Late Onset Neutropenia Develops Selectively in Only a Subset of Patients with T Large Granular Lymphocyte Proliferation After Rituximab Treatment for Lymphoma,

Abstract 3675

Emerging evidence suggests that B cells can actively modulate T cell immune responses by presenting antigen, providing co-stimulation and secreting cytokines. This has prompted investigation whether B cell depletion by monoclonal antibodies, including Rituximab, can alter the subset composition, activation or function of T cells. Rituximab-associated late onset neutropenia (R-LON) is increasingly recognized as a long-term adverse event of Rituximab. Although the etiology of R-LON is not fully elucidated, the syndrome seems to be multifactorial and likely caused by immune-mediated mechanisms. We have previously shown that at least a proportion of R-LON may develop in a setting of expanded cytotoxic T cell populations in peripheral blood (PB) with a large granular lymphocyte (LGL) phenotype (CD3+CD8+CD57+). Here, we extend our observations regarding PB lymphocyte subset composition in a cohort of 107 Rituximab-treated patients with available results from PB flow cytometry analysis performed at roughly similar intervals after the initial Rituximab administration. The present cohort included 107 patients, aged 16–83 (median 60), who received Rituximab for the treatment of chronic lymphocytic leukemia (CLL) (29), diffuse large B cell lymhoma (DLBCL) (20), marginal zone lymphomas (15), follicular lymhoma (FL) (15), mantle cell lymphoma (MCL) (8) and auto-immune cytopenias (20). Overall, we found: (i) increased (>1.0×10⁹/l) CD8+ cells in 45/107 (42%) cases; (ii) CD4+/CD8+ cell ratios <0.7 in 56/107 (52%) cases; and, (iii) T-LGLs >20% in 66/107 (63%) cases. Within this cohort, 33 cases (group A) developed R-LON, whereas the remainder (group B, n=74) did not develop this syndrome over a comparable observation period. Importantly, no patient with autoimmune cytopenia developed R-LON. R-LON was significantly more frequent in patients with lymphoma subtypes treated with intensive chemotherapy (CLL, DLBCL, MCL), as well as patients who underwent autologous transplantation (p\q0.001 for all comparisons). No significant differences were noted between groups A or B regarding PB lymphocyte subset composition. We next evaluated the findings from the histopathological study of bone marrow biopsy samples, available in 17 Group A and 19 Group B cases, all with a diagnosis of lymphoma. The morphological and immunohistochemical examination revealed a series of features common in both groups, summarized as follows: (i) mild-to-moderate small lymphocytic infiltration by CD20-CD79a-CD3+CD45RO+CD43+ (CD3>CD45RO) cells, predominantly nodular and/or interstitial (non intrasinusoidal); (ii) pronounced hyperplasia of the erythroid and megakaryocytic series with prominent dyserythropoiesis and dysmegakaryopoiesis, respectively, including abnormal paratrabecular localization, suggestive of myelodysplasia (MDS); and (iii) remarkable shift-to-the-left of the granulocytic series, often with abnormal localization of immature progenitors (ALIP), always with \q2% CD34+ cells. A proportion of cases showed hyperplasia of the granulocytic series. However, a major difference between the two Groups concerned hypoplasia of the granulocytic series, which was noted almost exclusively in group A. We conclude that lymphoma patients treated with Rituximab often develop cytotoxic T cell expansions than can have a variable impact on hematopoiesis, with R-LON perhaps representing the end of a spectrum of T-LGL-mediated autoimmune myelopathy/myelodyplasia. The selective development of R-LON in only a proportion of cases with expanded cytotoxic T cells associated with prominent hypoplasia of the granulocytic series and MDS-like changes of the hematopoietic marrow post Rituximab raises several questions regarding the underlying (genetically determined?) immunopathogenetic mechanisms.