Regadenoson for the treatment of COVID-19: A five case clinical series and mouse studies

Abstract

Background: Adenosine inhibits the activation of most immune cells and platelets. Selective adenosine A2A receptor (A2AR) agonists such as regadenoson (RA) reduce inflammation in most tissues, including lungs injured by hypoxia, ischemia, transplantation, or sickle cell anemia, principally by suppressing the activation of invariant natural killer T (iNKT) cells. The anti-inflammatory effects of RA are magnified in injured tissues due to induction in immune cells of A2ARs and ecto-enzymes CD39 and CD73 that convert ATP to adenosine in the extracellular space. Here we describe the results of a five patient study designed to evaluate RA safety and to seek evidence of reduced cytokine storm in hospitalized COVID-19 patients.

Methods and findings: Five COVID-19 patients requiring supplemental oxygen but not intubation (WHO stages 4-5) were infused IV with a loading RA dose of 5 μg/kg/h for 0.5 h followed by a maintenance dose of 1.44 μg/kg/h for 6 hours, Vital signs and arterial oxygen saturation were recorded, and blood samples were collected before, during and after RA infusion for analysis of CRP, D-dimer, circulating iNKT cell activation state and plasma levels of 13 proinflammatory cytokines. RA was devoid of serious side effects, and within 24 hours from the start of infusion was associated with increased oxygen saturation (93.8 ± 0.58 vs 96.6 ± 1.08%, P<0.05), decreased D-dimer (754 ± 17 vs 518 ± 98 ng/ml, P<0.05), and a trend toward decreased CRP (3.80 ± 1.40 vs 1.98 ± 0.74 mg/dL, P = 0.075). Circulating iNKT cells, but not conventional T cells, were highly activated in COVID-19 patients (65% vs 5% CD69+). RA infusion for 30 minutes reduced iNKT cell activation by 50% (P<0.01). RA infusion for 30 minutes did not influence plasma cytokines, but infusion for 4.5 or 24 hours reduced levels of 11 of 13 proinflammatory cytokines. In separate mouse studies, subcutaneous RA infusion from Alzet minipumps at 1.44 μg/kg/h increased 10-day survival of SARS-CoV-2-infected K18-hACE2 mice from 10 to 40% (P<0.001).

Conclusions: Infused RA is safe and produces rapid anti-inflammatory effects mediated by A2A adenosine receptors on iNKT cells and possibly in part by A2ARs on other immune cells and platelets. We speculate that iNKT cells are activated by release of injury-induced glycolipid antigens and/or alarmins such as IL-33 derived from virally infected type II epithelial cells which in turn activate iNKT cells and secondarily other immune cells. Adenosine released from hypoxic tissues, or RA infused as an anti-inflammatory agent decrease proinflammatory cytokines and may be useful for treating cytokine storm in patients with Covid-19 or other inflammatory lung diseases or trauma.

Fig 1. Subject flow diagram.

Fig 2. Changes in blood pressure before, during and after regadenoson (RA) infusion.

LD: Loading dose, 2.5 μg/kg over 30 minutes. MD: Maintenance dose,1.44μg/kg/hour for 6 hours. Data shown are means ± SEM from 5 RA treated COVID-19 patients. Repeated measures analysis applied to changes from baseline shows no significant changes during or after RA infusion (SBP, p = 0.28, DBP, p = 0.28).

Fig 3. Effects of regadenoson (RA) infusion on oxygen saturation and C-reactive protein (CRP) levels in plasma of COVID-19 patients.

(A) SpO2 was determined by pulse oximetry just prior to and 24 hours after the initiation of 6.5 hours of RA infusion. There was a statistically significant difference in in SpO2 based on paired T-test (p = 0.03,n = 5). (B) CRP before and 24 h after RA infusion, P = 0.075 (n = 4). (C) D-dimer before and after RA infusion (p = 0.043,n = 5).

Fig 4. Decrease in circulating iNKT cell activation during regadenoson (RA) infusion into patients with COVID-19.

(A) Flow cytometry gating strategy used to identify CD3+6B11+iNKT cells and their activation-state based on CD69 expression. (B) Comparison of percentages of circulating CD69+CD3+6B11- conventional T cells and CD69+6B11+ iNKT cells at baseline and 30 minutes into RA infusion. (C) Changes in the activation state of iNKT cells in blood of COVID 19 patients before, during and after RA infusion. PBMCs prepared from blood samples of 5 subjects were thawed and analyzed by flow cytometry at the same time. The color scheme used in Fig 4B & 4C denote the same individuals as shown in Fig 3A.

Fig 5. Changes of plasma proinflammatory cytokines/chemokines during and after RA infusion.

Cytokine levels were normalized to baseline just prior to RA infusion. Cytokine levels below and above baseline values are depicted by green and red, respectively. Blood was collected before RA infusion (baseline) and 3 times after the start of infusion: (A) at the end of loading dose (5μg/kg/hour, 30 minutes), (B) 4 hours into the maintenance dose (1.44μg/kg/hour, 4.5 hours) and (C) 17 hours post RA infusion, n = 5. Green dots and lines indicate cytokine levels lower than baseline, *P<0.1; **P≤0.05 relative to baseline by Paired T-Tests.

Fig 6. Regadenoson increases survival of SARS-CoV-2-infected K18-mice.

Seven day 0.5 μl/h Alzet pumps loaded with 80 μg/ml Regadenoson (N = 19), or saline (N = 11) were implanted subcutaneously into 2–3 month old B6.Cg-Tg(K18-ACE2)2Prlmn/J mice just prior to infection with 1250 Pfu of Hong Kong VM20001061/2020. Mice were euthanized based on criteria described in Methods. Data are combined from three independent experiments. Survival curves are statistically different based on the Log rank Mantel Cox test, P<0.001. The time of RA infusion is indicated by a horizontal bar.

Fig 7. Diagram showing some of the factors that mediate cytokin storm following infection with COVID-19.

The SARS-CoV-2 virus gains entry into lung type II epithelial cells by binding its spike protein to Angiotensis Converting Enzyme 2. CD1d is expressed by epithelial cells. Viral entry causes stressed cells to release alarmins such as IL-33 which is elevated in the blood of COVID-19 patients. IL-33 and possibly CD1d-restricted lipid antigens activate NF-κB in iNKT cells and induce the rapid transcription of cytokines such as IFN-γ, IL-4 and IL-17. The release of IFN-γ propagates cytokine storm by stimulating the release from stromal and other cells of T cell attractants, CXCL9, CLCL10 and CXCL11. The activation of NF-κB in iNKT cells also causes a slow induction over 24–48 hours of A2ARs that enhance over time the anti-inflammatory potency of adenosine.