extremities, trunk and penis.13 A subset of EH shows a more solid growth pattern with increased cellularity, mild nuclear pleomorphism and areas of necrosis, which are more worrisome histological traits. This subset of EH reveals more often an FOSB gene rearrangement resulting in FOSB-upregulation and therefore fall under the category of FOSB gene rearranged atypical EH.13 21 Immunohistochemistry (IHC) for FOSB (figure 2B) shows strong and diffuse expression in ± half of the EH cases with a low rate of expression in the cellular subtype. This could be explained by the underlying FOS gene rearrangement in this subtype instead of FOSB rearrangements. In ALHE cases, there was a strong and diffuse nuclear immunoreactivity for FOSB, even though no FOSB gene rearrangements were identified. This observation could be attributed to potential cross-reaction with other FOS proteins or the possibility that the upregulation of FOSB in ALHE may be driven by a mechanism independent of translocations.2 26 FOSB staining can be helpful for differentiating EH from its histological mimics but is not specific, since epithelioid sarcomas, epithelioid angiosarcomas and nodular and cellular benign fibrous histiocytomas show a typically focal weak FOSB immunoreactivity.2 Papke et al showed a strong FOS expression in 5 out of 10 cases of epithelioid and spindle cell haemangioma, a morphological variant of EH, which correlated with an FOS gene rearrangement detected by fluorescence in situ hybridisation (FISH).27 Furthermore, Tsui et al observed that FOSB and FOS was expressed in 68% and 46% of their EH cases, respectively. Interestingly, coexpression of FOSB and FOS was seen in 37% of the EH cases. However, it is worth noting that they also found FOS and/or FOSB expression in lobular capillary haemangioma (32%) and papillary endothelial hyperplasia (58% for FOS and 33% for FOSB).28
Pseudomyogenic haemangioendothelioma
PHE, also known as epithelioid sarcoma-like haemangioendothelioma, is a vascular neoplasm of intermediate biological potential. The name PHE was coined by Hornick and Fletcher in 2011. The term ‘pseudomyogenic’ is used to describe the myoid-appearing spindle cell morphology of this tumour, despite its true endothelial origin.29 PHE occurs mostly as a multifocal lesion in the lower limbs of young male adults. The lesions involve skin, superficial and deep soft tissues as well as bone (mostly due to direct involvement). The upper limbs and trunk are less commonly affected and it rarely occurs on the head or neck.26 29–31 PHE typically lacks vasoformative features and is composed of loose fascicles or sheets of plump spindled and epithelioid cells with abundant eosinophilic cytoplasm (figure 3A,B), with a subset resembling rhabdomyoblasts. Intracytoplasmatic vacuoles are not seen. Nuclear atypia is mild and mitotic activity is low. Immunohistochemically, PHE shows coexpression of endothelial (CD31, ERG but not CD34) and keratin markers (AE1/AE3 and focally for CAM5.2 and MNF116).26 29 30 32 The differentiation with epithelioid sarcoma can be difficult, due to shared clinical (multifocal lesion of soft tissue in extremities in young adults), histological (epithelioid and spindle cell morphology) and immunohistochemical characteristics (ERG and diffuse keratin expression). However, a key distinguishing factor is the loss of INI1 expression in epithelioid sarcoma, while it is retained in PHE.26 33 Additionally, specific genetic abnormalities, namely the SERPINE1::FOSB fusion and ACTB::FOSB fusion, have been identified as genetic hallmarks of PHE.14 26 34 Morphological features do not differ between PHE with different gene fusion, however it has been observed that PHE with an ACTB::FOSB fusion occurs more frequently as a solitary lesion.34 FOSB gene rearrangement and it subsequent upregulation of FOSB can be picked up by IHC showing a strong and diffuse reactivity (figure 3C). However, FOSB immunoreactivity is not specific since it can also be seen in EH. Also, other lesions as epithelioid sarcoma, nodular fasciitis, epithelioid angiosarcoma and cellular benign fibrous histiocytomas show focal and weak immunoreactivity for FOSB. Additionally, FOSB staining is common in endothelial cells and reactive myofibroblasts in granulation and scar tissue. This further underscores the non-specific nature of FOSB immunoreactivity.26 Furthermore, despite the presence of FOSB gene rearrangements with FOSB upregulation in PHE and EH, there is no significant morphological or extended immunohistochemical overlap (PHE shows immunoreactivity for cytokeratin AE1/AE3 and negativity for CD34) with EH and PHE. PHE shows a small risk of distant metastasis.2 26 No staining for FOS has been observed in PHE.35
Osteoid osteoma/OB/cementoblastoma
OB and osteoid osteoma (OO) are benign tumours found predominantly in men between 10 and 30 years old.24 36 37 These are both benign bone-forming lesions, with OB mostly occurring in the spine and OO in the long bones. OO presents with nocturnal pain, which can be relieved by NSAIDs.24 37 38 In 1935, Jaffe identified and described five cases of OO as a distinct pathologic entity.39 In 1956, the term ‘OB’ was independently coined by both Lichtenstein and Jaffe.40 41 OB and OO show morphological overlap and were, therefore, considered as variants within the same family. However, distinction between OB and OO should be made since OB has a tendency to recur.37 38 The distinction is made clinically and is based on growth pattern and size, with OB showing local aggressive behaviour and measuring larger than 2 cm.24 39 42 Histologically, OB and OO consist of connected woven bone trabeculae rimmed by a layer of osteoblasts (figure 4A). Different levels of bone mineralisation can be seen. Between the bone trabeculae there is loose, vascularised stroma containing fibroblast-like stromal cells. Osteoclast-like giant cells are diffusely spread throughout the lesion. The interface between the central part of the lesion (nidus) and the surrounding reactive bone is abrupt and circumscribed.37 Differentiation from (OB-like) osteosarcoma can be difficult, especially in small biopsies and when it shows local aggressive growth.37 43 Histological signs of malignancy include poorly circumscribed margins with infiltration into the bone, a lack of peripheral bone maturation, the presence of sheets of osteoblast-like cells without osteogenic matrix production, and immature bone lining mature trabeculae. Furthermore, cytological features of osteoblast-like cells with large and centrally located nuclei and the presence of atypical mitotic figures favour malignancy.43 A specific subset of OB exhibits an epithelioid morphology and has been previously referred to as ‘aggressive OB’ due to its potentially more aggressive clinical course. However, a study conducted by Della Rocca and Huvos did not find a consistent relationship between the behaviour of OB and its histological appearances.44 Therefore, it is preferred to use the term ‘epithelioid OB’ instead, and the term ‘aggressive OB’ should be avoided.24 45 In their study, Della Rocca and Huvos noted a more aggressive course in lesions occurring in short tubular or flat bones compared with those occurring in long tubular bones.44 FOS gene rearrangements can be found in OO and OB, where rearrangements are restricted to the osteoblastic cells that line immature bone and osteoid within conspicuous granulation-like reactive tissue. Translocations are mostly found in FOS (87%) and rarely in FOSB (3%).10 FOS::ANKH, FOS::KIAA1199, FOS::MYO1B, FOS::IGR, PPP1R10::FOSB and FOS::RUNX2 gene fusions have been described.10 36 IHC can be used for diagnostic purposes, as strong staining of FOS may suggest an underlying FOS gene rearrangement, though staining of FOSB lacks specificity in this regard (figure 4B).42 46 However, it is important to note that strong and diffuse staining for FOS was also observed in osteosarcoma (14%), but they did not show an underlying FOS gene rearrangement.10 46
Cementoblastoma (CB) is a benign rare odontogenic neoplasm associated with the apical third of the roots of teeth.47 It was first described by Dewy in the year 1927. It occurs most frequently in the posterior mandible and causes painful expansion of the jaw.48 The radiologic appearance is almost pathognomonic with a well-defined radiopaque mass extending from the root of a tooth, obliterating the periodontal space and generally showing a radiolucent rim. Histologically, CB exhibits an immature and dense matrix, resembling cementum, which is typically attached to the root of a tooth, although this is usually not present in biopsies. The reversal lines within the lesion can appear irregular and may bear resemblance to Paget disease of bone. Surrounding the matrix, there are activated cementoblasts, along with a well-vascularised fibroblastic stroma. The central areas of the lesion are hypocellular and highly mineralised, while the periphery often contains areas that closely resemble OB (since cementoblasts are morphologically indistinguishable from osteoblasts) (figure 4C,D). There has been a hypothesis suggesting that CBs might initially develop as ‘conventional’ OB within the tooth-bearing regions of the jaws and subsequently become connected to a tooth.49 Furthermore, Lam et al proposed that CB and OB/OO represent different histogenic variations of the same disease spectrum due to their shared clinical features (predilection for the second and third decades of life), histological characteristics and similar prognostic outcomes (OB and CB can recur). They also demonstrated that these lesions share a common molecular pathogenesis, as evidenced by the presence of FOS gene rearrangements. In their study, they observed FOS overexpression in CBs through IHC, and FOS gene rearrangements were detected using FISH.50
Proliferative fasciitis/myositis
In 1960, Kern reported a rapidly growing pseudosarcomatous lesion of the muscle and named it proliferative myositis (PM), in analogue to myositis ossificans (MO).51 In 1967, PM was further characterised by Enzinger et al and introduced the term proliferative fasciitis (PF). PM is a poorly demarcated soft tissue lesion involving the epimysium, perimysium and endomysium. Also, dermal and subcutaneous variants have been described with the lesion arising from or in close proximity to the fascia and therefore called PF.52 53 PF is seen most frequently in the upper extremities and PM in the trunk, shoulder and upper arms. It often presents as rapidly growing masses, usually <5 cm in size and may be accompanied by tenderness or pain.54 PF/PM is composed of a (myo)fibroblastic proliferation with scattered or groups of epithelioid ganglion-like cells (round to polygonal cells with abundant basophilic cytoplasm and a vesicular, eccentric nucleus with prominent nucleolus) (figure 5A–C). In the background, there are variable amounts of collagen and myxoid stroma. Mitotic figures are common, however presence of atypical mitoses is never seen. Distinguishing PF/PM from nodular fasciitis (NF), a benign (myo)fibroblastic lesion that develops on the fascia, can be challenging. However, in PF/PM, the presence of ganglion-like cells and their primary involvement in the interfascicular connective tissue (checkerboard pattern) differentiates them from NF, where the muscle fibres are more affected.53 PF/PM is more commonly seen in an older population (age >45 years) then NF (young adults <30 years).53 54 Additionally, it is noteworthy that PF/PM sometimes contain metaplastic bone, suggesting a potential association with conditions such as MO, a benign neoplasm composed of spindle cells and osteoblasts but lacking ganglion-like cells, or OO.54 Both NF and MO are distinguishable by the presence of an underlying USP6 gene, classifying them as USP6-associated neoplasms. This characteristic can be particularly useful in challenging diagnostic cases.55
A subtype of PF/PM has been described in children, with range of 2.5 months to 13 years. This paediatric subtype of PF/PM is more well circumscribed, lobulated and extremely cellular with diffuse sheets of ganglion-like cells. It shows a higher mitotic activity. Acute inflammation or foci of necrosis can be seen. These lesions could be misdiagnosed as a rhabdomyosarcoma.56
Over the years the pathogenesis of PF/PM remained unclear. In 2020, Makise et al observed a diffuse strong expression of FOS, primarily in the ganglion-like cells, along with the detection of an FOS gene rearrangement in these cells using FISH. Additionally, in one case, they performed RNA sequencing, which revealed an FOS::VIM fusion. These findings strongly suggest a neoplastic nature of the lesion. However, in their paediatric case, they did not observe FOS expression or FOS gene rearrangement, indicating this could be a distinct subtype.35 The finding of strong FOS expression in PF/PM, was confirmed by Vargas et al.57 Also Hung et al, in a separate study, observed a strong immunohistochemical staining for FOSB in 2 out of 20 tested cases of PF (figure 5D).26 In the differential diagnosis, it is important to notice that cases of ischaemic fasciitis can display FOSB positivity, however, further molecular investigations through FISH and RNA sequencing studies did not reveal underlying FOS(B) gene rearrangements.58
Practical approach of using FOS/FOSB IHC
Positive immunohistochemical staining for FOS and/or FOSB is considered significant when observed in more than 50% of the tumour cells, showing a strong and diffuse expression.26 35 This staining pattern is evident in the previous mentioned lesions. However, it is essential to be aware that other lesions and histological mimics may also show focal and weak positivity for FOS/FOSB or in some cases, even strong and diffuse positivity may be observed. Therefore, all findings must be interpreted together with the clinical, radiological, morphological and other immunophenotypic findings.26 It is also important to note that IHC may not be reliable after long decalcification. In challenging cases without a clear diagnosis, a low threshold for performing FISH or RNA sequencing to detect underlying FOS/FOSB gene rearrangements is recommended. However, it is essential to be aware that these molecular analyses may produce false-negative results, especially in instances of low tumour cellularity or when dealing with decalcified tissue. Additionally, it should be noted that overexpression does not always correspond to the presence of an underlying gene fusion (eg, AHLE).
Take home points
FOS and FOSB are part of the FOS gene family, which encodes for leucine zipper proteins that can dimerise with proteins of the JUN family, thereby forming the transcription factor complex activator protein 1.
Fos proteins play important roles in the regulation of cell proliferation, differentiation and transformation, making them significant players in both normal physiological processes and pathological conditions, including cancer.
FOS gene rearrangements have been observed in epithelioid haemangioma, osteoid osteoma/osteoblastoma/cementoblastoma and proliferative myositis/fasciitis. FOSB gene rearrangements have been observed in epithelioid haemangioma, pseudomyogenic haemangioendothelioma and osteoid osteoma/osteoblastoma. Despite the shared molecular features, there is no significant morphological overlap between these entities.
FOS and FOSB immunohistochemistry (IHC) can be used to support the diagnosis in the clinical and morphological context, together with a broader IHC panel to exclude alternative differential diagnoses.
Osteoid osteoma, osteoblastoma and cementoblastoma belong to the same spectrum of lesions since they exhibit clinical, histological and molecular overlap.
Ethics statements
Patient consent for publication
References
Footnotes
Handling editor Vikram Deshpande.
Contributors FC performed the writing of the paper and made the figures. DC performed the study concept, design and review of the paper. All authors read and approved the final paper.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Commissioned; externally peer reviewed.