Epstein-Barr virus-associated natural killer-large granular lymphocyte leukemia

Case Studies EPSTEIN-BARR

VIRUS-ASSOCIATED

NATURAL

KILLER-LARGE

GRANULAR

LYMPHOCYTE

LEUKEMIA

B. GELB, MD, MATTHIJSVAN DE RIJN, MD, PHD, DONALD P. REGULA, JR, MD, JOANNE P. CORNBIXET, MD, ONSI W. KAMEI., MD, DKRAN S. HOROUPIAN, MD, MICHAEI. L. CLEARY, MD, AND ROGER A. WARNKE,MD

ARNOLD

We describe thejirst case of an Epstein-Barr virus (EBV)-associated natural killer-large granular lymphoqte (NK-LGL) leukemia in the United States to the best of our knowledge. A 29year-old woman of Japanese descent developed EBV infection after a blood transfusion as indicated by a rise in serum antibody titers. Peripheral blood and bone marrow aspirate smears demonstrated increased LGLs. Flow qtometly showed that these cells expressed NK-associated sulfate antigens. Cytogenetic analysis of the bone marrow aspirate showed two distinct but related clones with multiple copies of a modified 7 marker chromosome. Death followed colonicperfortion. Findings at necropsy included bone marrow lymphoqtosis and mthrqbhagocytosis, a mononucleosis-lib lymphadenitis, atypical hepatitis with a mixed, jn-edominantly T-cell infiltrate, interstitial pneumonitis, and multiorgan system vasculitis with perforation of the transverse colon. Epsta’n-Baw virus transcripts were identijied in lymphocytes infiltrating liver and peripheral nerve by in situ hybridization. In addition, Southern blot anatyses showed monoclonal bands superimposed on oligoclonal ladders ofEBV &mini in liver and lymph node. The identical episomal form of EBV was found in the bone mawow, lymph node, and liver. No immunoglob ulin (I& T-cell receptor beta, or T-cell receptor gamma chain gene rearrangements were identified. These studies support the hypothesis that the LGL population was a neqblastic EBV-related clonal proliferation of NK cells. HUM PATHOL25:953-960. Co#yright 0 I994 4 W.B. Saunders Company

Large granular lymphocytes (LGLs) are defined morphologically by virtue of their size, abundant cytoplasm, and azurophilic granules on Wright-Giemsa-stained smears.‘,’ A variety of functional and immunologic characteristics of LGLs have been reported in the literature. A significant subset of LGLs demonstrate natural killer (NK) cell functional activity characterized by the ability to mediate nonmajor histocompatibility complex-restricted lysis of tumor cell targets, such as the K562 cell line and antibody-dependent cellular cytotoxicity (ADCC) .‘..1-5Activity of ADCC is mediated by receptors for the Fc fragment of immunoglobulin (Ig) G (CD16), which the majority of LGLs express. However, LGLs are heterogeneous with respect to expression of other cell surface antigens.”

From the Department of Pathology, Stanford University, Stanford, CA. Accepted for publication March 28, 1994. Dr Gelb’s present address is the Department of Laboratory Medicine, University of California, San Francisco, L518, San Francisco, CA 941430134. Supported in part by grants no. CA 34233 and 33119 from the National Institutes of Health, Bethesda, MD. KS, w&.x natural killer cell, leukemia, lymphoproliferative disorder, Southern blot, immunohistochemistry, in situ hybridization, large granular lymphocyte. Address correspondence and reprint requests to Roger A. Warnke, MD, Department of Pathology, Room L235, Stanford University Hospital, 300 Pasteur Dr, Stanford, CA 943055324. Copyright 0 1994 by W.B. Saunders Company 0046-8177/94/2509-0018$5.00/0

953

Using flow cytometry two major immunophenotypic categories of LGLs and lymphoproliferative disorders thereof have emerged as defined by expression of CD3.7-“’ Most LGLs are T-cell type (CD3+), often demonstrating a cytotoxic/suppressor immunophenotype (CD2+, CD3+, and CDS+) with variable expression of NK cell-associated markers, such as CDll, CD16, CD56, or CD57. ‘z’~ Less frequently the immunophenotype is indicative of “true” NK cells in that the LGLs fail to express CD3 or CD5 but are CD2+, CD16’, CD56+, or CD57’. CD3positive LGLs usually exhibit T-cell receptor beta chain gene rearrangements,” whereas CD3- LGLs usually lack such rearrangements. A caveat is that several cases have been reported with unusual phenotypes but strong NK cell activity.” Conceptually, proliferative disorders of LGLs may be thought of as part of a spectrum ranging from reactive to neoplastic, although the distinction may not be clear-cut.‘” Examples of conditions associated with reactive LGL proliferations include chronic infections such as tuberculosis or hepatitis, specific viral infections such as Epstein-Barr virus (EBV), human T-cell leukemia virus-1 or human immunodeficiency virus (HIV) I, and autoimmune disorders such as rheumatoid arthritis, nephrotic syndrome, and neoplasms.‘c” So-called “unclassified” LGLs, including Ty lymphoprc+ liferative disorder, lymphoproliferative disease of granular lymphocytes, or T lymphocytosis with neutropenia, are not clearly reactive or neoplastic. ” Rheumatoid arthritis has been reported in association with some cases of CD3’ LGLs. Despite a proliferation of mature LGLs predominantly of cytotoxic/ suppressor phenotype and infrequent clinical progression with time, a clonal T-cell receptor (TCR) p or y gene rearrangement is commonly identified. At the malignant end of the spectrum there are LGL lymphomas and leukemias. Although both entities may infiltrate solid tissues, leukemias by definition have peripheral blood cell counts of at least 2 X 10”/L.2”2”The relationship between LGL lymphoma and leukemia may be analogous to small lymphocytic lymphoma/chronic lymphocytic leukemia or lymphoblastic lymphoma/acute lymphoblastic leukemia. We are aware of only two previous reports of EBV-associated NK cell leukemia.‘,‘” In this report we describe the first case to occur in the United States to the best of our knowledge. The case is unique in that a chronic active EBV infection evolved into an NK-LGL leukemia containing clonal EBV genome and resulted in a systemic vasculitis with colonic perforation. CASE REPORT A 29-year-old woman of Japanese descent was well until her second pregnancy when she was noted to have proteinuria. She received a 1-U transfusion of packed red blood cells after delivery because of bleeding from the episiotomy site. Before discharge she developed low grade fevers and recurrent nephrotic range proteinuria. She subsequently was noted to have abnormal liver function tests but normal blood urea

HUMAN

TABLE 1.

CD10 CD19 CD22 Kappa/CD19 lambda/CD19 T lineage CD2 CD3 CD4

CD5 CD7 CD8 CD16 CD56 CD57 CDS/CD56 Myeloid lineage CD11 C CD14 CD33

X6 95

CA.lA,