JPED Journal

Introduction

Rhabdoid tumors (RT) are rare pediatric solid cancers with characteristic morphology and specific driver mutations affecting SMARCB1, less typically SMARCA4. Depending on localization, RT are classified into atypical teratoid RT of the central nervous system, RT of the kidney, and extracranial extrarenal RT of soft tissues. NTRK genes rearrangements as drivers in RT are highly improbable. Here we describe a SMARCB1-inactivated congenital soft tissue RT with an extremely aggressive clinical course, harboring a chimeric NTRK3 transcript, in a male infant

Case Presentation

A three-day-old male patient born with a visible lump in the right zygomatic region was hospitalized in a multi-profile facility for diagnosis and treatment planning. The lump, bluish-purple in color, 1.5 cm in diameter, wide and indistinctly outlined at the base, protruding under stretched, glossy skin, adherent to the dermis and painless on palpation. The child was examined by a maxillofacial surgeon who suspected a hemangioma. Subsequent ultrasound scan of facial soft tissues and MRI of the brain with contrast enhancement revealed a mass in soft tissues of the right zygomatic region, of heterogeneous consistency, with a relatively sharp and even outline, spreading to the right eyelid and infiltrating the right temporal muscle, accumulating the contrasting agent at margins. By the age of 14 days, the tumor increased in volume to 3.0×2.5×2.0 cm.

At the age of 15 days, the patient underwent surgical resection of the mass. Cytology assay revealed medium-to-large nonhematopoietic malignant tumor elements with wide cytoplasm and eccentrically located nucleolated round nuclei, occasional binucleated and single multinucleated cells, suggestive of rhabdomyosarcoma. A vincristine-doxorubicin-cyclophosphamide scheme was commenced immediately, authorized by a tumor board on the basis of a life-threatening emergency in advance of histological diagnosis.

Histopathological examination revealed polymorphic neoplastic cells of medium size, with eccentric nucleolated nuclei and high mitotic and apoptotic indexes, infiltrating the parotid gland and dermal tissues (Figure 1). The antibody panel for immunohistochemistry was expanded considering the neonatal age of the patient and the superficial localization of the tumor. The tumor was positive for Vimentin, focally positive for EMA, Pancytokeratin, SMA, MSA, PanTRK, with solitary cells positive for Desmin and Calponin. Staining for s100, CK5/6, Synaptophysin, Chromogranin, p63, and SOX-10 proteins revealed no expression. The loss of nuclear expression of INI1 was evident. The antibodies, clones, dilutions, antigen retrieval methods, and vendors are listed in (Table 1). Appropriate positive and negative controls for each antibody were run concurrently. The findings confirmed the diagnosis of extrarenal extracranial rhabdoid tumor (Figure 1).

Figure 1: Proliferation of polymorphic medium-sized cells with eccentric nucleolated nuclei and abundant cytoplasm and high mitotic and apoptotic indexes in hyalinized stroma. H&E, x200. Positive reaction with anti-PanTRK antibody, diffuse strong positivity for Pancytokeratin and the loss of nuclear expression of INI1.

Table 1: Antibody panel used in this case.

Fluorescent in situ hybridization (FISH) test was performed and no NTRK fusions were detected (NTRK1, NTRK2, NTRK3 Break Apart FISH Probes, CytoTest Inc., MD, USA). Considering this, further molecular tests were performed to comprehensively address the diagnosis. The purified tumor DNA was probed for multi-exon deletions in SMARCB1 using the multiplex ligation-dependent probe amplification (MLPA) assay with SALSA® MLPA® Probemix P258 SMARCB1 (MRC Holland, the Netherlands). The analysis revealed a heterozygous deletion of SMARCB1 (NM_003073.5) exons 2-4 and homozygous deletion of SMARCB1 exons 5-9. MLPA assay of the germline material revealed retained SMARCB1 alleles, thus confirming the somatic origin of both deletions. Furthermore, targeted high-throughput sequencing of the tumor and normal DNA showed no additional genetic variants in SMARCB1 gene.

Considering the positive signal with panTRK antibody, the material was further tested for NTRK family gene rearrangement using a multimodal OncoScope™ Non-Small Cell Lung Cancer Solution panel (Parseq Lab, Russia). The panel allows identification of chimeric transcripts containing an NTRK1/2/3, ALK, ROS1, or RET gene fragment. The high-throughput sequencing revealed an ETV6 (NM_001987.5) exon 5 – NTRK3 (NM_002530.4) exon 15 fusion. The finding was validated by reverse transcription polymerase chain reaction (PCR) assay with oligonucleotide primers published by Vokuhl et al. (2018) [1]. Verification by Sanger sequencing showed unambiguous alignment of the product to ETV6 exon 5 and NTRK3 exon 15 (Figure 2).

Figure 2: Top: schematic representation of ETV6::NTRK3 fusion. The catalytic tyrosine-kinase domain of NTRK3 is entirely preserved in the chimeric oncoprotein and highlighted in yellow. Bottom: Chromatogram obtained at fusion transcript validation by RT-PCR and

Fluorescent in situ hybridization (FISH) test was performed and no NTRK fusions were detected (NTRK1, NTRK2, NTRK3 Break Apart FISH Probes, CytoTest Inc., MD, USA). Considering this, further molecular tests were performed to comprehensively address the diagnosis. The purified tumor DNA was probed for multi-exon deletions in SMARCB1 using the multiplex ligation-dependent probe amplification (MLPA) assay with SALSA^® MLPA^® Probemix P258 SMARCB1 (MRC Holland, the Netherlands). The analysis revealed a heterozygous deletion of SMARCB1 Sanger sequencing. The reading frame is retained.

Chemotherapy was initiated according to EU-RHAB V2.2010 protocol with extra bevacizumab between the second and third cycles. Unfortunately, despite the intensive treatment, the negative clinical dynamics with continued growth of the tumor persisted ultimately turning into rapid progression of the disease with a fatal outcome at the age of 5 months and 21 days.

Discussion

RT are relatively rare pediatric cancers with unfavorable prognosis. The incidence rate per 1,000,000 children under 15 years is 0.89 for atypical teratoid RT, 0.19 for renal RT, and 0.32 for extrarenal extracranial RT [2,3].

The morphology of RT is characteristic: rhabdoid cells show clear boundaries; the nuclei are large, eccentrically located, vesicular, bean-shaped, or rounded, with prominent nucleoli, occasionally cells are binucleated; the cytoplasm is abundant, eosinophilic, may contain hyaline inclusions. Cell morphology can also be spindleshaped, epithelioid, or neuroectodermal, and may vary within one tumor; the architecture may be layered, papillary, acinar, or trabecular. Other features include infiltrative growth patterns, high mitotic activity with abnormal mitotic figures, areas of necrosis, and hemorrhages [3].

Immunohistochemically, the tumors are positive for Pancytokeratin, EMA, CD99, Synaptophysin, SMA, Vimentin, GFAP, NFP, MSA, S100, Sall-4, Glypican-3. A characteristic, diagnostically decisive feature is the loss of signal for INI1 (less typically, BRG1) [3]. RT pathogenesis has been associated with inactivating mutations in SWI/SNF subunit-encoding genes. The highly conserved SWI/ SNF chromatin remodeling complex is a crucial coactivator of gene expression in the nucleus. The core of the SWI/SNF complex binds chromatin to promote a release of energy sufficient for the dissociation of genomic DNA from nucleosome histones. The loops of DNA expand by nucleosome repositioning and expelling of histone octamers, creating the free chromatin zones in promoter and coding regions of genes to enable the RNA-polymerase-IImediated transcription [4]. RT pathogenesis is primarily driven by inactivating genetic events affecting this complex, most often SMARCB1 gene encoding a SWI/SNF core subunit INI1. An immunohistochemically test for INI1 is mandatory for differential diagnosis in suspected RT.

The loss of INI1 causes widespread dysregulation of the gene expression program. The effects are profound enough to promote oncogenic transformation; accordingly, RT are minimally burdened by extra genetic events and genomically balanced. About 5% of RT are driven by inactivating mutations in SMARCA4 encoding another SWI/SNF subunit, BRG-1 [3]. The impacts of SMARCB1 and SMARCA4 pathogenic variants are similar, as both genes encode core subunits in SWI/SNF and determine specific binding of the complex to chromatin.

The spectrum of tumors involving SMARCB1 or SMARCA4 inactivation is not limited to RT but also includes small-cell carcinoma of the ovary (hypercalcemic type), SMARCA4-deficient thoracic sarcoma and cribriform neuroepithelial tumor (CriNET). These tumors share rhabdoid morphology, unfavorable clinical course, and poor response to chemotherapy (except CriNET) [5-7].

Chimeric oncoproteins produced by chromosome breakdown and abnormal fusion are central to many cancers. Exemplary in this regard are NTRK rearrangements that involve NTRK1, NTRK2, and NTRK3 genes mapped to, respectively, 1q21-q22, 9q22.1, and 15q25 chromosomal regions. The corresponding proteins TRKA, TRKB, and TRKC are receptor tyrosine kinases (TRKs) with a conserved LRR1-3 motif. TRK ligands are secretory proteins called ‘neurotrophins’ including nerve growth factor NGF, brain-derived neurotrophic factor BDNF, and neurotrophins-3/4. The ligand-bound TRK phosphorylate their substrates to trigger RAS/RAF/MEK/ERK, PI3K/AKT/mTOR and PLCγ signaling cascades with multiple biological effects. Normal TRK-mediated signaling plays major roles in the embryonic development of the nervous system, synaptogenesis, axon and dendrite outgrowth, differentiation of sensory ganglia, nociception, proprioception, memory formation, as well as in cardiovascular, reproductive, and immune functionalities. Most significantly in the context of cancer, TRK-mediated signaling can support proliferation and suppress apoptosis.

The pathogenic aberrations comprise a 3′ kinase domain-encoding region of NTRK fused to a 5′ region of another gene, typically encoding a dimerization domain and controlled by a strong promoter. The unhinged, constitutive catalytic activity of TRK decoupled from the ligand binding represents a crucial driving force for oncogenesis [8].

NTRK fusion oncogenes are found in colorectal cancer, papillary thyroid carcinoma, non-small cell lung carcinoma, spitzoid tumors and melanomas, infantile fibrosarcoma, gastrointestinal stromal tumor, secretory breast carcinoma, mammary analog secretory carcinoma of the salivary glands, acute myeloid leukemia, high-grade glial tumors, congenital mesoblastic nephroma and other malignancies [9]. However, NTRK rearrangements in RT are highly improbable. To the best of our knowledge, the only published case of NTRK-rearranged RT (specifically, atypical teratoid RT) was identified by high-throughput RNA sequencing of tumor samples biobanked at the Institute Curie, France [10]. The simultaneous incidence of NTRK and SMARCB1/A4 aberrations in other cancers is equally low. Boulanger et al. (2023) describe an aggressive, therapy-resistant case of squamous cell lung carcinoma harboring aberrations of both NTRK1 and SMARCA4 [11]. Thyroid carcinomas [12] and sinonasal tract melanomas can harbor NTRK rearrangements and SMARC genes variants, but not in the same tumor [13]. The combination of an alteration in the epigenetic modifier and a mutation leading to kinase activation is non-typical for tumors, however L. Auffret et al. described a new subtype of diffuse midline glioma, H3 K27 and BRAF/FGFR1 coaltered. Authors stated that the mutation in epigenetic modifier (H3 K27M) occurs prior to the kinase activating genetic event (BRAF V600E) [14]. Consistently, in the described case we consider the SMARCB1 as a primary oncogenic driver and the ETV6::NTRK3 fusion as concurrent genetic alteration.

Regarding the negativity of the FISH probe for NTRK fusions in the discussed case, the information about sensitivity and false negative rates is limited and inconsistent to date. While some authors report a low false negative rate of 2,3% [15], other studies demonstrate limited sensitivity of 78% [16]. Considering this, one can assume that this case illustrates the probability of a false negative FISH result. However, major reasons for false negative FISH results are noncanonical breakpoints and NTRK partner genes [17,18], and the exact reasons for the inability to detect a canonical ETV6::NTRK3 fusion using FISH in this case remain unclear and may lay in the area of technical imperfection.

Conclusion

The described case of extrarenal extracranial rhabdoid tumor in a newborn has a unique tumor genetic landscape combining characteristic inactivating deletion of SMARCB1 to a gain-offunction NTRK3 rearrangement. The emergence of NTRK and SMARC oncogenic variants within one tumor is highly improbable. This is the first reported case of extrarenal extracranial rhabdoid tumor harboring this combination.