Reply to Zhou et al. ‘A novel approach for characterization of KSHV ‐associated multicentric Castleman disease from effusions’

We read with great interest the study by Zhou et al. entitled ‘A novel approach for characterization of KSHV-associated multicentric Castleman disease from effusions’ published in The British Journal of Haematology on the 4 October 2022.1 Kaposi sarcoma-associated herpesvirus (KSHV) + multicentric Castleman disease (MCD) is a life-threatening condition requiring prompt diagnosis that currently relies on a lymph node biopsy showing the presence of immunoglobulin (Ig)M+CD38+lambda+ KSHV-infected cells located in the mantle zone associated with plasma-cell rich features of CD. We previously reported the isolation of circulating IgM+CD38+lambda+ cells in the peripheral blood (PB) of three patients with active KSHV+ MCD.2 In this report, the authors first showed that flow cytometry could help differentiating KSHV+ MCD from primary effusion lymphoma (PEL), a KSHV-related malignant disorder carrying a dismal prognosis.3 In all, 11 patients presenting with effusions were analysed using a systematic antibody panel, and lambda-restricted CD38+ cells were isolated in eight of the 11 cases. No IgM staining was performed. These findings raise the question of the specificity of this analysis as PEL phenotype is extremely variable from one individual to another, as underlined by the authors. PEL are large cells expressing CD38, with a variable expression of the lambda light chain (three of 19 positive cases reported in Ref. [3]), CD138 (12/19 negative cases in Ref. [3]) and CD45 (one of 11 bright positivity in Ref. [1]) and the designed panel appears to be insufficient to formally differentiate KSHV-MCD from PEL. However, PEL is usually not a challenging diagnosis, excepted in cases with aberrant expression of T-cell markers that could suggest the possibility of T-cell lymphoma. The very peculiar cytological features and phenotype associated with very high human herpesvirus 8 (HHV8) DNA levels and complex karyotype easily prompt the diagnosis. The inconstant and unnecessary association with Epstein–Barr virus (EBV) is an additional argument for the diagnosis.3, 4 In contrast, KSHV-related diffuse large B-cell lymphoma, not otherwise specified often associated with MCD lesions5 could represent a challenging differential diagnosis as malignant cells express CD38 and exhibit an IgM lambda-restricted phenotype similar to that observed on MCD plasmablasts. This issue has not been addressed in this study. The authors then showed that lambda-positive cells were KSHV-infected using an antibody directed against the latent nuclear antigen (or LNA-1) by flow cytometry. The experimental conditions have not been detailed in the Material and Methods nor in the Data S1 sections and the antibody has not been referenced. Staining of an isotype control antibody has not been shown. As shown in Figure 1, we did not find any commercially available LNA antibody suitable for flow cytometry, notably because of a lack of specificity on KSHV+ cell lines (KSHV+EBV− PEL cell lines BC-3 and BCP1) when compared to KSHV− cell line (BJAB cell line, a KSHV−EBV− Burkitt-like lymphoma cell line). Immunohistochemistry prevents using multiparametric staining and we therefore developed a flow-fluorescent in situ hybridisation technique allowing the identification of latent and lytic KSHV transcripts coupled with an extracellular multiparametric staining.6 This method should therefore be preferred when assessing KSHV infection on liquid samples. The authors finally showed that KSHV-infected cells could be detected from PB or bone marrow of patients with active KSHV+ MCD (n = six) and KSHV inflammatory cytokine syndrome (n = three). The number of samples is relatively small, but the results are in line with our findings on the PB of 18 patients with KSHV+ MCD flare6 where we could detect KSHV-infected cells in up to 80% of the samples. It would have been interesting to analyse the PB of KSHV+ patients who had lambda-restricted cells within effusions. The lambda-restricted cells described in this work were called plasmablasts (or lambda-restricted plasmablasts [LRPs]), as KSHV+ MCD was formerly called the ‘plasmablastic’ variant due to the morphology of the large cells infected by KSHV/HHV8 found in the lesions. We performed a comparative phenotypic analysis between KSHV-infected cells found in KSHV+ MCD and conventional plasmablasts found in various reactive conditions and showed these cells differed from conventional plasmablasts and, therefore, should be better denominated as KSHV-infected viroblasts (KIVs) rather than plasmablasts. Of note, conventional plasmablasts were mainly IgG+ and uniformly expressed CD27 and CD86, whereas KIV were IgM+ and had a variable expression of CD27 and CD86.6 Altogether these novel findings require further technical investigations and validation on larger cohorts but represent interesting future directions in the understanding of KSHV-related lymphoproliferative disorders. A. Vanjak and D. Boutboul wrote the manuscript. A. Dossier, L. Galicier, R. Bertinchamp, E. Oksenhendler, and D. Boutboul contributed to the patient recruitment and management. A. Vanjak, M. Garzaro, S. Knapp, M.-A. Silvestrini, G. Martin de Fremont and D. Boutboul performed the experiments and analysed the data. V. Meignin and J. Calvani performed pathological reviewing of MCD samples. D. Boutboul supervised the project. All the authors critically reviewed the manuscript. Eric Oksenhendler is a consultant for Eusapharma. The other authors have no potential conflict of interest to declare. Data S1 Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.