We read with interest and some disappointment the paper by Dupin et al on HHV-8 in Castleman disease (CD).1 Regarding their observations concerning the evolution of clonal populations of HHV-8-positive B cells in this study, Dupin et al raise questions concerning the histogenesis of these cellular proliferations and fail to recognize that these questions were previously addressed in an article we published 4 years ago. In that paper, we immunophenotypically and genotypically analyzed an unprecedented number of cases of CD (n = 63) without clinical risk factors for HIV infection and demonstrated exclusively lambda-restricted plasma cells, immunoglobulin gene rearrangements, and non-Hodgkin lymphoma.2 We also postulated that an immunophenotypically distinctive B-cell subset accumulated in the mantle zone of plasma-cell (PC) CD, which was the precursor cell of the lambda-restricted neoplasms in these patients. The KiB-3 negative phenotype of the mantle zone lymphocytes in CD was consistent with the phenotype of Ly-1b lymphocytes,3 which preferentially demonstrate lambda immunoglobulin light chain rearrangement3 and express lambda immunoglobulin light chains.4 Ly-1b lymphocytes, although characterized as fetal (Ly-1) B lymphocytes, are found in adult patients and are produced in bone marrow like adult (Ly-2) B lymphocytes.5
Dupin et al's finding of latent nuclear antigen expression in the scattered plasmablastic cells in the mantle zone of patients with CD suggests that HHV-8 may selectively transform a subset of Ly-1b lymphocytes that reside in mantle zones in CD and may subsequently develop interfollicular microlymphomas in HIV infected patients with CD. We speculate that, because Ly-1b cells are important in the control of immune reactions involving recognition of self and non-self antigens, they may be associated with the development of anti-myelin-glycoprotein autoantibody formation and neuropathy in CD patients with the POEMS (polyneuropathy, organomegaly, endocrinopathy, monoclonal paraprotein, and skin changes) syndrome. Because interleukin-6 (IL-6) induces plasma cell differentiation from B-cell precursors, is produced by follicular dendritic reticulum cells and scattered interfollicular cells in CD,6-8 and is responsible for systemic symptoms in CD patients,9 IL-6 may also induce plasma-cell differentiation from mantle-zone Ly-1b lymphocytic precursors in PC CD. Because only a subset of HIV-negative CD patients are infected with HHV-8,10,11 because HHV-8 negative patients develop CD associated with the POEMS syndrome, and because systemic symptoms in HIV/HHV-8–infected CD patients resolve with human IL-6 neutralizing monoclonal antibodies,12 it is likely that human IL-6 rather than viral IL-6 is the principle pathogenic cytokine in PC CD.
The importance of Ly-1b (fetal) lymphocytes in the pathogenesis of hyaline vascular CD is suggested by the concentric accumulation of Ki-B3 mantle-zone lymphocytes in hyaline vascular disease, the similarities in anatomical distribution of Ly-1b lymphocytes, the clinical distribution of hyaline vascular CD,13-16 and the reported series of CD in children.17 Similarities in pathogenesis between hyaline-vascular and plasma-cell CD are suggested by Ly-1b type of mantle-zone immunophenotype in both, by the histologic identification of CD showing plasma-cell and hyaline-vascular histology or a transformation from one type of CD to another in sequential biopsies, and by the recognition of hyaline-vascular CD that produces IL-6 and is associated with systemic symptoms (POEMS syndrome).
Discussion of the paper by Dupin et al at the spring 2000 International Academy of Pathology meeting in New Orleans suggested that the single transformed mantle-zone plasmablasts and the associated plasmablastic microlymphomas lack somatic hypermutation of immunoglobulin variable regions. Because Ly-1b cells also lack hypermutation of immunoglobulin variable regions,18 it would be interesting to know whether the plasmablastic lymphomas or their transformed single-cell counterparts in the mantle zone also lack N-region diversity or restricted V-region gene usage characteristic of Ly-1b cells.19 The lack of somatic hypermutations in the plasmablastic lymphomas would confirm that they are not of germinal center derivation, unlike multiple myeloma, and may offer an explanation for the clinical differences in dissemination and behavior of the plasmablastic lymphomas arising in CD as compared with plasmablastic multiple myeloma. Microdissection experiments of the lambda-restricted plasma cells in PC CD are needed to determine whether they also lack evidence of somatic hypermutation, limited N-region diversity, or restricted V-region usage, as do Ly-1b cells.
HHV-8 positive cells in Castleman disease
Dr Menke takes us to task for not quoting his paper1-1 in which he claims to have discussed the same abnormal mantle-zone cells in Castleman disease (CD) as those that we have described. We were, of course, aware of Dr Menke's report of Ki-B3 negative mantle-zone cells in CD. He described these Ki-B3 negative cells in all cases of both hyaline-vascular- and plasma-cell-variant CD. It is difficult to ascertain whether he was referring to the entire mantle, as is suggested in figure 1A of his paper,1-1 or to isolated cells, and he did not illustrate the cytological features of the cells. By contrast, we have described cytologically distinct HHV-8 positive, IgMλ-positive plasmablasts occurring within an otherwise normal mantle zone. These cells were restricted to a subset of cases of plasma-cell-variant CD that were HHV-8 positive by PCR. In this context, we saw no reason to quote Dr Menke, and nothing in his letter has caused us to change our view.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal