Background: Idiopathic multicentric Castleman disease (iMCD) is a rare, atypical lymphoproliferative disorder with significant morbidity and mortality and unknown etiology. While anti-interleukin-6 (IL-6) therapy with siltuximab is effective in 34-50% of iMCD patients, disease pathophysiology remains largely unknown. A fundamental question is whether iMCD is most appropriately considered an infectious disease caused by an as-yet-unknown pathogen, an autoimmune disease caused by autoantibodies, an autoinflammatory disease caused by germline mutations, or a neoplastic disease caused by somatic mutations in a clonal cell population. While limited evidence exists for each of these hypotheses, the pathogen hypothesis has been the most systematically evaluated to date. In this study, we quantified the levels of autoantibodies that may be involved in iMCD pathology.
Methods: We constructed two custom bead-based protein arrays to screen serum samples for IgG autoantibodies from iMCD patients (n = 101) and healthy controls (HC, n = 30). The Connective Tissue Disease (CTD) array consisted of 52 antigens associated with traditional CTDs including in Scleroderma, Myositis/Overlap Syndromes, Systemic Lupus Erythematosus (SLE)/Sjögren's, GI/Endocrine, DNA-Associated, and Inflammation/Stress. The Anti-Cytokine Autoantibody (ACA) array included 38 cytokines, chemokines, and cell surface proteins. Serum samples came from the University of Pennsylvania (n = 38), University of Arkansas for Medical Sciences (n = 45), and Osaka University (Japan, n = 18). A “positive” autoantibody result in any sample was defined as greater than five standard deviations (SD) above the average mean fluorescent intensity (MFI) of the HC and at least 3,000 MFI. Fisher's exact tests were used to determine associations between autoantibody positivity and disease status.
Results: To evaluate assay precision, we investigated anti-IL-6 antibodies given that a large portion (38%) of iMCD patients were on anti-IL-6 therapy. As expected, 40 iMCD samples were positive for anti-IL-6 antibodies, the most common autoantibody detected in iMCD subjects. IL-6 was then removed from further analyses. Next, we evaluated the proportion of CTD autoantibody positive iMCD patients. Overall, 46 (46%) iMCD subjects were positive for a CTD autoantibody compared to five (17%) in HCs (OR 4.1, P = 0.005) (Figure 1A). When investigating autoantibody prevalence in CTD subcategories, Myositis/Overlap Syndromes (OR 7.5, P = 0.026) and SLE/Sjögren's (OR 6.2, P = 0.073) autoantibodies trended higher in iMCD patients, but these associations did not survive multiple testing correction. The most common Myositis/Overlap Syndrome autoantibodies were anti-Mi2 (iMCD: 7/101 vs HC: 0/30), and anti-SRP54 (7/101 vs 0/30). The most common SLE/Sjögren's autoantibodies were anti-La/SSB (9/101 vs 0/30) and anti-Ro60 (5/101 vs 0/30). For the custom ACA array, there were significantly more ACA positive iMCD subjects (n = 40, 40%) than HC (n = 3, 10%; OR 5.8, P = 0.002) (Figure 1B). The most common ACAs in iMCD were anti-OSM (9/101 vs 0/30), anti-TNF (6/101 vs 0/30), anti-ITM2B (6/101 vs 0/30), and IFNe (5/101 vs 1/30).
Conclusions: IgG autoantibodies associated with CTDs, such as anti-Mi2, anti-SRP54, anti-La, anti-Ro, and anti-Histone 3; and ACAs, such as anti-OSM, anti-TNF, and anti-ITM2B were common in iMCD patients. While autoantibodies suggest autoimmune involvement, the presence of autoantibodies does not directly implicate autoimmunity as a pathological mechanism. For example, Hodgkin's disease and other post-infectious etiologies can present with elevated autoantibodies.
Although not surviving multiple hypothesis correction, myositis and SLE autoantibodies trended higher in iMCD patients. While most of these autoantibodies have not been reported, anti-Ro and and anti-La antibodies have been reported previously in iMCD case reports. The presence of anti-OSM is interesting, as OSM is a member of the IL-6 family. Two hypothesized mechanisms include that autoantibodies may be non-neutralizing leading to increased half-life and cytokine activity or autoantibodies are neutralizing and lack of negative feedback from OSM may lead to increased signaling via other IL-6-family cytokines. More research is needed to understand the role of these autoantibodies and, if pathogenic, whether iMCD should be considered an autoimmune disease.
Disclosures
Utz:Board of Directors of the Physician Scientist Support Foundation: Membership on an entity's Board of Directors or advisory committees; Arthritis National Research Foundation: Membership on an entity's Board of Directors or advisory committees; 4DMT: Membership on an entity's Board of Directors or advisory committees; SeraNova: Membership on an entity's Board of Directors or advisory committees; Yolo Immune, Inc.: Membership on an entity's Board of Directors or advisory committees. Fajgenbaum:Medidata, a Dassault Systemes company: Consultancy; EUSA Pharma/Recordati Rare Disease: Consultancy, Research Funding.
