Clinical and pulmonary function analysis in long-COVID revealed that long-term pulmonary dysfunction is associated with vascular inflammation pathways and metabolic syndrome

Abstract

Introduction: Long-term pulmonary dysfunction (L-TPD) is one of the most critical manifestations of long-COVID. This lung affection has been associated with disease severity during the acute phase and the presence of previous comorbidities, however, the clinical manifestations, the concomitant consequences and the molecular pathways supporting this clinical condition remain unknown. The aim of this study was to identify and characterize L-TPD in patients with long-COVID and elucidate the main pathways and long-term consequences attributed to this condition by analyzing clinical parameters and functional tests supported by machine learning and serum proteome profiling.

Methods: Patients with L-TPD were classified according to the results of their computer-tomography (CT) scan and diffusing capacity of the lungs for carbon monoxide adjusted for hemoglobin (DLCOc) tests at 4 and 12-months post-infection.

Results: Regarding the acute phase, our data showed that L-TPD was favored in elderly patients with hypertension or insulin resistance, supported by pathways associated with vascular inflammation and chemotaxis of phagocytes, according to computer proteomics. Then, at 4-months post-infection, clinical and functional tests revealed that L-TPD patients exhibited a restrictive lung condition, impaired aerobic capacity and reduced muscular strength. At this time point, high circulating levels of platelets and CXCL9, and an inhibited FCgamma-receptor-mediated-phagocytosis due to reduced FcγRIII (CD16) expression in CD14+ monocytes was observed in patients with L-TPD. Finally, 1-year post infection, patients with L-TPD worsened metabolic syndrome and augmented body mass index in comparison with other patient groups.

Discussion: Overall, our data demonstrated that CT scan and DLCOc identified patients with L-TPD after COVID-19. This condition was associated with vascular inflammation and impair phagocytosis of virus-antibody immune complexes by reduced FcγRIII expression. In addition, we conclude that COVID-19 survivors required a personalized follow-up and adequate intervention to reduce long-term sequelae and the appearance of further metabolic diseases.

Keywords: COVID-19; chemokines; metabolic syndrome; pulmonary dysfunction; sequelae; vascular inflammation.

FIGURE 1

Study design flowchart. (A) A total of 89 patients with confirmed diagnosis of COVID-19 were invited to participate in the study, from which 29 were not included, resulting in a study cohort of 60 patients with different severity degree. Clinical and demographic data during acute phase and 4-months after COVID-19 was collected. (B) A computer tomography (CT) scan and diffusing capacity of the lungs for carbon monoxide (DLCO) exam were performed 4-months after acute COVID-19 defining abnormal CT scan total severity score (TSS) >1 and abnormal DLCO exam DLCOc <80%. Ordinary one-way ANOVA tests; ****p < 0.0001, ***p < 0.005, **p < 0.01, *p < 0.05. (C) The DLCO exam was reevaluated 12-months after acute infection in patients with abnormal CT scan and abnormal DLCO 4-months post infection. Before–after symbols and lines graph comparing the percentages of DLCOc in patients with L-TLD at 4- and 12-months post-infection; paired t-test ***p = 0.0003. Pie graph showing the percentage of patients with DLCOc <80%, DLCOc = 80%, DLCOc >80% and a missing value without follow-up due to pregnancy. (D) Sankey diagrams representing networks between COVID-19 severity during the acute phase and the level of pulmonary sequelae 4-month after infection according to the CT and DLCO exam.

FIGURE 2

Reduced aerobic capacity and handgrip strength in L-TPD. (A) Shapley Additive exPlanations (SHAP) graph showing the contribution of functional features in the definition of lung sequelae according to a SHAP-value assigned by the algorithm. (B) Scatter plots of spirometry tests forced vital capacity (FVC), forced expiratory volume (FEV₁), and FEV₁/FVC ratio were compared between patient groups pre and post treatment with bronchodilator salbutamol. (C) Scatter plots of physical and mental 12-item short form survey scores between patient groups. (D) Scatter plots of distance walked in 6 min and hand grip test between patient groups. For the 6MWT, (E) oxygen saturation and (F) fatigue scores were measured before and after the test and compared between patient groups. For panels (B–F), ordinary one-way ANOVA tests; ****p < 0.0001, ***p < 0.005, **p < 0.01, *p < 0.05.

FIGURE 3

Inflammatory parameters sustained in L-TPD. (A) Shapley Additive exPlanations (SHAP) graph showing the contribution of circulating features in the definition of lung sequelae according to a SHAP-value assigned by the algorithm. (B) Scatter plots of anaphylatoxins C3a, C4a, and C5a between patient groups. (C) Scatter plots of CXCL10, CXCL9, and IL-6 levels between patient groups. (D) Scatter plots of lymphocyte, monocyte, granulocyte, and platelet cell counts between patient groups. For panels (B–D), ordinary one-way ANOVA tests; **p < 0.01 and *p < 0.05.

FIGURE 4

Metabolic syndrome in L-TPD a year post-COVID-19. (A) Scatter plots of SF-12 physical, distance walked in 6MWT and hand grip test between patient groups at 12-months post-COVID-19. (B) Scatter plots of CXCL10, CXCL9, and IL-6 levels and platelets counts between patient groups at 12-months post-COVID-19. (C) Heatmaps representing individual patients in the x-axes and metabolic syndrome parameters in the y-axes for the different patient groups at 4- and 12-months post infection. Colored squares represent that the patient exhibited. Waist circumference (WC >102 cm for male and WC >88 cm for female), blood pressure (BP ≥140/90 mmHg), triglycerides (TG ≥150 mg/dl), HDL-cholesterol (HDL ≤40 mg/dl for male and HDL ≤50 mg/dl for female), and fasting blood glucose (BG ≥100 mg/dl). Then, pie charts compare the distribution of patients that exhibit <2, 3, 4, and 5 altered metabolic syndrome parameters between 4- and 12-month post infection for the different groups. (D) Pair comparison of body mass index and (E) triglycerides in patient groups between 4- and 12-months post-COVID-19. For panels (A,B), ordinary one-way ANOVA tests; ***p < 0.005, **p < 0.01, *p < 0.05. For panels (D,E), two-way ANOVA with Sidak multiple comparison tests; **p < 0.01 and *p < 0.05.

FIGURE 5

Cardiac dysfunction and chemotaxis are the main predicted annotations in L-TPD during acute COVID-19. (A) From the study cohort of 60 patients, 16 patients were selected from the CT (n = 8) and CT + DLCOc (n = 8) group and healthy controls (n = 11) without COVID-19. Serum from patients were collected during the acute phase and during the 4-month follow-up. (B) Serum samples were processed and acquired with TIMS-TOF Pro and the data was analyzed with R and IPA. (C) Principal component analysis of the protein profiling analyzed in samples that passed quality control, obtaining data from 11 healthy controls and the 16 patients during the acute (T0) and at 4 months post infection (T1) and (D) heatmap showing the proteins from serum differentially present between the different groups and their respective association with canonical pathways. (E) Overview of the main biological themes and (F) network regulators during the acute phase (T0) and 4-month follow-up (T1) after COVID-19 between patients who exhibited only CT scan abnormalities versus L-TPD (CT + DLCOc), considering canonical pathways, upstream regulators, diseases, and biological functions, showing a positive z-score in orange and a negative z-score in blue.

FIGURE 6

CXCL9 and monocyte chemotaxis are associated with myocardial infarction, however monocytes from L-TPD exhibited reduced expression and function of CD16. (A) Flowchart of coronary and peripheral blood samples obtained from patients suffering myocardial infarction. (B) Chemokine levels in plasma from coronary and peripheral blood samples. (C) Percentages of total CD14⁺ monocytes and lymphocyte (T, B, and NK) present in coronary and peripheral blood samples. (D) Percentage of CXCR3⁺ monocytes in the presence of plasma from coronary and peripheral blood samples for 72 h. (E) Representative dot plots and percentage of migrated monocytes to media, plasma from coronary and plasma from healthy control samples. The percentage of migration for each subset was calculated as (number of cells in the bottom chamber after 1 h × 100)/initial number of cells in the top chamber. (F) Ingenuity pathway analysis graphical representation of FCgamma receptor mediated phagocytosis in macrophages and monocytes in the CT (top) and CT + DLCO (bottom) groups at 4-months post infection. (G) Scatter plots and representative dot plots of CD16 expression in CD14⁺ monocytes from the different patient groups. For panels (B–D), paired t-tests and for panels (E,G), ordinary one-way ANOVA tests; ***p < 0.005, **p < 0.01, *p < 0.05.