Cytotoxic T-Lymphocyte Differentiation and Cytogenetic Alterations in γδ Hepatosplenic T-Cell Lymphoma and Posttransplant Lymphoproliferative Disorders

To the Editor:

We read with great interest the recent reports on hepatosplenic γδ T-cell lymphoma (γδ HSTCL)1 and T-cell posttransplant lymphoproliferative disorders (T-PTLD).2 Since 1991 we have encountered 5 patients with γδ HSTCL: 2 occurred in chronic renal transplant patients (Salhany et al, unpublished data, November 1996), but the other 3 patients were not immunosuppressed.3 Our cases were clonal CD4⁻, CD8⁻ T-cell lymphomas with hepatosplenic involvement, which expressed γδ T-cell receptors (TCR) and natural killer (NK) cell-associated antigens (4/5 CD16⁺ and CD56⁺, 1/5 CD11c⁺). Most presented with fever, anemia, and thrombocytopenia, but 2 also presented with severe neutropenia.3 Three of 5 patients died within 5 to 10 months of progressive lymphoma with only minimal response to CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) chemotherapy; 2 terminally developed overt leukemia. However, one nontransplant patient had a complete response to CHOP, but died of aplastic anemia 22 months later with no evidence of lymphoma at autopsy.3 

We were gratified to see that Cooke et al1 have confirmed our previous observation that γδ HSTCL are derived from cytotoxic T lymphocytes (CTL).4 We have subsequently demonstrated CTL differentiation in 4 additional γδ HSTCL3 (and unpublished data, November 1996). All cases expressed cytolytic effector and cytolytic granule-associated proteins, including Fas ligand (5/5), TIA-1 (5/5), perforin (3/5), and granzyme B (3/5). Moreover, we have recently demonstrated TCR-mediated cytolytic activity and antibody-dependent cellular cytotoxicity (ADCC) in a CD16⁺, CD56⁺ γδ HSTCL.3 Cooke et al1 suggested that γδ HSTCL represent NK-like T cells based on frequent coexpression of the NK cell-associated antigens CD16 and CD56; however, we did not find NK-like, non–major histocompatibility complex restricted, spontaneous cytolysis of K562 cells in one CD16⁺, CD56⁺ γδ HSTCL studied thus far (Salhany and Peritt, unpublished data, November 1996). Interestingly, however, we did demonstrate interferon-γ (IFN-γ) production by one γδ HSTCL in a patient presenting with severe neutropenia.3 IFN-γ–induced suppression of granulocytopoiesis has been implicated in the pathogenesis of severe neutropenia in a patient with subcutaneous γδ T-cell lymphoma,5 suggesting a similar mechanism in our case.

Isochromosome 7q has been espoused by Cooke et al and others as a cytogenetic abnormality that may define γδ HSTCL as a distinct entity.1,6 However, we found i(7q) in only 1 of 4 γδ HSTCL studied by standard cytogenetics (Salhany and Nowell, unpublished data, November 1996).3 Our cytogenetic studies and those of others7,8 suggest that i(7q) is not a consistent cytogenetic abnormality in γδ HSTCL. Moreover, i(7q) is not specific for γδ HSTCL; i(7q) has also been reported in acute myeloid leukemia, acute lymphoblastic leukemia, prolymphocytic leukemia, and Wilms' tumor.3,9 Interestingly, trisomy 8 has accompanied i(7q) in 5 of 7 reported cases, including our case and one from Cooke et al.1,3 6 

We agree that γδ HSTCL is a distinct clinicopathologic entity, but have noted more morphologic, immunophenotypic, cytogenetic, and clinical heterogeneity than observed by Cooke et al.1 Moreover, we have shown that γδ HSTCL can be derived from different γδ T-cell subsets; 4 of 5 were derived from Vδ1 γδ T-cells, whereas the other was derived from neither Vδ1 nor Vδ2 subsets3 (and unpublished data). Gaulard et al have reported similar findings.10 Interestingly, our Vδ1⁻ γδ HSTCL was different from the Vδ1⁺ cases, including presence of a small cell component, negative CD16 and CD56, lack of i(7q), severe neutropenia, massive hemophagocytosis, IFN-γ secretion without cytolytic activity, and complete response to CHOP chemotherapy.3 

Our 2 posttransplant γδ HSTCL cases share similarities to the 6 T-PTLD cases described by Hanson et al,2 except for CD8 expression by their cases. CD8 is more commonly expressed by αβ T cells, but CD8 can be expressed by some γδ HSTCL,1 suggesting possible γδ T-cell origin for Hanson's T-PTLD cases.2 Unfortunately, they did not evaluate their T-PTLD for αβ or γδ TCR expression,2 but retrospective study with βF1 might allow direct or indirect determination of the TCR subtypes. Regardless of αβ or γδ TCR subtype, expression of CD16 or CD56 and ultrastructural demonstration of type I cytolytic granules11 support CTL origin of their T-PTLD²; this could be confirmed by immunohistochemical staining for TIA-1, perforin or granzyme B.1,3 Interestingly, both of our γδ T-PTLD and all 6 of their T-PTLD2 developed in chronic renal transplant patients, suggesting that chronic CTL proliferation may be important in the pathogenesis of T-PTLD. Neither of our γδ T-PTLD patients responded to reduction in immunosuppression, but both had partial responses to chemotherapy; however, the response in one patient was short-lived, terminating in overt leukemia and death within 6 months. Follow-up in the other patient is only 1 month. Hanson et al2 reported similar responses. Thus, it appears that, unlike B-PTLD, T-PTLD does not respond to reduction in immunosuppression alone, and should be treated with chemotherapy at diagnosis.