Pre-existing antibody-mediated adverse effects prevent the clinical development of a bacterial anti-inflammatory protein

Abstract

Bacterial pathogens have evolved to secrete strong anti-inflammatory proteins that target the immune system. It was long speculated whether these virulence factors could serve as therapeutics in diseases in which abnormal immune activation plays a role. We adopted the secreted chemotaxis inhibitory protein of Staphylococcus aureus (CHIPS) as a model virulence factor-based therapeutic agent for diseases in which C5AR1 stimulation plays an important role. We show that the administration of CHIPS in human C5AR1 knock-in mice successfully dampens C5a-mediated neutrophil migration during immune complex-initiated inflammation. Subsequent CHIPS toxicology studies in animal models were promising. However, during a small phase I trial, healthy human volunteers showed adverse effects directly after CHIPS administration. Subjects showed clinical signs of anaphylaxis with mild leukocytopenia and increased C-reactive protein concentrations, which are possibly related to the presence of relatively high circulating anti-CHIPS antibodies and suggest an inflammatory response. Even though our data in mice show CHIPS as a potential anti-inflammatory agent, safety issues in human subjects temper the use of CHIPS in its current form as a therapeutic candidate. The use of staphylococcal proteins, or other bacterial proteins, as therapeutics or immune-modulators in humans is severely hampered by pre-existing circulating antibodies.

Keywords: C5aR chemotaxis; CHIPS; Clinical trials; Humanized mouse; Immune complex.

Fig. 1.

CHIPS binds and inhibits hC5aR1^KI murine neutrophils at levels comparable to those with human neutrophils. Quantification of hC5aR1 expression in hC5aR1^KI mice showed similar expression levels compared to human leukocytes (Tromp et al., 2018). Furthermore, hC5aR1^KI murine neutrophils responded normally to both murine C5a (mC5a) and human C5a as measured by Ca mobilization (Tromp et al., 2018). (A) hC5aR1^KI bone marrow neutrophils and human blood neutrophils were isolated and incubated with 3 µg/ml histidine-tagged CHIPS followed by anti-histidine-fluorescein isothiocyanite (FITC) antibodies. Cells were analyzed by flow cytometry and the FITC fluorescent signal was depicted as histograms. (B) As our hC5aR1^KI murine model generates mC5a, the assessment of CHIPS inhibition was performed by mC5a stimulation. Bone marrow neutrophils of hC5aR1^KI, wild-type (WT) mice and human neutrophils were pre-incubated with CHIPS at the indicated concentration and subsequently stimulated with murine C5a (10-8M). The basal fluorescence level was first measured for each sample before the addition of murine C5a. The C5a-mediated calcium influx was analyzed by flow cytometry using Fluo-4AM. The average Fluo-4AM fluorescent signal was used to calculate CHIPS-mediated inhibition of C5a responses. One experiment representative of two independent experiments is shown.

Fig. 2.

CHIPS inhibits neutrophil migration in vivo. (A) CHIPS (60 μg, n=10) was injected i.p., together with OVA i.v. in hC5aR1^KI mice 30 min before inducing the Arthus reaction. Samples were compared to mice that did not receive CHIPS (n=7). Control mice (n=4) received PBS i.v. and i.p. Peritoneal cavity lavage was performed 6 h post Arthus induction. The percentage of neutrophil influx was analyzed by flow cytometry by gating on a CD45⁺GR-1⁺F4/80⁻ population, and depicted as a percentage of total leukocytes (CD45⁺) retrieved after peritoneal lavage. All groups consisted of equal numbers of female and male mice. The median with interquartile range of the combined data from two independent experiments is shown. (B) The presence of anti-OVA and anti-CHIPS antibodies in the rabbit anti-OVA IgG fraction was determined by ELISA. (C) To detect neutralizing anti-CHIPS antibodies in the rabbit anti-OVA IgG, CHIPS (500 ng/ml) was incubated with 10 µg/ml rabbit anti-OVA IgG or PBS. Subsequently, Fluo-4AM-labeled human neutrophils were incubated with CHIPS/Rabbit IgG or CHIPS/PBS and challenged with human C5a. Ca mobilization was determined using flow cytometry and normalized to human neutrophils that did not receive CHIPS. Data are mean±s.d. Significance was calculated using ANOVA, and when needed, followed by Kruskal–Wallis post-test for multiple comparison and displayed as *P<0.05, ****P<0.0001 and NS (not significant).

Fig. 3.

CHIPS and anti-CHIPS antibodies in humans. (A) Frequency distribution of IgG anti-CHIPS titer in healthy human donors (n=168). The titer was defined as the log dilution that gives an absorbance of OD 0.300 after the subtraction of background value. Titers were depicted relative to the mean human pooled serum (HPS) titer (3.75). The anti-CHIPS antibody titer of the six subjects before study entry are depicted in the same graph for comparison. The ▪ represents subjects that had low anti-CHIPS antibodies (anti-CHIPS low), ▴ represents subjects with high anti-CHIPS antibodies (anti-CHIPS high) and the ● represents subjects in the placebo group. Open and closed symbols differentiate between receivers in each group. (B) Pharmacodynamics of CHIPS detected in the sera of the volunteers. CHIPS was measured by a specific capture ELISA at various time points after intravenous injection of CHIPS. (C) CHIPS is recovered on the surface of peripheral blood neutrophils. At various time points after intravenous injection, the presence of CHIPS bound to the surface of neutrophils was detected with rabbit-anti-CHIPS antibodies. Values are expressed as mean fluorescence (MFL) of gated neutrophils in EDTA whole-blood samples. The background MFL value for the secondary FITC-labeled conjugate was 6. (D) Immunogenicity of CHIPS in healthy human subjects. Specific IgG titers towards CHIPS were determined in all subjects before trial start, 7 and 42 days after close of trial and are depicted relative to HPS.

Fig. 4.

CHIPS possibly induces leukocytopenia and increased CRP levels in humans. (A,B) Levels of circulating peripheral WBCs (A) and serum inflammation marker CRP (B). At various time points after intravenous injection of CHIPS, WBC counts and CRP measurements were performed (1.1 and 1.6 indicate 1 day and 1 h or 1 day and 6 h, respectively). The data for WBCs are expressed relative to the value at T=0 and data for CRP are expressed in mg/ml. The ▪ represents subjects that had low anti-CHIPS antibodies (anti-CHIPS low), ▴ represents subjects with high anti-CHIPS antibodies (anti-CHIPS high) and the ● represents subjects in the placebo group. Open and closed symbols differentiate between receivers in each group.

Fig. 5.

Adverse effects of CHIPS as measured by levels of CICs, and mast cell marker tryptase. (A,B) At various time points after intravenous injection of CHIPS, specific assays were performed for CICs (A) and mast cell marker tryptase (B). The ▪ represents subjects that had low anti-CHIPS antibodies (anti-CHIPS low), ▴ represents subjects with high anti-CHIPS antibodies (anti-CHIPS high) and the ● represents subjects in the placebo group. Open and closed symbols differentiate between receivers in each group.