Remember me
Research Article
Open Access | 
10.1172/JCI178813
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Xi, G. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Mclnnis, E. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Lardinois, O. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Hu, P. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Poulton, J. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Free, M. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Chen, D. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Zeitler, E. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by
Wu, E.
in:
JCI
|
PubMed
|
Google Scholar
|
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Orzechowski, N. in: JCI | PubMed | Google Scholar
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by
Derebail, V.
in:
JCI
|
PubMed
|
Google Scholar
|
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by
Jennette, J.
in:
JCI
|
PubMed
|
Google Scholar
|
1University of North Carolina Kidney Center, Division of Nephrology and Hypertension, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
2Mass Spectrometry Research and Support Group, National Institute of Environmental Health Sciences, NIH, Research Triangle Park, North Carolina, USA.
3Department of Pathology and Laboratory Medicine,
4Division of Pediatric Rheumatology, Department of Pediatrics,
5Division of Rheumatology, Allergy, and Immunology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Address correspondence to: Ronald J. Falk, 125 MacNider Hall, CB # 7005, Chapel Hill, North Carolina 27599, USA. Phone: 919.966.4468; Email: ronald_falk@med.unc.edu.
Find articles by Falk, R. in: JCI | PubMed | Google Scholar
Published February 28, 2025 - More info
Published in Volume 135, Issue 8 on April 15, 2025
Related article:
Abstract
A wide variety of medications can induce adverse immune events and autoimmune responses such as vasculitis. Mechanistically, small molecule drugs known as haptens bind and modify endogenous proteins, triggering such immune reactions. In this issue of the JCI, Xi and colleagues investigated the immunological mechanism of autoimmune vasculitis associated with hydralazine. Notably, hydralazine-based haptenization modified myeloperoxidase (MPO), inducing the enzyme conformational change. The hydralazine-modified MPO induced IgM antibody specific for the modified enzyme, followed by immune complex precipitation, tissue deposition, and complement activation. These findings provide a mechanism by which hydralazine induces a type III hypersensitivity reaction associated with mild to severe vasculitis. The study serves as an example for understanding haptenation and may inform the development of diagnostics for determining susceptibility to drug-induced allergic or autoimmune responses.
Authors
× AbstractDrug-induced autoimmune diseases are increasingly recognized, although mechanistic insight into disease causation is lacking. Hydralazine exposure has been linked to autoimmune diseases, including antineutrophil cytoplasmic autoantibody (ANCA) vasculitis. Our hypothesis posits that hydralazine covalently binds to myeloperoxidase (MPO), triggering the autoimmune response in ANCA vasculitis. In vitro, we observed formation of carbonyl derivatives on amine groups in the presence of acrolein. This facilitated the subsequent binding of hydralazine to heme-containing proteins, including MPO, via a Michael addition. Our studies demonstrated that carbonyl derivatives and hydrazone adducts induced conformational changes in the MPO heavy chain, potentially changing its immunogenicity. We identified hydrazone adducts on circulating MPO in patients with hydralazine-associated ANCA vasculitis. These patients exhibited elevated anti-MPO IgM levels, while anti-MPO IgG levels were comparable between hydralazine-associated and nonhydralazine-associated vasculitis patients. IgM isolated from patients with hydralazine-associated MPO ANCA demonstrated a heightened affinity to hydralazine-modified MPO and activated neutrophil-like HL-60 cells. Hydralazine-modified MPO was pathogenic, as demonstrated by splenocyte transfer in a mouse model of ANCA vasculitis. Our findings unveil a mechanism of drug-induced autoimmunity wherein stepwise chemical modifications of MPO lead to conformational changes and hydrazone adduct formation, producing a neoantigen that generates pathogenic autoantibodies.
Graphical Abstract
Introduction
Since the initial report of sulfadiazine-associated lupus-like symptoms in 1945 (1), drug-induced autoimmunity has been a well-recognized phenomenon. Among the many implicated drugs, hydralazine is a common offender. Despite recognition of this phenomenon for decades, the underlying mechanisms that lead to drug-induced autoimmunity remain poorly understood. As drug-induced autoimmunity can be viewed as an “experiment of nature,” understanding the mechanisms of disease induction are key to also understanding other idiopathic autoimmune disease.
The etiology of autoimmune disease has been attributed to many factors that include, but are not limited to, genetic and environmental factors (2–5). An extensive array of potential contributors poses a formidable challenge in unraveling the intricacies leading to specific immune dysfunction and disease manifestation. Drug-induced autoimmunity provides a window into how an endogenous protein may become an autoantigen. Many theories have been proposed to explain this phenomenon (6). In hydralazine-induced antineutrophil cytoplasmic antibody (ANCA) vasculitis, one theory suggests that hydralazine accumulates in neutrophils, binds to myeloperoxidase (MPO), and prompts cytotoxic product generation, cell death, and the exposure of typically sequestered antigen (7). However, these hypotheses lack direct evidential support.
ANCA vasculitis is an autoimmune disease characterized by pathogenic autoantibodies against MPO or proteinase 3 (PR3). MPO and PR3 are abundant in neutrophils and monocytes (8). When ANCAs bind to autoantigens on the surface of activated neutrophils, degranulation is stimulated, which ultimately causes severe small vessel damage (9). In a subset of patients with ANCA vasculitis, disease onset has been linked to the use of drugs, such as hydralazine (10). Although hydralazine is a widely prescribed medication for hypertension and heart failure, the occurrence of associated ANCA vasculitis is rare, but well described.
To understand potential mechanisms of drug-induced autoimmune disease, the pathways of known or potential chemical modifications of the drug in question must be elucidated. Hydralazine is a known carbonyl scavenger that reacts with proteins under oxidizing conditions (11, 12). When neutrophils degranulate in response to events such as microbial infections, released MPO produces reactive oxygen species (ROS), causing cytotoxicity, protein oxidation, DNA damage, and lipid peroxidation. Lipid peroxidation yields molecules such as acrolein (13), a highly reactive unsaturated aldehyde. Through Michael addition, acrolein reacts with an amine group containing amino acids, such as lysine, histidine residues, and the N-terminus of proteins. The Michael addition adducts have an aldehyde group that further reacts with nucleophilic groups to generate additional adducts on the protein, which can serve as a neoantigen that triggers an immune response and generation of autoantibodies (14). Among Michael-type addition adducts, Nε-3-formyl-3,4-dehydropiperidino lysine (FDP-lysine) is a major product (15). The mature form of MPO is a disulfide-linked homodimer that contains multiple residues susceptible to react with acrolein (e.g., 12 Lys, 5 His and 1 N-terminus on the heavy chain). It is reasonable to predict that environmentally available acrolein may form carbonyl derivatives with these amino acid residues. Additionally, hydralazine has been shown to react with acrolein-modified and lysine-containing proteins to form hydrazone adducts, which was reported to reduce acrolein toxicity (12).
Our hypothesis posits a sequential process: (a) carbonyl derivatives form on MPO under oxidative conditions (i.e., in the presence of acrolein); (b) hydralazine reacts with these carbonyl derivatives to form hydrazone adducts; and (c) these hydrazone adducts serve as haptens, inducing an autoimmune response through antibody generation. To test this hypothesis, we conducted in vitro studies using commercially prepared myoglobin and MPO to explore chemical reactions, including carbonyl derivatives and hydrazone adduct formation on both proteins. We also explored MPO protein conformational changes after chemical modifications and assessed whether hydrazone adducts could be detected on MPO that was isolated from peripheral blood of patients with hydralazine-associated ANCA vasculitis. We purified IgG and IgM from patients or healthy controls to determine the presence of antibodies specific for hydralazine-modified MPO in individuals with hydralazine-associated ANCA vasculitis. Finally, we observed that antibodies to hydralazine-modified MPO antibodies are pathogenic in an in vivo murine model. Our results suggest that hydralazine modification induces MPO conformational changes and/or hydrazone adduct formation on MPO. This process generates neoantigens or reveals hidden epitopes, which facilitates neoautoantibody generation, thereby leading to the development of ANCA vasculitis.
ResultsCharacteristics of patients involved in the study and people in the healthy control group. Table 1 provides an overview of the participant cohort characteristics in this study. 5 separate clinicians examined medical records to identify patients with hydralazine-associated vasculitis based on predetermined criteria, which were compared with patients with ANCA vasculitis with no hydralazine exposure. The analysis includes 10 patients diagnosed with hydralazine-associated disease, with a range of hydralazine exposure spanning from 2–57 months prior to presentation. Among this cohort, 60% (n = 6) exhibited dual positivity for MPO- and PR3-ANCA, and the other 40% had high-titer MPO-ANCA. Consistent with previous studies (16, 17), all demonstrated high anti-histone antibodies when tested using a commercial histone ELISA (Immuno-Biological Lab). The majority of patients with hydralazine-associated disease had lung and kidney involvement at 60% and 90%, respectively. Conversely, upper respiratory and joint involvement were infrequent (10% and 0%, respectively). Notably, kidney involvement in the hydralazine-associated group was similar to the nonhydralazine-associated patient group, while joint involvement was significantly lower. At the onset of the disease, there were no significant differences between the 2 groups in serum creatinine (2.6 mg/dL versus 2.7 mg/dL) or estimated glomerular filtration rate (eGFR) (23.7 mL/min/1.73m2 versus 25.2 mL/min/1.73m2). Participants diagnosed with hydralazine-associated disease were significantly older than those without hydralazine-associated disease (72.5 versus 58.4, P < 0.006).
Table 1The demographic and clinical characteristics comparisons in groups
Carbonyl derivative is a prerequisite for hydrazone adduct formation. Previous studies have shown that hydralazine suppresses acrolein toxicity (12, 18) by forming hydrazones with Michael adducts generated by acrolein (12). We hypothesized that hydrazone adducts form on MPO in patients with hydralazine-associated ANCA. These hydrazone adducts occur through a series of reactions. One prominent adduct formed upon addition of hydralazine to acrolein-modified proteins is a bis-ACR-lysine hydrazone adduct (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/JCI178813DS1). Carbonyl modification can be assessed by mass spectrometry or by using 2,4-dinitrophenyl hydrazine (2,4-DNPH), a reagent that reacts with carbonyl groups to produce hydrazones.
Our initial proof-of-concept experiments were performed using myoglobin. Myoglobin, like MPO, is a heme-containing protein, but is significantly smaller in size (16.9 KDa versus 150 KDa, respectively). Coupled with the size difference and the absence of chemically bound carbohydrates, initial studies with myoglobin facilitated the identification and characterization of modifications on the globin moiety by mass spectrometry. Representative peptide maps with fluorescence detection at 360 nm (λex = 285 nm) of native myoglobin (the negative control), acrolein-modified myoglobin, and hydralazine plus acrolein-modified myoglobin are presented in Figure 1A. Comparison of the high-performance liquid chromatography (HPLC) traces revealed several peaks from the digest mixtures of the protein treated with acrolein or with hydralazine plus acrolein that were not present in controls (e.g., peaks 2, 3, and 4). The appearance of peaks 2, 3, and 4 was also accompanied by a significant decrease in intensity of peak 1 relative to control (Figure 1A). HPLC fraction corresponding to chromatographic peaks showing marked variation in fluorescence intensity relative to control were further concentrated and characterized by electrospray ionization mass spectrometry (ESI).
Figure 1In vitro investigation of carbonyl derivatives and hydrazone adduct formation on myoglobin and MPO. (A) C18 RP-HPLC separation of peptides produced by tryptic mapping of purified native myoglobin (Mb), acrolein-modified Mb, and hydralazine + acrolein-modified Mb. Chromatograms showed only the time window in which peptides exhibiting a strong fluorescence at 360 nm (λex = 285 nm) eluted. (B) Tandem mass spectra (MS/MS) fragmentation spectrum of a FDP derivatized peptide eluting at approximately 57.6 minutes (corresponding to peak 2 of A). The deconvoluted tandem mass spectra (MS/MS) spectrum was acquired from a parent ion of m/z 955.50 (+2), which corresponded in mass to N-terminal tryptic peptide plus FDP. (C) Tandem mass spectra (MS/MS) fragmentation spectrum of a hydrazone-bis-acrolein derivatized peptide eluting at approximately 59.0 minutes (corresponding to peaks 3 and 4 of A). The deconvoluted tandem mass spectra (MS/MS) spectrum was acquired from a parent ion of m/z 690.0 (+3), corresponding in mass to tryptic peptide plus a hydrazone-bis-acrolein adduct. (D and E) Detection of protein-bound carbonyl derivatives and hydrazone adduct by immunoblotting. Intact/unreacted MPO or acrolein-modified MPO were incubated with DNPH (D) or with hydralazine (E). Protein samples were separated by SDS-PAGE followed by blotting on a nitrocellulose membrane and detection with appropriate antibodies. Each lane contained 2.6 μg of MPO (n = 3).
Mass spectrometry analysis of 2 abundant ions of m/z 605.9 (+3) and 908.5 (+2) in peak 1 and 2 abundant ions of m/z 637.3 (+3) and 955.5 (+2) in peak 2, revealed a 94 dalton (Da) increase in mass of acrolein-modified peptide, compared with ions of unmodified peptide, which were consistent with the addition of 1 FDP group to the peptide. The tandem mass spectra (MS/MS) fragmentation spectrum acquired from the ion of m/z 955.5 (+2) is shown in Figure 1B. The spectra showed an almost complete series of both carboxy-terminal y ions (y2 to y15) and amino-terminal b ions (b2 to b10 and b13). The observed series of b ions showed a 94 Da difference relative to the unmodified N-terminal peptide, whereas the masses of the y ions were similar to those observed in the tandem mass spectrometry spectrum of peptide. The mass spectra of peak 3 and 4 exhibited 2 abundant ions of m/z 690.7 (+3) and 1,035.5 (+2) (data not shown), which corresponded in mass to the addition of 254 Da to the unmodified N-terminal peptide. The 254 Da mass increase was consistent with the addition of 1 bis-acrolein-hydrazone adduct to the peptide (Figure 1C). Altogether, these data indicate that Gly 1 is the site of formation of the hydralazine-bis acrolein adduct. Because the 2 chromatographic peaks (peaks 3 and 4) showed similar mass spectral data, 2 stereoisomers of the hydrazone adduct were likely formed.
In addition to hydralazine, multiple drugs have been implicated in drug-associated ANCA vasculitis (19). To understand the potential mechanism, we tested additional drugs with similar assays, showing that procainamide and aminoguanidine bound to MPO or myoglobin via carbonyl derivatives in a similar fashion to hydralazine (Supplemental Figure 2, A and B), but propylthiouracil (PTU) bound to myoglobin via sulfenic/sulfinic/sulfonic derivatives (Supplemental Figure 2C). Supporting these observations, aminoguanidine was able to competitively attenuate hydralazine binding to MPO (Supplemental Figure 2D) whereas levamisole, like PTU — bound to MPO via different functional groups — was unable to attenuate hydralazine binding to MPO (Supplemental Figure 2E).
After demonstrating the carbonyl derivative and hydrazone adduct formation on myoglobin, we examined whether the same reactions occurred on MPO in vitro. Using 2,4-DNPH, our in vitro data clearly demonstrated that carbonyl derivatives formed on MPO in the presence of acrolein (Figure 1D). Using an anti-hydralazine antibody, our data showed hydrazone adducts on MPO in the presence of acrolein (Figure 1E). The addition of acrolein was necessary for the formation of these adducts, thus a reactive carbonyl is necessary for hydralazine to react with MPO (Figure 1E).
Hydrazone adduct formation on MPO induces conformational changes. Following the identification of adduct formation on MPO, our objective was to explore whether these modifications induced conformational changes in MPO. To answer this question, we utilized antibodies specifically targeting various segments of the MPO heavy chain containing potential targets of acrolein. The native MPO heavy chain was uniformly detected by all test antibodies and exhibited comparable sensitivity (Figure 2A). However, the sensitivity of antibodies was more varied in detecting MPO with carbonyl derivatives. Antibodies with epitopes located at the distal C-terminus of the MPO heavy chain, such as the proteintech antibody (hMPO AA607-612, AA646-651; Figure 2A) and ABclonal antibody (hMPO AA617-622, AA640-645; Figure 2A), displayed reduced affinity to MPO with carbonyl derivatives (Figure 2A) (antibody epitope information is based on published epitope mapping data from our lab) (20). Conversely, antibodies with epitopes located at the distal N-terminus of the heavy chain, such as the Dako antibody (hMPO AA340-400, Figure 2A) and anti-KIV antibody (hMPO AA442-460, Figure 2A), or at the proximal N-terminus, such as the R&D System mouse antibody (hMPO AA295-302, Figure 2A) exhibited similar affinity to MPO with carbonyl derivatives compared with native MPO (Figure 2A). These findings suggest that the formation of carbonyl derivatives induced conformational changes of the C-terminus of the MPO heavy chain, which prevented antibody binding in that region (Figure 2B). Hydrazone adducts formed on MPO produced contrasting results regarding conformational changes (Figure 2A). This suggests that the C-terminus of MPO heavy chain reopened (epitopes accessible), while the distal N-terminus of the MPO heavy chain closed (epitopes inaccessible) (Figure 2B). Importantly, these conformational changes in hydralazine-modified MPO were found to be hydralazine dose dependent (Figure 2A). Intriguingly, the proximal N-terminus of the MPO heavy chain remained consistently unchanged under both conditions (Figure 2A).
Figure 2MPO conformational changes after acrolein or acrolein plus hydralazine modification. (A) The same amount of native MPO (1 μg), acrolein-modified MPO (cMPO), and 1 mM or 10 mM hydralazine-modified myeloperoxidase (HA-MPO) were loaded and separated by SDS PAGE gels. After transfer to nitrocellulose membranes, the membranes were immunoblotted (IB) with anti-MPO antibodies from Proteintech (gel 1), ABclonal (gel 2), Dako (gel 3), Alpha Diagnostic (custom ordered, gel 4) and R&D system (gel 5), respectively (n = 3). The corresponding antibody epitope information was listed under each membrane. Silver staining was performed on a separate membrane (gel 6) for loading control. (B) Based on the IB results, a model of MPO heavy chain conformational changes after carbonyl derivative (1st step) and hydrazone adduct formation (2nd step) were proposed.
Hydrazone adducts on MPO exhibit detectability exclusively in patients with hydralazine-associated ANCA vasculitis. Since all of the above conformational changes were deduced in a controlled in vitro setting, to better elucidate the underlying mechanism of hydralazine-associated ANCA vasculitis, our investigations then focused on the formation of hydrazone adducts on MPO ex vivo. We identified patients with hydralazine-associated ANCA vasculitis, nonhydralazine-associated ANCA vasculitis with high-titer MPO-ANCA serologies, and people who were in a healthy control group. Our hypothesis suggests the presence of hydralazine protein adducts in the circulation. Therefore, we employed immunoprecipitation with an anti-hydralazine antibody to enrich proteins containing hydrazone adducts. Subsequently, we utilized a biotinylated anti-MPO antibody to assess the formation of hydrazone adducts on MPO ex vivo with native MPO serving as a control. Our findings revealed the detection of hydrazone adducts on MPO (heavy chain) exclusively in the patients with hydralazine-associated ANCA vasculitis. In contrast, individuals with no known hydralazine exposure or healthy participants exhibited minimal MPO reactivity, indicating an absence of hydrazone adducts on circulating MPO (Figure 3, A and B). Remarkably, a patient with nonhydralazine-associated ANCA vasculitis, initially misclassified with the hydralazine-associated cohort, did not exhibit MPO reactivity in the same assay (Supplemental Figure 4). This study marks the demonstration of hydrazone formation on MPO ex vivo. To confirm the ex vivo formation of hydrazone adducts on protein, we examined circulating histones since previous studies (5,6) showed most patients with hydralazine-associated ANCA vasculitis were also anti-histone–antibody positive. Consistently, histone protein was detected among the proteins enriched for hydrazone adducts exclusively in patients with hydralazine-associated ANCA vasculitis (Supplemental Figure 3).
Figure 3Hydrazone adducts are detected on circulating MPO only from patients with hydralazine associated ANCA. (A) Plasmas from healthy participants (HC, n = 6 patient samples), patients with nonhydralazine-associated ANCA (ANCA, n = 6 patient samples) and patients with hydralazine-associated ANCA (HA-ANCA, n = 5 patient samples) were immunoprecipitated with an anti-hydralazine antibody. The immunocomplexes were immobilized with protein G and eluted with DTT containing 2× Laemmli sample buffer before loading SDS-PAGE gels. The separated proteins were transferred to nitrocellulose membrane before immunoblotted with biotinylated anti-MPO antibody. A native MPO was loaded for a positive control. (B) The densitometry data were obtained using Image J (NIH, 1.53K version) and the graph was drawn using GraphPad Prism (GraphPad Software, Version 9.5.1). *P < 0.05; **P < 0.01 assessed by Dunn’s multiple comparisons test.
IgM is the primary immunoglobin against hydrazine-modified MPO. Drawing on previous research demonstrating that an acrolein-lysine adduct can constitute an antibody epitope (12), we postulated that hydrazone adducts on MPO could similarly act as a hapten and elicit an autoimmune response in vivo. To test this theory, we measured the levels of anti-MPO IgG and IgM in our patient cohort. Measurement of anti-MPO IgG and IgM in plasma supported this hypothesis, revealing significantly higher levels of anti-MPO IgG in patients with both hydralazine- and nonhydralazine-associated ANCA compared with people who were in a healthy control group (Figure 4A). When examining the IgM subclass, both patient groups had higher anti-MPO IgM levels than healthy participants; however, the median level of anti-MPO IgM in hydralazine-associated patients was significantly higher than patients with nonhydralazine-associated ANCA (0.65 (0.54, 1.03) versus 0.06 (0.06,0.07), P < 0.0001) (Figure 4A). This suggests that the IgM subclass of ANCA may be a key immunoglobulin isoform resulting from hydralazine exposure and formation of neoantigen. Isolation of IgM and IgG from each patient group confirmed this hypothesis. IgM from hydralazine-associated ANCA vasculitis patients exhibited robust reactivity against hydralazine-modified MPO, consistent with ELISA results, indicating IgM as the main antibody isoform that can specifically react with hydralazine-modified MPO (Figure 4B). In contrast, IgG from patients with hydralazine-associated ANCA vasculitis showed similar affinity to both types of MPO (Figure 4C). Notably, no significant difference in IgG binding to native MPO was detected between patients with nonhydralazine-associated ANCA and healthy participants, although a trend of increased binding was detected in patients with ANCA (Figure 4C). In addition, IgM and IgG isolated from patients with nonhydralazine-associated ANCA vasculitis displayed similar affinity with both hydralazine-modified MPO and native MPO (Figure 4, B and C). To confirm these key results, a direct binding assay was performed. Consistent with Western blots results, IgM purified from patients with hydralazine-associated ANCA vasculitis had significantly higher binding to hydralazine-modified MPO than to native MPO. IgM purified from patients with nonhydralazine-associated ANCA had very low binding ability to both types of MPO (Supplemental Figure 5A). In addition, native MPO binding by IgG from patients with nonhydralazine-associated ANCA was significantly higher than that of IgG from people in a healthy control group (Supplemental Figure 5B). We also examined the ability of purified IgM from each group to induce degranulate neutrophils. Neutrophil-like HL-60 cells were treated with IgM purified from each group of patients and healthy controls. The results clearly showed that IgM isolated from patients with hydralazine-associated ANCA had the strongest ability to stimulate ROS generation while IgM purified from healthy participants and patients with nonhydralazine associated ANCA had no difference in their effect on ROS generation compared with controls (Figure 5). All these data suggest that IgM is the primary type of immunoglobin against hydralazine-modified MPO and may be pathogenic.
Figure 4IgM is a primary subtype of immunoglobins that is generated against hydralazine-modified MPO ex vivo. (A) Plasma anti-myeloperoxidase IgG and IgM from patients with hydralazine-associated ANCA (n = 10), patients with nonhydralazine-associated ANCA (n = 25), and participants in a healthy control group (n = 19) were measured following the protocol described in the Methods section. The multiple comparisons of IgM or IgG for any 2 groups of patients’ P values (*P < 0.05; ***P < 0.001; ****P < 0.0001) were calculated by Dunn’s multiple comparisons test. (B and C) IgG and IgM were purified from plasma obtained from healthy participants (HC, n = 5), patients with nonhydralazine-associated ANCA (ANCA, n = 4), and patients with hydralazine-associated ANCA (HA-ANCA, n = 6). The same amount of native MPO and hydralazine-modified MPO were separated using SDS-PAGE gels and transferred to nitrocellulose membranes. The membranes were immunoblotted with IgM (B) and IgG (C). Densitometry values were obtained using Image J (NIH, 1.53K version) and graph was drawn using GraphPad Prism (GraphPad Software, Version 9.5.1). *P < 0.05 assessed by Mann-Whitney test for comparison between 2 groups.
Figure 5IgM purified from patients with hydralazine-associated ANCA significantly induces oxidative burst in neutrophil-like HL-60 cells. Differentiated (neutrophils-like) HL-60 cells (1 × 105) were seeded to each well of a 96 well plate. Cytochalasin B (2 μM) was used to prime the cells before adding each treatment, including PMA (100 ng/mL), IgM (0.75 μg/well) isolated from individuals who were healthy controls (HC, n = 6), patients with nonhydralazine-associated ANCA (ANCA, n = 9),or patietns with hydralazine-associated ANCA (HA-ANCA, n = 8). Reactive oxygen species (ROS) was detected by DHE (2 μM) and the plate was read by a TECAN with EX 510 nm/EM 610 nm. Since the peak value for PMA and IgM treatment appeared at separate time points, control 1 (Ctrl 1) was control value for PMA treatment and control 2 (Ctrl 2) was control value for IgM treatment. Graphs were drawn using GraphPad Prism (GraphPad Software, Version 10.1.0) **P < 0.01; ****P < 0.0001 assessed by 1-way ANOVA multiple comparisons test.
Hydralazine-modified MPO induces glomerulonephritis in anti-MPO splenocyte-recipient mice. Based on our in vitro and ex vivo data demonstrating the association of hydralazine-modified MPO with development of ANCA-vasculitis, we sought direct evidence that hydralazine-modified MPO is pathogenic. Rag2−/− mice were used to test the pathogenicity of antibodies produced by transplantation of splenocytes from MPO–/– mice immunized with hydralazine-modified recombinant mouse MPO (HA-rmMPO) or unmodified rmMPO. Four days after the second boost with HA-rmMPO, splenocytes isolated from HA-rmMPO–immunized MPO–/– mice were transferred to Rag2–/––recipient mice. After 2 weeks, the mice were sacrificed, and kidney sections were examined histologically for glomerular necrosis and crescent formation (Figure 6A). As positive and negative controls, unmodified rmMPO was also used to immunize MPO–/– mice and WT mice. H&E staining of kidney sections clearly showed glomerulonephritis from splenocytes derived from rmMPO immunized MPO–/– mice (Figure 6B) but not from splenocytes derived from immunized WT mice (Figure 6C). Importantly, splenocytes from HA-rmMPO–immunized MPO–/– mice (Figure 6D) but not WT mice (Figure 6E) also induced glomerulonephritis. Glomerular crescents were quantified for each group (Figure 6F) with a similar pattern of glomerular necrosis (Figure 6G). In addition, like unmodified MPO, mice immunized with HA-rmMPO generated antibody (IgG but not IgM) only in MPO–/– mice but not in WT mice (Supplemental Fig
Comments (0)