Oligoclonal Expansion of Cd8+ T Cells during Idiopathic Multicentric Castleman Disease Flares Suggests an Antigen Driven Process

Castleman disease (CD) describes a group of heterogeneous diseases defined by shared characteristic lymph node histopathology and is classified based on the number of regions of enlarged lymph nodes. Multicentric CD (MCD) involves multiple regions of lymphadenopathy as well as systemic inflammation, cytopenias, and vital organ dysfunction due to a cytokine storm that often includes interleukin-6. In ~50% of patients, the pathogenic driver is Kaposi sarcoma-associated/human herpesvirus-8 (HHV-8) in the context of immunosuppression. In contrast, the etiologic driver in HHV8-negative MCD (idiopathic or iMCD) is unknown. To date, most research has focused on descriptive characterization of the enlarged lymph nodes, and the pathological cell types driving iMCD pathogenesis remain unidentified. Given that lymphoid cells circulate through the blood and lymph nodes, are able to produce high levels of cytokines upon activation, and are the primary cell types responsible for the enlarged lymph nodes in iMCD and other related diseases, we first performed a detailed immunophenotyping of peripheral blood mononuclear cells (PBMCs) obtained from iMCD patients in remission (n=16), iMCD patients during disease flare (n=6) and healthy donors (HD) (n=15). PBMCs were isolated by density gradient and either stained immediately or cryopreserved for future analyses. A HD sample was drawn at the same time as each experimental sample and processed and analyzed in parallel. Our initial hypothesis was that analysis of iMCD flare PBMCs would reveal an abnormal myeloid or lymphocyte subset. Thus, we stained and analyzed PBMCs for standard lineage markers: CD11b, CD15, CD19, CD3, CD56 and CD14. However, we observed no gross differences in population frequencies during either remission or flare compared to HD. Additionally, no differences in the proportions of natural killer T cells (CD3⁺CD56⁺), or CD4⁺ or CD8⁺ lymphocytes were observed. However, more refined examinations of the lymphocyte sub-sets based upon activation status revealed an increased proportion of activated memory (CD62L^lowHLA-DR⁺) CD8⁺ cells during iMCD flare compared to HD and iMCD patients in remission and a decreased proportion of naïve (CD62L⁺CD45RA⁺) CD8⁺ cells compared to HD (p<0.05 for each comparison, Bonferroni corrected 2-tailed t-test). Interestingly, for one patient followed over a 7-month period, during which time he experienced two disease flares, the proportion of activated memory CD8⁺ cells mirrored the re-emergence of clinical disease symptoms, rising to 80% of all CD8⁺ lymphocytes during disease flare and waning as flares subsided. We next questioned whether the expansion of memory CD8⁺ T cells represented a diverse population of T cells with unique TCRs or alternatively, reflected the expansion of one or a few dominate TCR clones. Thus, we performed bulk TCRβ sequencing on genomic DNA from peripheral CD8⁺ cells obtained during the patient's partial remission between his two flares and at the start of his second disease flare (flare 2). Interestingly, we found that TCRs from CD8⁺ cells showed an overrepresentation of only a few clonotypes, indicating increased oligoclonality; the top 20 clones accounted for 45% (partial remission) and 52% (flare 2) of bulk CD8⁺ reads. Furthermore, the top copy number rearrangement comprised 7.5% (partial remission) and 10% (flare 2) of the CD8⁺ cell repertoires (healthy subjects' CD8⁺ T cell top copy number rearrangement is typically below 5%). Sequencing of activated (HLA-DR⁺) and memory (CD45RO⁺) CD8⁺ populations also revealed increased clonality. In these populations, the top 20 clones accounted for 45% (CD45RO⁺) and 33% (HLA-DR⁺) of sequencing reads during the patient's partial remission and increased to 61% (CD45RO⁺) and 54% (HLA-DR⁺) of sequencing reads during flare 2. Further, the top TCRβ sequences persisted over time, being identified in both the partial remission and flare 2 samples. This study identifies quantitative and phenotypic differences in CD8⁺ cell populations during iMCD disease flare. Furthermore, we present evidence suggesting a role for oligoclonal T cells in iMCD, as our TCRβ sequencing findings reveal a substantial accumulation of only a few TCR clonotypes in CD8⁺ populations during disease flare. Cumulatively, these results suggest that TCR signaling, due to antigen stimulation or T cell dysregulation may be involved in iMCD pathogenesis.