Patient selection and clinical finding
Registered patients were retrieved retrospectively over the period from 1990 to 2020 from the Department of Pathology, Fukuoka University. Histological classifications were performed based on the WHO classification (2017) and detailed descriptions of EBV+ TNKLPDs [2, 10, 13]. Criteria for sEBV+ TCL of childhood were as follows: progressive clinical course with features of HLH, overt infiltrates of CD8+ and CD56− cytotoxic T-lymphocytes with small or occasionally medium to large-sized or large nuclei, and overt lymphoma or rearrangements of T-cell receptor (TCR) genes. EBV+ nodal cytotoxic TCL has a primarily nodal presentation with limited extranodal disease [11, 12]. sCAEBV presents mainly with IM-like symptoms persisting for 3 months, elevated peripheral blood EBV DNA (³ 102.5 copies/µg) and the presence of EBV+ T-lymphocytes with mild nuclear atypia [9, 14]. In this study, EBV+ HLH was excluded because of small amounts of Epstein-Barr virus-encoded RNAs (EBERs)-positive cells [15], and we excluded CD8+ and CD56+ EBV+ nasal type TNKCL [2, 3].
Histology, immunohistology, and detection of EBV-encoded RNA
Excised tissue specimens were fixed in 10% formalin to generate formalin-fixed paraffin-embedded samples, and were stained with hematoxylin and eosin. Bone marrow samples and related smears were examined in all 16 patients. Samples from involved lymph nodes, liver, and other organs were also examined. Autopsy examination was permitted in three older sEBV+ TCL patients. For immunohistochemistry, antibodies were applied to formalin-fixed tumor samples using a Leica Bond III-automated stainer (Leica Biosystems, Buffalo Grove, IL), and peroxidase reaction was developed using diaminobenzidine. The following antibodies were used: CD3 (PS1, Leica, Newcastle, UK), CD4 (4B12, Leica), CD8 (C81/44B, Leica), TIA1 (2GP, Beckman Coulter, Marseille, France), granzyme B (11F1, Leica), CD56 (1B6, Leica), TCRβF1 (8A3, Endogen, Rockford, IL), phosphate signal transducer and activator of transcription 3 (pSTAT3) (RB5791, Argent, San Diego, CA), CD30 (BerH2, Dako, Glostrup, Denmark), programmed cell death 1 (PD1) (NAT105, Abcam, Cambridge, MA), PD-ligand 1 (PD-L1) (E1L3N, Cell Signaling, Danvers, MA), CMYC (Y69, Abcam), p53 (DO7, Leica), Chemokine receptor (CXCR) 3 (1C6, Bioscience, San Diego, CA), CC chemokine receptor (CCR) 4 (1G1, Bioscience), CD20 (L26, Nichirei, Tokyo), latent membrane protein (LMP) 1 (CS1-4, Dako), and EBV nuclear antigen (EBNA) 2 (PE2, Leica). For all immunostains, reactions were considered positive when over 30% of tumor cells stained positively. Among them, ³50% of tumor cells stained positively for CMYC, p53, and PD-L1 were considered positive. Staining intensity of PD-L1 for non-neoplastic cells including histiocytes and dendritic cells, were scored as: R0 (no positive cells/HPF), R1+ (a few to 5% cells), R2+ (³ 5% to 20%) and R3+ (³ 20%) [16]. The presence of EBV infection was determined by in situ hybridization of EBERs+ nuclear signals in over 50% of atypical lymphoid cells. For EBERs stains, deparaffinized tissue sections were hybridized in a solution of 50% formamide containing fluorescein isothiocyanate-labeled EBERs oligonucleotides (EBERs probe, Leica). Double staining of lymphocyte markers and EBERs was performed to confirm the presence of EBV+ CD8+ T-cells.
Molecular analysis of the TCRγ gene locus
For polymerase chain reaction (PCR) amplification of the TCRγ locus, DNA was prepared from 11 available patient samples by standard proteinase K digestion and phenol-chloroform extraction. BIOMED-2 multiplex PCR assays (InVivoScribe Technologies, San Diego, CA) of TCRγ (Bottles A, B; V-J and D-J gene rearrangements) were performed using standardized protocols and primers [17]. Following amplification, TCR-PCR products were analyzed using GeneScan (MultiNA, Shimazu Co., Kyoto, Japan).