Picture of neutrophil extracellular trap by Dr Volker Brinkmann, Max Planck Institute for Infection Biology, Berlin, Germany (source)

Neutrophil Extracellular Traps in COVID-19

A key part of the innate immune system is neutrophil extracellular traps (NETs), formed by neutrophils as they eject their DNA from their cytoplasm. NETs are most commonly described as a mechanism for physically trapping bacteria so that they can’t spread and be more easily cleared by other phagocytes. However, NETosis also plays a role in viral immunity. In particular, lung damage from influenza is at least partially driven by NETosis.

There is evidence that a neutrophil-driven inflammatory response is involved in COVID-19. Neutrophil-to-lymphocyte ratio (NLR) is correlated with negative disease outcomes per an analysis of 5 studies and as-yet-unpublished results reported by Daniel Griffin (discussed at 5:15 here). A report on 4 autopsies from New Orleans found “increased neutrophil count with relative lymphopenia” (that is, high NLR), as well as significant neutrophil entrapment in thrombi in the lung vasculature. They also reported fibers that could be NETs in one (albeit immunocompromised) patient.

The two big disease states driving mortality appear to be ARDS (acute respiratory distress syndrome) and DIC (disseminated intravascular coagulation). NETs are involved in both. NETs also appear to be an important part of the cytokine storm feedback loop, since they have been shown to upregulate IL-1, IL-8, and TNF in human bronchial cells in vitro. Eliminating them in addition to blocking those cytokines directly may be a good treatment approach.

Since NETs are made from DNA, they can be degraded using DNase enzyme. In the case of COVID-19, we would want to deliver that enzyme to the lungs. It turns out there is already a drug on the market for this- dornase alfa (brand name Pulmozyme from Genentech, Tigerase from Generium). Dornase alfa is a nebulized DNase I enzyme used to treat cystic fibrosis.

Surprisingly given the amount of research on NETs in inflammatory and thrombotic responses, there are few trials using it for purposes other than cystic fibrosis.

One set of pediatric guidelines recommends use of dornase alpha on non-cystic-fibrosis patients only when they have an endotracheal tube section with heavy polymorphonuclear neutrophils. This is one of the defining features of ARDS. There is also a French RCT that just began on March 10 testing a protocol called TRAUMADORNASE- giving dornase alfa to ventilated trauma patients. The protocol was written last year, pre-COVID, but I imagine they won’t have much difficulty finding subjects now, and will hopefully release preliminary results as they go.

If the trials on chloroquine/hydroxychloroquine end up demonstrating efficacy, that may be in part due to the effect of chloroquine on NETs. In mouse models of acute pancreatitis and pancreatic cancer, administration of chloroquine reduced NET-associated biomarkers (cfDNA and CitH3). That said, another study involving hantavirus-induced NETosis in human cells found no effect of chloroquine.

Why might this not work

First off: most drugs don’t work. So no matter how much justification I put here, it’s still very unlikely to work.

Second, cystic fibrosis is characterized by heavy mucus whereas the WHO China report claims only 1/3 of COVID-19 cases had sputum production, while 2/3 had a dry cough. If nebulized dornase alfa only works on the mucus and it is not driving fatality, then it might not work. Alternatively, it may only work in the set of patients with sputum production. However, as we have seen above with the evidence on thrombosis, mucus production is not necessarily the only issue with excess NETosis.

Third, if reports of infection in other ACE2-expressing tissues (GI tract, kidneys, heart, T cells, etc) are true and are driving a significant amount of the damage, then anti-NET therapy delivered exclusively to the lungs will not help much. The prescribing information for pulmozyme is inconsistent in whether administration drives up serum concentrations of DNase (which could presumably increase efficacy of the drug in sites other than the lung). It cites two studies, one of which showed no increase in serum concentrations of DNase over 6 months of use, the other of which showed a modest increase over 14 days of use. It notes “dornase alfa is expected to be metabolized by proteases present in biological fluids.” However, if the inflammation is largely localized to the lung until the cytokine storm triggers a global response, then the targeted therapy should still help in preventing severe disease.

Fourth, even if the damage is isolated to the lungs, dornase alfa may be of limited use if it can only act in the airways. The New Orleans autopsy examination indicated that neutrophilic infiltration of the lungs only occurred in 1 of the 4 samples examined. In cases where there is sufficient damage that the blood vessels hemorrhage into the airways (possibly the cause of the characteristic “pink froth” in the lungs), the protein could enter the vessels directly. However, that would only be helpful if there’s a significant period of time between the beginning of hemorrhage and death.

Fifth, NETs have antiviral activity, so destroying them could be counterproductive. As with other immunological treatment approaches (IL-6 inhibitors, steroids, etc), it can be difficult to time the therapy directly so that the negative components of the inflammatory response are interrupted without hindering viral clearance.

What do to

The most obvious path forward is to set up a clinical trial for dornase alfa. Dr Jonathan Spicer announced on Twitter on March 19th that he was planning on adapting the French TRAUMADORNASE study for this purpose. I tried messaging him about it to see what has happened in the interim but he has not yet responded.

Another angle besides administering dornase alfa by itself is to administer it in conjunction with heparin (already being tested in COVID), or in patients that need to discontinue heparin use due to heparin-induced thrombosis/thrombocytopenia. A paper in Nature Communications shows that the immune response in HIT is driven by IgG antibodies binding to Fc-gamma receptors on neutrophils, triggering NETosis. Dornase alfa could potentially mitigate the unfortunate pro-thrombotic side effects of this anticoagulant drug.

Even without administering dornase alfa to patients, we can start testing the NETosis hypothesis by testing neutrophils, cfDNA, neutrophil elastase, and citrullinated histone H3 (NETosis biomarkers) in bronchoalveolar lavage, sputum and serum. Except for neutrophils in serum, I have not been able to find any papers reporting on these levels in COVID-19 patients. In addition to biomarkers, we could test NETosis in neutrophils directly. In October, there was a paper published describing a new assay for NETs in plasma that can predict disseminated intravascular coagulation and death in ICU patients. However, I was unable to find any information about commercial production of the assay so it’s unclear if it would be easily available for most hospitals.

If the NETosis hypothesis is true, it is worth being careful about bronchoalveolar lavage sampling- research in children, dogs, and monkeys indicates that lavage specifically induces neutrophil recruitment to the lungs.

In conjunction with clinical testing, Genentech and Generium need to be warned about the existence of the tests so that they can ramp up production of dornase alfa prior to seeing the results. Per the Cystic Fibrosis Foundation, there are only 60,000 people with cystic fibrosis worldwide. Given that target population, the total manufacturing capacity of both companies is likely small. Dornase alfa is manufactured from chinese hamster ovary cells, so I have hope that existing facilities also using CHO cells could be repurposed, but I don’t know how specific each facility is to each process.

If you find the dornase alfa approach compelling, reach out to me and let me know if I can help in any way, including (but not limited to) literature research, outreach or fundraising.