Distinct Human and Viral Interleukin-6 Profiles and Other Viral and Immunologic Abnormalities In KSHV-Associated Multicentric Castleman Disease: Relationship with Disease Activity and Individual Disease Manifestations

Multicentric Castleman disease (MCD) is a polyclonal B-cell lymphoproliferative disorder characterized by flares of severe inflammatory symptoms, including fever and cachexia, with cytopenias and biochemical abnormalities. In the idiopathic form, its pathogenesis is linked to overproduction of interleukin (IL)-6. A distinct form of MCD is caused by Kaposi sarcoma-associated herpesvirus (KSHV, also called human herpesvirus [HHV]-8). KSHV encodes a viral homolog of IL-6, vIL-6, and this has been hypothesized to be central to KSHV-MCD pathogenesis. However, viral and human cytokines have not been examined together in KSHV-MCD, and their contribution to disease activity and symptoms is not known.

Patients with pathologically-proven KSHV-MCD were enrolled on a prospective natural history study incorporating pilot evaluation of novel therapies (NCT00099073). Factors potentially important in pathogenesis were assayed at flare and complete clinical and laboratory remission: KSHV viral load (VL) in peripheral blood mononuclear cells, vIL-6, IL-1β, IL-5, human IL-6 (hIL-6), IL-8, IL-10, IL-12p70, IFN-γ and TNF-α. Paired analyses were performed for each patient comparing initial flare and remission. Associations between cytokines and individual disease manifestations were explored across all flares.

21 patients had at least one flare for analysis (19 [90%] male; med age 44 [range 29–52]). 34 flares were observed (range 1–3 per patient) and followed to remission (20 patients) or death (1). All were HIV infected: CD4 med 252cells/μL (range 24–1319), HIV VL med <48 copies/mL (<48–64,100), 94% receiving antiretrovirals. Clinical symptoms included fever (present in 62%, med 38°C [range 36.1–40.5]); fatigue (91%, med CTC grade 2 [0–3]), gastrointestinal (GI) symptoms (68%, med grade 1 [0–3]) and respiratory symptoms (61%, med grade 1 [0–2]). Laboratory findings included anemia (97%, 9.9g/L [6.8–14.4]), thrombocytopenia (68%, 100×10³cells/μL [6–567]), hypoalbuminemia (97%, 2.7g/L [1.2–3.9]) hyponatremia (68%, 133mEg/L [127–143]) and C-reactive protein (CRP) elevation (100%, 87.3g/L [6.3–339.5]).

Factors elevated during initial flares were: KSHV VL (med 14,700/10⁶ PBMCs [range 0–3,913,000] P <0.0001 compared with remission); vIL-6 (detected in 48%, <1560pg/mL [<1560-20,500] P= 0.0039); hIL-6 (15.9pg/mL [1.4-171.5] P= 0.0006); IL-10 (449pg/mL [2.8-85,900] P= 0.0007); TNF-α (29.0 pg/ml [7.9–90.8] P= 0.0083) and IL-1β (1.2 pg/ml [0.1–5.7] P= 0.0027). IL-5 was decreased compared with remission (0.6 [0.1–15.4] P= 0.016). Differences were most marked for vIL-6 (undetectable in all remissions), hIL-6 (med increase from remission 520%) and IL-10 (med increase from remission 5000%). KSHV VL was correlated with hIL-6 (R = 0.68, P= 0.001) and IL-10 (R=0.82, P= 0.001) across all flares and remissions; other correlations were weaker. There was also a strong association between KSHV VL and vIL-6 (P= 0.001 by Jonckheere-Terpstra). For hIL-6 and vIL-6 only, we observed distinct profiles across flares: vIL-6 elevation only (2 flares, 6%), hIL-6 only (17 flares, 50%), and hIL-6 with vIL-6 (13 flares, 38%). In 2 flares (6%), neither vIL-6 nor hIL-6 were detected at onset, but in both hIL-6 alone soon became elevated. In contrast, KSHV VL, IL-10, TNF-α and IL-1β were consistently elevated during flares. Compared with hIL-6-only flares, hIL-6 with vIL-6 flares exhibited significantly higher CRP (P =0.0009); worse hyponatremia (P =0.02); higher KSHV VL (P =0.016) and IL-10 (P= 0.012); and lower IL-5 (P= 0.009). In linear/logistic regression models of disease manifestations with cytokines, best predictors were: GI symptoms, respiratory symptoms, temperature, platelet count: hIL-6 alone; hemoglobin: hIL-6 and vIL-6; sodium: hIL-6, vIL-6 and IL-10; albumin: IL-1β and vIL-6.

This prospective analysis shows for the first time that vIL-6 and hIL-6 can independently or together lead to flares of KSHV-MCD. It shows novel associations of KSHV-MCD flares with IL-1β and TNF-α elevation and IL-5 depression while confirming associations with KSHV VL and IL-10. It further suggests that vIL-6 and hIL-6 may jointly contribute to the severity of some symptoms during flares. As vIL-6 can signal independently of the ligand-binding IL-6 receptor-a, these findings have implications for the development of novel KSHV-MCD therapies targeting IL-6 and its downstream signaling.

No relevant conflicts of interest to declare.