Nuclear protein in testis (NUT) carcinoma (NC), is a rare carcinoma characterized by a chromosomal rearrangement involving the NUT midline carcinoma family member 1 (NUTM1) gene, also known as NUT gene, located on chromosome 15q14 (1–5). This entity was first described in 1991 in two independent case reports of mediastinal carcinomas characterized by the t(15;19) translocation (6, 7). Since then, tumors harboring NUTM1 translocation have been increasingly recognized with numerous cases reported in the literature. Similarly, the mechanisms underlying the multiple NUTM1 gene rearrangements have opened the door to better understand tumor pathogenesis and the role of NUT and other proteins in the epigenetics of this rare neoplasm and multiple other cancers. Here, we review the clinicopathologic features, methods of diagnosis, and the molecular genetics and epigenetic alterations known to date in NC.

NC is a rare, poorly differentiated carcinoma characterized by an aggressive clinical behavior and advanced stage at diagnosis (8). Although the cell of origin is unknown, it has been speculated that NC may arise from primitive neural crest-derived cells (8). Despite lack of knowledge regarding a definitive anatomical site of origin, NC has been described to typically originate from midline structures of the thorax or from head and neck, hence the original term “NUT midline carcinoma”, and to predominantly affect young patients or adolescents. However, over the last decade, this entity has been identified in patients of all ages (0 - 81.7 years) with a median age varying from 16 to 24 years, observed in four meta-analysis studies (9–12), and affecting females and males almost equally (9, 10, 12). The most common location of primary NC has been found in the thorax (~50%) followed by the head and neck region (~40%) (12), but NC can also rarely arise outside of midline locations such as bladder (13), ocular globe (13), salivary glands (14–28), brain (29) kidney (29–31), stomach (29), adrenal gland (2, 20), pancreas (32), soft tissue (29), and bone (33).

By histology, NC is a poorly differentiated malignant neoplasm that usually grows as nests and sheets of primitive cells without an overlying in situ component (34), most of them (~55%) without squamous differentiation (12), and often with areas of confluent necrosis (34–36) ( Figure 1A ). The cells may have little or moderate amount of eosinophilic or amphophilic cytoplasm with indistinct borders, imparting the appearance of a cellular syncytium ( Figures 1A–C ). Focal cytoplasmic clearing or vacuolization ( Figure 1C ) (28, 34) as well as lumen or pseudo-lumen formation have been described (34). The tumor cells can be widely infiltrative and demonstrate a high mitotic rate (34, 35). The nuclei are usually large and quite monotonous, with vesicular chromatin and distinct nucleoli, lacking the pleomorphism typically encountered in high-grade carcinomas (3, 34, 35). A peculiar feature described in the literature as a clue to the diagnosis is the presence of keratinization in the form of large cells with intercellular bridges, focal squamous “pearls” or frank abrupt keratinization (i.e., poorly differentiated cells immediately adjacent to well differentiated squamous cells) ( Figure 1B ). However, this morphologic feature is only observed in about 30% of the cases (12) and it can also be seen in HPV-associated basaloid squamous cell carcinomas (3, 35). The background stroma varies from edematous, slightly myxoid to fibrous with variable amounts of desmoplasia (14, 36). The presence of an intratumoral neutrophilic infiltrate is common and can be very prominent ( Figure 1C ), and occasionally an intraepithelial and stromal lymphocytic infiltrate may also be observed (14, 28, 37).

The pathogenesis of NC is characterized by translocation-associated fusion oncoproteins that block cell differentiation and promote cellular growth (5). This distinct feature of single chromosomal translocation resembles those found in hematopoietic and mesenchymal malignancies and distinguishes NC from other epithelial tumors where multiple sequential mutations are required for tumorigenesis (38).

NC is defined by the rearrangement of NUTM1 gene on chromosome 15q14, which is frequently fused to the bromodomain containing protein 4 (BRD4) gene on chromosome 19p13.12, resulting in the most characteristic reciprocal translocation t(15;19) observed in 70-88% of the cases (9–13, 39). The predominant oncogenic variant involves the in-frame fusion of BRD4 exon 11 to the start of NUTM1 exon 2 (4, 6, 40, 41). However, variations of the specific exon fusions are known to occur including BRD4 exon 11 to NUTM1 exon 1b (41), BRD4 exon 15 to NUTM1 exon 2 (41), and BRD4 exon 15 to a partially deleted NUTM1 exon 2 (fusion at the last 124 nucleotides of NUTM1 exon 2) (42). The BRD4-NUTM1 fusion gene contains nearly the whole coding region for NUTM1 (exons 1b/2 to 7) and the three well characterized domains of BRD4 including the two bromodomains (BD 1 and BD2) and the extra-terminal (ET) domain, and a bipartite nuclear localization sequence (NLS) (4, 5). Its carboxyl-terminal domain (CTD), known to interact with the core positive transcription elongation factor b (P-TEFb), is absent (43) ( Figure 2 ). The NUTM1 gene encodes an unstructured protein with two acidic transcriptional activation domains (AD1 and AD2), a NLS, and a nuclear export signal (NES) (5) ( Figure 2 ). About 15-30% of NCs have a variant translocation involving exon 9 of the bromodomain containing protein 3 (BRD3) gene on chromosome 9q34.2 (5, 12). BRD3 gene encodes a protein similar to BRD4. The in-frame fusion involving BRD3 partner gene includes almost the entire NUTM1 structure (exons 2 to 7) along with the dual BDs, ET domain and the bipartite NLS of BRD3 (5) ( Figure 2 ).

Other rarer NUTM1 fusion variants have been reported in about 6% of cases (12) including the nuclear receptor binding SET domain protein 3 (NSD3) gene (38), the zinc finger-containing protein encoding genes ZNF532 (44) and ZNF592 (45), and other yet unknown genes. The NSD3 gene is located on chromosome 8p11.23 and is considered a BET-binding protein. The NSD3 portion of the fusion (exons 1 to 7) lacks the Su(var)3-9, Enhancer-of-zeste and Trithorax (SET) domain and contains only one of the two Proline-Tryptophan-Tryptophan-Proline motif (PWWP) domains, whereas nearly all NUTM1 (exons 2 to 7) is included in the fusion (38) ( Figure 2 ). The ZNF532-NUTM1 fusion gene encodes only the first 2 of 12 zinc finger domains from ZNF532 and almost the entire NUTM1 coding sequence (part of intron 1 of NUTM1 and its remaining exons 2 to 7) (44) ( Figure 2 ). The first zinc finger included in the ZNF532-NUTM1 fusion gene encodes a putative zinc-ribbon domain that is predicted to bind nucleic acids directly (44). The ZNF592-NUTM1 resultant fusion protein contains the coding sequence of ZNF592 up to exon 10 fused with exons 2 to 10 of NUTM1. The ZNF592 moiety of the fusion protein retains the first 11 of 13 zinc finger domains (45) ( Figure 2 ).

In a subset of malignant solid tumors from soft tissue and other organs, of uncertain relationship to NCs, NUTM1 has been reported to be fused with YAP1 (46, 47), MXD1 (29), MXD4 (39, 48, 49), CIC (50, 51), BCORL1 (29), ATXN1 (52), and MGA (39, 53, 54), in which most of them have been described to occur within the context of histologically defined high-grade sarcomas likely to be associated with a distinct pathogenetic pathway.

A definitive diagnosis of NC requires demonstrating the presence of NUTM1 gene rearrangement, which can be confirmed by immunohistochemistry with a NUT specific monoclonal antibody (clone C52) ( Figure 1D ). The immunohistochemical stain detects the presence of NUTM1 protein and has been reported as a relatively sensitive (87%) and highly specific (nearly 100%) tool for the diagnosis of NC (69). Diffuse (>50%) and strong nuclear positivity for NUTM1 is considered sufficient evidence for NUTM1 rearrangement, obviating the need of highly specialized genetic testing (1, 69). An alternative to NUT immunohistochemistry is molecular analysis to detect a NUTM1 gene rearrangement using fluorescence in situ hybridization, reverse-transcriptase polymerase chain reaction, cytogenetics, next generation sequencing, or whole-exome sequencing-based approaches. These methods should be considered if NUT immunohistochemistry is not available or if the result is negative or equivocal, and suspicion of NC is still high (29, 41, 51, 91–93).

The prognosis and outcomes of NC are very dismal with a median survival of 6.5 months (12) and poor response to conventional chemotherapeutic agents or radiotherapy. About 50% of patients present with lymph node involvement or distant metastatic disease (3, 9), frequently seen in lung and bones, and rarely in adrenal glands, brain, bone marrow, and liver (10, 14, 26, 36).

Although not required for diagnosis, molecular techniques can be used to determine the specific NUTM1 fusion partner which could be of potential prognostic and therapeutic significance. Recently, Chau et al. (12) proposed a prognostic risk classification model for NC survival outcomes based in the largest cohort of NC patients (n = 141) analyzed to date. In this study, they identified three distinct risk groups of patients based on anatomic site and NUTM1 fusion type, composed by the following: (1) Group A, patients with non-thoracic primary NC and presence of BRD3- or NSD3-NUTM1 fusion, (2) Group B, patients with non-thoracic primary NC and presence of BRD4-NUTM1 fusion, and (3) Group C, patients with thoracic primary NC regardless of the type of NUTM1 fusion. Interestingly, NSD3- or BRD3-NUTM1-positive tumors of non-thoracic origin are associated with significantly better overall survival, followed by the group of non-thoracic primary NC with BRD4-NUTM1 fusion. On the other hand, those patients with thoracic primary tumors, regardless of the NUTM1 fusion have worst prognosis than the other subgroups (12).

At the current time, there is no standard treatment for this rare and aggressive form of cancer. However, a multimodal approach with aggressive initial surgical resection, systemic chemotherapy, and radiation therapy is currently adopted in clinical practice (11). A variety of chemoradiation therapy regimens have been used including intensive treatments commonly applied in other carcinomas, sarcomas, germ cell tumors, and other solid neoplasms. Some of the chemotherapeutic agents that have been used with some success include cisplatin, taxanes and alkylating agents (9, 11, 24, 33, 94). However, despite rapid response, tumors become treatment-refractory with early progression and poor overall outcome (9, 13).

Targeted therapy using small-molecule BET inhibitors, which are acetyl-lysine histone mimetic drugs, result in depletion of megadomains, proliferation arrest, and cellular differentiation (1, 90). BET inhibitors (e.g., Birabresib aka OTX015/MK-8628, Molibresib aka GSK525762, RO6870810, ODM-207, and NEO2734) have shown activity but no obvious survival benefit (3, 95–100), most likely due to toxicity effects (i.e., severe thrombocytopenia, gastrointestinal symptoms, anemia, and fatigue), limiting its use (3, 96, 98). However, by the time of the writing of this manuscript, there is an ongoing clinical trial where pediatric patients with solid tumors, brain tumors and lymphoma are being enrolled on a Phase I research study to evaluate the use of BET inhibitors (known as BMS-986158 and BMS-986378) in those patients described above and as possible treatments for NUT carcinoma in children. Both of these drugs are currently still being studied in adult patients (101). Other preclinical studies have shown that the BRD4-NUTM1 fusion gene is associated with global decreased histone acetylation and transcriptional repression of genes required for differentiation. Some in vitro and xenograft models have shown that this acetylation can be restored with histone deacetylase (HDAC) inhibitors such as Vorinostat, resulting in global increase in histone acetylation, squamous differentiation, and growth arrest (72, 102). However, the use of HDAC inhibitors has been limited due to the toxicity effects like those seen with BET inhibitors (72). Both of these novel targeted agents hold great promise, either alone or in combination with chemotherapy (37). In particular, preclinical studies have highlighted that BET inhibitors show synergism with immune checkpoint modulators (103–105).

Currently, patients can be enrolled into the International NUT Midline Carcinoma Registry (http://www.nmcregistry.org) which follows patient’s outcomes and may direct them to the institution running these trials (1, 35, 37). This international registry was originally established in 2010 and was created to raise awareness and disseminate the most updated information about NC, provide pathologic review to assist in the diagnosis of NC, and collect clinical data and response to treatment. This has allowed the creation of a repository of clinical specimens that will support future research (106).

In summary, this an overview of NUT carcinoma with a brief discussion of the main epidemiologic and clinicopathologic features, along with the main molecular genetics/epigenetics findings and therapeutics for this tumor. Although several NUTM1 translocations have been found to be associated with NUT carcinoma, to the best of our knowledge, up to this day there is no known specific etiology for the NUTM1 translocation. In addition, given the rarity and relatively recent description of this entity, studies about the complete mutational landscape of NUT carcinoma are not yet available. Therefore, additional mutations that could play a role in oncogenesis are not well studied or known at this point in time. The awareness of this deadly tumor and the understanding of the underlying molecular mechanisms, genetics, and epigenetics will be helpful in future research for the development of novel targeted therapies.

Conceptualization: VM and SP-O. All authors VM, KS, and SP-O drafted, revised, and approved the submitted version of this manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.