Anti-CD20 (rituximab) therapy for anti-IFN-γ autoantibody-associated nontuberculous mycobacterial infection

Abstract

Patients with anti-IFN-γ autoantibodies have impaired IFN-γ signaling, leading to severe disseminated infections with intracellular pathogens, especially nontuberculous mycobacteria. Disease may be severe and progressive, despite aggressive treatment. To address the underlying pathogenic IFN-γ autoantibodies we used the therapeutic monoclonal rituximab (anti-CD20) to target patient B cells. All subjects received between 8 and 12 doses of rituximab within the first year to maintain disease remission. Subsequent doses were given for relapsed infection. We report 4 patients with refractory disease treated with rituximab who had clinical and laboratory evidence of therapeutic response as determined by clearance of infection, resolution of inflammation, reduction of anti-IFN-γ autoantibody levels, and improved IFN-γ signaling.

Figure 1

Timeline for rituximab therapy. Each timeline represents the treatment course for 1 patient. Panel A reflects IgG levels with normal range in gray, panel B has dots indicating months in which rituximab was given and a number indicating how many doses at 375mg/m² were given in that month. Panel C reflects total absolute B-cell numbers by CD19 staining with normal range in gray. Black arrows are the times at which plasma was collected for Ab titers and functional studies for Figures 2 and 3.

Figure 2

Anti–IFN-γ autoantibody titers in relation to treatment with rituximab. (A) IFN-γ labeled beads were incubated with subject plasma at 10-fold serial dilutions to generate an Ab titration curve, as a function of fluorescence intensity, and determine a common dilution factor within the dynamic range of the assay. A representative example of the titration curve for patient 4 during rituximab therapy with the 1:5000 dilution indicated (black arrow) that lies within the linear phase of the dilution curve. (B) Each patient's Ab titer over the course of rituximab therapy was plotted at 1:5000 dilution.

Figure 3

Patient plasma inhibition of IFN-γ–induced pSTAT-1 production and RNA expression. Normal PBMC incubated in 10% control or patient plasma at time points specified in Figure 1 were left unstimulated or stimulated with IFN-α or IFN-γ for 15 minutes. CD14⁺ cells were assayed for intracellular pSTAT-1 by flow cytometry. (A) pSTAT-1 in CD14⁺ cells, solid gray, unstimulated; red line, IFN-γ stimulated; blue line, IFN-α stimulated. Representative example of 1 of 3 experiments performed for each patient. (B) To determine the relative inhibitory effect of plasma on IFN-γ induced pSTAT-1 production, a stimulation index for each plasma (ratio of the geometric mean channel for stimulated to unstimulated) was calculated and then graphed as a percentage of the stimulation index seen in the presence of normal plasma. Error bars represent the SEM seen for 3 separate experiments performed for each series of patient plasma. (C) PBMC obtained from healthy donors (n = 4 experiments per series of patient plasmas) were cultured in the presence of normal or patient plasma (10%) and stimulated with IFN-γ (400U/mL) for 3 hours. Target gene expression was evaluated by real time PCR and values are mean fold induction (± SD) relative to the unstimulated cells. GAPDH was used as normalization control. Fold-induction of IFN-γ–induced gene expression in presence of subject plasma was calculated as a percentage of the fold-induction expression seen for the same PBMCs in the presence of normal plasma. Plasma from patient 4 inhibited IFN-γ induced gene expression when stimulated with IFN-γ 400U/mL at all time points, and so the experiment was repeated using IFN-γ at a higher concentration of 3000U/mL.