Immunological Subpopulations Within Critically Ill COVID-19 Patients.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections cause coronavirus disease 2019 (COVID-19). Most SARS-CoV-2 infections are self-limiting and pauci-symptomatic. However, a minority of SARS-CoV-2 infections develop pulmonary and extrapulmonary organ dysfunction (such as hypoxemic respiratory failure, acute kidney injury, and thrombotic complications) that require organ support (COVID-19 critical illness or severe COVID-19, equivalent to World Health Organization Clinical Progression Scale of ≥6 points).

FOR RELATED ARTICLE, SEE PAGE 1884

Dysregulated immune responses are key to the pathogenesis of COVID-19. , Briefly, as an intracellular pathogen, the unique nucleic acid structures and viral replication intermediates of SARS-CoV-2 are sensed by endosomal Toll-like receptors in innate immune cells and the cytosolic retinoic acid-inducible gene-like receptors present in most cells. This sensing of danger signals results in the production of pro-inflammatory cytokines through the nuclear factor-kB transcriptional program and inhibition of viral replication through interferons (IFNs) activating the interferon-stimulated genes program. , The cell-mediated effector immune responses to SARS-CoV-2 infection consist of the transcription factor T-bet and IFNγ dependent type 1 effector immune responses by innate lymphoid cells, natural killer cells, helper cells, and cytotoxic T cells. These immune responses result in viral clearance and illness resolution, especially in patients with self-limiting infections.

However, certain SARS-CoV-2 characteristics and host factors can adversely influence these responses to generate the complex dysregulated immune responses seen in severe COVID-19 illness. SARS-CoV-2 encodes viral proteins capable of evading recognition by immune cells, reducing IFN production, impairing IFN signaling, and impairing IFN-stimulated genes effector function program, all of which impair SARS-CoV-2 clearance. Host factors such as old age and genetic defects can result in delayed IFN responses, leading to persistence of virus and exaggerated systemic inflammation, resulting in severe disease. Severe COVID-19 illness is also associated with inborn errors in the IFN pathway, and antibodies to IFNs. , Furthermore, the effector immune responses associated with helminth infections and with extracellular pathogens (bacterial and fungal infections) appear activated, and they persist in patients with severe COVID-19 illness.

In this context, let us consider the cohort study by Dupont and colleagues in this issue of CHEST. The authors performed immunological assessments in 96 adults with severe COVID-19 illness. This cohort included 26 patients with immune comorbidities (history of malignancy or active malignancy or solid organ transplant), and approximately 60% (16/26) of these patients were receiving immunosuppressant medications. The authors performed Ward’s Hierarchical Agglomerative Clustering, using the following variables: D-dimers, cytokines (IL-6), IL-1β, and tumor necrosis factor-alpha, complement proteins (C3, sC5b-9), gamma globulin levels, and counts of the cytotoxic T cells (CD8), helper T cells (CD4), B cells (CD19), and natural killer cells. The authors identify three phenotypes. First, humoral response deficiency phenotype, characterized by B cell lymphopenia and hypogammaglobulinemia, was most prevalent in patients with immune comorbidities. Second, hyperinflammatory phenotype, characterized by pan T cell lymphopenia and highest cytokines levels, was most prevalent in patients receiving mechanical ventilation. Third, complement-dependent phenotype, characterized by the highest complement protein levels. The overall critical care mortality was 31%, with the highest mortality in the hyperinflammatory phenotype and the least in the complement-dependent phenotype. The authors conclude that these phenotypes should inform eligibility criteria for clinical trials testing immunomodulation.

When contextualizing this work, the key limitations to consider include are a single-center study, without an independent validation cohort and with immunological assessments only at a single time point—particularly because the average duration of symptoms at the time of sampling was 8 days. Additionally, the sensitivity analysis for testing cluster allocation excluded only 16 of the 26 patients with immune comorbidities, namely, those receiving immunosuppressive medication. The cytokine profile measured in this study appears limited, particularly when compared with the extended cytokine profile assessed longitudinally to identify COVID-19 phenotypes previously. However, the cytokines measured include those being considered as potential treatment targets in COVID-19, such as IL-6 (with IL-6 receptor antagonists such as tocilizumab) and IL-1β (with IL-1 receptor antagonist such as anakinra), giving the study context relevance. Similarly, in patients with COVID-19, despite the overall lymphopenia, there is a brisk plasmablast response, profoundly altered T cell subsets, and differential changes in B and T cell subsets that contribute to COVID-19 immune phenotypes, none of which are measured in this study.11, 12, 13 Although C-reactive protein, IL-8, ferritin, and human leukocyte antigen-DR isotype were measured and reported, these variables were not considered in the unsupervised clustering analyses. The relevance is that IL-6 levels are associated with high C-reactive protein and decreased human leukocyte antigen-DR isotype expression, and there is a strong positive correlation between ferritin and D-dimers in severe COVID-19.

COVID-19 immune phenotypes have been reported previously. , Lucas and colleagues performed longitudinal immunophenotyping (using cell subsets, cytokine, chemokines, and other markers) in patients with moderate (n = 80) and severe (n = 33) COVID-19 illness, using SARS-CoV-2 negative health-care workers (n = 108) as control subjects. In this study, the four immune signatures identified correlated with three distinct immune trajectory clusters. Two clusters represented severe COVID-19 illness, and one, moderate illness with better outcomes. The two severe COVID-19 illness clusters had increased levels of inflammasome-associated cytokines (IL-1α, IL-1β, IL-6, IL-18, and tumor necrosis factor), type 1 (IL-12, chemokines linked to monocyte recruitment, and IFNγ); type 2 (chemokines linked to eosinophil recruitment, IL-4, IL-5); and type 3 (IL-23, IL-17A) effector immune response markers. Mathew and colleagues performed high dimensional phenotyping of lymphocyte subsets and integrated the immune and clinical features in 125 COVID-19 patients with different illness severity. The authors report three groups of patients referred to as immunotypes. Two of these clusters were identified using principal component analyses, and the third immunotype was identified using additional information on clinical characteristics. Immunotype 1 signature included CD4 T cell activation, brisk plasmablast response, and lower frequencies of proliferating effector/exhausted CD8 T cells. The immunotype 2 signature included conventional effector CD8 T cell subsets, less CD4 T cell activation, and less proliferating plasmablasts and memory B cells. The immunotype 3 signature was characterized by minimal lymphocyte activation. Given the differences in biological measurements between these studies, head-to-head comparisons between phenotypes are inappropriate.

Severe COVID-19 illness immunology is complex. Immunological subpopulations reported depend on immunological measurements used, timing of measurements, and the analytic methods. Given the wealth of open-source data, it is essential that the research community engages in deriving and validating immunological subpopulations of COVID-19. Until then, studies such as these are hypothesis generating and have limited direct impact on clinical care.