Editorial: Gazing forward while looking back

Abstract

Discussion of how transcriptional responses of neutrophils contribute to the resolution of inflammation, and direct studies of human innate immune responses.

The findings of Basran et al. [1], reported in this issue of Journal of Leukocyte Biology, highlight how our understanding of innate immunity has evolved in the past two decades and prompt simultaneously reflection on recent insights and anticipation of where future investigations should focus.

For millennia, humans have been fascinated by the phenomenon of inflammation in the context of infection and trauma [2], and since the observations of Metchnikoff on starfish phagocytes, studies have focused on how neutrophils provoke and promote inflammation as a hallmark of the acute innate immune response. Until relatively recently, the repertoire of neutrophil responses, such as exocytosis and NADPH oxidase activity, was limited to rapid responses that reflect engagement of signal transduction pathways, which trigger reorganization of cytoskeleton and membrane, redistribution of subcellular components, and recruitment of cytosolic proteins to membranes. Given that neutrophils are terminally differentiated cells, the absence of transcriptional activity directly related to innate immune response was neither news nor a surprise. Seminal studies describing production and release of CXCL8 in vitro [3] and expression of IL-1 genes by neutrophils [4] contributed to the then novel notion that neutrophils are versatile and plastic cells that not only produce newly formed cytokines to mediate crosstalk with cells of the innate and adaptive immune systems but also condition the evolution of the inflammatory process. Instead of passive elements that simply undergo apoptotic death followed by rapid and silent elimination by resident macrophages via efferocytosis, neutrophils actively participate in “inflammation resolution”, a tightly controlled, coordinated series of events whose successful execution culminates in termination of neutrophil influx and promotion of monocyte recruitment, programmed death by apoptosis and rapid clearance of infiltrating neutrophils that otherwise might propagate additional tissue damage, neutrophil and macrophage release of anti-inflammatory lipid mediators and reparative cytokines, and ultimately, regeneration of disrupted tissue structures [5]. Although it is clear that neutrophils contribute at several critical points in the cellular network that orchestrates the resolution of inflammation, how the sequential steps coordinate at the molecular level to affect resolution of inflammation and a return to the baseline state is incompletely defined. In this context, the work by Basran and colleagues [1] aims to elucidate the mechanisms whereby neutrophil recruitment is rapidly terminated before the inflammatory response has resolved.

After establishing that intradermal injection of endotoxin into the forearm of healthy volunteers prompts early recruitment of neutrophils in vivo and concomitant local accumulation of CXCL8 mRNA, the authors performed in vitro experiments to investigate how neutrophils modulate the amounts of local extracellular CXCL8, thereby preventing a continuous neutrophil recruitment into an active inflammatory site. Neutrophils primed with TNF-α or GM-CSF specifically scavenge CXCL8 by sequestering >75% at 8.3 nM after exposure for 24 h to varied concentrations of the chemokine. Under similar experimental conditions, however, both untreated and primed neutrophils remove neither CCL2 nor IL-1β from the culture media. CXCL8 clearance is dramatically, albeit incompletely, inhibited by the concomitant blockade of CXCR1 and CXCR2—the two CXCL8 neutrophil receptors constitutively displayed by neutrophils—thus demonstrating that CXCL8 sequestration is, in part, receptor-dependent. On the other hand, CXCL8 scavenging by neutrophils is greatly reduced if pretreated with TNF-α in combination with GM-CSF or in an in vitro model of endotoxin-induced inflammation [1]. Although it is plausible that the lack of CXCL8 clearance by TNF-α plus GM-CSF-pretreated neutrophils could reflect a down-regulation of CXCR1/2 expression, data suggest that CXCL8 scavenging does not occur under all conditions, but rather is a tightly regulated event. It would have been informative for Basran and colleagues [1] to have investigated LPS-primed neutrophils for CXCL8 clearance, given the recent report of transient down-regulation of CXCR1, CXCR2, and of FcγRII and FcγRIII in an established model of human endotoxemia [6]. Of course, measurements made in vitro enjoy precisely defined and controlled conditions—features that contrast with the marked spatial and temporal variability seen in vivo and undermine confidence in drawing correlations. Basran and colleagues [1] conclude that neutrophils may contribute in vivo to the regulation of the size of the inflammatory response by receptor-mediated clearance of CXCL8.

CXCL8 is not the only neutrophil-specific chemoattractant, and mechanisms that target CXCL8 itself and other neutrophil-specific chemoattractants and related receptors could contribute to the arrest of neutrophil recruitment during acute inflammatory responses, especially given that CXCL8 clearance was not fully neutralized by CXCR1/2 blockade [1]. For example, CXCL8 or other neutrophil-specific chemoattractants could silence cognate receptors via receptor desensitization or by receptor internalization, which in turn, may dramatically reduce receptor functionality. Alternatively, the in vivo biological activity of released CXCL8 could be compromised as a consequence of a number of cellular-derived actions, including its i) degradation by neutrophil-derived proteases; ii) post-translational modification, such as citrullination [7] by neutrophil-derived peptidylarginine deaminase; and iii) neutralization by specific neutrophil-derived soluble inhibitors. Together with the eventual down-regulation of CXCL8 expression by locally produced anti-inflammatory cytokines, such as IL-10, the ultimate consequence of these mechanisms would be a dampening of inflammation by reducing the local influx of neutrophils. Superimposed on these potential regulatory networks are tissue-specific mechanisms that may also modulate the local persistence of CXCL8. For example, experiments in rabbit models demonstrate that the clearance of CXCL8 differs dramatically in lungs and skin, with slow clearance from lungs and rapid clearance from skin [8]. Remarkably, in these latter experiments, migrating neutrophils shortened the CXCL8 half-life in skin but not lungs [8], an observation that could be related to the data by Basran and colleagues [1] on TNF-α plus GM-CSF-primed neutrophils or their in vitro model of endotoxin-induced inflammation. Irrespective of these considerations, the data by Basran and colleagues [1] suggest that neutrophil-mediated clearance of CXCL8 has the potential to be an important regulator of inflammation in vivo. Considered collectively, these data demonstrate that the cessation of inflammation is not a consequence of cellular exhaustion but rather, the result of successful execution of processes that are initiated just as proinflammatory responses begin. Orchestrated production and processing of secreted proteins and novel lipids culminate, when successful, in restoration of the resting, unstressed state.

Like Janus—the Roman deity of beginnings and transitions (Fig. 1)—neutrophils simultaneously look to the future as they embark on pathways to initiate inflammation, while looking to the past to engage programs to restore homeostasis. The report by Basran et al. [1] likewise directs the reader in opposite directions, focusing attention on principles gleaned from studies in the recent past (see above) and pointing forward to how investigations into innate immunity may proceed in the future. A cursory read of Basran et al. [1] might prompt an unobservant reader to wonder what is “new” here; weren't many of these events already defined? In fact, many of the same cells and cytokines identified herein were found earlier in studies using murine models, human cells in vitro, or tissue from patients. However, the experimental model used by Basran et al. [1]—intradermal injection of purified endotoxin into normal humans—provides a novel tool for other investigators to probe fundamental aspects of the human innate immune response. Their experimental system can be used to examine the constituents and the kinetics of neutrophil responses to locally administered agonists, such as endotoxin, in a safe fashion. As noted previously [9, 10], murine and human immune responses often differ, especially in the context of innate immunity [11, 12], and murine models of human disease can fail at the mechanistic level to mirror human disease pathogenesis (e.g., ref. [13]). To apply optimally and accurately putative principles of innate immunity to understanding clinical disease and to alleviating human suffering, findings in models need validation in human systems, either in vivo (where feasible, as in Basran et al. [1] or ref. [14]) or by exploiting bioinformatics and in silico systems more fully (see ref. [15]). In some situations, the use of updated, previously described methods, such as skin windows, can provide unexpected and informative results, exemplified by discovery of a rich transcriptional program committed to wound repair expressed by human neutrophils migrating into skin windows [16] or by the results reported herein.

Figure 1. Janus Bifrons, the Roman God of beginnings and transitions, as depicted in a statue from the Vatican Museum.