Lymphoproliferative disease of granular T lymphocytes presenting as aplastic anemia

Lymphoproliferative disease of granular T lymphocytes (T-LDGL), also referred to as T-cell large granular lymphocyte leukemia, is a disorder of cytotoxic T lymphocytes, characterized by the presence of clonal proliferation of granular lymphocytes in the peripheral blood or bone marrow and clinically associated with cytopenias, lymphocytosis, frequent infections, and autoimmunity.1 The natural history of the disease is usually prolonged and the prognosis relatively favorable. In many patients, the disease is mild and requires little or no therapy.2 The cytopenia is usually manifested as neutropenia or anemia and much less frequently as thrombocytopenia. Although not well described, pancytopenia can be the presenting finding at diagnosis. Of the patients in our series of 203 T-LDGL patients, 14% had pancytopenia at presentation.3 A subset of these patients fit the diagnostic criteria for aplastic anemia. Here, we describe 9 cases of T-LDGL presenting as aplastic anemia.

After approval was obtained from the institutional review board, our database of T-LDGL patients was reviewed for presentation of aplastic anemia at diagnosis. Criteria for T-LDGL include unexplained abnormal blood counts (cytopenias, macrocytosis, and/or lymphocytosis) and expression of a clonal T-cell–receptor gene rearrangement. Detection of an increase of granular lymphocytes in the peripheral blood or bone marrow and supportive immunophenotypic studies by flow cytometry or histochemistry were considered relevant but not essential. The diagnosis of aplastic anemia was made according to the criteria published by the International Agranulocytosis and Aplastic Anemia Study group.4 We prepared a Wright-Giemsa–stained blood smear for each patient and counted the absolute number and percentage of granular lymphocytes as previously reported.2Multiparametric flow cytometry of the peripheral blood or bone marrow was used to study lymphocytic immunophenotype.

T-cell–receptor gene rearrangement study was performed by means of either Southern analysis or polymerase chain reaction. Complete response (CR) was defined as normalization of blood counts, including granular lymphocytes. Partial response (PR) was defined as more than 50% improvement in the deviation of blood counts from normal, but less than a CR, and by symptomatic improvement or, in the case of transfusion dependency, a 50% reduction in transfusion requirements. Any response less than partial was considered no response (NR).

Nine patients with T-LDGL presented with aplastic anemia. Of these, 6 patients were males and the median age was 61 years (range, 39-72). Pertinent associated medical conditions were rheumatoid arthritis in one patient and concurrent inflammatory bowel disease and giant cell arteritis in another. Presenting symptoms included fatigue, easy bruising, mucosal bleeding, and fever. Lymphadenopathy, hepatomegaly, or splenomegaly were not observed. Laboratory findings are summarized in Tables 1 and2. Three patients had severe aplastic anemia. None of the patients had absolute lymphocytosis. The absolute granular lymphocyte count, percentage of granular lymphocytes in the peripheral blood, and bone marrow granular lymphocytes were increased in 7, 8, and 6 cases, respectively. One patient did not have morphologic evidence of increased granular lymphocytes in the peripheral blood or bone marrow. Immunohistochemical staining showed interstitial marrow involvement by CD3⁺, CD8⁺, and CD57⁺ cells in most cases. Rheumatoid factor and antiplatelet antibody were tested in only 1 patient each, and were both negative. Antineutrophil antibody was not tested in any patient.

Two patients died before specific treatment was instituted (1 and 12 weeks after diagnosis). Initial treatment consisted of prednisone in 5 patients, resulting in 1 unmaintained CR, 1 steroid-dependent PR, 1 transient PR, and 2 NR. The 2 remaining patients received a combination of oral cyclophosphamide and prednisone. Both did not have a response. After initial prednisone failure, 2 patients received cyclophosphamide as second-line therapy without benefit. One of them subsequently received antithymocyte globulin with cyclosporine and achieved a transient PR. This patient, however, died of complications from allogeneic bone marrow transplantation. The other patient received cyclosporine and then azathioprine but did not have a response. Prednisone was tried again, and a PR was achieved for 2.75 years. Other second-line therapies included methotrexate (1 of 1 unmaintained PR) and antithymocyte globulin with cyclosporine (1 of 1 unmaintained CR). Time to treatment response varied from 1 to 3 months. For those patients who achieved CR, treatment was generally continued for 6 months to a year. After a median follow-up of 49 months (range, 15-106 or higher), 5 patients had died from the disease or treatment-related complications. Two patients died of sepsis, 1 of intracranial bleed, and 1 of hepatic veno-occlusive disease after allogeneic bone marrow transplantation. One other patient died of progressive disease, but the terminal event was unclear. The treatment results are summarized in the Figure 1. The median survival for the group was 40 months.

Most patients with idiopathic aplastic anemia are now known to have an immune-mediated destruction of hematopoietic cells.5Cytotoxic T cells play an important role in this regard by release of cytokines, which leads to inhibition of hematopoiesis and expression of Fas receptors on progenitor cells. The interaction of Fas receptor with the Fas ligand present in cytotoxic T cells triggers apoptosis. While the exact pathophysiology of T-LDGL is not clear, it may be similar to aplastic anemia.1 

Compared with patients with idiopathic aplastic anemia, patients with T-LDGL–associated aplastic anemia have similar demographic characteristics, presenting symptoms, and physical findings. Evidence of autoimmunity was observed in about half of all the patients. All but one patient had evidence of granular lymphocytosis, either relative or absolute, in the peripheral blood or bone marrow. We did not use granular lymphocyte count as an absolute criteria for diagnosis because some patients may have low granular lymphocyte count. This is consistent with a recent report updating the criteria for LDGL.6 

Because of limited antibody panels studied with the use of flow cytometry, we were not able to further characterize the immunophenotype of the granular lymphocytes. The presence of a clonal T-cell receptor gene rearrangement in all cases, however, pointed to a T-cell origin. The diagnosis of T-LDGL can easily be missed if peripheral blood smear and bone marrow are not carefully reviewed for granular lymphocytes, because absolute lymphocytosis is usually not observed in an automated blood count report.

With the exception of increased interstitial infiltration by CD3⁺/CD8⁺/CD57⁺ cells, the bone marrow findings were not different from acquired aplastic anemia. In our experience, bone marrow cellularity in T-LDGL was normal in 41%, increased in 39%, and decreased in 20% of the cases.3 It is likely that when T-cell–mediated suppression of hematopoietic cells is the dominant pathology, bone marrow will be hypocellular. When B-cell dysfunction predominates, then a more cellular marrow may be observed. We did not find any characteristic abnormality in the analysis of bone marrow cytogenetics.

Three patients had loss of Y chromosome, and another patient had translocation (1;7). Loss of Y chromosome is considered a normal age-related phenomenon.7 Translocation (1;7) has been associated with myeloproliferative and myelodysplastic disorders.8 The patient with translocation (1;7) did not have any evidence of these disorders at diagnosis and at subsequent follow-up. The use of cyclophosphamide, which is one of the preferred treatments of T-LDGL, was not associated with a clinical response, while noncytotoxic immunosuppressive agents being used for aplastic anemia produced a response in most cases. Further follow-up is necessary to assess the long-term outcome of these patients.

We thank Janice Hodnefield, Barb Todd, and Bill Wittrock for their laboratory assistance.