A Rare Partner of TFE3 in the Xp11 Translocation Renal Cell Carcinoma: Clinicopathological Analyses and Detection of MED15-TFE3 Fusion

Xp11 translocation renal cell carcinoma (RCC), a member of the microphthalmia-associated transcription factor (MiTF) family, is a rare renal tumor characterized by different translocations involving the TFE3 gene. Here, we reported a case of Xp11 translocation RCC with a rare MED15-TFE3 gene fusion by RNA sequencing. Morphologically, the tumor cells were arranged in a solid and small nest pattern. The cytoplasm was voluminous, flocculent eosinophilic, and vacuolated. The nuclei were round or polygon with fine granular chromatin, and the nucleoli were unconspicuous. Psammoma bodies were observed in mesenchyma. Immunohistochemically, the tumor cells were diffuse moderately or strongly positive for CD10, P504S, vimentin, PAX8, RCC, AE1/AE3, and SDHB and focally positive for CK7 and CA IX while negative for cathepsin K, HMB45, Melan-A, Ksp-cadherin, and CD117. The Ki67 proliferation index was approximately 3%. However, TFE3 labeling showed an uncertainly weak nuclear staining and been considered negative. Fluorescence in situ hybridization (FISH) demonstrated a positive result that splits signals with a distance of > 2 signal diameters. Subsequently, RNA sequencing confirmed a fusion of MED15 gene exon 11 on chromosome 22 with TFE3 gene exon 6 in the tumor. The patient was alive with no evidence of recurrence. Our report contributes to the understanding on MED15-TFE3 RCC.


Introduction
Microphthalmia-associated transcription (MiT) family translocation renal cell carcinoma (tRCC), consisted of Xp11 tRCC and t(6; 11) RCC, is a distinct cancer subtype named in 2016 WHO classification of renal tumor [1]. Xp11 tRCC is featured by a short arm of X chromosome translocation with other chromosomes, leading to TFE3 gene fusion. It is commonly reported in children and young adults, especially in young women, accounting for about 40% of RCC in children and 1.6-4% in adults [2][3][4]. e morphological manifestations of Xp11 tRCC are diverse. Additionally, they are occasionally confused with other common types of RCCs. TFE3 translocation-associated RCC was divided into different genotypes according to the target genes of the translocation. e morphological, immunophenotype, and prognosis of different subtypes were different and distinct. However, inadequate description has been given on it by WHO histological classification of RCC due to rarity. erefore, its pathological diagnosis is still a challenge. To make a definitive diagnosis, several techniques are required including detection of TFE3 antibody based on the immunohistochemical method, fluorescence in situ hybridization (FISH) detection of TFE3 fusion gene, and RNA sequencing. e relatively common fusion couples of TFE3 gene included ASPSCR1, PRCC, SFPQ, and NONO [5][6][7][8][9]. On rare conditions, TFE3 may fuse with other couples including CLCT, RBM10, PAPR14, MATR3, LUC7L3, FUBP1, DVL2, KHSRP, GRIPAP1, and MED15 [10][11][12][13][14][15][16][17][18][19].
In this study, we reported a case of MED15-TFE3 RCC confirmed by high-throughput RNA sequencing. Also, FISH involving the utilization of TFE3 break-apart probes was performed. Our data confirmed that the histological morphology and immunophenotype were different from the features described in the previous literatures [17,18].

Materials and Methods
2.1. Immunohistochemistry Staining. Tissue samples were obtained from a female patient with left-sided renal cancer. For the immunohistochemical analysis, tissues were fixed on 10% formalin, followed by embedding using paraffin. en, the tissue sections (4 μm) were treated with the Ventana BenchMark XT automated IHC stainer (Roche, Basel, Switzerland). Sections treated with PBS served as the negative control. As a positive control for TFE3, we selected one renal carcinoma associated with the Xp11.2 translocation presenting the TFE3 gene translocation by FISH. While the positive control of others was using the specific tissues according to the manufacturer's instructions. e protein antibodies utilized in this analysis were as follows: TFE3 TFE3 nuclear immunoreactivity was scored from 0 to 3 + referring to the criteria proposed by Argani [20]. e criteria were as follows: 2 + and 3 + were positive; 1 + was considered negative.

FISH.
Hematoxylin and eosin-stained tissue slides were prepared and were observed under a microscope to count the tumor cells in the selected fields. Upon deparaffinization with xylene thrice, the slides were washed twice with absolute ethanol. Digestive enzyme K was incubated with 40 ml 2 × SCC at 37°C for 20 min. e mixture was rinsed thrice at room temperature and was dehydrated with precooling gradient alcohol. TFE3 break-apart probe (Empir-eGenomics, Buffalo, NY) was used. e centromere side was labeled with green fluorescence, while the telomere side was labeled with red fluorescence. e 10 μl probe was dropped on the slide and then was quickly covered with a 2,222 mm cover glass. Afterwards, the section was slightly pressed for the even distribution of the liquid, followed by sealing with rubber cement. Denaturation was conducted by incubating the slides at 85°C for 5 min in a humidified box followed by hybridization at 37°C overnight. After removal of cover glass, the slide was washed with 0.1SSC/1.5M urea at 37°C for 10 min. Subsequently, the slide was washed again using 2 × SSC/0.1% NP-40 for 5 min at 37°C. e slides were put into 70% ethanol thrice and were air dried. e nuclei were counterstained with 4, 6-diamidino-2-phenylindole. After hybridization, all slides were maintained at 4°C in the dark.
At least 100 tumor nuclei with clear boundaries and no overlap were counted by fluorescence microscopy under a magnification of 1,000×. Positivity was defined as a distance of separation between red and green signals of more than 2 signal diameters in more than 10% tumor cell nuclei.

2.3.
High-roughput RNA Sequencing. RNA was extracted from the paraffin-embedded cancer tissues, and the RNA concentration quantified with Qubit precisely; RNA purity detected by Nanodrop, Agarose gel electrophoresis and Agilent 2100 Bioanalyzer was used to detect RNA integrity. Magnetic beads with oligo-dT were utilized to purify the mRNA. Agencourt SPRIselect Reagent kit was used for the purification of libraries and selection of fragments. Library concentration and library fragment length distribution were accessed using the Qubit and Agilent 2100 Bioanalyzer, respectively. High-throughput sequencing was performed on Illumina Hiseq platform with 2 × 150 bp double-end sequencing mode to get FastQ data. FusionCatcher and BLAT aligner algorithms were applied for the detection of any potential TFE3 fusion of RNA-sequencing data.

Clinical Information.
e patient was a 35-year-old woman presented to our department due to fatigue without other symptoms. She showed a history of hepatitis B for more than 10 years. ere was no family history of tumor. Abdominal computed tomography (CT) scan revealed a cystic solid density lesion at the lower pole of her left kidney, which protruded the renal parenchyma with a clear boundary. Imaging diagnosis confirmed RCC with a Bosniak grade of 3-4 ( Figure 1). Radical nephrectomy was performed. Macroscopy showed a well-circumscribed cystic solid mass at the lower pole of the kidney with a tumor size of 5 × 4 cm in a grayish red color of a tough texture. For the TNM staging, the patient was in a stage of T1N0M0, clinical stage I. e patient was still alive without occurrence after nephrectomy.

Histopathology.
Microscopically, the tumor was well defined with a pseudocapsule invasion into the capsule locally at a low magnification of 0.25×. Most parts of the tumor were solid and cystic structures were visible in some areas (Figure 2(a)). Under a higher magnification of 10×, the capsule wall was lined with flattened tumor cells or fibrocystic wall (Figure 2(b)). e tumor cells were arranged in a profile of solid and small nests (Figure 2(c)). e cytoplasm was voluminous, flocculent eosinophilic, and vacuolated ( Figure 2(d)). e nuclei were round or polygon with fine granular chromatins, and the nucleoli were unconspicuous. Normal renal tubules were involved in the tumor focally ( Figure 2(e)). In the mesenchyma of cancer tissues, there was a lack of capillary network, and psammoma bodies were observed (Figure 2(f )).

FISH.
Because of the indeterminacy of TFE3 immunohistochemistry stain, dual-color FISH using a TFE3 break-apart probe was examined. e result indicated that the distance of green and red signals surpassed the width of more than 2 signals. In total, 129 cells were counted and signals split were detected in 65 cells (50.39%), indicating the fusion of TFE3 gene (Figure 4).

RNA Sequencing.
To further investigate the gene partner fused with TFE3, high-throughput RNA sequencing was performed. Using both the FusionCatcher and BLAT aligner algorithms, we could detect 10 reads on MED15 and 5 reads on TFE3, suggesting a fusion of MED15 gene exon 11 on chromosome 22 with exon 6 of TFE3 gene on chromosome X. Fusion abundance was 38% ( Figure 5).

Discussion
Renal carcinoma associated with Xp11.2 translocations/ TFE3 gene fusions is an unusual renal tumor, which is recognized as an entity in 2004 WHO classification of tumors of the urinary system [1,21]   [23] showed the tumor presented papillae and granular cytoplasm with clearing; however, no psammoma bodies were identified. Immunohistochemically, all cases in documents were strongly positive for TFE3, six cases were positive for PAX8, as well as diffusely positive for melanocytic markers cathepsin K and Melan-A. In this study, we presented a case with cystic and solid construction under macroscopic condition and the capsule wall was lined with flattened tumor cells or fibrocystic wall under the microscope. e tumor was composed of clear or eosinophilic cells arranged in a solid and small acinar to nest pattern, which showed a relative simple structure compared with the mixed structures of papillary, acinar, and cystic. e cytoplasm of the tumor cells was voluminous, flocculent eosinophilic, and vacuolated mimicking SDHB-deficient RCC and with small nuclei without nucleoli (WHO/ISUP grade 1). Psammoma bodies were observed in mesenchyme. Interestingly, there were normal renal tubules involved in tumors, which had not been mentioned previously and showed some differences in histomophology. In addition, the immunotype of the case was also different from that reported previously [18,23], as the neoplastic cells were negative for cathepsin K, HMB45, Melan-A, and TFE3 which showed an uncertainly weak nuclear staining and considered negative. Cathepsin K is a cysteine protease from the papain family playing an important role in osteoclast function. Expression of cathepsin K in osteoclasts is regulated by MITF [24]. Cathepsin K was demonstrated to be a transcriptional target of the microphthalmia-associated transcription factor family, which revealed that Cathepsin K was differentially expressed depending on the fusion partner of the TFE3 gene [25,26]. In the known genotypes, PRCC-TFE3 gene fusion showed a positive reaction with antibody to cathepsin K, while ASPL-TFE3, SFPQ-TFE3, and NONO-TFE3 gene fusion showed negative staining [25]. Martignoni et al. [26] speculated that different expression of Cathepsin K between subtypes of Xp11 translocation renal cell carcinomas might be contributed to some of clinical, pathological, and biological differences. Cathepsin K expression was strongly and diffusely positive in all the 6 rare MED15-TFE3 gene fusion cases previously reported [18]; however, it was negative in our case. We suspected this difference in expression was related to the simplicity of the morphological structure.  TFE3 antibody is a characteristic marker of Xp11 translocation RCCs for the overexpression of the functional TFE3 fusion protein. However, immunohistochemical methods are susceptible to false-positive, false-negative, and uncertain results due to various factors, such as tissue fixation time, antigen retrieval mode, antibody clonal number, and data interpretation [27]. In this study, TFE3 showed a weak nuclear staining and was considered negative, while the positive control was strong positive, indicating the reliability of staining results and excluded false negatives caused by antibody clonal number. Rao et al. [27] compared the sensitivity and specificity of TFE3 immunohistochemical staining and TFE3 break-apart FISH assay in the diagnosis of Xp11 translocation RCCs. eir data showed that TFE3 break-apart FISH assay was a useful complementary method for confirming the diagnosis of Xp11 translocation RCC, especially when the morphologic or clinical suspicion was high but TFE3 immunostaining was negative or equivocal. TFE3 immunohistochemical staining was negative in our case; thereupon, TFE3 break-apart FISH assay was performed and the presence of TFE3 translocation was confirmed.
To further investigate the gene partner fused with TFE3, we performed high-throughput RNA sequencing. Our data confirmed that MED15-TFE3 RCC was an extremely rare fusion genotype. RNA sequencing showed MED15 exon 11 fused with TFE3 exon 6. MED15 was a part of the multiprotein mediator complex, which functioned as a bridge between regulatory proteins and RNA polymerase II (Pol II), thereby regulating the Pol II-dependent transcription [28,29]. According to the previous studies, MED15 was overexpressed in 35% of primary head and neck squamous cell carcinoma and was known to be involved in castrationresistant prostate cancer [30][31][32]. MED15-TFE3 fusion was also obtained in one case of melanotic Xp11 neoplasm [18]. Recently, it was reported to be a promoter of tumor progression and metastatic spread in renal cell carcinoma [33]. In future, further studies are required to investigate the role of MED15 in Xp11 translocation RCCs. e clinical manifestations of Xp11 translocation RCCs included hematuria, abdominal mass, abdominal pain, and weight loss. ese conditions showed no obvious specificity compared with other renal tumors. e survival time for patients with Xp11 translocation RCCs was similar to that with clear cell RCCs [34]. Compared with the patients with papillary RCCs, their survival time was significantly shorter [35]. Caliò et al. [36] reviewed 403 cases of Xp11 translocation RCCs described in the literatures and considered that there was no statistical difference of age between aggressive and nonaggressive cases, while a larger tumor size correlated with aggressive behavior [36]. More patients with ASPSCR1-TFE3 RCCs tended to show metastasis than those with PRCC-TFE3 RCCs (75% vs. 36%) [37]. However, most of the node-positive ASPSCR1-TFE3 RCC patients remained disease-free without adjuvant therapy [2,34]. Hence, locally advanced stage may not predict adverse outcomes. To our best knowledge, only 8 cases of MED15-TFE3 RCC were reported including our case. Five cases were symptom free in the 2-48 months follow-up, one case showed lung metastases after 15 years, and one case had no follow-up data [17,18,23]. e present case showed a low nuclear level in histology and followed up for 5 months with no evidence of recurrent disease. As the sample size was indeed small and the follow-up duration was short, long-term follow-up was still required to investigate the prognosis and true biological behaviors. Radical resection is one of the optional methods similar to that for conventional RCC. Adjuvant therapy is feasible, such as immunotherapy using cytokines, including interleukin-2 and interferon-alfa, but the curative effect is different [3]. In recent years, TFE3/IRS-1/PI3K/AKT/mTOR, as a potential dysregulated pathway in TFE3-tRCC, may serve as a therapeutic potential for vertical inhibition of such axis by using a dual PI3K/mTOR inhibitor for TFE3-tRCC patients [38].
ere are indeed some limitations in this study. Firstly, the sample size is not large due to disease rarity. Secondly, we cannot bring new information to the molecular mechanism of the disease. Our study is merely a report of a unique disease with a different subset.

Conclusions
In this article, we described MED15-TFE3 RCC, a rare gene subtype of Xp11 translocation RCCs, that was confirmed by FISH and RNA sequencing. e tumor demonstrated different morphological features and immunophenotypic characteristics with the cases reported in literatures, expanding our understanding on heterogeneity of MED15-TFE3 RCC. FISH analysis is an accurate and effective approach for the screening and confirmation of this tumor in the presence of uncertain TFE3 expression. RNA sequencing will help to identify this specific gene fusion subtypes, leading to more accurate diagnosis and better understanding of such type of tumor.