A clinical, histological and transcriptomic characterization of a selected series of Castleman disease’s cases

Castleman disease (CD) is a rare lymphoproliferative disorder that includes various clinico-pathological subtypes. According to clinical course CD is divided in unicentric CD (UCD) and multicentric CD (MCD). MCD is further distinguished based on the etiological driver in herpes-virus 8 related MCD (that can occur in the setting of HIV), in MCD associated with POEMS syndrome (polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, and skin changes) and idiopathic MCD (iMCD). The latter can also be divided in iMCD-TAFRO (thrombocytopenia, anasarca, fever, reticulin fibrosis, organomegaly) and iMCD-not otherwise specified. To date, CD pathogenesis is still uncertain, but it may represent the histological and clinical result of heterogeneous pathomechanisms, that mainly determines cytokines activation. A key role is played by IL-6, as demonstrated by efficacy of anti-IL6 therapy in some patients. To date, transcriptome investigations in CD lymph nodes documented expression and up-regulation of different cytokines and, specifically, have been made providing contrasting results about IL-6 trascription; furthermore few recent studies showed alterations of different T-cell subsets in CD patients, suggesting a possible role of the nodal microenvironment in CD development. On such basis, our study was aimed to investigate the distribution of T-cell subsets and to assess the expression and quantification of IL-6 RNA sequence (RNAseq), in the clinico-pathological spectrum of CD. We evaluated the CD4/CD8 ratio and the number of T regulatory (T-reg) FOXP3+ cells in 28 CD cases. 32% of cases showed a decreased CD4/CD8 ratio due to increased CD8+ T cells, including both UCD, iMCD and HHV8+ MCD cases. T-reg subset analysis revealed a statistically significant (p<0,0001) lower mean number of FOXP3+ T-reg cells in CD cases when compared with non specific reactive lymph nodes. We did not find statistically significant differences in T-reg number between the different CD subtypes. About IL-6 RNAseq we could enroll 5 of our CD patients (1 UCD, 2 iMCD, 1 HHV8+HIV- associated MCD, 1 HHV8+HIV+ associated MCD) and a non CD reactive control. All cases were tested to assess IL-6 RNAseq expression and quantification (held on 5 fields at 20x magnification using HISTOQUANT software). We used dual in situ hybridation-immunohistochemistry protocol to correlate the production of IL-6 with different nodal cellular population (CD3+ T-cell lymphocytes, CD68R/PGM1+ macrophages and CD31+ endothelial/lymphatic cells). All cases showed expression of IL-6 with a higher expression in CD cases compared to the control. Particularly, post-hoc analyses showed statistically significant differences (p<0.05) in IL-6 expression between HHV8+ cases and control. In CD cases IL-6 expression was noted in CD31+ endothelial/lymphatic cells whereas CD3+ T-cells and CD68R/PGM1+ macrophages were negative. Our findings about T-cell subsets in CD lymphonodes may suggest that alterations in T-cell subpopulations that can lead to disruption of immune system control may contribute to the numerous changes in different cellular compartments that characterize CD. Furthermore our findings about IL-6 RNAseq seem to suggest that despite CD heterogeneity some similarities may exist in the lymph node microenvironment of different CD types. In this respect, our analysis seems to indicate that the major source of IL-6 production may be CD31+ endothelial cells.