Metabolomics reveals the metabolic map of procainamide in humans and mice

Abstract

Procainamide, a type I antiarrhythmic agent, is used to treat a variety of atrial and ventricular dysrhythmias. It was reported that long-term therapy with procainamide may cause lupus erythematosus in 25-30% of patients. Interestingly, procainamide does not induce lupus erythematosus in mouse models. To explore the differences in this side-effect of procainamide between humans and mouse models, metabolomic analysis using ultra-performance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight mass spectrometry (UPLC-ESI-QTOFMS) was conducted on urine samples from procainamide-treated humans, CYP2D6-humanized mice, and wild-type mice. Thirteen urinary procainamide metabolites, including nine novel metabolites, derived from P450-dependent, FMO-dependent oxidations and acylation reactions, were identified and structurally elucidated. In vivo metabolism of procainamide in CYP2D6-humanized mice as well as in vitro incubations with microsomes and recombinant P450s suggested that human CYP2D6 plays a major role in procainamide metabolism. Significant differences in N-acylation and N-oxidation of the drug between humans and mice largely account for the interspecies differences in procainamide metabolism. Significant levels of the novel N-oxide metabolites produced by FMO1 and FMO3 in humans might be associated with the development of procainamide-induced systemic lupus erythematosus. Observations based on this metabolomic study offer clues to understanding procainamide-induced lupus in humans and the effect of P450s and FMOs on procainamide N-oxidation.

Fig. 1

PCA and PLS-DA analysis of the procainamide treatment and pretreatment groups in human. (A) Scores plot of a PCA model between procainamide treatment and pretreatment in the low, normal, and high salt groups. Each point represents an individual human urine sample. (B) Scores plot of a PLS-DA model between procainamide treatment and pretreatment in the low, normal, and high salt groups. Each point represents an individual human urine sample. The t[1] and t[2] correspond to principal components 1 and 2, respectively. Closed triangles represent procainamide treated-samples; open triangles represent pretreatment samples.

Fig. 2

Identification of procainamide metabolites in human and mouse urine using UPLC-ESI-QTOFMS-based metabolomics. (A) Scores plot of an OPLS model and OPLS loadings S-plot of urinary ions from pretreatment and procainamide-treated human innormal salt group. Each point represents an individual human urine sample (top) and a urinary ion (bottom). Procainamide and its metabolites are labeled in the S-plot (I, II, III, IV, V, VII, VIII, IX, and XII), and “f” indicates a fragment ion. (B) Scores plot of an OPLS model and OPLS loadings S-plot of urine ions from control and procainamide-treated mice. Each point represents an individual mouse urine sample (top) and a urine ion (bottom). Procainamide and its metabolites are labeled in the S-plot (I, II, III, IV, V, VI, and VIII), and f indicates a fragment ion. The t[1]P and t[2]O correspond to principal components 1 and 2, respectively. The p(corr)[1]P values represent the interclass difference and p[1]P values represent the relative abundance of the ions. All the data were obtained in positive mode (ESI⁺).

Fig. 3

Tandem MS and chemical structures of metabolites V, VIII, IX, and XI. (A) Procainamide N-oxide (V). (B) N-Acetylprocainamide N-oxide (VIII). (C) N-Formylprocainamide (IX). (D) N-Propionylprocainamide (XI).

Fig. 4

Chemical reduction of procainamide N-oxide (V) and N-acetylprocainamide N-oxide (VIII). (A) Procainamide N-oxide in human urine. (B) N-Acetylprocainamide N-oxide in human urine. (C) Procainamide N-oxide was lost from human urine following treatment with TiCl₃. (D) N-Acetylprocainamide N-oxide was lost from human urine following the treatment with TiCl₃.

Fig. 5

Representative chromatograms of major procainamide metabolites in urine samples of humans, CYP2D6-humanized mice (hCYP2D6), and wild-type mice. Ions within the 20 ppm range of theoretical accurate mass ([M + H]⁺) of procainamide metabolites (278.189, 250.156, 208.145, 252.169, 294.183, 264.170, 430.223, 292.203, 308.159, and 306.219) were extracted from each 10 min LC–MS run.

Fig. 6

Relative intensity of procainamide and its metabolites produced in vivo and in vitro. (A) Relative intensity of procainamide and its metabolites in humans, CYP2D6-humanized mice, and wild-type mice urine. (B) Relative intensity of procainamide and its metabolites in HLM and MLM from wild-type and CYP2D6-humanized mice. The metabolite codes correspond to those in Table 1. *, significant differences (P < 0.05) in human or CYP2D6-humanized mice (hCYP2D6) and wild-type mice; **, significant differences (P < 0.01) in human and wild-type mice.

Fig. 7

Relative N-oxidation activities of recombinant human P450 and FMO isozymes, and enzyme kinetics of CYP2D6 and FMO3. (A) Relative activities of recombinant human P450s for the generation of procainamide hydroxylamine (VI). The enzymatic activity of the P450 form with the highest yield of VI was set as 100%. The relative activity was presented as the mean ± SD (n = 3). (B) Kinetics of formation of procainamide hydroxylamine (VI) production by CYP2D6. The abundance of VI in the incubation with 200 μM procainamide was set as 100%. (C) Relative activities of recombinant human FMOs for the generation of procainamide N-oxide (V). The enzymatic activity of the FMO with the highest yield of V was set as 100%. (D) Kinetics of procainamide N-oxide (V) production by FMO3. The abundance of V in the incubation with 200 μM procainamide was set as 100%.

Fig. 8

Summary of differential metabolism of procainamide in humans and mice. The acylation of procainamide in humans is responsible for the protective metabolism, consistent with procainamide acylation in mouse. However, more acylated metabolites of procainamide can be formed by acyltransferase in humans than in mice. The differential metabolism of procainamide is N-oxidation and N-hydroxylation associated with the side-effects of procainamide in vivo. In humans, N-oxidation mediated by FMO1 and FMO3 is the major route of metabolism of procainamide, while N-hydroxylation regulated by CYP2D6 and CYP1A1 is a minor metabolic pathway. In contrast, N-hydroxylation is the major metabolic route of procainamide in mice, while N-oxidation is a minor pathway in this species.

Fig. 9

Major in vivo procainamide metabolic pathways. The primary amine in procainamide (I) can conjugate with different types of acyl group to generate several N-acylation metabolites (II, XI, XII, XIII, and XIV). The metabolites II and XII can be further metabolized to VIII and IX, respectively. Procainamide (I) can also be converted to N-hydroxylation metabolite VI and N-oxide V under the effect of oxidation. VI can be transformed into VII by esterification. In addition to desethylation of I to IV, procainamide also can be conjugated with a hexuronic acid to X.