BackgroundSevere Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) induced Coronavirus Disease 2019 (COVID-19) has posed a global threat to public health. The immune system is crucial in defending and eliminating the virus and infected cells. However, immune dysregulation may result in the rapid progression of COVID-19. Here, we evaluated the subsets, phenotypic and functional characteristics of natural killer (NK) and T cells in patients with COVID-19 and their associations with disease severity.MethodsDemographic and clinical data of COVID-19 patients enrolled in Wuhan Union Hospital from February 25 to February 27, 2020, were collected and analyzed. The phenotypic and functional characteristics of NK cells and T cells subsets in circulating blood and serum levels of cytokines were analyzed via flow cytometry. Then the LASSO logistic regression model was employed to predict risk factors for the severity of COVID-19.ResultsThe counts and percentages of NK cells, CD4(+) T cells, CD8(+) T cells and NKT cells were significantly reduced in patients with severe symptoms. The cytotoxic CD3(-)CD56(dim)CD16(+) cell population significantly decreased, while the CD3(-)CD56(dim)CD16(-) part significantly increased in severe COVID-19 patients. More importantly, elevated expression of regulatory molecules, such as CD244 and programmed death-1 (PD-1), on NK cells and T cells, as well as decreased serum cytotoxic effector molecules including perforin and granzyme A, were detected in patients with COVID-19. The serum IL-6, IL-10, and TNF-alpha were significantly increased in severe patients. Moreover, the CD3(-)CD56(dim)CD16(-) cells were screened out as an influential factor in severe cases by LASSO logistic regression.ConclusionsThe functional exhaustion and other subset alteration of NK and T cells may contribute to the progression and improve the prognosis of COVID-19. Surveillance of lymphocyte subsets may in the future enable early screening for signs of critical illness and understanding the pathogenesis of this disease.