Methylprednisolone therapy induces differential metabolic trajectories in severe COVID-19 patients

Abstract

The SARS-CoV-2 virus infection in humans induces significant inflammatory and systemic reactions and complications of which corticosteroids like methylprednisolone have been recommended as treatment. Our understanding of the metabolic and metabolomic pathway dysregulations while using intravenous corticosteroids in COVID-19 is limited. This study will help enlighten the metabolic and metabolomic pathway dysregulations underlying high daily doses of intravenous methylprednisolone in COVID-19 patients compared to those receiving placebo. The information on key metabolites and pathways identified in this study together with the crosstalk with the inflammation and biochemistry components may be used, in the future, to leverage the use of methylprednisolone in any future pandemics from the coronavirus family.

Keywords: COVID-19; SARS-CoV-2 virus; corticosteroid; metabolomics; methylprednisolone; profile.

Fig 1

Plasma metabolic signatures in the placebo and MP groups from D1 to D14. (A) Manhattan plot showing the distribution of significant metabolite features after a moderated F-test of metabolites in the patients’ plasma samples. The dashed line indicates P < 0.05. (B) Hierarchical clustering of identified significant metabolite features in the two patient groups. (C) Differential abundance analysis of significant metabolites common in the two treatment groups. (D) Mummichog pathway analysis of significant features per group. A multiple comparison test was used for between-day comparisons. *P < 0.05.

Fig 2

Longitudinal comparison between placebo or MP-treated COVID-19 patients. (A) Differential abundance of the significant metabolite features found from D1 to D14. (B) Venn diagram of differentially abundant metabolites on D5 to D14. The numerical values are indicative of the number of different metabolites detected at different time points. The color code is illustrative of the time points: D5-Orange, D7-Indigo, D11-Green, and D14-Purple. (C) Pathway enrichment analysis of significant metabolite features for each of the sampling time points. (D) Temporal analysis of the relative abundance of the significant metabolite features detected in the two patient groups at D1 to D14. *P < 0.05, **P < 0.01.

Fig 3

Favorable or detrimental metabolic effects of MP treatment. (A and B) Volcano plots comparing two sets of patients at D1 and D5, demonstrating significantly different metabolite features differing of treated survivors and fatalities. The dashed line represents P < 0.05. The blue dots represent downregulated, while the red dots represent upregulated features in each comparison. (C) Mummichog pathway analysis of significant metabolite features on D1 and D5 for survivors and fatalities in the MP and placebo groups. Pathways that were significant for comparisons of survival and death in the same group of treatment were removed. (D) Differential abundance of annotated metabolites identified in treated individuals that survived compared to placebo who died. *P < 0.05, **P < 0.01.

Fig 4

Survival after MP treatment is associated with differences in lipid and vitamin metabolism. (A) Differential abundance of metabolite features between MP-treated patients who died before or survived after 14 days at D1 and D5. (B) Venn diagram comparing significant metabolite features at D1 and D5. (C) Significant metabolic network associated with steroid hormone metabolism enriched by differentially abundant metabolite features at D1. (D) Mummichog pathway analysis of significant metabolite features at D1 and D5 for MP-treated patients who died before or survived after 14 days. (E) Differential abundance of annotated metabolites identified in MP-treated patients that survived compared to those who died before 14 days of treatment. *P < 0.05, **P < 0.01.

Fig 5

Integrative network analysis of placebo and MP-treated COVID-19 patients. (A) Integrative network between laboratory parameters (biochemical and hematological), cytokines, and metabolite clusters on D1. The red lines indicate positive associations and the blue lines represent negative associations. (B) Examples of the most significant correlations found in the integrative network. (C) Cluster enrichment analysis comparing methylprednisolone versus placebo at D1. (D) Cluster enrichment analysis comparing methylprednisolone versus placebo at D5, D7, and D14. (E) Pathway enrichment analysis of significant clusters in items (C) and (D).