Insights into the etiology of Castleman disease

In this issue of Blood, Chan et al report on in-depth studies of identical twins with idiopathic multicentric Castleman disease (iMCD), which provides a unique opportunity to study a genetic background for this enigmatic disorder.¹ 

Chan et al describe 2 germ line mutations that could play a role in the development of iMCD. Homozygous mutations in the nuclear receptor coactivator 4 (NCOA4) and heterozygous mutations in tumor necrosis factor receptor–associated factor TRAF genes were found in each of the twins. Variants in NCOA4 have been previously described in 5 of 22 patients with iMCD (23%) studied by You et al.² Although You et al characterized those variants as likely to be somatic, the variant allele frequency of approximately 50% in every case and their unusual methodology of utilizing external patient controls for calling somatic variants suggest that they are likely germ line heterozygous variants in that study. It has been suggested that the interaction between androgen receptor–associated proteins like NCOA4 and MAPK may enhance this pathway, which also can be stimulated by interleukin-6 (IL-6).² TRAF mutations have been linked to B-cell dysregulation and hyperactivity, increased inflammatory response, abnormal immunity, and increased risk for B-cell malignancies. Of interest is that the timing of iMCD development and clinical phenotype was not the same in both twins, suggesting that there is an interplay between genetic background and unidentified environmental triggers.

Castleman disease (CD) encompasses several hematologic disorders that have similar lymph node histopathological features. Broadly, CD can be divided into unicentric and multicentric depending on the extent of lymphadenopathy. Many multicentric CD cases are driven by the human herpes virus type 8 (HHV8) typically in the setting of immunosuppression (HHV8-associated multicentric CD), although other cases occur together with a clonal plasma cell disorder called POEMS (polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, skin changes) syndrome (POEMS syndrome–associated multicentric CD). However, in approximately 50% of cases no cause is identified, and these patients are now referred to as having iMCD. iMCD is associated with a cytokine-driven inflammatory syndrome, often involving IL-6, causing fevers, night sweats, constitutional symptoms as well as laboratory abnormalities such as elevated C-reactive protein and/or erythrocyte sedimentation rate and hypergammaglobulinemia. In severe cases, there can be organ dysfunction including renal impairment or even death. Recently, progress has been made by the formulation of international consensus criteria for the diagnosis of Castleman by an expert panel under the auspices of the Castleman Disease Collaborative Network (CDCN).³ Similarly, the CDCN has established international consensus treatment guidelines in which neutralization of the cytokine IL-6 with the monoclonal antibody therapy siltuximab is recommended as first-line treatment.⁴ 

Less progress has been made in the etiology of iMCD, which has remained elusive thus far. A comprehensive search with a virome capture sequencing platform found no known or unknown virus associated with iMCD. Thus, the hypothesis that a viral infection leading to constitutive inflammatory cytokine release has been largely unsupported to date.⁵ It has also been proposed that autoimmune mechanisms may trigger iMCD. Low-level autoantibodies can be present in approximately one-third of patients, but these are virtually always nonspecific. In fact, using the new diagnostic criteria, the presence of a clearly defined autoimmune syndrome excludes a diagnosis of iMCD. Other proposed causes of iMCD include germ line mutations in genes involved in the inflammatory cascade, as found by Chan et al in this study. Alternatively, somatic mutations in cells contained in the lymph nodes may also be contributory. In unicentric CD (UCD), activating mutations in platelet-derived growth factor receptor-β have been in found in 17% of patients with CD45⁻ stromal cells.⁶ 

It has been long established that IL-6 plays an important role in iMCD. Studies of the cellular source of IL-6 production have not been conclusive, and IL-6 was initially thought to originate from plasma cells. A recent transcriptomic analysis of paraffin-embedded lymph node samples complemented by RNA in situ hybridization found IL-6 overexpression in CD31⁺ endothelial or lymphatic structures.⁷ Chan et al likewise found overexpression of IL-6 pathway genes in endothelial cells and in nodal fibroblastic cells, suggesting a possible role of stroma in iMCD as is the case in UCD.

Proteomic analysis, supplemented by immunohistochemistry and other studies, has provided interesting new insights into the pathogenesis of iMCD and pointed to new therapeutic opportunities. It is important to consider that 35% to 50% of patients with iMCD respond to blockade of the IL-6 pathway, suggesting that other drivers and signaling pathways are important. Recently, the JAK-STAT pathway has found to be activated even in patients who do not respond to the anti-IL-6 monoclonal antibody siltuximab.⁸ There have been incidental reports of patients responding to JAK-STAT inhibitor therapy. It is possible that a group of patients signal through JAK-STAT as result of other cytokines or ligands. In addition, phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin (mTOR) signaling occurs in anti-IL-6 refractory patients, and the presence of increased mTOR activation has led to a clinical trial with the mTOR inhibitor sirolimus, with some patients responding.

Overall, these considerations suggest that the picture of iMCD is complex but beginning to be resolved. Clinically different phenotypes of iMCD have been described including idiopathic plasmacytic lymphadenopathy; thrombocytopenia, anasarca, fever, reticulin fibrosis, and organomegaly; and not otherwise specified. Proteomic analyses suggest that there may be 5 distinct groups, 1 of which is enriched for responders to the siltuximab.⁹ 

The hope for deep sequencing studies has been to find the iMCD equivalent of the Reed-Sternberg cell in Hodgkin disease. However, it is perhaps not surprising that no common mutational profile has emerged from the limited studies performed thus far taking the clinical and biological heterogeneity of iMCD into account (reviewed by Butzmann et al).¹⁰ The broader context of the study by Chan et al emphasizes the urgent need for a comprehensive search for both germ line mutations as well somatic aberrations in lymph node tissue, which could uncover a variety of mutations affecting inflammatory pathways. The diagnosis of iMCD is notoriously difficult and is hampered by the absence of specific biomarkers in blood and lymph nodes. A future targeted sequencing panel may help clinicians and pathologists during the diagnostic process and possibly provide insights into novel therapeutics.

Conflict-of-interest disclosure: F.v.R. serves on the advisory boards of Adicet Bio, Bristol Myers Squibb, EUSA Pharma, GlaxoSmithKline, Janssen, Kite Pharma, Secura Bio, and Recordati. D.F. serves on the advisory boards of Recordati and EUSA Pharma.