Search

Flucloxacillin is a β-lactam antibiotic associated with a high incidence of drug-induced liver reactions. Although expression of HLA-B*57:01 increases susceptibility, little is known about the pathological mechanisms involved in the induction of the clinical phenotype. Irreversible protein modification is suspected to drive the reaction through the presentation of flucloxacillin-modified peptides by the risk allele. In this study, the binding of flucloxacillin to proteins of liver-like cells was characterized. Flucloxacillin was shown to bind to proteins localized in bile canaliculi regions, coinciding with the site of clinical disease. The localization of flucloxacillin was mediated primarily by the membrane transporter multidrug resistance-associated protein 2. Modification of multiple proteins by flucloxacillin in bile canaliculi regions may provide a potential local source of neo-antigens for HLA presentation in the liver.

Idiosyncratic drug-induced liver injury (iDILI), which is rare and often recognized only late in drug development, poses a major public health concern and impediment to drug development due to its high rate of morbidity and mortality. The mechanisms of DILI are not completely understood; both non-immune- and immune-mediated mechanisms have been proposed. Non-immune-mediated mechanisms including direct damage to hepatocytes, mitochondrial toxicity, interference with transporters, and alteration of bile ducts are well-known to be associated with drugs such as acetaminophen and diclofenac; whereas immune-mediated mechanisms involving activation of both adaptive and innate immune cells and the interactions of these cells with parenchymal cells have been proposed. The chemical signals involved in activation of both innate and adaptive immune responses are discussed with respect to recent scientific advances. In addition, the immunological signals including cytokine and chemokines that are involved in promoting liver injury are also reviewed. Finally, we discuss how liver tolerance and regeneration can have profound impact on the pathogenesis of iDILI. Continuous research in developing in vitro systems incorporating immune cells with liver cells and animal models with impaired liver tolerance will provide an opportunity for improved prediction and prevention of immune-mediated iDILI.

Hypersensitivity reactions occur frequently in patients upon treatment with sulfamethoxazole (SMX). These adverse effects have been attributed to nitroso sulfamethoxazole (SMX-NO), the reactive product formed from auto-oxidation of the metabolite SMX hydroxylamine. The ability of SMX-NO to prime naïve T-cells in vitro and also activate T-cells derived from hypersensitive patients has illustrated that T-cell activation may occur through the binding of SMX-NO to proteins or through the direct modification of MHC-bound peptides. SMX-NO has been shown to modify cysteine residues in glutathione, designer peptides, and proteins in vitro; however, the presence of these adducts have not yet been characterized in vivo. In this study a parallel in vitro and in vivo analysis of SMX-NO adducts was conducted using mass spectrometry. In addition to the known cysteine adducts, multiple SMX-NO-derived haptenic structures were found on lysine and tyrosine residues of human serum albumin (HSA) in vitro. On lysine residues two haptenic structures were identified including an arylazoalkane adduct and a Schiff base adduct. Interestingly, these adducts are labile to heat and susceptible to hydrolysis as shown by the presence of allysine. Furthermore, SMX-modified HSA adducts were detected in patients on long-term SMX therapy illustrated by the presence of an arylazoalkane adduct derived from a proposed carboxylic acid metabolite of SMX-NO. The presence of these adducts could provide an explanation for the immunogenicity of SMX and the strong responses to SMX-NO observed in T-cell culture assays. Also, the degradation of these adducts to allysine could lead to a stress-related innate immune response required for T-cell activation.

The HLA class I allele HLA-A*33:03 is a risk factor for ticlopidine-induced liver injury. Herein, we show HLA class I-restricted ticlopidine-specific CD8+ T-cell activation in healthy donors expressing HLA-A*33:03. Cloned CD8+ T-cells proliferated and secreted IFN-γ in the presence of ticlopidine and autologous antigen presenting cells. A reduction of the T-cell response after blocking with HLA-class I and HLA-A*33 antibodies indicates that the interaction between drugs and the HLA allele detected in genetic association studies may be important for T-cell activation.

Amoxicillin-clavulanate (AC) is one of the most common causes of drug induced liver injury (DILI). The association between AC-DILI and HLA alleles and the detection of drug-specific T cells in patients with AC-DILI indicate that the adaptive immune system is involved in the disease pathogenesis. In this study, mass spectrometric methods were employed to characterize the antigen formed by AC in exposed patients and the antigenic determinants that stimulate T cells. Amoxicillin formed penicilloyl adducts with lysine residues on human serum albumin (HSA) in vitro, with K190 and K199 being the most reactive sites. Amoxicillin-modified K190 and K199 have also been detected in all patients, and more extensive modification was observed in patients exposed to higher doses of amoxicillin. In contrast, the binding of clavulanic acid to HSA was more complicated. Multiple adducts were identified at high concentrations in vitro, including those formed by direct binding of clavulanic acid to lysine residues, novel pyrazine adducts derived from binding to the degradation products of clavulanic acid, and a cross-linking adduct. Stable adducts derived from formylacetic acid were detected in all patients exposed to the drug. Importantly, analysis of hapten-protein adducts formed in the cell culture medium revealed that the highly drug-specific T-cell responses were likely driven by the markedly different haptenic structures formed by these two drugs. In this study, the unique haptenic structures on albumin in patients formed by amoxicillin and clavulanic acid have been characterized and shown to function as chemically distinct antigens which can stimulate separate, specific T-cell clones.

The risk of developing hypersensitivity to alternative antibiotics is a concern for penicillin hypersensitive patients and healthcare providers. Herein we use piperacillin hypersensitivity as a model to explore the reactivity of drug-specific IgG against alternative β-lactam protein adducts. Mass spectrometry was used to show the drugs (amoxicillin, flucloxacillin, benzyl penicillin, aztreonam, and piperacillin) bind to similar lysine residues on the protein carrier bovine serum albumin. However, the hapten-specific IgG antibodies found in piperacillin hypersensitive patient plasma did not bind to other β-lactam protein conjugates. These data outline the fine specificity of piperacillin-specific IgG antibodies that circulate in patients with hypersensitivity.

It is hypothesized that lapatinib-induced liver injury is caused by HLA-mediated antigen presentation to CD4 positive T cells. However, analysis of PBMC and cloned T-cells from patients with HLA-DRB1*07:01-restricted lapatinib-induced liver injury revealed no evidence for drug-specific activation. T cells were exposed to lapatinib, the M11 aldehyde, and quinone imine [oxidized form of hydroquinone amine M1] metabolites. Reactivity of the quinone imine was confirmed by mass spectrometry.

Antituberculosis (TB) drug exposure is associated with a mild elevation of liver enzymes that occasionally develops into severe liver injury. Herein, we identify ethambutol- and rifampicin-specific CD4+ and CD8+ T-cell clones in a patient with fatal anti TB-related liver injury. The clones were activated to proliferate and secrete IFN-γ, Il-13, and granzyme B following drug treatment. Drug-responsive T-cells may contribute to the pathogenesis of antituberculosis-related liver failure.

Drug hypersensitivity remains a major concern, as it causes high morbidity and mortality. Understanding the mechanistic basis of drug hypersensitivity is complicated by the multiple risk factors implicated. This study utilized sulfamethoxazole (SMX) as a model drug to (1) relate SMX metabolism in antigen presenting cells (APCs) to the activation of T-cells and (2) characterize covalent adducts of SMX and myeloperoxidase, which might represent antigenic determinants for T-cells. The SMX metabolite nitroso-SMX (SMX-NO) was found to bind irreversibly to APCs. Time- and concentration-dependent drug-protein adducts were also detected when APCs were cultured with SMX. Metabolic activation of SMX was significantly reduced by the oxygenase/peroxidase inhibitor methimazole. Similarly, SMX-NO-specific T-cells were activated by APCs pulsed with SMX, and the response was inhibited by pretreatment with methimazole or glutaraldehyde, which blocks antigen processing. Western blotting, real-time polymerase chain reaction (RT-PCR), and mass spectrometry analyses suggested the presence of low concentrations of myeloperoxidase in APCs. RT-PCR revealed mRNA expression for flavin-containing monooxygenases (FMO1-5), thyroid peroxidase, and lactoperoxidase, but the corresponding proteins were not detected. Mass spectrometric characterization of SMX-NO-modified myeloperoxidase revealed the formation of N-hydroxysulfinamide adducts on Cys309 and Cys398. These data show that SMX's metabolism in APCs generates antigenic determinants for T-cells. Peptides derived from SMX-NO-modified myeloperoxidase may represent one form of functional antigen.

Abacavir (ABC), a nucleoside-analogue reverse transcriptase inhibitor, is associated with severe hypersensitivity reactions that are thought to involve the activation of CD8+ T cells in a HLA-B*57:01-restricted manner. Recent studies have claimed that noncovalent interactions of ABC with HLA-B*57:01 are responsible for the immunological reactions associated with ABC. However, the formation of hemoglobin-ABC aldehyde (ABCA) adducts in patients exposed to ABC suggests that protein conjugation might represent a pathway for antigen formation. To further characterize protein conjugation reactions, we used mass spectrometric methods to define ABCA modifications in patients receiving ABC therapy. ABCA formed a novel intramolecular cross-linking adduct on human serum albumin (HSA) in patients and in vitro via Michael addition, followed by nucleophilic adduction of the aldehyde with a neighboring protein nucleophile. Adducts were detected on Lys159, Lys190, His146, and Cys34 residues in the subdomain IB of HSA. Only a cysteine adduct and a putative cross-linking adduct were detected on glutathione S-transferase Pi (GSTP). These findings reveal that ABC forms novel types of antigens in all patients taking the drug. It is therefore vital that the immunological consequences of such pathways of haptenation are explored in the in vitro models that have been used by various groups to define new mechanisms of drug hypersensitivity exemplified by ABC.

β-Lactam antibiotics provide the cornerstone of treatment for respiratory exacerbations in patients with cystic fibrosis. Unfortunately, approximately 20% of patients develop multiple nonimmediate allergic reactions that restrict therapeutic options. The purpose of this study was to explore the chemical and immunological basis of multiple β-lactam allergy through the analysis of human serum albumin (HSA) covalent binding profiles and T-cell responses against 3 commonly prescribed drugs; piperacillin, meropenem, and aztreonam. The chemical structures of the drug haptens were defined by mass spectrometry. Peripheral blood mononuclear cells (PBMC) were isolated from 4 patients with multiple allergic reactions and cultured with piperacillin, meropenem, and aztreonam. PBMC responses were characterized using the lymphocyte transformation test and IFN-γ /IL-13 ELIspot. T-cell clones were generated from drug-stimulated T-cell lines and characterized in terms of phenotype, function, and cross-reactivity. Piperacillin, meropenem, and aztreonam formed complex and structurally distinct haptenic structures with lysine residues on HSA. Each drug modified Lys190 and at least 6 additional lysine residues in a time- and concentration-dependent manner. PBMC proliferative responses and cytokine release were detected with cells from the allergic patients, but not tolerant controls, following exposure to the drugs. 122 CD4+, CD8+, or CD4+CD8+ T-cell clones isolated from the allergic patients were found to proliferate and release cytokines following stimulation with piperacillin, meropenem, or aztreonam. Cross-reactivity with the different drugs was not observed. In conclusion, our data show that piperacillin-, meropenem-, and aztreonam-specific T-cell responses are readily detectable in allergic patients with cystic fibrosis, which indicates that multiple β-lactam allergies are instigated through priming of naive T-cells against the different drug antigens. Characterization of complex haptenic structures on distinct HSA lysine residues provides a chemical basis for the drug-specific T-cell response.

The non-nucleoside reverse transcriptase inhibitor nevirapine (NVP) is widely used for the treatment of human immunodeficiency virus type 1 (HIV-1), particularly in developing countries. Despite its therapeutic benefits, NVP has been associated with skin and liver injury in exposed patients. Although the mechanism of the tissue injury is not yet clear, it has been suggested that reactive metabolites of NVP may be involved. The detection of NVP mercapturate in the urine of patients undergoing standard antiretroviral chemotherapy indicates that NVP undergoes bioactivation in vivo. However, covalent binding of drug to protein in patients remains to be determined. In this study, we investigate the chemical basis of NVP protein adduct formation by using human serum albumin (HSA) and glutathione S-transferase pi (GSTP) as model proteins in vitro. In addition, HSA was isolated from serum samples of HIV-1 patients undergoing NVP therapy to measure NVP haptenation. Mass spectrometric analysis of 12-sulfoxyl-NVP-treated HSA revealed that the drug bound selectively to histidine (His146, His242, and His338) and a cysteine residue (Cys34). The reaction proceeds most likely by a concerted elimination-addition mechanism. This pathway was further confirmed by the observation of NVP-modified Cys47 in GSTP. Importantly, the same adduct (His146) was detected in HSA isolated from the blood of patients receiving NVP, providing direct evidence that NVP modifies protein in vivo, via the formation of a reactive metabolite.