According to the study team lead by scientists from the University of São Paulo Medical School, unmetabolized ATP (adenosine triphosphate), a key source of energy for cellular processes causes the dysregulation of the immune system leading to cytokine storms that contributes to the disease severity.
The study findings were published in the peer reviewed journal: Frontiers Of Immunology.
ATP dephosphorylation into ADO mediated by ectonucleotidases is a key regulatory mechanism of immune responses, promoting the shift from ATP-driven inflammation to immunosuppression induced by ADO (41). Given the pronounced inflammatory characteristics of severe COVID-19 and the role of purinergic signaling in immunosuppression, alterations in the metabolism of extracellular nucleotides could contribute to the immunopathogenesis of the disease.
Herein, we demonstrate that impaired expression of nucleotidases and lower ADO concentration in the blood is associated with a worse prognosis of COVID-19, while in vitro administration of exogenous ADO helps to prevent inflammatory responses in the leukocytes of patients.
Expression profiles of ectonucleotidases can be modified under pathological conditions. Higher expression of CD39 in lymphocytes, for example, has been reported in solid tumors and chronic viral infection by HIV and HCV, and more recently during acute infection by SARS-CoV-2 (21, 42, 43).
Induction of CD39 expression occurs upon cellular activation and is regulated by hypoxia, oxidative stress, and inflammatory cytokines such as IL-6 and TNF-α (41, 44–46), which are frequently increased in COVID-19 patients (47–50). Although higher expression of CD39 in T cells from COVID-19 patients might indicate a possible mechanism to counterbalance the inflammatory responses via consumption of ATP, its expression in B cells, the major leukocyte population that expresses CD39, is diminished.
The compromised expression of ENPP1, ENPP2, ENPP3, and NT5E (CD73) in the peripheral blood of COVID-19 patients shown here, negatively correlates to plasma levels of the inflammatory marker CRP and the neutrophil-to-lymphocyte ratio, supporting the hypothesis of a direct contribution of the purinergic metabolism to the pathogenesis of the disease.
In accordance with our findings, Ahmadi et al. reported that loss of CD73 expression in CD8+ T and NKT cells of COVID-19 patients negatively correlates with serum levels of ferritin, another inflammatory marker of the disease (20). It is important to mention, though, that gene expression analysis in whole blood of COVID-19 patients should be interpreted with caution due to the mono-lymphopenia associated with the disease (51, 52).
In this context, it is hard to distinguish whether the altered gene expression results from transcriptional events or whether it is a consequence of the imbalance proportion of lymphocytes and monocytes that express such enzymes in the blood of patients. In any case, reduced expression of nucleotidases in peripheral blood may be possibly responsible for the lower concentrations of ADO in the plasma of COVID-19 patients shown here.
Deep and cell-specific analysis evidence that the surface expression of CD73 is impaired in CD4+ T cells and CD8+ T cells in patients with COVID-19, corroborating previously published data (20, 21). In addition, we verified, for the first time, lower expression of CD73 also in CD19+ B cells. Loss of surface CD73 can be explained by alterations at the transcriptional level or also by the shedding of the enzyme from the cell membrane upon cellular activation (53).
Complementary, our single-cell RNA-sequencing data analysis from PBMCs exposed to SARS-CoV-2 indicate that the expression of nucleotidases is not likely to be directly affected by the virus (Figure 2E). However, it has been shown that incubation with plasma from COVID-19 patients inhibits the expression of CD73 in lymphocytes from HDs (21), suggesting the contribution of soluble factors in these alterations besides the virus itself.
Moreover, matrix metalloproteinase (MMP-9), which is released by neutrophils during acute lung damage and is elevated in COVID-19 patients’ blood (54, 55), can cleave CD73 from the cellular membrane generating a soluble protein (56, 57). In our cohort, however, no differences regarding soluble CD73 in the plasma of COVID-19 patients and HDs were observed (Figure 2D).
Regardless of the specific mechanism behind this alteration has not been elucidated in detail here, loss of CD73 could contribute to the maintenance of the effector function of T cells by preventing ADO-mediated immunosuppression (53, 57). In support of our findings, lymphocyte activation markers such as CD38, CD69, and CD44 are highly expressed on CD4+ and CD8+ T cells of COVID-19 patients (58), and CD73 absence in CD8+ T cells induces granzyme production in these individuals (20).
Different from other lymphocytes, the majority of B cells express CD39 and CD73, contributing significantly to the generation of ADO and inhibiting proliferation and cytokine production in T cells (36). Herein we evidence that B cells from patients with severe COVID-19, which have lower expression of CD39 and CD73, show an impaired capacity to hydrolyze ATP.
Similar results were already reported in patients infected with HIV and HBV, where the compromised generation of ADO is claimed to favor inflammatory responses and immune activation (15, 16, 59). Therefore, we have evidences to support the hypothesis that the marked reduction of CD39+CD73+ B cells, together with the impaired expression of CD73 in T cells and the lack of other nucleotidases in the blood, lead to the lower concentrations of plasmatic ADO in COVID-19 patients and might exacerbate innate immune activation.
In addition, the absence of CD73 and defective activation of the ADO downstream PKA-mediated phosphorylation of activation-induced deaminase (AID) impairs immunoglobulin class switching in human B cells (60, 61). Whether these alterations compromise humoral responses in COVID-19 patients remains uncertain.
Like us, others have suggested that increased systemic levels of ATP are likely to be involved in the immunopathogenesis of COVID-19 (62, 63), possibly as consequence of the altered expression of nucleotidases. Indeed, the accumulation of ATP has been shown the trigger inflammatory responses such as the activation of the inflammasome pathway (64, 65).
Interestingly, there is a higher activation of the NLR family PYRIN domain containing-3, NLRP3, inflammasome in COVID patients (66, 67). In addition, extracellular ATP can contribute to lung local inflammation by recruiting eosinophils, dendritic cells, and neutrophils via P2Y2 receptor (68–70), implying that imbalanced metabolization of this nucleotide could directly contribute to the immunopathogenesis of COVID-19.
Apart from the defective hydrolysis of ATP, ADO signaling itself is apparently compromised in COVID-19, where lower expression of ADO receptors and reduced activity of PKA were observed. Mechanistically, activation of the cAMP/PKA pathway via ADO receptors inhibits the production of TNF-α, IFN-γ, and IL-2 and T cell proliferation (71, 72).
At the same time, ADO receptor signaling, more specific A2AR and A2BR, has been shown to trigger IL-10 production via CREB activation (73–75). Therefore, the compromised generation of ADO by leukocytes, especially B cells, associated with impaired signaling mediated by ADO receptors could exacerbate inflammatory responses systemically.
Although ADO signaling is compromised in leukocytes of severe COVID-19 patients, we verified that in vitro treatment can attenuate the production of TNF-α and IL-6 in CMNs after activation. Both cytokines are increased in the blood of COVID-19 patients (76). In fact, it has been reported that ADO reduces NF-κB activation in T cells and monocytes of COVID-19 patients in vitro (21), and the administration of an A2aR agonist attenuated the production of pro-inflammatory cytokines in SARS-CoV-2 mice infection model (77).
Moreover, preliminary data suggest that inhaled ADO reduces the levels of CRP in the blood while improving oxygenation rates and reduces hospitalization time in patients with COVID-19 (78, 79). Here we observed that, at the dose used, extracellular ADO can partially overcome the impaired expression of its receptors and imbalanced induction of PKA activity (Figures 4, 5), attenuating inflammatory responses.
It is important to mention though that co-treatment with other cAMP inducers, such as the neuropeptide PACAP (80), could result in a more pronounced immunosuppressant effect. Taken together, these findings evidence the potential of ADO as a therapeutic strategy to overcome the exacerbated inflammatory responses in COVID-19.
In summary, our findings indicate that alterations in the purinergic signaling contribute, at least in part, to the immune activation and worse prognosis in acute COVID-19 and reveal the therapeutic potential of ADO-mediated responses in this disease. Whether these alterations persist in recovered patients or impact the development of post-COVID-19 syndrome remains to be elucidated.