An NMR-Based Model to Investigate the Metabolic Phenoreversion of COVID-19 Patients throughout a Longitudinal Study

Abstract

After SARS-CoV-2 infection, the molecular phenoreversion of the immunological response and its associated metabolic dysregulation are required for a full recovery of the patient. This process is patient-dependent due to the manifold possibilities induced by virus severity, its phylogenic evolution and the vaccination status of the population. We have here investigated the natural history of COVID-19 disease at the molecular level, characterizing the metabolic and immunological phenoreversion over time in large cohorts of hospitalized severe patients (n = 886) and non-hospitalized recovered patients that self-reported having passed the disease (n = 513). Non-hospitalized recovered patients do not show any metabolic fingerprint associated with the disease or immune alterations. Acute patients are characterized by the metabolic and lipidomic dysregulation that accompanies the exacerbated immunological response, resulting in a slow recovery time with a maximum probability of around 62 days. As a manifestation of the heterogeneity in the metabolic phenoreversion, age and severity become factors that modulate their normalization time which, in turn, correlates with changes in the atherogenesis-associated chemokine MCP-1. Our results are consistent with a model where the slow metabolic normalization in acute patients results in enhanced atherosclerotic risk, in line with the recent observation of an elevated number of cardiovascular episodes found in post-COVID-19 cohorts.

Keywords: COVID-19; atherosclerotic risk; inflammation; lipidomics; long COVID; metabolomics.

Figure 1

The COVID-19 disease time evolution, showing the approximate duration of the different lineages of the virus and the times when serum samples from the different cohorts were collected. The COVID-19 status and general characteristics of the cohorts are described in Table 1 and Tables S1–S3, respectively.

Figure 2

(A) Score plot from O-PLS-DA model that discriminates between COVID-19 patients (red circles, AC0 and AC1 cohorts) and healthy individuals (green circles, HC cohort). Dashed line corresponds to the value that maximizes Youden’s index. (B) Receiver operating characteristic curve for the O-PLS-DA model shown in Figure 2A. The area under the curve (AUC) is 0.998. In red it is indicated the t_pred value that maximizes the Youden’s index and the resulting specificity and sensitivity between parentheses. (C) Circular bar plot with the 38 metabolic parameters used for the discrimination model. The bars are proportional to the weight in the discrimination. (D) Forest plot with some inflammation markers under consideration. Filled circles represent statistically significant differences (p-value < 0.05). Horizontal lines are standard errors.

Figure 3

(A) Projection of the recovery samples (blue triangles, RE0 and RE1 cohorts) in the O-PLS-DA model that discriminates between COVID-19 patients (red circles, AC0 and AC1 cohorts) and healthy individuals (green circles, HC cohort), as a function of the recovery time for the sample, as indicated in the plots. (B) Projection of the recovery trajectory on the O-PLS-DA model of COVID-19 discrimination for hospitalized patients (blue circles) and non-hospitalized recovered (NHR) (yellow circles). (C) distance to recovery (D) as a function of the days from the infection onset for hospitalized patients (circles) and NHR individuals (triangles). The color and the size of the circles or triangles is proportional to the fraction of recovered people at a given time, as indicated in the legend. (D) Histogram of the individual recovery times for the metabolic phenotype assuming a linear recovery model. The distribution fits well to a GEV function (blue line), whose fitting parameters are enclosed. The green and red lines show the same type of model but using only a subset of patients according to the hospital’s severity criteria: mild-moderate (green) or severe (red). The color code for the histogram bars displays the average age, as indicated in the legend. (E) Boxplots showing the time evolution of serum concentration corresponding to the four cytokines/chemokines that were found significantly upregulated in hospitalized patients. Statistical significance of the difference between the two groups was estimated by the p-value: 0.01 (**) and 0.0001 (****). (F) distance to recovery (D) as a function of the days from the infection onset for hospitalized patients as a function of age: people above/below 65 years old are represented by circles/triangles. The color and the size of the circles or triangles are proportional to the fraction of recovered people at a given time, as indicated in the legend. (G) Correlation between the variation of MCP-1 over time (assuming a linear decay from day 8) and the individual recovery time (assuming linear recovery). The blue line corresponds to the best linear fit, with the confidence levels depicted in gray.

Figure 4

(A) O-PLS-DA score plot representation for the discrimination between the first 115 samples collected during 2020 (115_AC0, red circles) and the last 115 samples collected during 2021 (AC1_115, blue circles). (B) Receiver operating characteristic curve for the O-PLS-DA model shown in Figure 2A. The area under the curve (AUC) is 0.92 In red is indicated the t_pred value that maximizes Youden’s index and the resulting specificity and sensitivity between parentheses. (C) Circular bar plot with the 14 metabolic parameters used for the discrimination model. The bars are proportional to the weight in the discrimination. (D) Triglyceride variation between 115_AC0 and AC1_115.