Introduction

Synovial sarcoma (SS) is an aggressive tumor that most often affects young adults, and accounts for 5–10% of soft-tissue sarcomas. The majority of tumors arise in the deep soft tissues of extremities, with the remaining occurring in the head, neck, trunk wall, and internal trunk, including the thoracic cavity [1]. SS can be categorized as a more common monophasic spindle cell type and biphasic type with epithelial nests/glands. A poorly differentiated pattern may be present in a subset. It is genetically defined by SS18 gene fusions including SS18–SSX1, SS18–SSX2, and SS18–SSX4 [1]. The SS18–SSX protein exerts oncogenic activity through various mechanisms that disrupt epigenetic control. For example, the fusion protein binds to the SWI/SNF chromatin remodeling complex, resulting in the displacement of native SS18 and eviction of BAF47 (SMARCB1). It also interacts with KDM2B to bring together the SWI/SNF complex and PRC1.1 on the unmethylated CpG islands to aberrantly reactivate the expression of developmentally regulated genes that are otherwise repressed by PRC2 [2, 3]. The SMARCB1 eviction from the SWI/SNF complex can be visualized as an immunohistochemical reduction of SMARCB1 staining, which is a diagnostically helpful finding because it is specifically seen in 90% of SS samples [4, 5]. SS is otherwise genomically silent and additional genetic alterations are rare in the primary tumors, with uncommon secondary mutations including TP53, PTEN, CTNNB1, APC, SETD2, and FBXW7 [6].

Primary intrathoracic SS is uncommon, but SS might represent the most common sarcoma type in the thoracic cavity [7, 8]. The management of intrathoracic SS has numerous challenges, including late detection, large tumor size, high histological grade, high patient age, and difficulty to obtain adequate surgical margin [7, 9]. In addition, correct diagnosis can be delayed by a small sample size and a variety of histological mimics. Consequently, intrathoracic SS is associated with poorer outcome than its soft-tissue counterparts, with frequent recurrence and metastasis [7,8,9,10,11]. The prognosis of recurrent/metastatic SS remains poor [12, 13], highlighting the need for a novel therapeutic strategy.

Herein, we describe two patients with SS that harbored BRAF V600E mutation, for the first time to our knowledge. The tumors in both patients primarily involved the thoracic cavity and one of them responded well to the combined BRAF/MEK inhibition, until it ultimately recurred with an additional NRAS mutation.

Materials and methods

The study was approved by the Institutional Review Board of the National Cancer Center, Tokyo, Japan (2012–374, 2014–089).

Next generation sequencing

We performed targeted next generation sequencing (NGS) using the NCC Oncopanel to capture coding exons and the reported translocated introns of ~100 genes. The genes targeted by the assay are listed in Supplementary Table 1. This assay is currently used in clinical practice and a detailed protocol has been published previously [14, 15].

Immunohistochemistry

A paraffin section of 4-μm thickness was cut from the representative block of each tumor. Heat-induced epitope retrieval was performed using Targeted Retrieval Solution pH9 (Dako, Glostrup, Denmark). The endogenous peroxidase activity was blocked using 3% hydrogen peroxide. The primary antibodies used were BRAF V600E mutation-specific antibody (clone VE1, dilution 1:200; Spring Bioscience, Pleasanton, CA, USA) and phosphorylated ERK (pERK, #4370, dilution 1:400; Cell Signaling Technology, Danvers, MA, USA). The slides were incubated for 1 h at room temperature with the primary antibodies and subsequently labeled using the EnVision system (Dako). Mouse LINKER (Dako) was used for BRAF V600E staining. BRAF V600E staining was considered positive when diffuse cytoplasmic staining was observed. pERK staining was considered positive when nuclear staining was observed in ≥10% of cells.

Results

Case summary of BRAF-mutant SS

Case 1: A 32-year-old Japanese woman presented with chest pain resulting from a 12-cm mass in the right thoracic cavity (Fig. 1a). A thoracoscopic biopsy revealed a monophasic SS showing a classic spindle cell morphology (Fig. 1b). The immunohistochemical analysis showed focal-positive staining for epithelial membrane antigen and cytokeratin, whereas SMARCB1 expression was reduced. RNA sequencing revealed an SS18–SSX2 fusion transcript, which was validated by fluorescence in-situ hybridization using the SS18 break-apart probe (Fig. 1c). The patient received chemotherapy with doxorubicin and ifosfamide, which resulted in some tumor shrinkage. She underwent tumor resection and adjuvant chemotherapy with doxorubicin and ifosfamide, leading to a complete remission. After 18 months, a pulmonary metastasis appeared and the resected specimen showed a similar histology. The patient was enrolled in a phase I trial of trabectedin, which led to transient disease control. Following tumor progression, she was treated with ifosfamide monotherapy and pazopanib. Clinical NGS using NCC Oncopanel v2 on the resected pulmonary metastatic tumor revealed a BRAF V600E mutation (Fig. 1d), with a variant allele frequency of 24.8% (280/1129 reads) in case of a histologically estimated tumor cell ratio of 75%. The tumor showed diffuse intense immunohistochemical reactivity using BRAF V600E mutation-specific antibodies decorating virtually all tumor cells, both in the primary and metastatic sites (Fig. 1e). Immunohistochemistry for pERK was also positive in both primary and metastatic tumors (Fig. 1f). The patient was unable to receive BRAF inhibitors because such trials were not open at that time. She received supportive care and died of the disease 43 months after the initial presentation.

Fig. 1: Case 1.
figure 1

Computed tomography image showing a large mass in the right thoracic cavity (a, arrow). The tumor showed a classic histology of monophasic synovial sarcoma consisting of a fascicular growth of uniform spindle cells (b). RNA sequencing revealed an SS18–SSX2 fusion, which was confirmed by positive evidence of SS18 rearrangement using FISH (c, isolated red signals indicate SS18 rearrangement). Clinical next generation sequencing revealed a BRAF V600E mutation (d), which was supported by diffuse intense immunohistochemical reactivity using BRAF V600E mutation-specific antibody (e). Phosphorylated ERK immunohistochemistry was positive (f).

Full size image

Case 2: A 23-year-old Japanese woman presented with Horner’s syndrome due to a 4.3-cm tumor in the superior mediastinum (Fig. 2a). The tumor was complicated by a hemothorax, which required immediate tumor resection. The tumor displayed classic monophasic SS histology (Fig. 2b), which was supported by RNA sequencing that revealed an SS18–SSX2 fusion transcript. This was further validated by reverse transcriptase polymerase chain reaction and Sanger sequencing (Fig. 2c). Five months later, she presented with pain in the right arm and shoulder due to local recurrence. She received a range of therapies including doxorubicin and ifosfamide, local irradiation, pazopanib, and ifosfamide monotherapy, all of which induced only transient tumor response with the subsequent regrowth. Clinical NGS using NCC Oncopanel v4 performed on the resected tumor revealed a BRAF V600E mutation (Fig. 2d), with a variant allele frequency of 53.8% (276/513 reads) in case of a histologically estimated tumor cell ratio of 100%. Immunohistochemically, the tumor showed diffuse intense positivity using BRAF V600E mutation-specific antibodies (Fig. 2e) decorating all tumor cells, and pERK staining was also positive (Fig. 2f).

Fig. 2: Case 2.
figure 2

Computed tomography image showing a mass in the superior thoracic cavity (a, arrow). The tumor showed classic histology of monophasic synovial sarcoma (b). RNA sequencing revealed an SS18–SSX2 fusion, which was validated by Sanger sequencing (c). Clinical next generation sequencing detected a BRAF V600E mutation (d), which was supported by diffuse intense immunohistochemical reactivity using BRAF V600E mutation-specific antibody (e). Phosphorylated ERK immunohistochemistry was positive (f).

Full size image

BRAF-mutant SS responded to BRAF/MEK inhibition

Patient 2 received a combination therapy of dabrafenib (BRAF inhibitor, 150 mg BID) and trametinib (MEK inhibitor, trametinib 2 mg QD), which led to a partial response according to RECIST version 1.1 and the tumor was not detectable in 3 months (Fig. 3a, b). The tumor showed continuous remission until 7.5 months after BRAF/MEK inhibition, when it locally recurred (Fig. 3c). The biopsy of the recurrent tumor showed similar monophasic SS histology and was immunoreactive to BRAF V600E and pERK. NGS of this recurrent specimen detected the BRAF V600E mutation; in addition, it revealed an NRAS Q61K mutation, which was undetectable in the primary specimen.

Fig. 3: Efficacy of BRAF/MEK inhibition against synovial sarcoma with BRAF V600E mutation (Case 2).
figure 3

Computed tomography image prior to the combination therapy showing a 2.1-cm recurrence in the right superior thoracic cavity (a, arrow). Following the administration of dabrafenib and trametinib, the mass shrank to 1.4 cm (−33% reduction, partial response) at 41 days. At 90 days, the mass was unmeasurable (b, arrow). The tumor recurred after 7.5 months (c, arrow). Biopsy of the recurrent tumor showed monophasic synovial sarcoma histology with BRAF V600E and a newly acquired NRAS Q61K as a mechanism of resistance.

Full size image

BRAF V600E immunohistochemical screening of archival SS tissues

The recurrent identification of BRAF mutation in the two patients with SS prompted us to undertake a retrospective survey of archival SS tissues. We performed BRAF V600E immunohistochemical analysis of 67 SS tumor tissues originating from 67 patients. The diagnosis of all tumors was reviewed by a soft-tissue pathologist (AY) and confirmed by histological analyses in addition to the positive evidence of SS18 rearrangement by FISH and/or reduced immunoexpression of SMARCB1. The tested cases were enriched for thoracic primary tumors (N = 23) arising from the lung, pleura, mediastinum, or chest wall. No additional tumors positive for BRAF V600E immunoreactivity were found, and the estimated prevalence of BRAF V600E mutation in SS was up to 2.9% (2/69) in all anatomical sites and up to 8% (2/25) in the primary thoracic tumor.

pERK immunohistochemistry in the cohort lacking BRAF V600E

With a hypothesis that some SSs lacking BRAF V600E mutation might harbor other activating mechanisms of the mitogen-activated protein kinase (MAPK) pathway, we examined pERK immunohistochemistry in 53 available SS tumor tissues that lacked BRAF V600E immunoreactivity. We found that 17 cases showed positive staining with a range of 10–80%, whereas the remaining 68% showed negative (0% or <10%) staining. Of note, pERK-positive and -negative cohorts were not significantly different with regard to patient age, sex, tumor site (thoracic vs. non-thoracic), histological type (monophasic vs. biphasic), or overall survival. Among the 17 pERK-positive tumors, the amount and quality of six specimens were adequate for the NCC Oncopanel v4 assay. One of the six tumors showed an activating mutation of FGFR2 (c.C758G, p.P253R) with an allele frequency of 20.5% in case of a histologically estimated tumor cell rate of 60%, whereas the remaining five tumors harbored no mutations in the genes that were covered by the panel. The FGFR2-mutant tumor was a biphasic SS in the forearm of a 29-year-old woman.

Discussion

To the best of our knowledge, this is the first report to describe a BRAF V600E mutation in SS. BRAF mutation has not been detected in >180 SS tissues sequenced previously [2, 16,17,18,19,20,21,22,23]. BRAF is a member of the RAF family of serine/threonine kinases that plays key roles in the canonical MAPK cascade, which conveys signals from the surface receptor tyrosine kinase through RAS toward the downstream MEK and ERK. The oncogenic BRAF mutation is reported in a wide variety of human neoplasms, of which BRAF V600E is the most common type accounting for >90% of all reported mutations. The estimated frequency of BRAF V600E mutation in SS is similar to other solid tumors in which small subsets of cases harbor this alteration, according to historical and recent observations [24,25,26]. From a diagnostic standpoint, the BRAF V600E mutation in spindle cell “sarcoma” is often suspected as a clue for misdiagnosed sarcomatoid (“dedifferentiated”) malignant melanoma [27, 28]. Although this suspicion is reasonable for many cases, our study shows that the mutation can also rarely occur in bona fide sarcomas. Of note, SW982 “synovial sarcoma” cell line with BRAF V600E mutation [29] likely did not originate from SS because it lacks SS18–SSX fusion [30].

Rare (<5%) but recurrent presence of BRAF V600E mutations in SS might open a potential avenue to targeted therapy. Diffuse intense immunoreactivity using BRAF V600E-specific antibodies, observed in all tested samples (i.e., primary, recurrence, and/or metastasis) of both patients, suggests an early driver role of BRAF mutation, rather than being acquired later in a small subclone, although the mutation is likely a secondary event to initiating oncogenic SS18–SSX fusion. Combined therapy using dabrafenib and trametinib has been proven effective for advanced malignant melanoma, anaplastic thyroid carcinoma, and non-small cell lung carcinoma carrying BRAF mutations [31,32,33,34], and is expected to be promising in other tumor types [35]. Patient 2 indeed showed response to combined BRAF/MEK inhibition; however, the tumor ultimately recurred 7.5 months later, with an additional activating NRAS mutation. This pattern of initial response and subsequent progression to targeted therapy is similar to observations in other tumor types. NRAS mutation is a known cause of resistance to BRAF/MEK inhibition in melanoma and non-small cell lung carcinoma with BRAF mutation [36, 37], and this particular NRAS Q61K has been reported in 2–8% of BRAF-mutant melanoma cases resistant to single-agent BRAF inhibition [38,39,40,41] or combined BRAF/MEK inhibition [37]. Although it is presently unclear how NRAS mutation confers resistance to BRAF/MEK inhibition [36], the proposed hypotheses based on preclinical NRAS-mutant melanoma models include compensatory signaling via other RAF family members [42] and/or activation of the PI3K pathway [43, 44].

In the present report, both BRAF-mutant SSs displayed a considerable degree of clinicopathological similarity, including monophasic subtype, SS18–SSX2 fusion, primary intrathoracic location, and patient age and sex with both of them being young adult women. Whether these shared features indicate any significance remains to be determined, but focusing on these particular clinicopathological aspects might be worthwhile to identify additional cases belonging to this molecular class.

Both BRAF-mutant SSs displayed high pERK immunoexpression. Because ERK phosphorylation is widely considered as a marker of MAPK pathway activation, this finding is consistent with the aberrantly enhanced signal transduction via this pathway induced by BRAF mutation. The finding is similar to some previous studies, in which pERK immunohistochemistry was uniformly positive in tumors with known genetic abnormalities that activate the MAPK pathway, such as those involving BRAF and MEK1 [45, 46]. We thus hypothesized that SSs lacking BRAF V600E mutation yet overexpressing pERK might harbor mutations in genes that encode other members of the MAPK pathway. We first showed by immunohistochemistry that pERK is variably expressed in a third of clinical tumor samples of SS, and, in one of the six such cases tested, we identified an activating FGFR2 mutation as a potential cause of ERK activation. Although FGFR2 mutation has not been reported in SS [2, 16, 17, 21, 22], FGFR2 P253R is a known oncogenic mutation that has been recurrently detected in human tumors including endometrial and lung carcinomas, and it is likely a gain-of-function mutation that alters extracellular domain and enhances the affinity to ligands [47, 48].

Our study is limited by a small number of patients with BRAF-mutant SS. Although one patient responded to targeted therapy, whether the response is specific for this small molecular subset remains unknown. In addition, subsequent BRAF screening was performed primarily by immunohistochemistry owing to a limited number of materials available. Hence, cases with alternative BRAF mutations might have been missed, although the lack thereof was confirmed by NGS in six cases with the highest pERK expression. Finally, it is unclear whether screening by pERK staining is effective to identify cases with actionable gene abnormalities within the MAPK pathway, because the correlation between gene mutation and pERK immunoexpression has not been consistent in the literature, and there are reports of BRAF-mutant colorectal carcinomas or melanomas that were negative for pERK immunohistochemistry [49, 50]. In addition, we noted that pERK staining tended to be more enhanced at the periphery than the center of the tissue section, as seen in a previous report [45], suggesting a contribution of pre-analytic parameters such as fixation time and time to fixation. Further studies are required to estimate the actual prevalence of BRAF-mutant SS and to accurately measure the therapeutic effect of targeted therapy.

In conclusion, we discovered that BRAF V600E mutation is present in a small subset of intrathoracic SS, and reported that one of such tumors showed transient response to BRAF/MEK inhibition. Our study suggests a role of MAPK pathway and potential clinical implication of BRAF mutation screening in SS.