Abstract:

Background: NUT carcinoma is an aggressive malignancy characterized by translocations in the NUTM1 gene. There are currently no consensus treatment recommendations for NUT carcinomas.

Methods: Here we describe the case of a previously healthy male diagnosed with NUT carcinoma after presenting with sinus pressure and found to have a sinonasal mass and distant metastatic disease in the lungs. While pathologic evaluation and immunohistochemistry were consistent with NUT carcinoma, initial genomic profiling did not demonstrate a NUTM1 translocation.

Results: Whole transcriptomic RNA sequencing of the tumor revealed a YAP1-NUTM1 fusion. Based on an in vitro drug sensitivity screen, the patient was treated with gemcitabine and nab-paclitaxel, achieving a partial response that is ongoing at 6 months.

Conclusions: Unbiased transcriptomic sequencing may identify previously uncharacterized NUTM1 fusion partners. Gemcitabine and nab-paclitaxel is a well-tolerated combination regimen and could offer a novel treatment approach for NUT carcinoma.

Case Presentation

The patient is a now 40-year-old  now 4039-year-old, previously healthy male who initially presented with sinus and ear pressure. He was treated with three courses of antibiotics for presumed sinusitis, with only minimal relief. Ultimately, he saw an otolaryngologist and a computed tomography (CT) scan was performed showing only sinusitis, per report.  He underwent endoscopic sinus surgery and turbinate reduction, resulting in some relief, but his symptoms of congestion continued.  The patient sought a second opinion from another otolaryngologist and had a repeat CT scan.  By report, it again showed opacification of the sinus cavity, but no discrete mass.  The second surgeon took him back to the operating room and a mass was identified during surgical exploration.  The biopsy was interpreted as a poorly differentiated non-small cell carcinoma with features consistent with squamous cell carcinoma.  Positron emission tomography (PET) imaging showed a hypermetabolic nasopharyngeal mass, bilateral cervical lymphadenopathy, and pulmonary nodules. He was referred to our institution for a medical oncology consultation.

As part of his workup, pathology materials were requested for in-house review, which demonstrated a poorly differentiated non-small cell carcinoma with indistinct cell borders, enlarged, hyperchromatic, vesicular nuclei with prominent nucleoli, and frequent mitoses and apoptotic debris (Figure 1A).  Immunohistochemistry (IHC) showed positive staining with cytokeratins AE1/AE3 and CK5, and focal P40. There was negative staining with TTF-1, Napsin-A, synaptophysin, chromogranin-A, smooth muscle actin, p16, S100 and EBER (EBV-ISH). PD-L1 22C3 was expressed with a combined positive score (CPS) of 25. Based on the pathologic features, absence of risk factors for nasopharyngeal cancer, and the midline nature of the tumor, an additional immunostain for NUT (nuclear protein in testis) was requested. The NUT stain revealed strong, diffusely positive nuclear expression in approximately 95% of tumor nuclei, establishing the diagnosis of NUT carcinoma (Figure 1B). A transbronchial fine needle aspiration (TBNA) of the right upper lung mass also demonstrated poorly differentiated non-small cell carcinoma with identical morphologic features on hematoxylin and eosin (H&E) stain, confirming metastatic disease (Figure 1C). 

A targeted DNA and RNA-based next generation sequencing (NGS) panel (STRATA oncology) demonstrated a CDKN2A deletion, but did not identify a NUT rearrangement.  Subsequently, whole transcriptomic RNA sequencing (Tempus xT Panel) demonstrated a YAP1-NUTM1 rearrangement (Figure 2). Additional genomic variants identified include BAP1 p.Q665* (variant allele frequency (vAF) 52.3%), NOTCH1 pL2203fs (vAF 11.4%), CDKN2A/B copy number loss, and MTAP copy number loss. Tumor mutational burden was 4.7 mutations per megabase. The patient has a family history of non-melanoma skin cancer and was referred to medical genetics after the identification of the BAP1 mutation.

Treatment options were discussed. Given his young age, the patient strongly valued any chance at durable control and was very open to the risks or side effects of aggressive trials.  He was consented to a clinical trial combining a PD-1 checkpoint inhibitor and a TLR7 agonist.  Unfortunately, his cancer grew during treatment. Symptomatically the patient noted increased headaches in the temples, obstructed nasal passages, and intermittent blurry and double vision. MRI brain showed interval progression of the nasal, sinus, and nasopharyngeal mass, with new intracranial extension through the right cribriform plate and likely involvement of the clivus, as well as increasing extension through the right lamina papyracea with mass effect on the medial rectus muscle posteriorly (Figure 3A).  Chest imaging demonstrated interval progression of diffuse bilateral pulmonary metastases with pleural nodules, left pleural effusion, and bilateral hilar lymph node enlargement (Figure 3B).  For palliation of his symptoms, he was treated with 45 Gy in 15 fractions of radiation therapy to the nasopharynx, resulting in significant symptomatic benefit and partial response. 

Additional treatment options were considered.  We had previously treated a patient initially thought to have lung adenocarcinoma, who was subsequently re-diagnosed with NUT carcinoma based on the identification of a BRD4-NUTM1 translocation on genomic profiling.  Prior to sequencing, This patient was initially treated for stage III disease with cisplatin, pemetrexed, and radiotherapy, followed by durvalumab. Upon progression, carboplatin plus paclitaxel resulted in a partial response. When genomic profiling demonstrated a NUT translocation, a BET inhibitor (on single patient protocol) resulted in response but only transient disease control.  In searching for other options for our current patient, the clinical literature (reviewed below) did not suggest a high probability treatment option with any confidence.  We therefore chose gemcitabine (1500mg/m2 day 1, 15 of 28-day cycle) and nab-paclitaxel (175mg/m2 day 1, 15 of 28-day cycle) based on an in vitro drug screening study while, in parallel, generating a single patient protocol for a BET inhibitor to be used upon progression.1 The treatment resulted in complete response at his primary site and a partial response in lung lesions. Treatment was briefly interrupted to allow recovery after the patient was hit by a car while biking, and has been complicated by a single episode of culture-negative febrile neutropenia, grade 1 neuropathy, and a soft-tissue infection. The response is currently ongoing at 6 months

Discussion

NUT carcinoma (NC) is a rare, highly aggressive malignancy defined by the presence of rearrangements in the NUTM1 (aka NUT) gene. Previously referred to as NUT midline carcinoma because it often arises in midline structures of the head/neck and thorax, subsequent reports have noted that NC can arise from a number of anatomic sites.2 Histologically, NUT carcinomas have a monomorphic, poorly differentiated appearance with focal squamous differentiation seen in a subset of cases. As these histopathologic features can overlap with other poorly differentiated carcinomas, the diagnosis is dependent on staining or molecular testing for NUT.  In normal human tissues, the NUT protein is expressed only in post-meiotic spermatids. Strong diffuse NUT immunostaining is highly specific and diagnostic of NUT carcinoma in the appropriate setting. Currently the WHO guidelines state that only strong and diffuse (greater than 50%) nuclear staining should be considered a positive result.3 Pathologists need to be aware, however, that variable staining has also been reported in some germ cell tumors.4

Molecular testing: Early cytogenetic characterization indicated the majority of NCs (70%) harbor a clonal, reciprocal translocation between the NUTM1 and BRD4 genes t(15;19), generating an in-frame fusion gene encoding BRD4-NUTM1. Subsequently, multiple additional translocation partners for NUTM1 have been identified, including BRD35 and NSD36.  Recent DNA and RNA-based next generation sequencing (NGS) has demonstrated additional fusion partners in NC7 and revealed NUTM1 translocations in other solid malignancies (sarcomas, poromas, and CNS tumors) and leukemia.8  While identifying the specific translocation is not necessary for diagnosis, analysis of clinical outcomes suggests there may be differences in prognosis based on the binding partner.9  Moreover, identification of the translocation partner may also have implications for treatment as discussed below.

Several techniques are available for molecular characterization of NUTM1 rearrangements, each with unique test characteristics. While cytogenetics was used to initially describe the classical t(15;19) rearrangement, this assay is difficult to perform in practice due to the need for fresh tissue and would not identify novel NUTM1 fusions. Fluorescence in situ hybridization (FISH) using break apart probes for NUTM1 is another sensitive method to identify rearrangements, but cannot identify the translocation partner and is not a widely available test. With the broader use of NGS in oncology, multiple commercially available multigene panels now include NUTM1. However, the ability of a targeted DNA-based sequencing panel to identify fusions and particularly novel partners can be limited as we observed in our patient. While NUTM1 was included on the initial NGS assay performed in our patient, that assay was designed to detect fusions specifically with BRD3, BRD4, CIC and WHSC1L1, thus a translocation involving a novel partner was not identified. In contrast, “partner agnostic” RNA-based sequencing using assays such as the Archer FusionPlex or whole transcriptome RNA sequencing can identify novel fusions.

In our patient, a YAP1-NUTM1 fusion was identified by whole transcriptome RNA sequencing. YAP1 is a transcriptional co-activator, functioning downstream of the evolutionarily conserved HIPPO signaling pathway.10 To our knowledge, YAP1-NUTM1 fusions have not been seen in NUT carcinoma but have recently been described in poromas and certain porocarcinomas.11  Our case did not appear morphologically consistent with a porocarcinoma. Sekine and colleagues demonstrated YAP1-NUTM1 fusions activate TEAD-dependent transcription, suggesting a mechanism for their oncogenic function.

Therapeutic Options: Given the rarity of the disease, there are no prospective studies examining treatment options for NUT carcinoma. A registry has been established (NMCregistry.org) to collect pathologic and outcomes data. Clinical series have demonstrated NC patients treated with multi-modality therapy including surgery, radiation, and chemotherapy have better outcomes but prognosis is still very limited.9,12,13  For patients with metastatic disease, an optimal chemotherapy regimen has not been established.14  Case series have described success with multi-agent pediatric sarcoma treatment regimens; however, these can have significant toxicity.13

BRD4 is a bromodomain (BRD) and extra-terminal domain (BET) family member that plays an important role in gene transcription by binding to acetylated histones. BET family proteins have been shown to contribute to carcinogenesis and treatment resistance in multiple malignancies. Specifically in NUT carcinoma, preclinical studies with BET inhibitors lead to growth arrest and differentiation with the MYC oncoprotein being an important target of BRD4-NUTM1.15,16 Interestingly, BRD4 appears to play an important role in YAP mediated transcription and BET inhibitors have preclinical activity in YAP-dependent malignancies.17  This suggests BET inhibitors may be a therapeutic option for our patient despite the presence of a YAP1-NUTM1 translocation. Two recent early phase clinical trials have examined the activity of BET inhibitors in NC patients. Birabresib (MK-8628/OTX015) is a BRD2, BRD3 and BRD4 inhibitor.  In the phase Ib study, of the nine evaluable patients with NC three partial responses were observed with an additional three patients having stable disease. Duration of response ranged from 1.4 to 8.4 months.18  The results of a phase I/II study of molibresib (GSK525762) were recently reported.19 In the NC cohort, the overall response rate was 11% (2/19 confirmed partial responses). Overall 4 patients remained on study drug for longer than 6 months. Interestingly, these patients all had non-thoracic primaries and the NUTM1 fusion partner was BRD3, which may suggest differential activity against BRD family proteins. While these studies demonstrate proof of concept for BET inhibition in NC, responses were not durable. Moreover, there may be differences in the activity of BET inhibitors based on the specific breakpoints of the translocation.1 Combination strategies targeting the unique transcriptional dependencies of this aggressive malignancy are also being studied, with hope for improving outcomes for this challenging disease.20,21

Given the limited response observed with BET inhibitors, we sought alternate therapeutic approaches. Stirnweiss et al had previously reported genomic profiling and high throughput drug screening of a series of NUT carcinoma cell lines.1 Their analysis demonstrated recurrent RECQL5 alterations and deficiency in DNA-repair. Screening for drug sensitivity demonstrated increased sensitivity to anthracyclines, topoisomerase, and microtubule inhibitors prompting our decision to treat with gemcitabine and nab-paclitaxel.  The tolerability and encouraging treatment response seen in our patient suggest this combination is worth further study in NUT carcinoma. More detailed genomic analysis of the nature of the NUTM1 translocation, fusion partner, and co-occurring alterations may also play an important role in treatment selection.