Prolonged indoleamine 2,3-dioxygenase-2 activity and associated cellular stress in post-acute sequelae of SARS-CoV-2 infection

Abstract

Background: Post-acute sequela of SARS-CoV-2 infection (PASC) encompass fatigue, post-exertional malaise and cognitive problems. The abundant expression of the tryptophan-catabolizing enzyme indoleamine 2,3-dioxygenase-2 (IDO2) in fatal/severe COVID-19, led us to determine, in an exploratory observational study, whether IDO2 is expressed and active in PASC, and may correlate with pathophysiology.

Methods: Plasma or serum, and peripheral blood mononuclear cells (PBMC) were obtained from well-characterized PASC patients and SARS-CoV-2-infected individuals without PASC. We assessed tryptophan and its degradation products by UPLC-MS/MS. IDO2 activity, its potential consequences, and the involvement of the aryl hydrocarbon receptor (AHR) in IDO2 expression were determined in PBMC from another PASC cohort by immunohistochemistry (IHC) for IDO2, IDO1, AHR, kynurenine metabolites, autophagy, and apoptosis. These PBMC were also analyzed by metabolomics and for mitochondrial functioning by respirometry. IHC was also performed on autopsy brain material from two PASC patients.

Findings: IDO2 is expressed and active in PBMC from PASC patients, as well as in brain tissue, long after SARS-CoV-2 infection. This is paralleled by autophagy, and in blood cells by reduced mitochondrial functioning, reduced intracellular levels of amino acids and Krebs cycle-related compounds. IDO2 expression and activity is triggered by SARS-CoV-2-infection, but the severity of SARS-CoV-2-induced pathology appears related to the generated specific kynurenine metabolites. Ex vivo, IDO2 expression and autophagy can be halted by an AHR antagonist.

Interpretation: SARS-CoV-2 infection triggers long-lasting IDO2 expression, which can be halted by an AHR antagonist. The specific kynurenine catabolites may relate to SARS-CoV-2-induced symptoms and pathology.

Funding: None.

Keywords: Apoptosis; Autophagy; IDO2; Kynurenine; Mitochondrial_activity; PASC.

Fig. 1

Kynurenine pathway metabolites in plasma from PASC patients, and recovered but prior hospitalized COVID-19 patients. For comparison we included values of plasma from patients with fatal/severe COVID-19 (reference; these separate data have not been published before). Statistical analysis of parametric variables were performed using analysis of variance and post hoc tests with Tukey HSD. Non-parametric data were analysed using the Kruskal–Wallis H test with post hoc pairwise Kruskal–Wallis test with Benjamini-Hochberg correction. The area between the dotted lines represent the normal reference values as determined in reference 22 for plasma and reference 23 for serum. Metabolites in the fatal/severe group had several measurements at the lower limit of quantification for quinolinic acid, 3OH-anthranilic acid, kynurenic acid and xanthurenic acid, which were set at 1/10 of the lower limit of quantification. ns: not significant, ∗: p-value <0.05, ∗∗: p-value <0.001. Recov: Prior hospitalized COVID-19 patients PASC: Post-acute sequelae of COVID-19.

Fig. 2

Immunohistochemical analyses of peripheral blood mononuclear cells (PBMC) from PASC patients and SARS-CoV-2-infected participants fully recovered (Recov). Cytospins were prepared directly after isolation of PBMC. Absent IDO1 stain in PASC and Recov (A and B; a positive control was done separately: not shown). IDO2 stain in PASC and recover (C and D). IDO2-expressing cells are predominantly monocytes, CD4 lymphocytes and B lymphocytes (C, E–H). 3OH-anthranilic acid was shown in PBMC from PASC patients and not in those from healthy individuals (I, J). Nuclear localization of AHR in IDO2-expressing PBMC from PASC patients (K).

Fig. 2

Immunohistochemical analyses of peripheral blood mononuclear cells (PBMC) from PASC patients and SARS-CoV-2-infected participants fully recovered (Recov). Cytospins were prepared directly after isolation of PBMC. Absent IDO1 stain in PASC and Recov (A and B; a positive control was done separately: not shown). IDO2 stain in PASC and recover (C and D). IDO2-expressing cells are predominantly monocytes, CD4 lymphocytes and B lymphocytes (C, E–H). 3OH-anthranilic acid was shown in PBMC from PASC patients and not in those from healthy individuals (I, J). Nuclear localization of AHR in IDO2-expressing PBMC from PASC patients (K).

Fig. 3

Dose-dependent inhibition of IDO2 expression, autophagy and apoptosis by an aryl hydrocarbon receptor (AHR) antagonist. Peripheral blood mononuclear cells were cultured for 8 days in the presence of the AHR antagonist at the indicated concentration or with vehicle (DMSO) only. Immunostains for two PASC patients (A and B) are shown. Cells were evaluated for viability daily by microscopy. Cas3 is activated caspase 3, which is a marker of apoptosis. LC3b is a marker of autophagy.

Fig. 4

Metabolomics data for peripheral blood mononuclear cells from PASC patients compared to that from SARS-CoV-2-infected patients who recovered in full (recover). Volcano plot with colored symbols for significantly altered metabolites (A). Correlation plot for metabolites (B). As these analyses were explorative and with a relatively low number of samples, no correction for multiple testing was performed.

Fig. 4

Metabolomics data for peripheral blood mononuclear cells from PASC patients compared to that from SARS-CoV-2-infected patients who recovered in full (recover). Volcano plot with colored symbols for significantly altered metabolites (A). Correlation plot for metabolites (B). As these analyses were explorative and with a relatively low number of samples, no correction for multiple testing was performed.

Fig. 5

Normalized respirometry data showing the oxygen consumption rate and spare respiratory capacity for peripheral blood mononuclear cells from PASC patients and participants fully recovered from a SARS-CoV-2 infection without hospitalization (Healthy). The blue dashed line represents the mean respiration value of participants who had not yet experienced a SARS-CoV-2 infection (non-COVID-19). All values were normally distributed and analysed with a t-test, which was followed with Benjamini-Hochberg (BH) correction. ∗: p-value<0.05.