A GRIM fate for human neutrophils in airway disease

Nowhere in biomedical research does a cell type seem as simultaneously vital and vexing as the neutrophil. Human immunity deteriorates rapidly in the absence of neutrophils whereas their dysfunctional abundance can cause misery and harm, as exemplified by cystic fibrosis (CF), neutrophilic asthma, and chronic obstructive pulmonary disorders (COPDs). The goal of treating the underlying causes of these diseases without impairing immunity has been elusive, leading to calls for improved understanding of neutrophil plasticity and the possibility that it enables generation of pathological neutrophil subsets.^1,2 To this end, Forrest, Tirouvanziam, and colleagues report development of a tissue culture method that recreates several of the pathological phenotypes observed in airway neutrophils from CF patients, summarized by the memorable acronym “GRIM” for their granule-releasing, immunoregulatory, and metabolic attributes. Among the most intriguing and important aspects of their study is its demonstration of microenvironment as a determinant of functional responsiveness in human neutrophils, an epigenetic plasticity previously suspected as a contributor to CF² and other inflammatory diseases.³

New assays involving cultured human cells are a welcome addition to the existing animal models used in preclinical drug testing. This is because most, 80%, of all drugs that show promise in mouse or other animal studies go on to fail in phase I or II human clinical trials.⁴ Numerous conceptual advances in immunology have been made using mice,⁵ but it seems our species are not sufficiently alike to consistently predict clinical success. One recent comparison of neuro-inflammatory pathways undertaken to understand the high rate of failure of anti-inflammatory treatments for Alzheimer’s disease⁶ found that whereas up to 60% of inflammatory cell or cytokine interactions were similar in mouse and humans, 10–15% were “reversed.” The remaining interactions that were evaluated were unique to humans, with no known counterparts in the mouse. That additional methods of preclinical testing are needed should not be surprising given the estimated 100 million year evolutionary distance between rodents and humans.⁷ The report from Forrest et al. thus adds a potentially important tool to the drug development toolbox, as well as further evidence of neutrophil plasticity that is reshaping perceptions about their functionally dynamic responsiveness in health and disease.

Neutrophil accumulation in airway disease is somewhat unusual in that it involves transepithelial migration into the air-bearing lumen of the lung in addition to transendothelial extravasation from the circulatory system. To replicate the additional transepithelial step, Forrest et al. used a clever twist of standard 3D tissue culture methods in which a human cell line derived from secretory club cells, H441, was initially grown at an air-liquid interface on top of a porous, polystyrene scaffold that had been coated with collagen. After an airway-like epithelial surface had formed, the sealed cell layer and its supporting scaffold were turned upside down into culture medium containing dissociated sputum from CF patients (Fig. 1). Purified blood neutrophils loaded into the top of the newly exposed matrix of the scaffold were observed to migrate from within the matrix, through the sealed H441 cell layer and into the lower culture medium where they expressed phenotypes similar to those found in airway neutrophils collected from CF patients, such as increased intracellular reactive oxygen species (ROS), pinocytosis and cell surface CD63 (a marker of elastase exocy- tosis) and decreased CD16 (FcγRIIIB, the low affinity receptor for IgG that is shed from activated neutrophils). These “GRIM” neutrophil phenotypes did not appear if transepithelial migration was instigated by a control chemoattractant, the leukotriene LTB₄; nor did it appear if neutrophils were cultured directly with dissociated CF sputum. Hence, it was only the combination of transepithelial migration and exposure to factor(s) present in patient sputum that elicited GRIM phenotypic changes in human neutrophils.

FIGURE 1. A tissue culture model of neutrophil plasticity engendered by transepithelial migration.

A) The natural microenvironment adjoining lung airways in vivo. Circulating blood neutrophils extravasate through microvascular endothelium into fluid-filled interstitial spaces formed by a reticular network of collagen fiber bundles. Further migration into the mucus fluid lining of small airways requires additional transmigration through an epithelial layer containing proliferative basal cells, ciliated epithelial cells, and secretory club cells. B) An in vitro model of human neutrophil plasticity in obstructive airway disease. H441 cells are used as surrogates for club cells found in small airway bronchioles. After growth at an air-liquid interface the cell layer and its supporting scaffold are inverted into tissue culture medium containing dissociated sputum from patients with obstructive airway disease. Neutrophils that transmigrate through the H441 cell layer after being loaded into the upper scaffold, a porous polystyrene matrix coated with collagen, acquire pathological phenotypes and functions useful for drug discovery and preclinical evaluation

Another phenotypic component of the GRIM neutrophil hypothesis is increased expression of immunoregulatory effectors, arginase-1 and PD-L1, known to downregulate T cell responses. These immunoregulatory outcomes are not necessarily exclusive to CF disease because they eventually appeared in neutrophils that had transmigrated toward sputum from healthy control participants. Surface display of CD16 and CD63 also shifted after transmigration to control sputum, but again was markedly delayed. At least some phenotypes of GRIM neutrophils may therefore need to be considered as kinetic rather than absolute markers of pathological conditioning. Given this dynamic, and the fact that ex vivo systems often fail to achieve or maintain the same cellular steady states observed in vivo, selection of appropriate time points will be needed when deploying the new culture method for mechanistic studies of neutrophil plasticity or evaluation of drug candidates. As can often be said after a first report, “the proof will be in the tasting of the pudding” with respect to determining the value of the Tirouvanziam transepithelial culture system in basic neutrophil research or drug development.

Many interesting parameters beg to be assessed in the course of vetting the GRIM phenotype hypothesis. For example, which factors in CF sputum are responsible for pathological reprogramming of neutrophils? The authors note that CF sputum samples are easily collected because they are (unfortunately) coughed up in abundance and easily processed such that identification and reconstitution of the responsible factors is not needed for drug testing to begin. Systematic evaluation of factors in CF sputum to determine which drive GRIM phenotypic changes is nevertheless worth pursuing as a source of new therapeutic targets. Indeed, the knowledge that Gram-negative Pseudomonas aeruginosa establishes chronic infection in most CF patients led the authors to test CF sputum for evidence that LPS contributes to generation of GRIM neutrophils. They found that addition of an MD-2/TLR4 antagonist to the assay delayed and reduced generation of intracellular ROS by neutrophils but had no effect on cell surface CD16 or CD63. P. aeruginosa actively synthesizes LPS variants unique to the CF lung environment, each with distinct properties as TLR4 agonists. Some LPS variants primed superoxide release without affecting CD63 levels,⁸ which is consistent with the authors’ findings that TLR4 signaling was required for ROS generation, but not CD63 display, in GRIM neutrophils. Identification of CF sputum components that rapidly trigger exocytosis of azurophilic granules (CD63 display) or loss of FcγRIIIB (CD16) is far more feasible, now that pathological conditioning and generation of GRIM neutrophils can be performed routinely in ex vivo tissue culture.

Some attributes of GRIM pathological conditioning were observed when the authors tested sputum from patients with severe asthma or COPD, suggesting the transepithelial assay may useful in the context of other airway diseases. However, similar patterns were evident after neutrophil transepithelial migration toward sputum from healthy volunteers, if given enough time, indicating this important but preliminary finding needs further evaluation. Ongoing comparisons between outcomes observed in vivo with those generated in the ex vivo culture model are thus important to continue. Various parameters can be tweaked and tested to maximize recapitulation of the in vivo disease state, and may point the way to interesting new research directions as well. As noted, neutrophil accumulation in airway diseases involves transepithelial migration into airway spaces; however, extravasation from capillaries is also required, an element not included in the current culture system (Fig. 1). Could an endothelial vascular cell layer be introduced opposite the scaffold support for epithelial cells, and if so would it have any effect on neutrophil plasticity? It will similarly be important to understand the extent to which epithelial cell-type affects neutrophil plasticity and whether or not it is fully recapitulated by the H441 cell line. Lung epithelium contains multiple cell types, including regenerative basal cells, mucin-producing Goblet cells, secretory club cells, and ciliated epithelial cells, each of which have distinct capabilities as sensors of microbial infection and tissue damage.⁹ Primary lung epithelium will need to be tested in the assay, if not routinely then as a benchmark for initial evaluation of H441 and other potentially useful epithelial cell lines to determine which best recapitulate in vivo outcomes.

Finally, how does passage through a collagen-coated, porous matrix contribute to neutrophil plasticity? This component of the authors’ culture system is reminiscent of intriguing interstitial structures recently discovered in a wide variety of human mucosal tissues, including lung bronchioles.¹⁰ The structures were revealed by use of nonstandard tissue fixation and visualization techniques and consist of columns of collagen fibers framing fluid-filled spaces (Fig. 1). They are situated between mucosal epithelial surfaces and blood vessels, where neutrophils would encounter them immediately following extravasation. Hence, the authors’ use of a collagen-coated artificial matrix into which to load purified neutrophils may be beneficial because it replicates a more natural microenvironment for neutrophils to interact with each other or with other cell types that remain after purification, or by fostering multiple cycles of neutrophil adhesion and release. If so, understanding the interactions that set the stage for GRIM phenotypic changes in molecular detail could identify targets for therapeutic interventions that prevent neutrophils from becoming pathologically reprogrammed in the first place.

As for the “M” in GRIM, neutrophil metabolism indicative of aerobic glycolysis was elicited by transmigration to CF sputum but not by transmigration alone (to LTB₄) as revealed by increased glucose uptake, oxygen consumption, and lactate acidification. GRIM neutrophils had weak bactericidal activity despite their elevated levels of intracellular ROS, a key hallmark of CF disease. This pattern prompted the authors to test the novel hypothesis that pharmacological manipulation of glucose metabolism could restore a healthy balance between the inflammatory and antibacterial functions of neutrophils. Metformin is an agonist of the metabolic regulator AMP kinase (AMPK) used to moderate glucose metabolism in type II diabetes and has been proposed in the past as a potential treatment for the inflammatory consequences of ion channel impairment in CF.¹¹ The authors found that metformin was indeed effective in reversing or blocking oxygen consumption, intracellular ROS generation, and exocytic release of elastase. However, bactericidal activity became weaker—not stronger—indicating metformin is not likely to be successful as a CF therapeutic. Far from representing a failure, these metformin experiments illustrate the value of predicting pharmacological outcomes early in the drug development process, prior to costly and time-consuming failures in clinical trials.

The longstanding perception of neutrophils as terminally differentiated cells with a limited array of preset effector functions has been steadily eroded by evidence that they have longer lifespans and are more functionally responsive to changes in their environments than had been thought. One recent genomic analysis reached the astonishing conclusion that inter-individual variability in neutrophils is far greater than monocytes or naïve T cells, as indicated by a comparison of gene expression and DNA methylation patterns in FACS-purified cells from 125 healthy blood donors.¹² The extensive epigenetic plasticity revealed by this study probably means that neutrophils will continue to vex biomedical researchers for a long time to come. That’s the bad news. But the good news is development of innovative culture systems, such as that of Tirouvanziam and colleagues, shows progress is being made toward the goal of avoiding GRIM outcomes in late-stage CF and other obstructive airway diseases.