‘A novel approach for characterisation of KSHV‐associated multicentric Castleman disease from effusions’: Response

We appreciate the insightful comments by Vanjak et al. regarding our recent publication ‘A novel approach for characterization of KSHV-associated multicentric Castleman disease from effusions’.1 In our report, we used a broad definition of ‘plasmablasts’ to describe the Kaposi sarcoma-associated herpesvirus (KSHV)-infected, lambda-restricted B-lineage cells in KSHV-associated multicentric Castleman disease (KSHV-MCD), referred to as the lambda-restricted plasmablastic population (LRP) in the publication. As we have previously discussed,2, 3 the term ‘plasmablasts’ has been widely used in the literature to describe the KSHV-infected B-lineage cells in the lymph nodes of KSHV-MCD, although it may be debatable whether these cells truly represent plasmablasts. Historically, the term was used as a histological descriptor, reflecting the plasmablastic morphology of these KSHV-infected cells, rather than by their phenotype or ontogeny.4 As Vanjak et al. noted, the phenotype of these KSHV-infected cells is somewhat different from that of conventional plasmablasts, which have historically been defined as CD19low, CD38+, CD27+, immunoglobulin G (IgG)+ cells that have undergone somatic hypermutation and class switch recombination during germinal centre reactions. However, there is noteworthy controversy and ongoing debate surrounding the characterisation of these cells. Previous research suggests that KSHV-infected B cells may differentiate into plasmablasts through a non-standard extrafollicular pathway, bypassing the germinal centre reaction, expressing IgM instead of IgG, and lacking somatic hypermutation in their IgV genes.5, 6 These processes may conceivably contribute to a different immunophenotype observed in these cells.7 While we feel it may be preferable to use conventional terminology when describing these cells until more conclusive evidence is generated, we concur with the Vanjak et al. points and appreciate their discourse in highlighting this topic for discussion. Regarding the latency associated nuclear antigen-1 (LANA-1) flow cytometric assay, we used the rat anti-ORF73 monoclonal antibody clone LN53 (Sigma-Aldrich/Millipore) in our study. The LN53 antibody was able to detect the BC3 primary effusion lymphoma (PEL) cell line spiked into normal blood in our cell line experiment (Figure 1). At the time of our study, a directly conjugated LN53 antibody was not commercially available; therefore, a directly conjugated AF647-LN53 antibody was obtained through a custom conjugation service provided by the vendor. The gating in our LANA-1 flow cytometric assay was established by using ‘Fluorescence Minus One’ (FMO) and normal cell populations within the blood, such as lymphocytes, served as an internal negative control, which are advantageous over isotype controls in the analysis and reporting of clinical specimens.8, 9 The assay was rigorously validated on samples from patients with PEL and KSHV-MCD by comparing the LANA-1 flow cytometry results to concurrent LANA-1 immunoperoxidase staining performed on cytology slides, which served as the ‘gold standard’. We observed that although the levels of expression varied, LANA-1 was reliably detected in the KSHV-infected cells by flow cytometry. We speculate that the discrepant results presented by Vanjak et al. may be related to the utilisation of fluorochrome-conjugated secondary antibodies for detection of their unconjugated primary antibodies in their staining procedures. In our experience, an indirect staining protocol using unconjugated primary antibody and fluorochrome-labelled secondary reagents gives rise to increased non-specific labelling for detection of surface antigens, especially in clinical specimens such as blood, bone marrow and body fluids; furthermore, the issue is compounded when attempted for detection of intracellular antigens, due to the difficulty in effectively removing excess reagents during washing. Fortunately, directly conjugated LN53 antibodies have recently become commercially available (e.g., Novus Biologicals). These advancements will further facilitate the use of the LANA-1 flow cytometric assay as a streamlined and valuable diagnostic tool. Vanjak et al. raised an important point regarding the specificity of the flow cytometry-based analysis in differentiating between KSHV-MCD and other KSHV-associated lymphoproliferations, including PEL and KSHV/human herpesvirus 8 (HHV8)-positive diffuse large B-cell lymphoma (KSHV/HHV8+ DLBCL). We concur that PEL has a highly variable phenotype and can occasionally show overlapping expression of one or more individual markers included in our panel. However, when the entire antigen expression pattern is evaluated as a whole, the PEL cases in our limited series were readily distinguished from the LRP, which demonstrated a consistent and reproducible phenotypic profile relative to the normal plasma cells usually present in the background. While we agree that morphological evaluation remains the cornerstone of PEL diagnosis, we believe the flow cytometry-based panel nonetheless can serve as a useful adjunct, especially in the setting of residual PEL detection after treatment, where remaining tumour cells may be scarce. In this situation, the variable cytological atypia of LRP may be misinterpreted as PEL; however, its characteristic phenotype will point to a diagnosis of an underlying or co-existing KSHV-MCD, which is not uncommon in patients with PEL. As for the diagnosis of KSHV/HHV8+ DLBCL, our available data are currently not sufficient to make conclusions on the utility of flow cytometry in this context. We have encountered cases of KSHV+ DLBCL that showed a marked expansion of atypical cells in the peripheral blood, with a phenotype indistinguishable from LRP (Figure 2). The findings suggest that flow cytometric phenotyping alone may have limitations distinguishing between the two. However, the presence of a significant expansion of LRP or LRP-like population should prompt further evaluation for the possibility of progression to KSHV/HHV8+ DLBCL, including radiographic studies as well as histological and clonality evaluations. We also agree with Vanjak et al. that it would be valuable to analyse the peripheral blood of KSHV-positive patients who have LRP within effusions. Due to the retrospective nature of this study, evaluation of peripheral blood in these patients was not routinely performed, resulting in limited data. Peripheral blood evaluation may provide insight into the question of whether the body cavities represent primary sites of the disease or sites of secondary involvement ‘seeded’ by circulating LRP. Detection of circulating LRP in patients with liquid-form MCD would suggest that these cells may be able to traverse between compartments and potentially act as a vector for viral dissemination to targeted tissue. This concept may prove to be a valuable avenue for further investigation in understanding the pathogenesis of KSHV-MCD and also inform the design of more effective therapeutic strategies. In conclusion, our study highlighted the usefulness of flow cytometric analysis in the diagnosis of KSHV-MCD and introduced a novel concept of liquid-form MCD. Admittedly, our study included a limited number of cases, and we look forward to these findings potentially being further validated on larger cohorts. The rarity of these diseases may necessitate multicentric studies. We hope that these results will serve as a catalyst for future research in this field.