Neutrophils during SARS‐CoV‐2 infection: Friend or foe?

Summary

Coronavirus disease 2019 (COVID‐19) is caused by the severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) and has resulted in more than 6 million deaths worldwide. COVID‐19 is a respiratory disease characterized by pulmonary dysfunction leading to acute respiratory distress syndrome (ARDs), as well as disseminated coagulation, and multi‐organ dysfunction. Neutrophils and neutrophil extracellular traps (NETs) have been implicated in the pathogenesis of COVID‐19. In this review, we highlight key gaps in knowledge, discuss the heterogeneity of neutrophils during the evolution of the disease, how they can contribute to COVID‐19 pathogenesis, and potential therapeutic strategies that target neutrophil‐mediated inflammatory responses.

1. INTRODUCTION

Neutrophils, the most abundant immune cells in human blood, are the effector arm of the innate immune system. They are often the first cells recruited to tissues when homeostasis is disturbed by a pathogen but are also summoned in sterile inflammation. ¹ The activation of resident cells by either pathogen‐associated molecular patterns (PAMPs) derived from microbes, or by danger‐associated molecular patterns (DAMPs) released by damaged cells, leads to production of a vast area of inflammatory mediators. Those mediators communicate with the circulating neutrophils in part by stimulating endothelial cells, that express adhesion molecules and additional chemokines culminating in the tethering rolling, adhesion and emigration (recruitment) to the locus of infection/damage. ² Once in an infectious nidus, neutrophils counteract an infection through phagocytosis, expression of proteolytic and oxidative mediators, and/or the release of neutrophil extracellular traps (NETs). ³ In sterile injury, it is unlikely these same anti‐microbial events occur and the neutrophil may actually provide important repair functions. This dichotomy will be discussed below.

The role of neutrophils in the defense against pathogens has been mainly studied during fungal and bacterial infections. Their importance is exemplified by the fact that neutropenia is the biggest risk factor for severe infection. ⁴ Indeed, the impairment of neutrophil migration to an infectious focus leads to pathogen dissemination and systemic infection. ⁵ Usually, neutrophils are the first to enter an infectious nidus and they phagocytose pathogens into phagosomes that combine with lysosomes wherein proteases and oxidants eliminate the invaders. In this scenario, very little untoward inflammation/injury occurs and the system returns to homeostasis. When faced with a bigger challenge ranging from an organism too large to phagocytose, a biofilm or perhaps intracellular pathogens, sufficient activation of neutrophils may begin to release oxidants, proteases, and NETs extracellularly. Neutrophils can in this exuberant state cause collateral tissue damage during infection. ³ , ⁶ This situation is usually still localized to a primary site, although it can cause injury to that tissue. Examples of this include the recurrent/ongoing infections that occur in the lungs of cystic fibrosis patients and the swelling of specific joints with gout. ⁷ , ⁸ It is when inappropriate activation of neutrophils beyond the primary afflicted site occurs, this than contributes to systemic inflammation observed in different pathologies such as acute respiratory distress syndrome (ARDs) and sepsis. ³ , ⁵ , ⁹ Below we will begin by giving a brief overview of neutrophils and sepsis and the lack of good therapies to warn against making similar mistakes with SARS‐CoV‐2. We will then go on to summarize some of the new work that is being done in neutrophil biology as it pertains to SARS‐CoV‐2 and highlight where gaps exist.

2. SEPSIS AND THE NEUTROPHIL

Many models of systemic neutrophil activation have over the years been examined. Depletion of neutrophils and/or inhibition of oxidants, proteases, or NETs provide some protection in these models. ¹⁰ , ¹¹ , ¹² , ¹³ The stimuli used include lipopolysaccharide (LPS), single strain bacterial or polymicrobial infections induced by cecal ligation and puncture. ¹⁴ Generally, in these models, regardless of whether the inciting agent is given intraperitoneally or intravascularly, it activates neutrophils in the blood stream, and these immune cells are then recruited to the lungs to cause significant injury, including increased vascular permeability (leakage of fluid into the parenchyma and alveoli) or even frank hemorrhage. This leads to poor oxygenation and potentially multi‐organ dysfunction. ⁵ Depressed cardiac function, inappropriate liver inflammation, and kidney failure are a few examples of the systemic dysfunction and multi‐organ failure that occurs. There are many, many examples of inhibition of a single inflammatory molecule reducing the multi‐organ failure and improving survival in rodent models. ⁹ , ¹⁵ Many of these interventions have been tested subsequently in larger animals including non‐human primates with some success. However, to date none of these many different approaches have ever been successfully translated into clinical practice. Usually, large multi‐centered clinical trials fail and if successful they subsequently are not reproduceable. Instead, improvements in sepsis survival have resulted from small, incremental advances in the life support offered in intensive care units (ICUs). As a result, deaths during the first weeks have declined from 40%‐50% to 20%‐25%; however, this is entirely due to antibiotics and fluid delivered within the first few hours of being admitted to hospital. ¹⁶ Sepsis affected 18 million patients/year worldwide, with 750 000 North Americans and 50 000 Canadians prior to COVID‐19. ¹⁷ , ¹⁸ For patients with SARS‐CoV‐2 infections (a form of sepsis) leading to intensive care unit stays, this has simply added to these huge mortality numbers.

Therefore, it is worth re‐stating that no single intervention including all the therapeutics with anti‐neutrophil properties have ever provided any benefit in sepsis. Although the failed trials are blamed on the mouse being a poor model for human sepsis, it is worth mentioning that mouse neutrophils have many more features in common (than different) with human neutrophils. ¹⁹ , ²⁰ Pre‐treatment with therapeutics in mouse studies is one problem as patients generally have sepsis for quite sometime before arriving in the ICU. Second, LPS is not a model of sepsis, which by definition requires an infectious pathogen. When these two parameters are removed, there are few studies where inhibition of neutrophil function results in benefit to septic mice. Finally, it is important to mention that septic patients in the ICU that qualify for trials range from 90 years old with broken hips to 20 years old gun shot victims to patients with various co‐morbidities. It is even more important to mention that most clinical trials are designed to take patients with gram positive, gram negative, polymicrobial, and fungal sepsis. Thus, it is no wonder with these less than stringent rules, that a molecule like Eritoran (TLR4 inhibitor) failed in this type of trial design. The challenge, however, for bacterial sepsis trials is identifying the inciting pathogen, and as such, clinical trials do not restrict the type of pathogen that causes the sepsis. As such, our meticulously specific inhibitors are doomed to failure.

Therefore, why would trials fair any better in SARS‐CoV‐2 sepsis. We believe there are advantages to potentially treating SARS‐CoV‐2 patients with novel therapies. The patient population is much more homogenous, the inciting pathogen is known, and even the length of time the patient has been septic can be determined. In fact, one could target patients on clinical wards with SARS‐CoV‐2 that have not yet reached “septic status.” As such, there is some value in considering very specific therapies including those that affect the neutrophil as a potential therapeutic target in SARS‐CoV‐2. However, learning from previous septic trials, it is important to take into account patients that have secondary bacterial or fungal infections and serious co‐morbidities.

3. THE ROLE OF NEUTROPHILS IN VIRAL INFECTIONS

Until recently, the neutrophil was not considered a major player in viral infections. However, there is growing evidence, that during viral infections, neutrophils may play both beneficial but also detrimental roles. ²¹ , ²² Firstly, neutrophils are recruited to sites of viral infections. Using imaging of the liver microcirculation, unveiled a significant recruitment of neutrophils following pox virus infection of liver cells. ²³ Despite the infection being primarily of Kupffer cells (intravascular liver macrophages) and perhaps some hepatocytes (never endothelium), there was clear recruitment, but also very profound activation of the neutrophils (very significant shape change etc.). In addition, platelet adhesion to the neutrophils led to very profound aggregate formation leading to NET release and all this could be prevented by depleting the liver Kupffer cells. A number of viral detectors (MAVs and MDA‐5) were activated within this system but accounted for <50% of the immune response suggesting other key unidentified molecules were also involved. ²³ A similar NET release mechanism was also observed with HIV. ²⁴ However, depletion of neutrophils or preventing NETs from forming led to increased numbers of infected liver cells. Similarly, the depletion of neutrophils in mice infected with the H3N2 influenza or during neurotropic hepatitis virus infection led to the increase of those diseases' severities ²⁵ , ²⁶ raising concerns about targeting neutrophil effector mechanisms perse. By contrast, limiting neutrophil influx during HSV ocular infection or after influenza A virus infection in mice reduced the severity of stromal keratitis lesions ²⁷ and lung injury, ²⁸ summarized elsewhere, ²¹ raising the possibility that these specific viruses could be hijacking the neutrophils for their benefit. In other words, depending on the virus, the neutrophil could help or harm the host. What about SARS‐CoV‐2?

4. NEUTROPHILS AND SARS‐COV‐2

Coronavirus disease 2019 (COVID‐19) is primarily a respiratory disease characterized mainly by pulmonary failure that can progress to acute respiratory distress syndrome (ARDs), thrombosis, disseminated coagulation, and death. ²⁹ These sequeli are not restricted to COVID‐19, and have also been reported for bacterial and fungal infections and therein neutrophils appear to be centrally involved. There is some evidence that neutrophils are involved in severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) the causative pathogen of the COVID‐19. SARS‐CoV‐2 infects host epithelial cells after binding angiotensin‐converting enzyme 2 (ACE2), and this somehow leads to the many multi‐organ problems. ³⁰ , ³¹ , ³² This begs the question whether SARS‐CoV‐2 can also infect other cells to cause some of the many complications associated with COVID‐19. The key receptor ACE2 is definitely expressed on numerous epithelial cells throughout the body, but with some of the pathology seen may also infect endothelial cells and perhaps some immune cells including neutrophils. ³³ , ³⁴ The incubation of neutrophils ³⁴ or platelets ³⁵ with SARS‐CoV‐2 leads to changes in effector function which could simply be the stimulation of membrane receptors for viral molecules. However, these same studies claimed that both neutrophils and platelets express ACE2, ³⁴ , ³⁵ making it plausible that these circulating cells could be infected by virus leading to some of the cardiovascular pathology described below. However, the neutrophils themselves would not need to be infected to be activated, especially if the surrounding endothelium was infected. From the epithelium, it is a very short distance of <2 μm to the capillary endothelium of the alveolar compartment to ensure effective diffusion of oxygen into the blood. As such, SARS‐CoV‐2 could potentially infect the alveolar capillary endothelium ³⁶ which could cause pronounced vascular disturbances including thrombosis, coagulation, and inflammation. ³⁷ There is some evidence that SARS‐CoV‐2 can be found in endothelium from some but not all patients ³⁸ in some vascular beds and in only some parts of the vascular tree (namely capillaries). While infection of endothelium could occur via the ACE‐2 pathway, because of the aforementioned proximity of the alveolar epithelium and lung endothelium, infection could also occur via extracellular vesicles as well as tunneling nanotubes to deliver virus from one cell type to another as has been shown for dissemination of other viruses independent of cognate host surface receptors. ³⁹ , ⁴⁰ , ⁴¹ It remains to be seen which cells get infected by SARS‐CoV‐2.

Severe cases of COVID‐19 have been associated with high levels of cytokines including IL‐6, IL‐8, TNF, and G‐CSF which is reminiscent of sepsis induced by many other pathogens. ²⁹ These molecules are all potent activators of neutrophils, as well as endothelium, as well as other immune cells and platelets. There are increased numbers of neutrophils in the circulation of patients ²⁹ with some elevation in the airways in some patients, and this has been a negative prognosticator for patients. ⁴² NETs have also been identified in autopsy specimens along with platelets binding to neutrophils. This latter observation was first made in bacterial septic patients where platelets bound neutrophils in septic blood and induced NET release. ⁴³ Inhibition of single inflammatory molecules in the bacterial septic setting has failed to reduce NET formation ⁴⁴ suggesting that pan inhibitors may be required. As such, there are changes in number and effector function of neutrophils in SARS‐CoV‐2 in part due to a pro‐inflammatory milieu. However, whether these neutrophils are benefiting or harming the host remains an area of huge contention. To address this issue, we would first like to summarize the literature regarding various types of neutrophils, some being more pro‐inflammatory and harmful, while others being more repair‐like. In other words, depending on the type of neutrophil could dictate patient outcome.

5. NEUTROPHIL HETEROGENEITY

Until recently, neutrophils were thought to be a homogeneous population of cells that mature in the bone marrow and are released into the circulation when fully differentiated, ready to react upon an encounter with a pathogen or during a tissue injury. Because of this terminal differentiation and the fact that blood neutrophils are quite homogenous, it is generally presumed that all neutrophils are the same. However, with the advent of single‐cell RNA sequencing, this concept has been changing in recent years, as different groups have shown the presence of different populations of neutrophils in the circulation and organs, mainly in mice, but also in humans, during homeostasis and following various perturbations. ⁴⁵ , ⁴⁶ , ⁴⁷ It is worth mentioning that even if neutrophils were homogenous in blood, tissue environment could certainly skew their phenotype. Indeed, a very good example of this is in cancer where neutrophils can be anti‐microbial and as such anti‐tumorigenic (N1) or more repair like and pro‐tumorigenic (N2) in their behavior. ⁴⁸ Similar paradigms have been proposed for wound healing where anti‐microbial neutrophils may be recruited to tissue to initially remove debris and then a second pro‐angiogenic neutrophil is either recruited or the initial neutrophil is modified so that repair can take place. ⁴⁹ , ⁵⁰ How this is all regulated is unclear; however, a simple explanation is that mature neutrophils in the circulation are anti‐microbial and respond almost immediately to infections and injuries whereas less mature neutrophils still residing in the bone marrow could be released as more repair like, less anti‐microbial in nature. Whether these are simply immature neutrophils or perhaps an immature neutrophil that has been skewed toward a more repair phenotype remains unclear. Tracking these neutrophils has not been easy and while there are molecules that might discriminate among, at the very least young and old neutrophils, a consistent marker remains elusive. Certainly, CD101, Siglec F, Ly6G (high vs intermediate), CXCR2, CXCR4, and numerous other molecules seem to be expressed at different levels in neutrophils; however, a unifying model has not been established.

One variable causing heterogeneous neutrophil phenotypes is their maturation state. Neutrophils are derived from granulocyte–monocyte progenitors (GMPs), and their maturation takes place in the bone marrow (BM). GMPs differentiate into different cell subsets, including a population of committed proliferative neutrophil precursors, termed pre‐neutrophils (CD117⁺ cells), which then differentiate into non‐proliferating immature pro‐neutrophils and then into mature neutrophils. ⁵¹ Mature and immature neutrophils can be differentially identified by distinct surface marker expression. While BM immature neutrophils are CD11b⁺CXCR4⁺CXCR2⁻CD101⁻, mature neutrophils are CD11b⁺CXCR4⁻CXCR2⁺CD101⁺. The expression of CD62L, CXCR2, and Ly6G is up‐regulated with maturation. ⁵¹ Once maturation is complete, mature neutrophils are released into the circulation. At homeostasis, immature neutrophils are found primarily in the bone marrow, while mature neutrophils are in circulation. The up‐regulation of CXCR2 on neutrophils is thought to be fundamental for their transit from the bone marrow to the circulation; however, a CXCR2‐negative neutrophil subset and/or an immature neutrophil subset has been noted by numerous groups following emergency granulopoiesis during infection, inflammation, or injury ⁵² suggesting other mechanisms of egress from bone marrow besides CXCR2.

The analysis of neutrophils in the circulation has also revealed different neutrophil phenotypes even at homeostasis, which is regulated in part by circadian signals. Over the course of 6 hours, mouse and human neutrophils lose CD62L expression and up‐regulate CD11b and CXCR4, which has been referred to as neutrophil aging. ⁵³ , ⁵⁴ , ⁵⁵ An important caveat is that L‐selectin is shed from the neutrophil surface and CD11b is up‐regulated during activation, which could give the false impression that newly activated neutrophils are aged. In fact, neutrophils can even up‐regulate annexin V during activation, appearing to be apoptotic. It is also important to remember that while neutrophils were thought to be short lived, some neutrophils may have extended lifespans of days. Indeed, neutrophils can be found in lymph nodes, spleen and other organs and so an aged neutrophil may be one that has spent significant time in various tissues. ⁵⁶ , ⁵⁷ , ⁵⁸

In an attempt to better characterize neutrophil subsets, some groups have proposed that neutrophils can be separated into subsets based on size and density via density gradient centrifugation. While normal density neutrophils (NDNs) are generally more mature with higher granule content and are found in the granulocyte fraction in a Ficoll density gradient, low‐density neutrophils (LDN) are more immature, contain fewer pre‐formed granules, and are found in the mononuclear fraction (PBMC layer) of gradients. ⁵⁹ Although low‐density neutrophils are commonly associated with diseases (and are thought to be enriched for younger or immature neutrophils), they were also isolated from the blood of healthy humans, showing phenotypic (surface marker expression) and functional differences when compared to normal density neutrophils isolated from the same individuals. ⁶⁰ Again, it is worth noting that activated neutrophils that have released some or all of their granules could be misconstrued as immature neutrophils. Clearly, there is no optimal way to characterize these subsets and caution must be exercised when making conclusions.

Finally, the bone marrow may not be the only source of neutrophils during perturbations. Neutrophils are also found in tissues, either marginated or in the parenchyma, showing tissue‐specific phenotypes. In the spleen, there are two populations of neutrophils in the parenchyma (mature and immature) and both seem to be critical to eradicate Streptococcus pneumoniae. ⁵⁷ Recent work has suggested that during brain injury, monocytes can be selectively released from spleen and indeed there may be a similar selective neutrophil recruitment from spleen. ⁶¹ Whether these neutrophils differ significantly from bone marrow is not entirely clear although one sub‐population appears to be intermediate (not high) in Ly6G, Ckit positive and as such an immature source of these myeloid cells. ⁵⁷ In the lungs, there are “resident” marginated neutrophils, which are poised to rapidly up‐regulate CD11b to crawl and phagocytose bacteria during systemic infection. ⁵⁸ In another study, Ballesteros et al. ⁴⁷ showed that neutrophils acquire heterogeneity at the chromatin, RNA, and receptor levels as they enter specific naive tissues, performing tissue‐specific functions. These examples provide evidence that neutrophils are not a homogenous population and also suggest that they can be differentially primed by tissue‐derived signals, as is the case for other cells, such as macrophages to perform an array of functions beyond just pathogen killing. ⁵¹

6. NEUTROPHIL HETEROGENEITY DURING SARS‐COV‐2 INFECTION

The number of studies using unbiased high‐resolution technologies, such as mass cytometry by time of flight (CyTOF), single‐cell RNA‐seq (scRNA‐seq), and single‐cell assay for transposase‐accessible chromatin (ATAC)‐seq, has increased exponentially in the last few years. These studies have been fundamental to the advancement of knowledge about neutrophil diversity during different contexts of disease including COVID‐19 pathogenesis. However, there is still a lack of knowledge regarding the function of each of these subtypes, and most studies propose rather than demonstrate potential functions. This remains a huge gap in understanding what neutrophil subtypes do. In this section, we will discuss the recent advances in understanding neutrophil heterogeneity in the context of SARS‐CoV‐2 infection.

6.1. Neutrophils in the circulation

Most of the initial studies analyzing blood samples from patients with COVID‐19 have reported an increase in the absolute numbers of circulating neutrophils in patients when compared to healthy donors. More recently, further studies have shown that this is mainly dependent on the severity of disease. While no significant differences were reported between healthy controls and individuals with mild COVID‐19, most of the studies with severe COVID‐19 patients reported a significant increase in neutrophil counts in the circulation. ²⁹ On the contrary, when COVID‐19 patients with severe disease were compared to patients with bacterial ARDS, the latter showed higher neutrophil counts when compared to COVID‐19 patients. ⁶² , ⁶³ These data indicate that, while the extent of neutrophilia observed during SARS‐CoV‐2 infection could be used as an indicator of disease severity, in its most severe forms, not all reserves of neutrophils are released. Whether this is intentional or a result of something perpetrated by the virus remains uncertain.

As mentioned above, neutrophils can be separated into low and high density based on their size and density. The low‐density neutrophils are found in the mononuclear cell (PBMC) layer of a density gradient (Ficoll or Percoll), while normal density neutrophils are found in the lower granulocyte layer, and therefore, this becomes a critical consideration to ensure that some neutrophils are not missed during harvesting. As such, it is not surprising that whole blood of patients might generate more neutrophil subtypes than studies where only the PBMC or granulocyte layers were harvested for studies. This must be kept in mind when trying to capture a full picture of circulating neutrophil subsets. Wilk et al., ⁶⁴ using scRNA‐seq of PBMCs, showed a “developing” population of neutrophils in COVID‐19 patients that developed ARDS. These neutrophils expressed genes encoding neutrophil granule proteins but did not express genes encoding canonical neutrophil markers such as CXCR2 consistent with these being immature neutrophils. Whether more mature neutrophils, normally found in the granulocyte layer, were missed in this study is not clear, but the study does highlight a novel emerging population of neutrophils in SARS‐CoV‐2 patients. Schulte‐Schrepping et al. combined single‐cell RNA sequencing and single‐cell proteomics of whole‐blood and peripheral‐blood mononuclear cells (PBMC), identifying distinct populations of low‐density immature neutrophils in the PBMC fraction from severe‐COVID‐19 patients but not from patients with mild disease. Their analysis revealed eight clusters of neutrophils during severe disease, which were divided into immature (pro and pre‐neutrophils) (CD81⁺SPN, FUT4, CD63, CEACAM8, ITGAM, CD10) or mature neutrophils (CXCR2, FCGR2A‐CD32, FCGR1A‐CD64, or MME‐CD10). Most clusters expressed high levels of genes encoding alarmins S100A8 and S100A9, genes that are involved in Neutrophil Extracellular Trap (NET) formation (MPO, ELANE, and PADI4), the latter being implicated in the pathogenesis of COVID‐19. There were also genes involved in the inhibition of T cell activation (ARG1 and CD274‐PD‐L1). ⁶⁵ Importantly, when they analyzed the whole blood of patients using CyTOF or scRNA‐seq, they found similar results: distinct clusters of immature neutrophils (pre and pro‐neutrophils; CD11b low expression), and mature activated neutrophils (low expression of CD62L and high expression of CD64), which mirrors findings from others groups. ⁶⁶ In addition, both immature and mature neutrophils showed elevated PD‐L1 expression, indicating a potentially suppressive function of those granulocytes. This was distinct from samples of SARS‐CoV‐2‐negative patients with flu‐like symptoms, who showed a decreased expression of PD‐L1 on mature neutrophils, with barely any immature neutrophils. ⁶⁵ Similar results were seen in a different study using whole blood and granulocyte samples, wherein the authors found signatures of activated neutrophils and potentially suppressive neutrophil function in severe COVID‐19 patients. ⁶⁷ Interestingly, in a longitudinal study, it was demonstrated that patients who recovered from COVID‐19, besides having neutrophils that returned to baseline levels of circulating neutrophils, also showed a reduction in PD‐L1 expression on neutrophils by the time of discharge. ⁶⁸ Whether the overall signature or program of neutrophils returns to that observed during homeostasis following recovery from COVID‐19 is still not entirely known and whether the immunosuppressive phenotype persists in long‐COVID also is unclear. Certainly in bacterial sepsis it is well known that a subset of patients develop a long sepsis phenotype and are much more likely to develop subsequent often unrelated infections suggesting some long‐term impact on neutrophil function.

Neutrophils with an IFN signature (IFN‐stimulated genes—ISG) have been reported in various infections with fungi, parasites, virus and even during rodent and human malaria, where they have been reported to be closely associated with liver pathology. ⁶⁹ , ⁷⁰ , ⁷¹ ISG neutrophils were also associated with severe COVID‐19 pathogenesis. In the study by Schulte‐Schrepping et al., ⁶⁵ the authors showed that ISG neutrophils are present in the circulation of patients with severe COVID‐19 disease and specifically expressed PD‐L1. The authors also suggested a possible immunosuppressive function of these cells during severe disease. In another study, Sinha et al. also identified a population of neutrophils with an IFN+ signature, and it was suggested that these cells were associated with the pathology observed in COVID‐19. ⁶² Indeed, patients that were treated with Dexamethasone showed a depletion of the population of IFN‐signature neutrophils and an expansion of immature neutrophils. In this case, the immature IFN negative neutrophils expressed high levels of arginase 1 (ARG1) and Annexin A1 (ANXA1), suggesting a possible immunomodulatory function for the IFN negative cells. Moreover, dexamethasone treatment also reduced protein levels of S100A8, S100A9, SERPINA1, SERPINA3, ORM1, LBP, VWF, PIGR, AZGP1, and CRP, which were previously identified as circulation‐based biomarkers distinguishing severe versus mild COVID‐19 cases. ⁶² It is intriguing to hypothesize a link between better outcomes in patients treated with dexamethasone (RECOVERY CR) and the demonstrated reduction in IFN‐signature neutrophils. However, to our knowledge, this causal effect has not been directly shown and the exact role of these cells in the context of the SARS‐CoV‐2 infection remains unclear. Although the study by Sinha et al. suggests that these cells are associated with COVID‐19 pathology, another study comparing mild with severe‐COVID‐19 showed that patients with mild COVID‐19 exhibited higher expression of ISG‐neutrophils when compared with neutrophils from severe patients. ⁷² Siwicki and Pittet interpreted these data to mean that severe disease could occur even when ISG‐neutrophils are absent and begged the question whether they played any causal role. ⁷³ We would submit that while these documentations of heterogeneity of neutrophils are important, more studies are now necessary to understand the role of these different subsets, for example, whether IFN‐neutrophils should be a potential target for the treatment of COVID‐19 pathology.

There are a few groups that have tried to elucidate the role of some of these subsets. In a study published by Morrissey et al., the authors reported the emergence of a low‐density population of neutrophils expressing intermediate levels of CD16 in the circulation of patients with COVID‐19. Functional assays revealed that these cells had an increased capacity to phagocytose bacteria when compared to CD16 high neutrophils; released spontaneous neutrophil extracellular traps (NETs); and produced TNF and IL‐6. In addition, platelets that interacted with CD16 intermediate neutrophils showed increased expression of activation markers, which was correlated with increased D‐dimer levels (activated thrombin levels) in severe COVID‐19 patients. ⁷⁴ In another study, mass cytometry was used to profile whole blood from patients with COVID‐19 ARDS, and compared this to blood from patients with non‐COVID‐19 ARDS. Neutrophils were the dominant immune cell in the blood of all patients, and the analyses revealed an augmented expression of CD11b, CD66b, and CD11a and reduced CD62L and CD107A on neutrophils from patients with COVID‐19 ARDS, consistent with a more activated phenotype. Furthermore, the results showed that neutrophils from patients with COVID‐19 displayed higher basal ex vivo NET release compared with neutrophils from healthy donors. A positive correlation was observed between the abundance of a cluster of mature and activated neutrophils and higher production of NETs, suggesting a possible pathogenic role for this population during COVID‐19. ⁶³ However, all of these studies are still only correlations and neutrophil subset modulation is needed to determine which dominant neutrophil subset would lead to improved health.

6.2. Neutrophils in the airways

Under normal conditions in healthy mice and humans, there are almost no neutrophils in the alveoli and the environment is dominated by alveolar macrophages. These cells move from alveolus to alveolus constantly “vacuuming” these compartments removing any debris and pathogens. In fact, intratracheal administration of up to one million Staphylococcus aureus or Pseudomonas aeruginosa was insufficient to recruit neutrophils as the alveolar macrophages efficiently removed these pathogens. ⁷⁵ Neutrophils are recruited to the lung tissue and to the alveolar space during significant infection to help the alveolar macrophages in the clearance of bacteria, and other pathogens, through phagocytosis and/or the generation and release of ROS and NETs. However, these neutrophils are often highly activated by PAMPs or DAMPs, and will cause inadvertent vascular and alveolar damage. It is therefore not surprising that increased levels of neutrophils in the lungs of COVID‐9 patients were associated with adverse symptoms including ARDS. Using single‐cell RNA sequencing, Liao et al. ⁴² demonstrated that in patients with severe COVID‐19, but not with mild or moderate disease, neutrophils were present at high frequencies in the bronchoalveolar space and these cells had a similar phenotype to circulating neutrophils, showing high expression of CXCR4, as well as S100A8 and S100A9, associated with an activated state. ⁷⁶ Conversely, in another study, neutrophils were also present in the alveolar space of patients with moderate COVID‐19, and again at lower levels than in samples from critical COVID‐19 patients. Critical patients showed an increase in the levels/signaling of chemokines and chemokine receptors involved with neutrophil recruitment to the airways. ⁷⁷ , ⁷⁸ The chemokine signature involved with neutrophil recruitment to the lungs was predominantly expressed in the macrophages and epithelial cells from the pulmonary tissue of COVID‐19 patients. In one study, in comparison with patients with bacterial pneumonia who were PCR negative for SARS‐CoV‐2, neutrophils were more frequent in the bronchoalveolar lavage of patients with severe COVID‐19. ⁷⁹ This is somewhat surprising as bacterial pneumonia is generally thought to be one of the most potent inducers of neutrophil recruitment. In the study by Wauters et al., five different clusters of neutrophils were found in the bronchoalveolar lavage of patients, and these were separated into “progenitor” neutrophils (cluster 1: CXCR4, CD63, VEGFA, and cathepsins), immature neutrophils (cluster 2: LTF, LCN2, MMP8/9, PADI4, and ARG1), inflammatory mature neutrophils (cluster 3: IFN‐induced genes, S100A8/9, S100A9, and S100A12), and cluster 4—(CXCL8, CCL3, CCL4), and hybrid neutrophils (cluster 5: MHC class II and complement activation genes). Progenitor and inflammatory neutrophils were the clusters most abundant in the lungs of patients and showed increased frequency in COVID‐19 versus non‐COVID‐19 patient samples. ⁷⁹ Although similar populations of neutrophils are also found in the circulation of SARS‐CoV‐2‐infected patients, it is still not clear whether there are specific signals that determine which neutrophil populations are recruited to the lungs, or whether in this critical situation all neutrophils are recruited en masse into the airways to try to reduce the infection.

Analysis of lung tissue from post‐mortem individuals also revealed neutrophil infiltration in the lungs of patients with severe COVID‐19. RNA sequencing and mass‐spectrometry showed an up‐regulation of alarmins, such as S100 A8, A9, A11, A12, and P, in the lungs of SARS‐CoV‐2‐infected patients in comparison with control samples (healthy lung regions from individuals with cancer). This could reflect the signature of the neutrophils that infiltrated the lungs, similar to signatures observed with single‐cell studies in the circulation and bronchoalveolar lavage of COVID‐19 patients. ⁸⁰ Moreover, analysis of lung transcriptomes from rhesus macaques and mice infected with SARS‐CoV‐2, but not other viruses, also revealed the up‐regulation of S100A8 and genes involved in neutrophil chemotaxis. ⁸¹ Interestingly, Guo et al. demonstrated that SARS‐CoV‐2 infection in mice resulted in infiltration into the lungs of what they termed “aberrant” neutrophils, which expressed variable levels of Ly6G and low levels of Cxcr2 and Fcgr3. It was proposed that S100A8 was important for the recruitment of these aberrant neutrophils in the lungs of mice. Inhibition of S100A8 binding to TLR4, or blocking TLR4 signaling, prevented the infiltration of these specific neutrophils into the rodent lungs, leading to improved survival of mice infected with SARS‐CoV‐2. ⁸¹ This is one of the few studies that provides causal data for a subset of neutrophils.

In summary, most studies show an increase in the number of activated immature or mature neutrophils and the number of subsets during COVID‐19. The presence of immature neutrophils in the periphery or in the lungs can be associated with emergency granulopoiesis, which occurs during severe insults as a means to release more neutrophils to the circulation to deal with increased demands. Although several differences are found between the different studies, it is important to keep in mind that this can be due to a difference in patient cohorts studied and in the stages of disease for each group of patients. It is known that the cellular landscape changes according to disease severity and over time during the course of COVID‐19 disease. ⁸² While much has been learned about the diversity of neutrophils during SARS‐CoV‐2 infection, more causal studies are necessary in order to understand the precise functions of diverse neutrophil states. This may reveal potential targets to provide new therapeutics for the treatment of COVID‐19.

7. NEUTROPHIL EXTRACELLULAR TRAPS (NETS) DURING COVID‐19 PATHOGENESIS

Neutrophils extracellular traps have become an attractive target to try to reduce some of the injury associated with SARS‐CoV‐2. NETs are a weblike structure of DNA coated with histones, elastase (NE), myeloperoxidase (MPO), cathepsin G, and other proteins, which are released by neutrophils upon activation and have tremendous killing capacity of any pathogen or host tissue they contact. ³ Initial studies demonstrated that a key role of NETs is to immobilize or trap and kill bacteria extracellularly due to the presence of anti‐microbial molecules. ⁸³ Indeed, in vitro studies showed that neutrophils activated with PMA release NETs, which are able to trap and kill microorganisms due to the delivery of a high concentration of anti‐microbial molecules. ⁸³ In vivo, it was possible to visualize NETs using intravital microscopy in liver sinusoids after the administration of bacteria and/or virus in mice. ⁹ , ²³ Interestingly, disruption of NETs by treatment with DNase resulted in increased dissemination of either pathogen. ⁴³ , ⁸⁴ PMA‐activated neutrophils were able to capture and neutralize the Chikungunya virus and HIV‐1 in vitro. ²⁴ , ⁸⁵ In vivo, DNase treatment to inhibit NETs resulted in the enhanced susceptibility to the Chikungunya virus‐infected mice. ⁸⁵ Therefore, tampering with NETs would seem detrimental to the host and perhaps enhancing NET production might provide benefit.

However, it is also known that an abundance NETs during infection or sterile inflammation can cause tissue damage and be highly detrimental to the host. ³ This runs beyond its direct toxic effects to tissues but when released in circulation, actually causes inappropriate activation of both coagulation and thrombosis occluding blood flow and oxygen, as well as nutrient supply to tissues. ⁸⁶ NETs injure endothelial cells which are important regulators of coagulation and thrombosis, and this likely further contributes to what is observed in SARS‐CoV‐2. Mechanistically, it has been proposed that histones have cytotoxic effects on host cells such as endothelium and that they are involved in formation of microaggregates in the circulation through the generation of thrombin and activation of platelets. ⁸⁷ Microaggregates can contribute to organ damage due to vascular occlusion of small or large vessels. ⁸⁶ In ARDS in the parenchyma, NETs also induce the death of epithelial cells, contributing further to the increase in inflammation and pulmonary injury. ⁸⁸ As such, this is the greatest problem in trying to treat patients who have systemic infections. The neutrophil effector mechanisms are critical to survival, but in these systemic inflammatory processes lead to major tissue injury. Interestingly, during ARDS induced by bacteria in mice, the partial reduction of NETs by DNase I treatment or only partial PAD4 deficiency (heterozygous) reduced acute lung injury and improved survival, whereas complete NET inhibition by PAD4 deficiency (homozygous deficiency) reduced lung injury but was counterbalanced by increased bacterial load and inflammation, without improving the survival of these mice. ⁸⁹

Many groups have suggested that NETs are involved in the pathogenesis of COVID‐19 but direct evidence is not readily available. Many studies have demonstrated that patients with severe COVID‐19 show elevated levels of NET markers in the circulation, including cell‐free DNA, MPO‐DNA complexes and citrullinated histone H3, when compared to healthy controls. ³⁴ , ⁹⁰ , ⁹¹ In addition, there is a positive correlation between serum level of these NET markers and the severity of COVID‐19. Indeed, MPO‐DNA complexes were correlated with the Sequential Organ Failure Assessment score (SOFA), and patients requiring mechanical ventilation (severe cases) showed higher levels of NETs in circulation when compared to patients with mild disease. ⁹¹ , ⁹² Moreover, neutrophils isolated from the circulation of COVID‐19 patients have been demonstrated to have an increased capacity for NET release under basal conditions when compared with neutrophils isolated from healthy controls ³⁴ , ⁶³ , ⁹³ ; the same increase was not observed in the case of non‐COVID‐19 ARDS. ⁶³ Altogether, these data suggest that the higher levels of NETs in the circulation of patients with COVID‐19 are perhaps due to increased neutrophil counts during severe disease, but more importantly due to neutrophils or subsets of neutrophils with increased capacity to release NETs. Yet the major question remains, are these NETs beneficial, detrimental or both?

Whether the increased capacity of neutrophils to release NETs is temporary or persists even after disease remains unclear. In one study with a small cohort (n = 4 patients) being analyzed, patients who recovered from COVID‐19 did not show a decrease in NET biomarkers relative to the time of disease onset, ⁹² which is in contradiction to another study, also with a small number of subjects (n = 2), showing that NET levels decreased 4‐6 weeks after hospitalization. ⁹³ Based on the small sizes of these studies, it is hard to conclude whether patients who survive COVID‐19 are at a higher risk to develop NET‐associated complications including cardiovascular and infectious disease. It would seem reasonable that after the patients have recovered, and infection is no longer a concern, some anti‐NET treatment might not be detrimental and could prevent other untoward events.

Besides NETs being found in the circulation of patients infected with SARS‐CoV‐2, their presence has also been reported in the airways. Tracheal aspirate obtained from patients with severe COVID‐19 under mechanical ventilation, and lung sections from post‐mortem COVID‐19 patients, showed high levels of NETs in comparison with healthy control subjects. ³⁴ , ⁹³ Not surprising, but similar to what have been observed with other respiratory infections, NET levels were higher in the airways of patients than in their circulation. ³⁴ , ⁹³ , ⁹⁴ Indeed, once neutrophils leave the circulation and enter tissues, it is thought that they increase their capacity to produce NETs. ⁹⁵ Because there is lots of damage in the tissue proper of the lung, NETs could be mediating this lung pathology that is observed during COVID‐19. Likewise, in another study using transcriptomics and proteomics, it was demonstrated that many genes involved in the generation of NETs were up‐regulated in the lungs of patients who succumbed to COVID‐19. However, low levels of SARS‐CoV‐2 RNA were found in the lungs of these patients, leading the authors to suggest that, rather than SARS‐CoV‐2 virus being responsible for NET formation, it is possible that NETs were triggered by some other host cell. ⁸⁰ In fact, platelet factor 4 expression was increased in those patients, and the authors hypothesized that platelets were involved in the NET formation. This is reminiscent of the original work of NET release in bacterial sepsis involving platelets in both humans and mice. ⁴³ , ⁸⁴ By contrast, levels of NETs have also been observed to correlate with the SARS‐CoV‐2 RNA load in sputum of patients, ⁹⁴ suggesting that high titers of the virus itself may stimulate increased NETosis. Furthermore, Veras et al. ³⁴ showed that viable, but not inactivated, SARS‐CoV‐2 was able to induce NET release in vitro when healthy neutrophils were incubated with the virus.

Interestingly, the authors found SARS‐CoV‐2 antigens within circulating neutrophils of COVID‐19 patients, and these SARS‐CoV‐2+ neutrophils were more efficient at producing NETs ex vivo in comparison with SARS‐CoV‐2‐negative neutrophils. ³⁴ Likewise, in another study using single‐cell deep‐immune profiling of bronchoalveolar lavage fluid from patients with COVID‐19, neutrophils were positive for the nucleocapsid protein (N) of SARS‐CoV‐2; pathway analysis on differentially expressed genes revealed that N was mostly present on a cluster of mature neutrophils, which showed enrichment for IFN‐signaling. Whether those cells were more prone to release NETs in the bronchoalveolar space was not investigated. ⁷⁹ Overall, although in vitro experiments are important for the understanding of some mechanisms, whether SARS‐CoV‐2, platelets, chemokines, cytokines, or a combination of these factors are responsible for triggering NET formation in the circulation and tissues of patients with COVID‐19 are still under debate. It should also be considered that the incubation of healthy neutrophils with plasma from patients with COVID‐19 was not able to induce NET release by these cells in a study published by our group, ⁶³ which is contradictory to studies from other groups. ³⁴ , ⁹³ Whether this is due to differences in the cohort of patients, severity of COVID‐19, or other variables is not clear. However, these studies ³⁴ , ⁹³ show that substances present in the circulation of some patients infected with SARS‐CoV‐2 can induce NET release by neutrophils.

The mechanisms involved with NET release have been extensively studied. NETs can be induced by several stimuli, such as different pathogen‐associated molecular patterns (PAMPs) and/or damage‐associated molecular patterns (DAMPs), that signal through Toll‐Like receptor (TLR) 2, TLR4, TLR7, or complement receptors, triggering the activation of protein‐arginine deaminase 4 (PAD4) in an NADPH oxidase‐independent or dependent manner. PAD4 citrullinates arginine on histones causing chromatin decondensation, while MPO and neutrophil elastase (NE), which are released from cytoplasmic azurophilic granules, translocate to the nucleus and also contribute to unfolding of chromatin. NE also cleaves gasdermin D (GSDMD), a factor that mediates pyroptosis in macrophages, ⁹⁶ in the cytosol to its active form (N‐terminal cleavage product:GSDMD‐NT ‐). GSDMD‐NT forms pores in the plasma membrane, but also mediates pore formation in nuclear and granule membranes, releasing chromatin to the cytosol and enhancing NE and other granular content release. Once chromatin enters the cytosol, it mixes with cytosolic proteins and then is released to the extracellular compartment. ³ Although this is a general view of the mechanism involved with NET release, it is important to mention that NETs can be formed via different pathways, for example without the need for PAD4 or NE. As such, in order to consider targeting NET formation in a therapeutic capacity, it is important to understand the mechanisms involved in NET formation within the unique contexts of specific diseases.

During COVID‐19, NET release was shown to be dependent on PAD4 and GSDMD. PADI4 gene expression was enhanced in neutrophils from patients with severe COVID‐19, ⁶⁵ and the release of NETs by circulating neutrophils harvested from COVID‐19 patients, or by healthy neutrophils activated with SARS‐CoV‐2, was reduced after incubation of these cells with a PAD4 inhibitor. ³⁴ Furthermore, the cleaved and active form of GSDMD (GSDMD‐NT) was found in circulating neutrophils from patients with severe COVID‐19, exceeding levels in neutrophils from patients with mild disease, which also produced lower levels of NETs. Interestingly, GSDMD‐NT was also found in association with NET structures in the lungs of patients who died from COVID‐19, and the inhibition of GSDMD in vitro abrogated NET release from neutrophils isolated from COVID‐19 patients. ⁹⁷ Altogether, these data indicate that PAD4 and GSDMD are potential targets for the inhibition of NET release by neutrophils during COVID‐19 pathogenesis.

8. NEUTROPHILS/NETS AND THROMBOSIS

In addition to pulmonary failure, one of the complications that affects patients with severe COVID‐19 is disseminated coagulation. In a process called thromboinflammation, cells of the innate immune system facilitate the activation of the coagulation cascade through the release of inflammatory mediators. ⁹⁸ Cytokine storm, neutrophils and NETs, platelet activation, and endothelial dysfunction have all been reported to induce abnormal clot formation and have been associated with this form of hypercoagulation. ⁹⁹ Indeed, neutrophils can contribute to thromboinflammation through the release of tissue factor, NETs, and other factors that have been associated with thrombosis in other diseases. NE and cathepsin G, for example, enhance tissue factor‐ and factor XII‐dependent coagulation in a process involving local proteolysis of the coagulation suppressor tissue factor pathway inhibitor (TFPI). ¹⁰⁰ The association of NETs with thrombosis was first reported around 15 years ago, when it was demonstrated that neutrophil extracellular traps in the vasculature could provide a scaffold and stimulus for thrombi formation. The authors showed that NETs or a component of NETs (histones) induced platelet aggregation in vitro, in a model where NETs were perfused with blood or platelets—an effect that could be inhibited by DNase. ¹⁰¹ Histones can activate platelets through the activation of TLR2 or TLR4 on platelets, leading to thrombin generation and microaggregate formation. ⁸⁷ Interestingly, TLR activation on platelets also enhances NET formation on neutrophils, triggering a positive feedback loop. In addition, in in vivo mouse models of sepsis, platelet aggregation, thrombin activation, and fibrin clot formation were shown to be dependent on NET release, and inhibiting NETs with DNase or by using PAD4‐deficient mice resulted in improved vascular perfusion. ⁹ Furthermore, NETs were found within microthrombi from septic patients, and were shown to cause vascular occlusion and damage to the lungs, liver, and other organs in experimental model of sepsis. ⁸⁶ Altogether, the results discussed above indicate that neutrophils and NETs can be involved in thrombosis formation both directly and indirectly.

During COVID‐19, around 60% of the patients with severe disease developed COVID‐19‐associated coagulopathy (CAC), which is characterized by elevated d‐dimer levels in the circulation, unchanged or decreased platelet count, decreased prothrombin time, and an increased risk of thrombosis. ⁷⁴ Moreover, non‐survivors presented higher d‐dimer levels when compared to survivors, and around 70% of non‐survivors met the criteria for disseminated intravascular coagulation during their hospital stay, in contrast to 0.6% of survivors. ¹⁰² Interestingly, COVID‐19 patients with thrombotic events presented higher levels of NETs and calprotectin (S100A8 and S100A9) in the circulation when compared to non‐thrombotic SARS‐CoV‐2‐infected individuals, and there was a positive correlation between the peak of these mediators and d‐dimer levels in the thrombotic patients. ¹⁰³ In addition, in lung autopsies of patients who died from COVID‐19, platelets were found to be colocalized with citrullinated histone H31 and neutrophils undergoing NET formation within pulmonary microthrombi. Also, there were higher levels of neutrophils‐platelet aggregates in the circulation of patients in comparison with healthy controls. ⁹³ In another study, besides the presence of clots containing neutrophils and NETs (identified by NE and citrullinated histone H3‐double positivity) in lung vasculature, analysis of liver and kidney revealed similar NET‐containing microthrombi in damaged tissue areas. ¹⁰⁴ Likewise, platelets, fibrin, neutrophils, and NETs were also found in cardiac micro vessels from COVID‐19 patients, ¹⁰⁵ and neutrophil numbers were increased in cerebral thrombi from COVID‐19 patients in comparison with uninfected stroke patients. ¹⁰⁶ While these are associative studies and do not show causality, the data suggest that neutrophil‐ and NET‐containing microthrombi may be driving multi‐organ damage in patients with severe COVID‐19. Based on these findings, it is tempting to propose that targeting neutrophils and NETs may be an effective strategy for reducing complications related to thrombosis in patients with severe COVID‐19. However, to definitively show whether neutrophils and NETs are driving microthrombus formation or whether their presence within microthrombi is deleterious, additional experimental studies are required. Figure 1 summarizes how neutrophils can contribute to the COVID‐19 pathology.

FIGURE 1.

Neutrophil contribution to the COVID‐19 pathology. (1a) SARS‐CoV‐2 infects and/or activates epithelial cells, resident macrophages, and (1b) endothelial cells surrounding the alveolar space. (1c) These events lead to the release of inflammatory mediators, such as chemokines and cytokines. (2) The activation of endothelial cells by inflammatory mediators or by the virus itself, and/or the presence of inflammatory mediators systemically leads to the activation of immature and mature neutrophils in the circulation. Immature and mature neutrophils were previously released from the bone marrow in a process of emergency granulopoieses. Activated neutrophils start releasing factors that can damage the host cells in the circulation. (3) Activated platelets can activate neutrophils, which can then release NETs. NETs can cause endothelial cell damage, which can then induce abnormal clot formation through the release of tissue factor and other mediators. (4) Histones in NETs can activate platelets through the activation of TLR receptors on platelets, leading to thrombin generation and microaggregate formation. (5) SARS‐CoV‐2 present in the circulation can also activate or infect circulating neutrophils and platelets, contributing further to NET formation. (6) NET can also be found within microthrombi in the circulation of patients. (7) Activated mature or immature neutrophils are recruited to the alveolar space. (8) In the alveolar space, neutrophils can also be infected or activated by SARS‐CoV‐2, (9) leading to NET formation

9. THERAPEUTIC APPROACHES TARGETING NEUTROPHILS

In light of the fact that neutrophils could potentially play a deleterious role in the pathology of COVID‐19, targeting their recruitment to infected tissues or their effector functions could represent a promising strategy for therapeutic intervention.

Neutrophil recruitment to different tissues is mainly mediated by selectins expressed on endothelial cells and their counter‐receptors on neutrophils, followed by the activation of integrins. Integrins are responsible for the firm adhesion of leukocytes on endothelial cells, transendothelial migration, and chemotaxis. It is well characterized that selectins (E‐ and P‐selectin) and integrins (CD18 and CD11b) participate in the recruitment of neutrophils in vascular beds present in the muscle, mesentery, and skin, for example. ² However, the inhibition of those molecules does not abrogate neutrophil recruitment to the lungs during pulmonary inflammation. Recently, a non‐canonical molecule, dipeptidase 1 (DPEP1), was described as an important adhesion molecule expressed in the pulmonary vasculature of mice and humans, and was shown to be involved in neutrophil recruitment to the lungs during ARDS in mice. ¹⁰⁷ Whether DPEP1 mediates neutrophil recruitment to the lungs during COVID‐19 is not known; however, if involved, it could represent a promising therapeutic target. Indeed, a clinical trial is now targeting DPEP1 in moderate‐to‐severe COVID‐19 patients (Clinical‐Trials.gov NCT04402957). If this approach has an impact, experimental studies should be carried out to determine the effects of DPEP1‐targeting on neutrophil recruitment to the lungs during COVID‐19 disease.

Chemokine receptors and chemokines mediate the chemotaxis of leukocytes to sites of inflammation and infection. Neutrophils express mainly receptors of the CXCR family of chemokine receptors, including CXCR1, CXCR2, and CXCR4, which respond to CXC chemokines, such as CXCL2, CXCL8, and CXCL12. ¹⁰⁸ , ¹⁰⁹ During COVID‐19, it has been demonstrated that increased numbers of neutrophils in the alveolar space of patients with severe COVID‐19 are associated with increased levels/signaling of chemokines/chemokine receptors involved with neutrophil recruitment to the airways. ⁷⁷ , ⁷⁸ More specifically, CXCL8 released from myeloid cells was more pronounced in severe cases of COVID‐19, being associated with elevated numbers of neutrophils in the circulation and in the lungs of patients. ¹¹⁰ Interestingly, CXCL8 may also induce the formation of NETs by human neutrophils, ⁸³ which could contribute further to lung damage. In this way, targeting CXCL8 or its receptors (CXCR1 and CXCR2) could potentially represent a therapeutic approach for mitigating COVID‐19 pathogenesis in severe cases of disease. Interestingly, a phase 2 clinical trial with a CXCR2 antagonist showed the potential to reduce inflammation and to improve lung function in chronic obstructive pulmonary disease (COPD). ¹¹¹ Currently, an ongoing phase 2 clinical trial is using anti‐CXCL8 on patients with severe COVID‐19 disease (NCT04347226).

Besides targeting neutrophil recruitment to the lungs, targeting their function is also a possible strategy to modulate neutrophil‐induced tissue damage during COVID‐19. As discussed above, one of the effector functions of neutrophils that has been associated with disease pathology in COVID‐19 is the release of NETs. Interestingly, the disruption of NETs with inhaled recombinant DNase treatment improved lung function in cystic fibrotic (CF) mice and helped to solubilize sputum from CF patients. ¹¹² In SARS‐CoV‐2‐infected K18 mice, the treatment with subcutaneous DNase I improved clinical score of disease and reduced lung injury (Veras et al., BioRxiv). ¹¹³ In COVID‐19 patients, clinical trials are being conducted with aerosolized DNase treatment (NCT04541979) as well as DNase in combination with Baricitinib (Janus Kinase inhibitor) and Tocilizumab (anti–interleukin‐6) (NCT05279391). The combination of DNase with other drugs already used for the treatment of COVID‐19, such as dexamethasone, or Baricitinib and Tocilizumab, represents a promising multi‐pronged strategy. Indeed, although treatment with dexamethasone during severe COVID‐19 shows a moderate benefit, patients treated with dexamethasone did not show reduction in NET production, ⁶³ suggesting that targeting NETs in addition to dexamethasone treatment could further improve disease pathology. Furthermore, dexamethasone may itself work in‐part by affecting neutrophil function: as discussed before, patients treated with dexamethasone showed a reduction in the population of IFN‐signature neutrophils and an expansion of immunosuppressive immature neutrophils, ⁶² raising the possibility that depletion of IFNactive neutrophils might be a mechanism by which dexamethasone is protective. Additional work will be required to determine the precise role of IFNactive neturophils during COVID‐19 disease and the ways in which dexamethasone treatment affects this biology.

Gasdermin D (GSDMD) is involved with NET formation during COVID‐19. Therefore, the inhibition of GSDMD could represent a strategy for the reduction of NETs in COVID‐19 patients. Disulfiram is a medication used in patients with alcoholism and inhibits the enzyme acetaldehyde dehydrogenase. In addition, disulfiram also inhibits gasdermin D. ¹¹⁴ During sepsis, disulfiram was shown to prevent the formation of NETs, protecting mice from organ damage and increasing survival. ⁹⁷ In the context of COVID‐19, incubating neutrophils isolated from patients with disulfiram inhibited the release of NETs by these cells, and this effect was associated with the abrogation of GSDMD activation. ⁹⁷ Furthermore, the treatment of SARS‐CoV‐2‐infected K18 mice ⁹⁷ or hamsters ¹¹⁵ with disulfiram reduced circulating levels of NETs, as well as inflammation and lung damage. Blocking GSDMD during COVID‐19 also inhibits pyroptosis and the consequent release of inflammatory cytokines by cells other than neutrophils, such as macrophages. ¹¹⁶ Interestingly, an observational study based on clinical records from the national US Veterans Affairs healthcare system revealed a reduced risk of SARS‐CoV‐2 infection and deaths in individuals treated with disulfiram. ¹¹⁷ Whether this effect was related to NET or cytokine release, or both, warrants further investigation. Currently, there is one ongoing clinical trial with disulfiram and COVID‐19 (NCT04485130).

Besides the up‐regulation of genes involved with NET formation, genes encoding alarmins S100A8 and S100A9 are also elevated in neutrophils from COVID‐19 patients. These molecules are generally up‐regulated during inflammation, being secreted by neutrophils and other cells, then acting as DAMPs via a TLR4 receptor‐dependent mechanism. This further increases inflammation and promotes additional neutrophil recruitment into the inflamed site. In a study published by Guo et al., ⁸¹ the authors showed that inhibition of S100A8/A9 using a drug called Paquinimod in mice infected with SARS‐CoV‐2 protected mice from lung damage and decreased weight loss. Mechanistically, it was proposed that Paquinimod inhibited the accumulation of aberrant neutrophils into the lungs of mice—a cell type that was associated with disease pathology. Altogether, the inhibition of alarmins represents a promising strategy to prevent the accumulation of deleterious neutrophils in the lungs.

10. PERSPECTIVES AND SUGGESTIONS OR DISCUSSION

The outbreak of the COVID‐19 pandemic at the beginning of 2020 prompted the scientific community at large to develop and create tools for the study of this disease, and led to the publication of many studies characterizing patients infected with SARS‐CoV‐2 in a very short timespan. Much has been accomplished regarding the characterization of neutrophils in patients; however, there is still little known about the roles for different subtypes of neutrophils during COVID‐19 pathology. For this, additional, larger studies involving both associative and ex vivo functional readouts may ultimately illuminate the likely roles for different neutrophil subtypes in this disease. This, in combination with in vivo mouse models of COVID‐19, may enable development of interventions that target specific populations of neutrophils that contribute to disease pathology. Another important point to be considered is that many of the studies used healthy individuals or individuals with no severe disease to compare with severe COVID‐19 patients. Although those studies do not belittle the findings about the individuals that are infected with SARS‐CoV‐2, the conclusions that are made can sometimes overrate the role of a determined cell or pathway as a particular characteristic of the COVID‐19 pathogenesis when it is not different from what happens for example with patients with sepsis of a different origin.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

Castanheira FVS, Kubes P. Neutrophils during SARS‐CoV‐2 infection: Friend or foe? Immunol Rev. 2022;00:1‐14. doi: 10.1111/imr.13175

DATA AVAILABILITY STATEMENT

Data sharing not applicable ‐ no new data generated.