The immune response of inbred laboratory mice to Litomosoides sigmodontis: A route to discovery in myeloid cell biology

Abstract Litomosoides sigmodontis is the only filarial nematode where the full life cycle, from larval delivery to the skin through to circulating microfilaria, can be completed in immunocompetent laboratory mice. It is thus an invaluable tool for the study of filariasis. It has been used for the study of novel anti‐helminthic therapeutics, the development of vaccines against filariasis, the development of immunomodulatory drugs for the treatment of inflammatory disease and the study of basic immune responses to filarial nematodes. This review will focus on the latter and aims to summarize how the L sigmodontis model has advanced our basic understanding of immune responses to helminths, led to major discoveries in macrophage biology and provided new insights into the immunological functions of the pleural cavity. Finally, and most importantly L sigmodontis represents a suitable platform to study how host genotype affects immune responses, with the potential for further discovery in myeloid cell biology and beyond.


| INTRODUC TI ON
Filarial nematodes are thread-like tissue-dwelling roundworms that are transmitted by arthropod vectors and infect over 120 million people worldwide. [1][2][3] The most well-known of human filarial nematodes, Wuchereria bancrofti and Brugia malayi, reside in the lymphatics and are thus referred to as lymphatic filaria. The other major filarial nematode of human consequence, Onchocerca volvulus, forms nodules in subcutaneous tissue. However, all filaria including the focus of this review, Litomosoides sigmodontis, migrate through the lymphatics as part of their life cycle. 4 A clinically important feature of filarial life cycles is that sexually mature female worms produce microfilaria (mF) a larval form that are acquired by the arthropod vector via the blood or skin. 5 Infections which lead to detectable circulating microfilaraemia in the mammalian host are said to be patent.
Filarial nematodes are best known as the causative agents of the disfiguring and disabling lymphatic filariasis (lymphoedema, hydrocele and elephantiasis) and onchocerciasis (river blindness/ sowda), as well as the less severe loasis. 1,6,7 Clinical symptoms are generally a consequence of damage to lymphatic vessels or chronic immunopathology in infected tissue. 6 Infected individuals also suffer from periodic attacks of fever. 6 However, the simplistic causal relationship between filarial nematodes and immunopathology fails to capture the reality of infection on a population level. 8 In areas endemic for filarial diseases, there are a variety of outcomes to infection. Some 'resistant' individuals clear worms prior to the onset of patency, while others develop long-lasting infections with or without microfilaraemia. Asymptomatic (or subclinical) infection is typically associated with high levels of circulating microfilariae. Progression to the characteristic clinical pathologies is also highly variable. In lymphatic filariasis, pathology can occur in the presence or absence of detectable parasitaemia. In onchocerciasis, pathology is more common in infected individuals which do not have mF. 1

| L sigmo dontis
L sigmodontis is a filarial nematode of the Onchocercidae family first isolated from the cotton rat Sigmodon hispidus. 11 The site of infection is the pleural cavity which resembles the relatively rare human diseases caused by Mansonella ozzardi and M perstans. 3 L sigmodontis is transmitted to the primary host via an arthropod vector, the tropical mite Ornithonyssus bacoti, which acts as the intermediate host and a reservoir for infective L sigmodontis L3 larvae. The life cycle and maintenance of L sigmodontis in a laboratory setting have been described in detail elsewhere. 12 For experiments with laboratory mice, infection can be achieved by allowing infected mites to feed naturally, with L3 larvae entering the skin when the mite performs a blood meal (natural infection). Alternatively, a known number of L3 larvae can be isolated from mites and injected directly into mice (subcutaneous infection). 13 Although natural infection includes the activation of innate immune responses in the skin that follow mite feeding, subcutaneous infection allows the infective dose to be known, and the immunological findings from laboratories that use the different routes have not fundamentally differed.
Many L3 larvae are destroyed in the skin. 14,15 Surviving larvae forcibly enter lymphatic vessels 15,16 and migrate to the pleural cavity by about day 4 post-infection (p.i.). The exact route the worms take on their way to the pleural cavity is not well understood but may involve translocation through the lung. 17 At around day 8, the worm undergoes a moult to become an L4 larva and another final moult at circa day 28-30 to become an adult. However, it takes another 30 days for the worms to become sexually mature and produce mF.
It remains the only filarial nematode in which infection with larvae leads to circulating mF in the immunocompetent murine host.

| THE S ITE OF INFEC TI ON: THE PLEUR AL C AVIT Y
A major advantage of the L sigmodontis model is that the site of infection is simple in terms of tissue architecture. Isolation of worms and immune cells is performed by lavage of the pleural cavity, without the need for destructive tissue homogenization or digestion.
The pleura is a serous membrane with a two-layer membrane structure folded back on itself made up of a layer of mesothelial cells (the mesothelium). 18 Lining the lung is the visceral membrane, and lining the chest wall and diaphragm is the parietal membrane. 19 The space between the two membranes is only in potential a cavity. In reality, it is a thin layer of fluid kept at negative pressure which allows the lungs to remain 'attached' to the parietal membrane when the chest expands, thus catering for lung inflation. 19,20 The pleural fluid contains lysozyme, antibody, complement and proteins such as surfactants, that reduce the friction of breathing. 19,20 Many of these factors, including clotting factors and complement, are locally produced rather than entering the cavity via the serum. Pleural fluid enters via capillaries lining the parietal membrane and is drained via lymphatics within the cavity (Figure 1). 20 The pleural fluid is highly

| S TR AIN DEPENDEN C Y
L sigmodontis was first used to infect 'albino' mice in 1946, but it was not widely used as a murine model until the early 1990s. 22 The establishment of L sigmodontis as a mouse model to study filarial nematode infection and strain differences owes much to the pioneering work of Odile Bain's laboratory at Muséum National d'Histoire Naturelle, Paris. Most notably, their paper from 1992 which investigated the susceptibility of various strains of mice to L sigmodontis | 3 of 17 FINLAY ANd ALLEN using a number of quantitative readouts of worm health and sexual maturity at a later stage of infection. 23 We have compiled and reanalysed the parasitology data from that study to provide an over- Only about one third or less L3 survive the 4 day journey from the skin to the pleural cavity. 24,26 Prior to day 20 p.i., L4 larvae are significantly larger in susceptible BALB/c mice than C57BL/6 mice 26 with a small difference in worm number (Figure 3). 24 Despite these differences, there is little observable worm killing in the pleural cavity in either strain before the adult moult around day 28-30 p.i. (Figure 3). 24,26 Later in infection, obvious differences emerge. C57BL/6 mice begin to kill worms around day 35 p.i., and from day 40 p.i. onwards, dead and granulomatous worms are readily detectable in the pleural lavage. In contrast, worms in BALB/c mice continue to grow, undergo embryogenesis and by day 56 p.i. start to produce mF which are detectable in the bloodstream ( Figure 3). 24,26 Patency has a strong sexual dimorphism. 23,24 Depending on host sex, approximately 20% (male) to 70% (female) of BALB/c mice will develop patent infection. The presence of mFindividuals and the wide range in mF numbers is reflective of human filarial infections. Female mice harbour more adult worms than males and are over four times more prone to be mF + than males, independent of the number of adult worms borne. 24 Nonetheless, in all BALB/c mice, worms are eventually killed between days 90 and 120 p.i.

| ROLE OF LYMPHO C Y TE S
It is clear that mice require an adaptive immune system to control L sigmodontis infection and mF. 29 Infected RAG2 −/− IL2Rγ −/− mice which lack T, B and NK cells, harbour more worms than wild-type (WT) C57BL/6 controls and have microfilaraemia in the blood at day 72 p.i. which is never seen in WT C57BL/6 mice. 29 T cells are required for worm killing in both resistant and susceptible mice. Depletion of CD4 + T cells from infected BALB/c mice increases worm burden and blood mF. 30 We have completed similar experiments in C57BL/6 mice and found that depletion of CD4 + T cells resulted in a signifi-  32 This includes antibody-producing B2 cells and innate-like B1 cells. 33 Parasite-specific IgE and IgG1 are produced by both resistant F I G U R E 1 The Pleural Cavity. The pleural cavity is a fluid-filled space created by a two-membrane (pleura) structure that lines the lung and chest wall. The pleural fluid supports breathing by providing lubrication and by allowing close apposition of the inner pleura, covering the lung, with the outer pleura, covering the chest wall. Pleural fluid enters the cavity via the intercostal arteries and is drained by the lymphatics. The pleural fluid is cell dense with CD45 + immune cells. In mice, these include populations of B cells (B1 and B2 cells), macrophages (F480 high and F4/80 low ) and T cells with smaller numbers of mast cells and dendritic cells and susceptible L sigmodontis-infected mice. 34 Although differences in B-cell numbers between strains are not apparent in the pleural lavage, C57BL/6 mice exhibit sustained increases in all B-cell subsets within the fat-associated lymphoid clusters (FALCs) of the cavity. 33 In contrast, BALB/c mice exhibit only a transient increase and fail to maintain B1b and B2 cells, which produce the majority of antigen-specific IgM in the pleural cavity. The result is that BALB/c mice have far less L sigmodontis-specific IgM than C57BL/6 mice by day 18 p.i. 33 Release of IgM in the serous cavity by FALC B cells may represent a crucial source of protective antibodies, as serum IgM cannot diffuse into the body cavities. The potential importance of IgM for elimination of filarial larvae has been demonstrated in a closely related model of filarial infection, using sIgM −/− mice. 35 In an apparent contradiction, C57BL/6 'μMT' mice which lack B cells are still able to kill worms and control mF. 34 The same mutation backcrossed onto a BALB/c background showed no defect in latestage worm killing, but there were reduced mF counts and worm maturation, suggesting that B cells are somewhat protective. However, infected BALB/c μMT also exhibit reduced Th1 and Th2 response and macrophage accumulation. 31 In contrast, 'B-less', another B celldeficient mouse on the BALB/c background, actually has increased mF (but no change in worm numbers) suggesting B cells play a role in control of patent infections. 36 Total B-cell deficiencies may paradoxically obscure the role of antibodies. Inconsistent results between different genetically altered B cell-deficient mouse strains maybe explained by the differential role of distinct B-cell subsets. B cells, particularly B1 cells, have innate immune functions 37 which include immune regulation. 38 Thus, removing B cells from the C57BL/6 strain may alter worm killing even in the absence of host-protective antibody. Supporting this hypothesis, BALB.Xid mice which have reduced B1 cells, particularly CD5 + B1a which are responsible for the production of antibodies to T-independent antigens, harbour significantly more worms and higher blood mF than WT controls 39,40 (Tables 1 and 2).

| T cells: key role for IL-4
Immune responses to L sigmodontis are highly Th2 polarized in both resistant and susceptible strains but infected C57BL/6 mice exhibit dramatically higher local cytokine responses in the pleural cavity than BALB/c mice. 26 IL-4 is critical for resistance to L sigmodontis as IL-4-deficient C57BL/6 mice resemble susceptible BALB/c mice, with equivalent adult parasite survival and circulating mF 34 (Table 2).
The role of IL-4 in susceptible BALB/c mice is not as straightforward. IL-4-deficient or IL-4Rα-deficient BALB/c mice, harbour similar numbers of worms as WT BALB/c mice late in infection indicating that IL-4 is not responsible for killing of adult worms in this strain. 36,41,42 However, IL-4 greatly limits the mF counts in BALB/c mice. 36,41,42 Infected IL-4-deficient BALB/c mice also have reduced eosinophils but increased neutrophils. 36 In contrast to IL-4 deficiency, IL-5 deficiency on the BALB/c background does increase adult survival, 41 and a dramatic effect is seen in IL-4Rα −/− /IL-5 mice which harbour more worms and at least a hundred-fold increase in blood mF. 43,44 In contrast, C57BL/6 IL-5-deficient mice are equally resistant to WT controls, 45,46 highlighting another significant strain difference, as discussed below with regard to eosinophils. It is important to note that IL-13 has not been specifically investigated in L sigmodontis infection and that IL-4-deficient mice on the C57BL/6 background have a profound IL-13 defect while IL-4Rα-deficient mice do not distinguish the contribution of IL-4 and IL-13 ( Table 1).
Given that GWAS studies of onchocerciasis show association of IL-13 polymorphisms with the development of pathology, 47,48 an investigation of the role of IL-13 is warranted.
Patients with onchocerciasis and lymphatic filariasis typically present with a Th2-polorized phenotype, but many patients have a more mixed Th1/Th2 T-cell phenotype, especially those with pathology. 9-10,49 IFN-γ and Th1 cells increase along with Th2 cytokines in both resistant and susceptible strains of mice infected with L sigmodontis. 26,50 Interestingly, transcripts for IFN-γ peak when mF are first produced in BALB/c mice. 51 IFN-γ and IL-5 synergize to enhance F I G U R E 2 Resistance to L sigmodontis infection varies across host genotype. Data from Tables 2 (readouts of L sigmodontis infection in ten strains of mice) and III (readouts of morphology of L sigmodontis worms taken from ten strains of mice) from Petit et al 23 were reanalysed to create a relative susceptibility score. Only data for male mice were included. The following readouts were collated and weighted as follows: mF/10 μL of blood was scaled 0-100 with 100 being given to the strain with the highest mF numbers (weighted 2), percentage mF + (weighted 2), percentage worm + (weighted 1), percentage localization of worms in the pleural cavity (weighted 0.5), number of worms recovered as a percentage of injected L3 larvae (weighted 1.5), length worms scaled to longest worm recovered from any strain (weighted 0.25 each for male and female worms), width of worms scaled to widest worm recovered (weighted 0.25 each for male and female worms), uterine egg score for recovered worms (weighted 0.25), divided egg score for recovered worms (weighted 0.25), percentage of worms with internal mF (weighted 1) and mF score for recovered worms (weighted 0.5). Weighted scores were added together to give a total possible score of 1000, and this number was divided by 100 to give scores shown in the graph above. Bar colour represents MHC haplotype for each strain

| T-cell regulation and Hyporesponsiveness
Immune regulation and immune hyporesponsiveness are features of chronic helminth infection including filariasis. [52][53][54][55] The most common outcome in human filariasis is mild/asymptomatic disease despite the presence of adult worms and mF, reflecting a remarkable degree of tolerance by the host. 1,6,8,9 This can be explained by active immunoregulation on part of the parasite but also host regulatory networks, which act in concert to prevent immunopathology. [56][57][58] Pathological symptoms in onchocerciasis and lymphatic filariasis are often associated with immune responses to dead/dying worms which remain in the tissue, indicating that tolerance to established parasites can be more beneficial to the host than 'protective' worm killing. 1 to the appearance of mF, Th2 cytokine production is notably reduced. 51,60-62 Following the production of mF however, there is a recovery in IFN-γ and Th2 cytokines which coincide with worm killing in BALB/c mice ( Figure 4). 51 Mechanistically, T-cell hyporesponsiveness can be viewed as being cell-intrinsic (hyporesponsiveness of Th2 cells themselves) or cell-extrinsic (active regulation by other factors). A recent study has provided evidence that T-cell hyporesponsiveness is intrinsic to the Th2 cell. 62 By day 60 p.i., Th2 cells in infected BALB/c mice cells acquire a dysfunctional phenotype that is IL-21 + Egr2 + c-Maf + Blimp-1 + IL-4 lo IL-5 lo T-bet + IFN-γ + . Critically, this phenotype is maintained upon transfer of these cells to naïve mice. 62 An intriguing feature of these cells is the production of IL-21, which appears to contribute to mF survival while limiting germinal centre B cells and parasite-specific IgM and IgG1. 62 There is also evidence of extrinsic regulation of T cells in L sigmodontis-infected BALB/c mice. Th2 cells acquire PD1 expression, and signalling from PD-L2 reduces Th2 cytokine production, and increases blood mF and worm fecundity. 61 Monocyte-derived pleural cavity macrophages are the likely source of PD-L2. 32 Figure 4, Table 1). 65 Collectively, these results reveal that T-cell hy-

| ILC, NK cells, basophils and mast cells
The numbers of NK cells, basophils and type 2 innate lymphoid cells

| WORM K ILLING , G R AN ULOMA S , EOS INOPHIL S AND NEUTROPHIL S
We do not yet have a detailed understanding of the mechanisms by which L sigmodontis parasites are killed but it does involve the gradual encasement of the worm within a granuloma. Cuticles that are shed during moulting are rapidly encased in granulomas formed of eosinophils. 27 However, the living worm represents a more difficult target due to its high motility and selective pressure to avoid exposing residues that enable immune cell attachment. 27 In the fully permissive Mongolian jird, granulomas contain neutrophils as well as eosinophils. 27 In BALB/c mice granuloma formation around young adult worms is rare but develops later in infection around larger adult worms and contains neutrophils which form the immediate layer of the granuloma exposed to the worm. 27 In BALB/c mice neutrophils are recruited relatively late in infection prior to   (Table 1).
In both C57BL/6 and BALB/c mice, eosinophils increase in the pleural cavity as early as 11 days p.i., 32 and recruitment is much reduced in mice lacking CD4 + T cells. 29 These mice harbour more worms than even susceptible BALB/c mice at day 28 p.i. implicating eosinophil degranulation as an important factor in worm killing in BALB/c mice. 83 Notably, BALB/c mice deficient in Eotaxin-1 have increased late-stage worm counts without any change of mF counts. 84 (Tables 1 and 2). 44 Finally, administration of anti-histamine to infected BALB/c enhances clearance of adult worms in an eosinophil-dependent manner, indicating that parasite-induced histamine impairs anti-parasite immune responses by limiting eosinophil activation. 85 Granuloma formation is less apparent in BALB/c mice lacking both IFN-γ and IL-5 50 and neutrophil recruitment and activity is markedly decreased in these mice, with greater worm burdens than mice deficient in either cytokine alone. 50 GM-CSF a key survival and proliferative factor for both eosinophils and neutrophils has a greater inhibitory effect on worm killing than IL-5. 82 Interestingly, neutralization of IL-5 in infected BALB/c mice, which reduces worm killing, is associated with reduced neutrophil accumulation, 82 indicating an unexpected cooperation between neutrophils and eosinophils in late-stage worm killing in BALB/c mice ( Figure 4, Table 1).
Evidence for eosinophils mediating L sigmodontis worm killing in resistant C57BL/6 is less clear. In C57BL/6 mice, granulomas form along the lateral lines of small newly moulted adult worms. 27 Granulomas in these mice are predominantly eosinophilic (unpublished observation). However, there is no deficit in worm killing following primary infection of IL-5-deficient C57BL/6 mice ( Figure 5). 45,46 This finding, that infected IL-5-deficent C57BL/6 mice, which cannot attract eosinophils have no defect in worm killing is hard to square with the idea that granulomas are required to kill worms (Table 2). 45 This raises the possibility that granulomas in C57BL/6 may form around already dead worms, like they do around shed cuticles rather than being a key component in worm killing itself.
A far more straightforward role for eosinophils exists in killing of larval parasites during secondary infection. In both C57BL/6 and BALB/c mice, eosinophils are rapidly recruited to the skin and degranulate where they mediate larval killing; IL-5-deficent mice of both strains lose protection following vaccination. 45,86 This is consistent with a greater role for eosinophils in larval killing over adult worm killing seen for other helminths. 87 The evolutionary importance of eosinophils in parasite control is perhaps best illustrated by the finding that, in both BALB/c and C57BL/6 mice, L sigmodontis L3 larvae respond to the presence of eosinophils by accelerating their development. 46 Thus, the parasite appears able to detect cues that a host is mounting an effective immune response.
Taken together, these results show that killing of adult L sigmodontis is highly strain dependent. Both eosinophils and neutrophils contribute to worm killing and mF control in late-stage infection in BALB/c mice but killing of immature worms in resistant C57BL/6 is still poorly defined.

F I G U R E 5
Immune response to L sigmodontis infection in C57BL/6 mice leads early to worm killing in the pleural cavity. C57BL/6 mice mount an early and sustained Th2-biased immune response to L sigmodontis characterized by greater cell accumulation in the pleural cavity than in BALB/c mice. This is associated with eosinophilia, proliferation of B2 cells and the development of FALCs within the pleural and pericardial cavities and the production of parasite-specific IgE, IgG1 and IgM. The pleura supports resistance by the production of chemokines including CXCL12. F4/80 high macrophages proliferate and greatly expand in number and become M(IL-4) activated producing Ym1 and RELM-α. Monocytes, F480 low macrophages and PD-L2 + intermediate macrophages are present but outnumbered by F4/80 high macrophages. Dead worms encased in eosinophilic granulomas are detectable by day 35 p.i., and infection is rapidly cleared thereafter

| M(IL-4) activation and proliferation in L sigmodontis infection
Leaving aside lymphocytes, the most abundant cells in the pleural cavity of infected C57BL/6 mice are macrophages and eosinophils.
The fact that IL-5 was found to be dispensable for worm killing in C57BL/6 mice while IL-4 is required highlights the potential impor- In the pleural cavity of L sigmodontis-infected C57BL/6 mice, F4/80 high macrophages greatly expand in number, beginning early in infection and continuing throughout infection, eventually becoming the most abundant cell type. 32,111 This is accompanied by only limited recruitment of monocytes and relative loss in the proportion of F4/80 low macrophages. 111 The pleural cavity F4/80 high macrophages undergo a burst of proliferation peaking at 10 days p.i. which is dependent on IL-4 and occurs without the input of blood monocytes ( Figure 5). 111 This finding was the first demonstration that macrophages could expand dramatically in number through local proliferation rather than blood cell recruitment during an inflammatory response. The paradigm shift in macrophage biology that resulted was a direct consequence of the L sigmodontis model. Macrophages that can respond to IL-4 outcompete, through proliferation, macrophages which lack IL-4Rα 98 and proliferation is independent of the CSF1R highlighting the direct role of IL-4 in macrophage proliferation. 98 The cellular source of IL-4 is not known but given that macrophages fail to proliferate in L sigmodontis-infected RAG-deficient mice, an adaptive immune cell is involved. 98 Our unpublished data indicate that the prolif- macrophages to worm killing have been hampered by difficulties in effectively deleting IL-4Rα due to lack of penetrance using Cre recombinase technology. In the strong IL-4 environment of L sigmodontis infection, the few cells that still express the receptor outcompete the IL-4Rα-deficient cells by local proliferation. 113 However, in a study of BALB/c mice using LysMCre + X IL-4Rα fl/fl , mF numbers were significantly higher than controls, despite the fact that the macrophages had reverted to a WT phenotype by the time the infection became patent. 113 This suggests that early in infection M(IL-4) macrophages contribute to a pathway leading to containment of mF later in infection.

| THE PLEUR AL NI CHE
Given the presence of inflammatory infiltrate, oedema and motile worms, it is not surprising that the stroma of the tissue niche undergoes transformations during L sigmodontis infection. 17,43,114,115 Progressive fibrotic pathologies develop in the pleura of BALB/c mice infected with L sigmodontis 43,115 along with inflammatory foci in the lungs. 17,43,114 The effect of the pleura itself upon pleural immune cells and the outcome of infection is largely unknown. We have previously shown that tissue-specific niche-derived factors support IL-4Rα-dependent proliferation of tissue-resident macrophages. 116 C1q, a complement component, acts to enhance F4/80 high peritoneal macrophage proliferation and activation in the peritoneal cavity in response to IL-4, while surfactant protein A performs the same function in the lung.
Both factors act through the atypical myosin, Myo18A However, pleural cavity F4/80 high macrophages did not express Myo18A. 116 Nevertheless, the concept that the niche can provide factors which support macrophage proliferation and M(IL-4) activation is relevant to L sigmodontis infection.
The pleura and in particular the mesothelium play an active role in the maintenance of macrophages in the serous cavities. For example, retinoic acid (RA) produced by WT1 + mesothelial cells maintains GATA6 expression in F4/80 high macrophages, which is required for its 'residency' programme. 89 RA and IL-4 mediate the conversion of thioglycolate-elicited F4/80 low macrophages into F4/80 high macrophages. 11 In the absence of RA, macrophages do not fully convert into F4/80 high macrophages remaining in an intermediate PD-L2 + phenotype. 11 This phenotype resembles the phenotype of PD-L2 + F4/80 low macrophages that we find in BALB/c mice infected with L sigmodontis. 32 This suggests that RA signalling deficiencies may contribute to the accumulation of this immunoregulatory cell in infected BALB/c mice.
In evidence for the role of the tissue niche in L sigmodontis infection, mesothelium was shown to be critical for the recruitment of cells to the pleural cavity. CXCL12 is produced by the mesothelium in infected-C57BL/6 but not BALB/c mice. 117 Chemical blockade of the CXCL12-CXCR4 pathway significantly reduced immune cell recruitment and reduced worm killing in C57BL/6 mice but had no effect in BALB/c mice. 117 Another evident difference between strains is the expansion of the FALCs, which is significantly greater in infected C57BL/6 mice in comparison to BALB/c mice. 33 IL-33 is produced by FALC stroma, 33 and thus, IL-33 may be more abundant in C57BL/6 mice during L sigmodontis infection. IL-33, which is also expressed in the lining of the lung and by mesothelial cells in other organs, 77,118 is a diagnostic marker of pleural effusion. 119 Our unpublished data indicate that macrophages from C57BL/6 but not BALB/c mice upregulate ST2 during infection. Consistent with this finding and the potentially lower availability of IL-33 in BALB/c mice, 33 ST2 or IL-33-deficency in BALB/c mice does not affect worm killing in BALB/c mice, although ST2-deficient mice have less recruitment of eosinophils, CD4 + T cells and macrophages to the site of infection. 78,120 In BALB/c mice IL-33 is not required for macrophage proliferation with L sigmodontis infection. 120 However, BALB/c mice may not have been a relevant strain to investigate macrophage proliferation as they display reduced macrophage proliferation in comparison to C57BL/6. 32 Given its potential to produce cytokines, chemokines or other factors such as RA which support F4/80 high macrophages, the pleura certainly has the potential to impact upon the outcome of infection. Fundamentally, we do yet not know whether resistance to infection seen in different strains of mice is mediated by immune cells or if the tissue niche dictates the outcome.

| Beyond the pleural cavity
This review has focused on the local immune responses to L sigmodontis in the pleural cavity. Thus, we have neglected to mention responses which occur early in infection in the skin and evidence for L sigmodontis having effects on immune responses in distal sites.
Most L sigmodontis L3 larvae do not survive their journey from the skin and relatively little is known about the responses in the skin apart from the aforementioned contribution of mast cells in primary infection 79 and eosinophils in secondary infection. 9 (Tables 1 and 2). Combined data from two papers suggest that mast cells via IL-6 activation and CCL17 recruitment may contribute to killing of L sigmodontis larvae in the skin following primary infection. 79,128 In addition, neutrophils have the capacity to kill L3 larvae in the skin and are recruited to the skin following infection. 125,129 However, neutrophil depletion in WT C57BL/6 mice had no effect on adult worm numbers, 129 leaving open the contribution of neutrophils to early resistance. The contribution of the skin to innate resistance, and perhaps more importantly, the impact the immune response to the migrating larvae in the skin/lymphatics has upon the later protective immune responses in the different strains remains an unresolved question.
In keeping with the extensively studied immune regulatory/ suppressive effects of helminth infection, 53 L sigmodontis infection or products protects mice against E-coli-induced sepsis, 121 diet-induced glucose intolerance, 130 type-1 diabetes, 131,132 atherosclerosis 18 and allergic responses. 133,134 These studies provide insights into new helminth-mediated regulatory pathways such as the suppression T-cell response by interference with the development of cytotoxic CD8 + T cells, 135 follicular helper T cells, 136 or by TGF-β. 69,132 These results highlight the capacity of L sigmodontis to induce global immune alterations in the host best exemplified by the parasites ability to interfere with vaccine-induced protective immunity against, Plasmodium, 135 Friend virus 137 and influenza. 66 Fitting these findings, and perhaps partly explaining them, is the remarkable finding that peritoneal implantation of a single adult female (but not male) delays the elimination of exogenously delivered blood mF, suggesting that the parasite acts systemically to protect her offspring. 28

| FUTURE PER S PEC TIVE S
We believe that the unique features of the L sigmodontis model will reveal further insights into myeloid cell function, as well cellular communication across tissues and between immune and stromal cells. Results from our laboratory strongly suggest a key role for pleural cavity macrophages in L sigmodontis infection but the heterogeneity of these cells has yet to be described accurately. In both resistant and susceptible mice, we observe cells which fall under the banner of M(IL-4) cells but in fact these are an amalgam of cells of diverse origins and activation states. New technologies such singlecell sequencing, mass spectrometry and high-parameter cytometry will allow us to better describe complex myeloid populations without the bias of 'gates'. The pleural cavity is a relatively simple organ system which also represents a model to study niche-immune cell interactions such as ligand-receptor analysis that will provide a more holistic understanding of immune cell function and interactions. More importantly, we have yet to convincingly link these macrophages to worm killing. The use of novel genetically modified mice may help to answer these questions. In particular, Ms4a3-reporter mice will allow for fate mapping of monocyte-derived cells, 10 macrophage-specific GATA6-deficent mice will allow us to ascertain the role of the residency programme in F4/80 high cells in resistance, 15 and mice with a cell-specific deletion in Bhlhe40 can be used to ascertain the importance of F4/80 high macrophage proliferation 138 to infection outcome.
The variation in susceptibility across many mouse strains ( Figure 2) paints L sigmodontis infection as a suitable candidate for mapping susceptibility to genetic loci using haplotype lineage analysis or quantitative trait locus. The fact that only mice of the BALB lineage develop mF after infection limits susceptibility loci to that mouse. Many common mouse strains are historically derived from 19th-century mouse fanciers who bred mice for behavioural and aesthetic traits. Thus, rather than being truly separate, inbred mouse lines share much common ancestry and are predominantly derived from the Mus musculus domesticus subspecies. [139][140][141] The most common strains of mice only represent a small portion of the diversity in alleles amongst wild mice. 142 The resistance of relatively new wild-derived strains to L sigmodontis is of considerable interest, as this may provide more accurate determination of genetic loci, without the need for genotyping. 139,143 This would be achieved by the use of recombinantly inbred mouse lines which make use of wild-derived mice, such as the Collaborative Cross, that have been created to minimize the number of strains required to capture maximal genetic diversity. 142 There are now thought to be in the region of 450 mouse strains. 144 However, if anything the number of strains in common use has been in steep decline for decades. One reason is the wide availability of genetically altered mice only on certain backgrounds and the fact that the C57BL/6J mouse is the reference genome. 145 This gradual acceptance of the dominant position of C57BL/6 mice raises the issue that many basic discoveries in immunology may be mouse strain-specific. Indeed, it was the L sigmodontis model in C57BL/6 mice that allowed the paradigm-shifting discovery that macrophages could increase dramatically in number through local proliferation, without requirement for blood-derived cells during an inflammatory response. 111 Similar experiments in BALB/c mice would not have revealed the same myeloid cell dynamics. Nonetheless, the basic biology holds true in other strains but to widely varying extent 32 and the comparison itself is now proving powerful. It is thus imperative to be aware that recent breakthroughs in myeloid biology have largely been achieved using only C57BL/6 mice. [95][96][97]138 Since the first extension of the Th1-Th2 dichotomy to macrophages, 146 the genetic underpinnings of these phenotypes have remained largely elusive.
Despite their hybridized ancestry, inbred strains of mice harbour an astonishing amount of loss of function mutations. [147][148][149] Ultimately, cell identity and function in a type 2 immune environment will depend less on simple allelic differences and more upon transcription factor interactions. 150,151 The biological consequences provided by differences in cis-regulatory elements in different strains of mice have already been shown to have profound effects on macrophage transcription factor interactions and cell function. 152,153 We are unlikely to be able to model the functional consequences of such genomic differences any time soon. Until such time, we will need to pay attention to the repertoire of genetic diversity provided by extant inbred mouse strains.

ACK N OWLED G EM ENTS
We thank Brian Chan, Alistair Chenery and Pedro Papotto for their critical review of the manuscript. We thank the editor of this issue, Dominik Rückerl, and the anonymous referees, for very comprehensive and helpful review of the manuscript. Figures made with the aid of biorender.com.