Hepatosplenic T-cell lymphoma (HSTCL) is a highly aggressive T-cell neoplasm that arises from the proliferation of gamma/delta T cells (yöT) infiltrating the liver, spleen, and bone marrow sinusoids.1 It commonly emerges in individuals with chronic immune suppression, predominantly affecting young adults who manifest hepatosplenomegaly, cytopenias, and systemic symptoms.2 In contrast, large granular lymphocytic leukemia (LGLL) typically manifests as an indolent proliferation of cytotoxic alpha/beta T cells (a|3T), primarily afflicting older adults with neutropenia and concurrent autoimmune disorders.3 Nevertheless, splenomegaly is frequently observed in LGLL, and leukemic cells can carry a yöT-cell receptor (TCR), complicating the differentiation from HSTCL.4,5 This study, approved by our institution’s ethics committee (number 23.34), aims to establish novel diagnostic criteria for distinguishing HSTCL from yöT-LGLL. Herein, we report a distinctive oyster-shell morphology and identify stereotyped VD1-JD1 CDR3 sequences in HSTCL cells. The morphological description of HSTCL has focused primarily on histological aspects of splenic or hepatic biopsies, with minimal emphasis on cytological features.6 We centrally reviewed 23 bone marrow aspirate smears, ten blood smears and two biopsy touch preparations (spleen and liver). Bone marrow was involved in all HSTCL cases and blood samples were infiltrated in 42% of cases. The tumor burden in bone marrow ranged from 5% to 73% and varied from sparse neoplastic cells to a diffuse involvement with large pseudometastatic aggregates. Hemophagocytosis was observed in six patients (25%). Atypical lymphoid cells were monotonous, medium-sized, with irregular nuclei displaying fine chromatin and prominent nucleoli. Their cytoplasm had a jagged outline, peripheral basophilic enhancement, and frequent cytoplasmic projections, giving the tumoral cells a distinctive oyster-shell pattern (Figure 1A-C). These cells were predominant within the neoplastic infiltrate and were observed in all the blood and bone marrow samples from HSTCL cases. In touch preparations, tumoral cells often presented with rounder blastic nuclei and reduced cytoplasm, consistent with previous descriptions (Figure 1D).6 Conversely, LGLL cells had a round nucleus with dense chromatin and pale cytoplasm with azurophilic granules (Figure 1E). Thus, the distinctive oyster-shell morphology observed in both yöT and a|3T HSTCL subtypes may provide a new criterion for HSTCL diagnostics (Figure 1F, G). This characteristic could guide additional investigations early in this challenging diagnosis and potentially avoid more invasive procedures. The phenotypic and oncogenic characteristics of the two groups align (Table 1) with previously published data (Online Supplementary Figure S1).2,4 All HSTCL cases were positive for the pan-T-cell markers CD3 and CD2. CD7 was positive in 14 cases (93% of tested cases). Most HSTCL cases were double negative for both CD4 and CD8 (60%). As could be expected, 18 cases (90%) exhibited a yö-TCR while two cases expressed an a|3-TCR (10%). Several markers might be useful in distinguishing HSTCL from LGLL. A complete lack of CD5 was observed in 17 HSTCL samples (15% of positive cases), whereas LGLL displayed at least weak positivity for CD5 in 81% of cases (P<0.001). CD56 was positive in 13 HSTCL samples (87%), while only two LGLL cases were positive (13%) (P<0.001). Conversely, 14 cases of LGLL expressed CD57 (93%), while CD57 expression was detected in three of the five HSTCL cases tested. Lastly, in four bone marrow biopsies, HSTCL were positive for TIA-1 and negative for granzyme B staining, while this marker is typically expressed in cases of LGLL.3 Cytogenetic analysis was available for 14 patients with HSTCL. Isochromosome 7q was detected in ten cases (71%) and trisomy 8 in seven cases (50%). Four patients (29%) had a complex karyotype. Subsequently, we conducted a comparative mutational analysis of 21 HSTCL samples and 16 yöT-LGLL using targeted high-throughput sequencing (Online Supplementary Table S1). In the HSTCL group, 17 patients (81%) had at least one somatic mutation. Overall, we identified 30 clinically relevant variants in 13 different genes. The most frequent event was a mutation in the SH2 domain of STAT5B, detected in eight patients (38%), with five of them carrying the N642H hotspot. DNMT3A was mutated in four cases (19%), TET2 in three patients (14%), and EZH2, TP53 and SETD2 in two patients (10%). Among yöT-LGLL, 14 (88%) had at least one significant mutation. The STAT3 gene was mutated in 12 patients (75%). A DNMT3A mutation was found in three cases (19%), and a TET2 mutation in two cases (13%). Finally, TNFAIP3, STAT5B, and IDH2 mutations were detected in one patient. Consistent with previous studies, the presence of isochromosome 7q remains the most frequent cytogenetic aberration in HSTCL, and it is retained as a diagnostic criterion in the latest revision of the World Health Organization classification.1 ,7 The detection of STAT5B mutations can provide an additional molecular marker for HSTCL, in contrast to STAT 3 alterations which constitute the molecular hallmark of LGLL.8 Mutations in genes related to DNA methylation (TET2 and DNMT3A) were identified in both HSTCL and LGLL samples and have been commonly reported in the context of clonal hematopoiesis, making them unsuitable as specific molecular markers for HSTCL. Conversely, SETD2 appeared to be specifically associated with HSTCL, being present with a frequency of 10% in the cases in our cohort and in up to 25% in previous studies.9,10 Lastly, exome sequencing provided a more comprehensive view of the genomic landscape of HSTCL, revealing alterations in other genes involved in chromatin modification such as INO8, SMARCA2, and ARID1B, as well as in other signaling pathways (e.g., PIK3CD or KRAS), offering potential therapeutic targets for this rare disease.10,11 The originality of our study stems from the analysis of TCR rearrangement specificity in a sizable cohort of HSTCL patients using high-throughput sequencing. Recently, Teramo et al. investigated the TCR repertoire profile in LGLL, revealing stereotyped TRG rearrangements associated with clinical features.12 Our study provides complementary results on the immunogenetic profile of HSTCL, highlighting biased TRG and TRD gene usages that distinguish them from LGLL (Figure 2).
Figure 1.Comparative cytological analysis of hepatosplenic T-cell lymphoma and T-cell large granular lymphocytic leukemia. May-Grünwald-Giemsa staining, magnification x600. (A) Bone marrow aspirate smears from five representative cases of yb hepatosplenic T-cell lymphoma (HSTCL). Blue arrows indicate images of hemophagocytosis. (B) Schematic illustration depicting the oyster shell pattern observed in HSTCL cases. (C) Tumor cells on blood smears from two yö-HSTCL cases. (D) Liver biopsy touch preparation from one yö-HSTCL case. The black arrow points to a Küpffer cell. (E) Large granular lymphocytes (LGL) on blood smears (CellaVision images) from one yöT-LGL leukemia case. (F) Bone marrow aspirate smears from two a(3-HSTCL cases. (G) Blood smear (CellaVision images) from one case of a(3T-LGL leukemia.
Table 1.Characteristics of patients with hepatosplenic T-cell lymphoma and γδT-cell large granular lymphocytic leukemia.
Figure 2.Comparison of TRG and TRD rearrangement profiles in dominant clonotypes from hepatosplenic T-cell lymphoma, γδT-cell large granular lymphocytic leukemia and a polyclonal control group. (A) Heatmap showing frequency of segment usage for TRG and TRD loci in the hepatosplenic T-cell lymphoma (HSTCL) group, compared to the γδT-large granular lymphocytic leukemia (LGLL) group and the polyclonal control group. *TRGV4 is significantly overrepresented in HSTCL clonotypes versus LGLL and versus control clonotypes; TRDV1 is overrepresented in HSTCL and LGLL clonotypes versus control clonotypes; TRDV2 is overrepresented in LGLL and control clonotypes versus HSTCL clonotypes (P<0.05). (B) Sunburst representation of the utilization of TRG segments in the dominant clonotypes identified in the HSTCL, LGLL and control groups. (C) Sunburst representation of the utilization of TRD segments in the dominant clonotypes identified in the HSTCL, LGLL and control groups. (D) Comparison of the CDR3 amino-acid sequences of VD1-JD1 clonotypes in the HSTCL, LGLL and control groups. TRG: T-cell receptor gamma; TRD: T-cell receptor delta.
TRG and TRD gene rearrangements were sequenced according to the two-step polymerase chain reaction Euroclonality next-generation sequencing protocols on a MiSeq platform (Illumina) and analyzed with the Vidjil tool.13 We compared the dominant clonotypes from 20 HSTCL and 16 yöT-LGLL cases with an in silico polyclonal control composed of the top 100 clonotypes from five healthy donors (Figure 2A). Regarding the TRG locus (Figure 2B), 31 clonotypes were detected in 18 HSTCL cases (90%) with a predominance of rearrangements involving the TRGV4 gene in nine cases (50%). The TRGV4 gene was significantly more prevalent in HSTCL clonotypes (29%) than in yöT-LGLL (7%; P=0.043) and the control group (9%; P=0.002). In yöT-LGLL and the control group, the most represented gene was TRGV9, occurring in eight LGLL clonotypes (27%) and in 37% of control clonotypes, which is significantly higher than in the HSTCL group (10%) (P=0.002). Regarding the TRD locus (Figure 2C), we identified 20 clonotypes in 16 HSTCL samples and 22 clonotypes in 15 yöT-LGLL samples. We observed a higher number of clonotypes using the TRDV1 gene in HSTCL and yöT-LGLL cases than in the control group (P<0.001 for both). Additionally, 13 HSTCL cases (81% of yöT-positive cases) showed a VD1-JD1 rearrangement, while this occurred in eight yöT-LGLL cases (50%). To a lesser extent, HSCTL clonotypes used the TRDV3 gene more frequently than yöT-LGLL did (P=0.050). In contrast, TRDV2 was more frequently represented in the control group (80% of clonotypes) and in yöT-LGLL patients (47%) than in the HSTCL group (0%) (P<0.001 and P=0.004, respectively). Finally, only two HSTCL patients had TRB rearrangements, corresponding to the two samples expressing αβTCR identified by flow cytometry. Overall, we demonstrate a higher prevalence of TRGV4 and TRDV1 segment utilization in HSTCL patients. This observation aligns with previous phenotyping investigations that showed a predominant expression of V51 chain in HSTCL and in normal splenic yb T cells.14,15 Moreover, tumor infiltrating V51 T cells often exhibit immunosuppressive effects, in contrast to classical Vy9Vö2 T cells, which mainly exist in peripheral blood and have strong anti-tumor effects in various types of tumors.15 In line with the findings of Teramo et al., LGLL exhibited a higher frequency of TRGV9 and TRDV2 genes in their TRG and TRD rearrangements, respectively. Consistently, the preferential expression of the Vy9Vö2 phenotype in yöT-LGLL appears to mimic the spectrum of normal T cells in the peripheral blood of healthy subjects and is associated with less symptomatic presentation in LGLL.12 In HSTCL clonotypes, we observed similar peptide amino-acid sequences within the CDR3 region of the ö chain, particularly involving VD1-JD1 rearrangements (Figure 2D and Online Supplementary Table S2). These CDR3 sequences were shorter and showed less diversity than those observed in VD1-JD1 clonotypes from subjects with γδT-LGLL and healthy donors, suggesting a role for antigenic recognition as an early event in HSTCL lymphomagenesis.
In conclusion, HSTCL may originate from a clonally selected γδT population with stereotyped TCR, through the acquisition of somatic mutations leading to JAK/STAT pathway deregulation and chromatin modifications. We provide a more comprehensive characterization of HSTCL patients compared to those with γδT-LGLL, highlighting a specific oyster-shell morphology and restricted TRG and TRD segment usages in HSTCL. Together, these findings may serve as a valuable tool for distinguishing HSTCL from other γδT proliferations and could potentially reduce the need for more invasive procedures such as splenectomy or liver biopsy.
Footnotes
- Received July 31, 2023
- Accepted January 15, 2024
Correspondence
Disclosures
No conflicts of interest to disclose.
Contributions
AD, SB, and CP designed the research and wrote the paper. FT performed T-cell receptor sequencing bioinformatics analysis. MP described the cytological pattern. AD, SB, JG, and MP reviewed cytological data. FL-G reviewed histological data. FG, LC, LB, PL, CLL, PC-L, A-CG, CB, ER, ED, CF, FB-J, SW, and VB performed the initial investigations and addressed the cases for constitution of the cohort. TL managed the patients with large granular lymphocytic leukemia. MR provided flow cytometric data. AD, SB, TF, and CP analyzed the molecular data. All the authors critically reviewed the paper.
References
- Alaggio R, Amador C, Anagnostopoulos I. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: lymphoid neoplasms. Leukemia. 2022; 36(7):1720-1748. Google Scholar
- Weidmann E. Hepatosplenic T cell lymphoma. A review on 45 cases since the first report describing the disease as a distinct lymphoma entity in 1990. Leukemia. 2000; 14(6):991-997. Google Scholar
- Lamy T, Moignet A, Loughran TP. LGL leukemia: from pathogenesis to treatment. Blood. 2017; 129(9):1082-1094. Google Scholar
- Barilà G, Grassi A, Cheon H. Tγδ LGLL identifies a subset with more symptomatic disease: analysis of an international cohort of 137 patients. Blood. 2023; 141(9):1036-1046. Google Scholar
- Cooke CB, Krenacs L, Stetler-Stevenson M. Hepatosplenic T-cell lymphoma: a distinct clinicopathologic entity of cytotoxic gamma delta T-cell origin. Blood. 1996; 88(11):4265-4274. Google Scholar
- Vega F, Medeiros LJ, Gaulard P. Hepatosplenic and other gammadelta T-cell lymphomas. Am J Clin Pathol. 2007; 127(6):869-880. Google Scholar
- Wang CC, Tien HF, Lin MT. Consistent presence of isochromosome 7q in hepatosplenic T gamma/delta lymphoma: a new cytogenetic-clinicopathologic entity. Genes Chromosomes Cancer. 1995; 12(3):161-164. Google Scholar
- Nicolae A, Xi L, Pittaluga S. Frequent STAT5B mutations in γδ hepatosplenic T-cell lymphomas. Leukemia. 2014; 28(11):2244-2248. Google Scholar
- McKinney MS, Moffitt AB, Rempel RE. SETD2 functional loss through mutation or genetic deletion promotes expansion of normal and malignant γδ T cells through loss of tumor suppressor function and upregulation of oncogenic pathways. Blood. 2016; 128(22):1052. Google Scholar
- McKinney M, Moffitt AB, Gaulard P. The genetic basis of hepatosplenic T-cell lymphoma. Cancer Discov. 2017; 7(4):369-379. Google Scholar
- Travert M, Huang Y, de Leval L. Molecular features of hepatosplenic T-cell lymphoma unravels potential novel therapeutic targets. Blood. 2012; 119(24):5795-5806. Google Scholar
- Teramo A, Binatti A, Ciabatti E. Defining TCRγδ lymphoproliferative disorders by combined immunophenotypic and molecular evaluation. Nat Commun. 2022; 13(1):3298. Google Scholar
- Brüggemann M, Kotrová M, Knecht H. Standardized next-generation sequencing of immunoglobulin and T-cell receptor gene recombinations for MRD marker identification in acute lymphoblastic leukaemia; a EuroClonality-NGS validation study. Leukemia. 2019; 33(9):2241-2253. Google Scholar
- Przybylski GK, Wu H, Macon WR. Hepatosplenic and subcutaneous panniculitis-like gamma/delta T cell lymphomas are derived from different Vdelta subsets of gamma/delta T lymphocytes. J Mol Diagn. 2000; 2(1):11-19. Google Scholar
- Li Y, Li G, Zhang J. The dual roles of human γδ T cells: anti-tumor or tumor-promoting. Front Immunol. 2020; 11:619954. Google Scholar
Data Supplements
Figures & Tables
Article Information
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.