Chic-FILI · Definition of the chemical and immunological characteristics of flucloxacillin-induced liver injury

Definition of the chemical and immunological characteristics of flucloxacillin-induced liver injury

Drug-induced liver injury (DILI) is a major concern for public health as it is a severe and common form of adverse drug reaction. Furthermore, reactions have a socio-economic impact as they are a leading cause of drug withdrawal. The effective combination of amoxicillin and clavulanic acid (co-amoxiclav) provides broad-spectrum anti-microbial activity, which makes it one of the most frequently prescribed antibiotics throughout the world. Most individuals safely tolerate the drug; however, a small portion patients report severe and occasionally life-threatening liver injury. Several genetic studies have shown that individual susceptibility to co-amoxiclav induced liver injury is linked to a specific HLA haplotype, HLA-DRB1*1501-DQB1*0602 (Hautekeete ML et al. 1999 Gastroenterology 117: 1181–1186, Lucena MI et al. 2011 Gastroenterology 141: 338–347). Although co-amoxiclav induced liver injury is a well-recognized complication both in clinics and pharmacogenomics, the underlying immune mechanism is still poorly understood. Here, we examined drug-specific lymphocyte proliferation from patients with co-amoxiclav induced liver injury and investigated the HLA-restricted T-cell immune response to the drug combination using T-cell clones. Peripheral blood mononuclear cells from 7 patients were cultured in vitro with amoxicillin, clavulanic acid, or co-amoxiclav. All patients showed a strong lymphocyte proliferative response to amoxicillin. Increased interferon γ (IFN-γ) secretion was also found in the presence of amoxicillin using ELISpot as a readout for the antigen-specific T-cell response. Interestingly, there was no proliferative response to clavulanic acid. However, IFN-γ secretion was enhanced. Optimal T-cell stimulation was observed for both drugs in the range of 0.5mM to 2mM. T-cell clones against amoxicillin and clavulanic acid were generated from 3 patients by limiting dilution. Amoxicillin-specific T-cell clones showed a strong proliferative response and IFN-γ secretion to amoxicillin and in a dose dependent manner. Clavulanic acid-specific T-cell clones proliferated weakly in the presence of the drug, but high levels of IFN-γ secretion was observed (in the presence of clavulanic acid and co-amoxiclav). Importantly, amoxicillin and clavulanic acid-specific T-cell clones were not activated in the presence of the alternative drug-antigen (i.e., amoxicillin-responsive T-cell clones specifically recognized amoxicillin, but not clavulanate, and vice versa). Phenotype analysis using flow cytometry demonstrated that all T-cell clones had CD4+ /CD45RO+. HLA-restricted T-cell proliferation was examined using amoxicillin and clavulanic acid-specific T-cell clones and anti-MHC blocking antibodies. The proliferative response of the T-cell clones was strongly inhibited by an anti-HLA (anti-HLA-DR) class II antibody. IFN-γ secretion was also strongly inhibited in the presence of the anti-HLA-DR antibody. The amoxicillin-responsive clones were then used to define the HLA-DR allele(s) that the drug interacts with to stimulate a T-cell response. Amoxicillin-responsive clones were stimulated to proliferate with allogenic antigen-presenting cells expressing HLA-DRB1*1501. In contrast, antigen-presenting cells expressing HLA-DRB1*0701 didn’t stimulate the clones. IFN-γ secretion from amoxicillin-specific T-cell clones showed the same HLA-restriction. These data demonstrate that the drug-specific T-cell response is dependent on the drug-derived antigen interacting with specific MHC class II molecules. To study the priming of naïve T-cells from drug-naïve donors, we utilized our recently established dendritic cell/T-cell co-culture system (Faulkner et al., 2012 Toxicol Sci. 127:150-8). T-cells from a HLA-DRB1*1501 positive donor were effectively primed against amoxicillin when dendritic cells were used to present the drug-derived antigen. Following amoxicillin restimulation, the primed T-cells were found to proliferate and secrete cytokines such as IFN-γ and IL-13. To examine cross-reactivity of the amoxicillin-specific T-cells, clones were cultured with autologous APC and alternative ß–lactam antibiotics such as piperacillin, penicillin G, and flucloxacillin which modify similar lysine residues on albumin. These experiments revealed that the T-cell response was highly drug specific; clones were activated with amoxicillin, but not the other structurally-related drugs. Mass spectrometric techniques were used to define the chemistry of drug–protein conjugation using human serum albumin as a model target. Amoxicillin and clavulanic acid were found to bind in a time- and concentration-dependent fashion to specific lysine residues. Collectively, these findings suggest that a major susceptibility factor relates to the restriction of the fit of the antigen into particular immunological receptors in co-amoxiclav induced liver injury and the HLA-DRB1*1501 risk genotype relates the drug-specific immune response. Based on this study, we have designed a prospective clinical study to evaluate the HLA genotype on pharmacokinetics/pharmacodynamics after multiple administration of co-amoxiclav and explore representative biomarkers for drug-induced liver injury in Korean normal healthy controls. Socio-economic impact: There are no available methods to diagnose immune-mediated drug reactions in patients with drug-induced liver injury. This project indicates that in vitro assays characterizing drug-antigen-specific T-cells can be used to diagnose immunological drug-induced liver injury. This experimental approach can be applied to drug screening for candidate drugs and assist patient management by warning usage in susceptible patients.

Data: CORDIS, © European Union