Helminth infections affect almost 2 billion people worldwide, particularly in the developing world, and are a major cause of morbidity. Their high prevalence results in part from poor development of protective immunity despite ongoing exposure to these infections. Better understanding of the mechanisms that contribute to long-term protection is crucial for establishing more targeted approaches to deal with disease and support vaccine development. Although rodent models of hookworm infection have been used to study secondary protection since the early 1900s1, understanding of the cell types that contribute to this protection and the tissue-specific mechanisms involved are still poorly understood. The protective ability of granulocyte populations, including basophils, eosinophils and mast cells, has been well studied in infection with various helminths2. Whether neutrophils, another granulocyte population that is rapidly recruited to almost all sites of infection and injury3, also have a role in protection has been poorly explored. In this issue of Nature Immunology, Chen et al. describe a previously unknown and unexpected role for neutrophils in 'training' macrophages to acquire a long-term protective function against helminth larvae as they transit through the lungs4.

Neutrophil function during infection has been investigated most extensively in response to bacteria; during such infection, they have been long thought of as terminally differentiated cells with a direct protective role3. Protection is mediated by their phagocytic ability, along with their production of a plethora of bactericidal factors, including reactive oxygen species, elastase and neutrophil extracellular traps (NETs)3. The recruitment of neutrophils is not limited to bacterial infection, however, as they swiftly transit to the site of pathogen entry in response to various types of invader, including helminths2,5,6. Although they may serve a direct role in killing larvae, one possibility is that here too their dominant function is protection against bacteria, as helminths can carry bacteria into tissues as a result of barrier breach during invasion2,6.

A less-well-explored aspect of neutrophil function is their ability to influence the function of cell populations with which they are closely localized during infection, such as monocytes-macrophages. Perhaps the best described example of neutrophils' modulating macrophage function is the uptake of apoptotic neutrophils (efferocytosis), which leads to monocytes-macrophages' taking on a proresolution activity7. During infection with Staphylococcus aureus, however, neutrophils can secrete factors that are able to differentiate macrophages toward an alternatively activated (M2) state8; this suggests that there are multiple mechanisms by which neutrophils can influence their innate cell neighbors during infection.

Chen et al. demonstrate that during primary infection with Nippostrongylus brasiliensis (a rodent helminth with a lifecycle similar to that of the human hookworm), neutrophils can 'instruct' lung macrophages to take on altered functions that confers on them the ability to rapidly kill larval stages of the parasite during a secondary challenge4 (Fig. 1). This macrophage 'memory' is still evident at 45 days after infection and around 30 days after clearance of the helminth. Enhanced killing is dependent on signaling via the α-chain of the receptor for interleukin 4 (IL-4) and the production of arginase-1, which have been described as protective during infection with various helminths2. In addition, expression of integrin αM (CD11b) is increased on lung macrophages following primary infection with N. brasiliensis and, using an in vitro larval killing assay, the authors show that this integrin is important in augmenting macrophage-mediated killing4. Unexpectedly, upregulation of the expression of arginase-1 and CD11b on lung macrophages, and hence secondary protection against the helminth, is critically dependent on the presence of neutrophils during primary infection. In agreement with those findings, sorted lung macrophages from mice infected with N. brasiliensis but depleted of neutrophils are unable to mediate protection against helminth challenge following intravenous transfer into naive hosts, unlike macrophages from control mice infected similarly but treated with isotype-matched control antibody.

Figure 1: Neutrophils 'train' lung macrophages to take on features that protect against helminth larvae during secondary infection.
figure 1

Kim Caesar/Nature Publishing Group

In a primary infection (left), following entry through the skin, N. brasiliensis larvae are poorly impeded by the immune system and transit to the lungs, where neutrophils are rapidly recruited. At this site, neutrophils take on an N2 phenotype that is associated with production of IL-13, as well as other factors, that 'train' macrophage function to support the killing of parasites (blue, before training; red, after acquisition of the M2 phenotype). However, larvae are not killed at this stage, and they transit to the gut, where during primary infection worms are completely expelled, in large part by enhancement of goblet-cell secretion and gut peristalsis. The macrophages in the lungs remain in a 'trained' state for at least 45 days, until there is a subsequent invasion of N. brasiliensis, at which time they kill the parasite in a manner dependent on CD11b and arginase-1 (Arg1). Although in this setting much protection is mediated in the skin (in an antigen-specific manner), these reprogrammed lung M2 macrophages provide a second level of protection that is independent of the adaptive immune system. In contrast to primary infection, during secondary challenge (right) the gut is not a key mediator of protection, as small numbers of worms survive to reach this site. Vertical bars (right and left margins) indicate minor (yellow) or major (red) sites of protection.

Full size image

The ability of neutrophils to 'train' lung macrophages is associated with their acquisition of a distinct global transcriptional profile early after infection, compared with that of lipopolysaccharide-elicited neutrophils at the same site. In particular, they have augmented expression of genes associated with type 2 immunity, including Retnla, Chi3l3 and Il13. Furthermore, coculture experiments reveal that production of the cytokine IL-13 by neutrophils obtained from mice after primary infection with N. brasiliensis is critical for favoring the enhancement of factors associated with the alternative activation of lung macrophages and upregulation of CD11b expression. Thus, in a manner analogous to M2 macrophages, neutrophils are described in this setting as taking on an alternatively activated ('N2') phenotype. One reason for the importance of these N2 neutrophils in long-term 'training' may be their large numbers and close proximity to monocytes-macrophages during larva-driven inflammation.

Acquisition of N2 features by neutrophils has been described before during infection with S. aureus8. Here, as in infection with N. brasiliensis, neutrophils are able to induce M2 polarization of macrophages, although this was attributed to IL-10 and the chemokine CCL2 (the monocyte chemotactic protein MCP-1). How neutrophils take on an N2 phenotype rather than classically activated (N1) phenotype was not investigated during infection with S. aureus8 and is not fully described during infection with N. brasiliensis in the current study4. It is likely that type 2 cytokines produced early by innate lymphoid cells (such as ILC2 cells) or factors derived from antigen-presenting cells mediate this switch. Alternatively, products secreted directly from helminths can result in altered function of innate cells9. Further exploration will be needed to establish how neutrophils enter into this state and whether N2 neutrophils are a common feature of various infections.

Perhaps one of the most surprising observations by Chen et al. is that acquired immunity to the larval stage of the parasite trafficking through the lungs does not require the adaptive immune system4. It is intriguing that appropriate 'training' of the innate immune system alone is entirely responsible for the protection observed in this model. Such 'training' of the innate immune system can provide protection against Candida albicans in the absence of an adaptive immune response10, and lung macrophages can be maintained in an immunologically altered state following viral infection11. The longevity of specific populations of macrophages residing in the lungs may provide an explanation for the lack of necessity for adaptive immune responses to provide secondary protection at this site. Although it is unclear from this study4 which population of macrophages is targeted by neutrophils, the lung alveolar macrophage population represents a likely candidate. Alveolar macrophages are an extremely long-lived population that can expand locally after infection12. It is well established that epigenetic regulation is critical for the M2 polarization of macrophages13. Therefore, long-term epigenetic reprogramming of these cells may underlie their modified protective state.

Whereas the data in the present study underscore the importance of 'training' lung macrophages in acquired immunity, other tissue sites are also critical in protecting against helminths. During infection with N. brasiliensis, the parasite first invades through the skin, then transits through the bloodstream to the lungs and ultimately travels to the gut, where the adult worm produces eggs for transmission to new hosts (Fig. 1). Following a primary infection, the majority of the protection against the helminth is mediated in the gut, in which secretions initiated by cells of the immune system, especially goblet cells, and peristalsis (so-called 'weep and sweep' responses) are involved in mediating complete expulsion of the parasite2. In a secondary infection, on the other hand, the skin has been described as a major (if not complete) bulwark of immunity5. At this tissue site, the adaptive immune system is critical in providing antigen-specific protection, as basophils armed with immunoglobulin E are required in favoring the M2 polarization of macrophages. Regardless of the setting, it appears critical that macrophages take on the M2 phenotype and, in particular, produce arginase5. A hypothesis for the differences between the skin and lungs in their requirement for the adaptive immune system is that the macrophages in the skin turn over much more rapidly than those in the lungs and thus long-term 'training' of skin macrophages is not possible. Even in secondary infection, any parasites that do survive transit through both the skin and lungs are expelled by the gut (Fig. 1).

Overall, this study4 highlights the importance of neutrophils beyond their well-described role as a pathogen-killing population during bacterial infection. Future research based on this study by Chen et al.4 will no doubt aim to elucidate whether the N2 differentiation and function of neutrophils can be manipulated to support long-term protection against helminth infection. What remains to be explored is whether the 'training' of macrophages by neutrophils is a common feature of lung inflammation and if targeting of this interaction could provide a novel therapeutic avenue for some of the vast array of respiratory diseases afflicting the world today, such as chronic obstructive pulmonary disease or asthma.