The SMARCA subgroup of genes belong to the SWI1/SNF1 family that are responsible chromatin remodelling and repair. Inactivating mutations in the main SMARCA genes A2 and A4 lead to loss of expression of their respective proteins within the nucleus and, as such have characterised a set of malignancies that are underpinned by SMARCA-deficiency.
The morphology of these tumours ranges from small to large epithelioid cells, giant cells and rhabdoid cells. The rhabdoid cells are frequently present in these tumours but are not a sine qua non for the diagnosis. Most of these tumours are undifferentiated or dedifferentiated, high-grade pleomorphic carcinomas. Focally, areas of better differentiation can be seen. The initial description of a SMARCA4-deficient malignancy was the small cell carcinoma of the ovary, hypercalcaemic type. Subsequently, tumours fitting this characteristic morphology and immunophenotype have been described in the lung, thoracic cavity, endometrium and sinonasal tract, gastrointestinal tract and kidney. Immunohistochemical loss of SMARCA2 and SMARCA4 may occur concomitantly or independently of each other.
SMARCA-deficient malignant tumours represent a unique subset of tumours with typical morphological and immunohistochemical findings.
- cancer genetics
Statistics from Altmetric.com
Human DNA is exposed to a variety of noxious, damaging events on a daily basis from many internal and external events. Internal damaging sources result from by-products of intracellular metabolic activities producing reactive oxygen species; while external sources include ionising radiation, chemical carcinogens and other environmental deleterious events. These result in damage to DNA, including double-strand damage, which is particularly pernicious and can result in deleterious genetic events. If not corrected or if inappropriately repaired, double-strand damage can lead to molecular aberrations such as gene mutations and chromosomal translocations. The end result of this is an increased propensity to carcinogenesis.1 Human cells have evolved a mechanism called a DNA damage response that detects, signals and repairs DNA lesions induced by double-strand damage to prevent genomic damage and consequent human disease.1 Damaged DNA is repaired within the chromatin network by several mechanisms like DNA methylation, histone post-translational modifications and nucleosome remodelling by ATP-dependent chromatin remodelling complexes.2–6
All chromatin remodelling complexes contain an ATPase/helicase of the SWI2 (SWItch 2)/SNF2 (Sucrose Non-Fermenting 2) superfamily that generates the energy for chromatin remodelling through the hydrolysis of ATP.7 In addition, there are other discrete functional domains in the ATPase/helicase proteins that participate in chromatin remodelling complexes along with SWI/SNF: ISWI (Imitation SWItch), CHD (Chromodomain Helicase DNA-binding) and INO80 (Inositol 80).7
Several SWI/SNF2 family members belong to the SMARCA (SWI/SNF-related Matrix-associated, Actin-dependent Regulator Chromatin group A) class or subgroup of chromatin remodelers.7 The SWI/SNF2 group is further made up of: SMARCA1 (SNF2L; Sucrose Non-Fermenter 2 L), SMARCA2 (BRM BRahMa), SMARCA3 (HLTF; Helicase-like Transcription Factor), SMARCA4 (BRG1; Brahma-Related Gene 1), SMARCA5 (SNF2H; Sucrose Non-Fermenter 2 hour), SMARCA6 (HELLS Helicase Lymphoid Specific), SMARCAD1 and SMARCAL1.
The SMARCA2 gene is located on chromosome 9p24.3, and is one of the two mutually exclusive catalytic subunits of the switch/sucrose-non-fermenting (SWI/SNF) chromatin-remodelling complex. The protein encoded by this gene is a member of the SWI/SNF family of proteins and is highly similar to the brahma (BRM) protein of Drosophila. Members of this family have helicase and ATPase activities and are believed to regulate transcription of several genes by altering their chromatin structure. Inactivating SMARCA2 mutations are rarely detected in cancers, however, cell lines and primary cancers have been found to display little or no SMARCA2 protein expression.8 This implies some sort of mutational inactivation of the SMARCA2 gene. It is not known whether these low levels of genetic aberration in SMARCA2 are critical and result in genetic susceptibility leading to cancer, or whether in some cancers develop in the absence of both SMARCA2 and SMARCA4. 8
The SMARCA4 gene is located on chromosome 19p13.2 and encodes the BRG1 protein; the other catalytic subunit is BRM, encoded by the SMARCA2 gene (see above).
SMARCA2 and SMARCA4 were previously called as SNF2 and STH1 (SNF Two Homolog 1), respectively, when originally described in Saccharomyces cerevisiae.5 In investigating their role in S. cerevisiae, SMARCA2 serves as catalytic subunit for the multiprotein SWI/SNF and, SMARCA4 is involved in the remodelling of chromatin structure.
In humans, SWI/SNF and chromatin remodelling complexes are called BAF (BRG1-Associated Factors) and PBAF (Polybromo-BRG1-Associated Factors).5 SMARCA2 and SMARCA4 function as mutually exclusive ATPases within the BAF complex, while SMARCA4 only is present in the PBAF complex. SMARCA2 and SMARCA4 increase chromatin accessibility allowing for double strand damage repair.
Antibodies to both SMARCA2 and SMARCA4 are available. Retained staining is nuclear in location.
SMARC4 deficiency and malignancies
Small cell carcinoma of Ovary, hypercalcaemic type
Loss of SMARCA4 protein expression resulting from inactivating sporadic or germline mutations in the SMARCA4 gene was discovered as the initiating and driver molecular event in a distinctly uncommon but highly aggressive primary ovarian malignancy occurring preferentially in children and young females: small cell carcinoma of the ovary, hypercalcaemic type.9 10 The morphology of this tumour may not always strike one as the typical ‘small round blue cell tumour’ and sometimes larger cells with appreciable amounts of eosinophilic cytoplasm may be encountered. Furthermore, the morphology engenders a wide differential diagnosis. Conlon and colleagues evaluated the diagnostic utility of SMARCA4 immunohistochemistry and demonstrated that 16 of 17 small cell carcinomas of the ovary, hypercalcaemic type showed loss of SMARCA4 expression (94%), while only two of 279 tumours (an ovarian clear cell carcinoma and ovarian melanoma) showed loss of SMARCA4 expression.10 They concluded SMARCA4 immunohistochemistry is highly sensitive and specific for small cell carcinoma of the ovary, hypercalcaemic type and very helpful in distinguishing from morphological look-alikes.10
Non-small cell lung cancer
SMARCA4 (BRG1) gene has been shown to be mutated in up to 10%–35% of non-small-cell lung carcinoma.11 Genetic analyses of lung cancer have displayed mutations particularly in SMARCA4 and recently, immunohistochemistry has been employed in cohorts of lung cancer that have SMARCA mutations.12–14 There was concomitant loss of SMARCA4 and SMARCA2 in 10% of non-small cell lung cancers in the study conducted by Reisman et al.12 Furthermore, they observed that SMARCA4/SMARCA2-deficient carcinomas have a statistically significant lower survival compared with those without loss, independent of tumour stage.12 On the other hand, the study by Matsubara and colleagues showed loss of SMARCA4 and SMARCA2 protein expression immunohistochemically in 12% and 17% of 93 primary lung adenocarcinomas, respectively.13
In a larger study of 316 lung cancers, Herpel et al, found complete loss of SMARCA4 in 5.5% of pulmonary adenocarcinomas and 5.2% squamous cell carcinomas.14 Loss of SMARCA2 was seen in 6.4% of adenocarcinomas and only 1.7% of squamous cell carcinomas. Interestingly, concurrent loss of both markers was observed in only four cases (26%).14
Endometrial undifferentiated carcinoma and sarcoma
Initially, SMARCA4 undifferentiated, high-grade endometrial carcinomas were described.15–18 In the initial descriptions, these tumours were either undifferentiated or dedifferentiated carcinoma, invariably had a rhabdoid cell component. Tumour cells vary from small to larger more epithelioid (figure 1A,B). Ramalingam and colleagues then demonstrated that approximately one‐third of endometrial undifferentiated carcinomas show loss of SMARCA4 and SMARCA2 protein expression, and only a proportion exhibited a rhabdoid morphology.18 Furthermore, the majority of the SMARCA4‐deficient cases showed concomitant loss of SMARCA2 expression and, this group did not find any correlation between SMARCA4 or SMARCA2 protein expression with outcome.18 More recently, Kolin and colleagues described so-called ‘SMARCA4-deficient undifferentiated uterine sarcoma’, an entity with morphological, immunohistochemical and molecular overlap with small cell carcinoma of the ovary, hypercalcaemic type.19 These tumours were composed of rhabdoid cells, showed loss of SMARCA4 by immunohistochemistry and were highly aggressive with adverse outcomes.19
Given the morphological overlap with undifferentiated/dedifferentiated endometrial carcinomas, this same group undertook a comparison of undifferentiated endometrial cancers and SMARCA4-deficient uterine sarcomas.20 The latter tumour group occurred in younger patients, was characterised by a phyllodes, leaf-like pattern (‘phyllodiform’) with the tumour pushing into normal endometrial glands resembling a uterine adenosarcoma and, by inactivating mutations of SMARCA4 with consequent loss of SMARCA4 protein expression.20
Gastrointestinal undifferentiated carcinomas
Agaimy and colleagues examined a series of undifferentiated, pleomorphic carcinomas from the gastrointestinal tract.21 These tumours were characterised morphologically by sheets of medium to large undifferentiated cells, pleomorphic giant cells and many of the tumours contained rhabdoid cells (figure 1C,D).21 Glandular differentiation can be seen focally, but the dominant morphology is that of an undifferentiated tumour. They are cytokeratin positive and, in the series, examined 12/13 (92%) of cases showed loss of a SWI/SNF protein with SMARCA2 being the most common and lost in 10 cases and SMARCA4 lost in two cases.21 Interestingly, all SMARCB1 deficient cases were also SMARCA2 deficient.21
Thoracic undifferentiated sarcomas
These malignancies are characterised by sheets of monotonous ovoid cells with indistinct cell borders, abundant eosinophilic cytoplasm and prominent nucleoli. Scattered foci with rhabdoid morphology (eccentric nuclei, dense eosinophilic cytoplasm, dyscohesion of tumour cells) were present in all the cases.22–24
Concomitant loss of SMARCA4 and SMARCA2 immunoexpression was a consistent feature in these tumours. In addition, most tumours expressed epithelial markers (pan-keratin or epithelial membrane antigen). Interestingly, the vast majority of these tumours are TTF-1 (thyroid transcription factor 1) negative. A recent analysis of SMARCA4-deficient thoracic sarcomas and non-small cell lung cancers concluded that the former are cigarette smoking-associated undifferentiated/dedifferentiated carcinomas rather than primary thoracic sarcomas.25 Despite this histogenetic relationship with non-small cell lung cancer, the authors still felt that there was sufficient clinicopathological uniqueness warranting their recognition as a distinct entity.25
Sinonasal undifferentiated carcinoma
Agaimy and colleagues described 10 cases of sinonasal, undifferentiated carcinomas all of which were SMARCA4 deficient.26 These tumours were characterised morphologically by small basaloid cells or large epithelioid cells organised in nests as well as solid sheets with extensive areas of necrosis. No evidence of glandular or other forms of differentiation were observed in these cancers.26 In this particular cohort of cases, only one case showed concomitant loss of SMARCA4 and SMARCA2.26
Renal cell carcinomas with undifferentiated/rhabdoid components
Like in other sites, cases of undifferentiated or anaplastic renal cell carcinomas with rhabdoid cells lacking expression of SMARCA proteins have also been described.27 A differentiated component was noted in approximately two-thirds of the cases and the undifferentiated component was composed of pan-keratin-positive rhabdoid, large epithelioid and small monotonous anaplastic dyscohesive cells.27 In this series SMARCA2 was most frequently lost with concomitant reduction or loss of both SMARCA4 and SMARCA2 in six of the 21 cases.27
SMARCA deficiency has been noted in other tumours: 15% of Burkitt’s lymphoma, 5%–10% of childhood medulloblastoma and occasionally in pancreatic adenocarcinoma, ovarian clear cell carcinoma and melanoma.28–32
The SMARCA5 (SNF2H) gene is located on chromosome 4q31.21 and is a member of the ISWI family of chromatin remodelers. It functions in a catalytic capacity by sliding and displacing nucleosomes by various mechanisms, including coupling with poly ADP-ribosylation by poly (ADP-ribose) polymerase 1 (PARP1), ubiquitinylation, deacetylation and possible methylation.5
The SMARCAD1 gene is located on chromosome 4q22.3 and belongs to the INO80 family of chromatin remodelers. It plays important roles in transcription regulation, mitosis and double strand breakage repair.5
Role in malignancies
SMARCD1 overexpression is implicated in a variety of cancers including skin, bladder, breast and pancreatic.35
The SMARCA group of genes and proteins are implicated in several undifferentiated and differentiated cancers and, their role in carcinogenesis involves several pathways as summarised in figure 2.
Take home messages
SMARCA is a family of genes (SWI2/SNF2 family) who proteins play a role in chromatin remodelling, especially in double-strand damage repair.
Inactivating mutations in these genes leads to loss of protein expression in the nuclei of tumour cells.
Tumours associated with SMARCA deficiency are poorly to undifferentiated high-grade malignancies made up of small to large epithelioid cells, giant cells and rhabdoid cells.
They represent undifferentiated or dedifferentiated carcinomas as pan-keratin immunopositivity can be seen.
They are encountered mainly in the ovary, lung, thoracic cavity, gastrointestinal tract, sinonasal tract and uterus.
SMARCA4 and SMARCA2 proteins may be lost concomitantly or independently of each other.
Handling editor Cheok Soon Lee.
Contributors Both authors contributed equally to the concept, design, writing and editing.
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.
Patient consent for publication Not required.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.