Worms: Pernicious parasites or allies against allergies?

Summary Type 2 immune responses are most commonly associated with allergy and helminth parasite infections. Since the discovery of Th1 and Th2 immune responses more than 30 years ago, models of both allergic disease and helminth infections have been useful in characterizing the development, effector mechanisms and pathological consequences of type 2 immune responses. The observation that some helminth infections negatively correlate with allergic and inflammatory disease led to a large field of research into parasite immunomodulation. However, it is worth noting that helminth parasites are not always benign infections, and that helminth immunomodulation can have stimulatory as well as suppressive effects on allergic responses. In this review, we will discuss how parasitic infections change host responses, the consequences for bystander immunity and how this interaction influences clinical symptoms of allergy.


| K E Y ELEMENTS IN T YPE 2 IMMUNE RE SP ONSE S
Dendritic cells (DCs) are an innate immune population absolutely required for the development of optimal effector Th cell immune responses, including Th2 responses. 4,5 DCs are intimately associated with barrier sites such as the lungs and will take up antigens from within and beyond the epithelial barrier. Upon detection of a helminth infection or an environmental allergen, 6 DCs become activated and migrate to the draining lymph nodes, presenting antigens to T cells and potentially inducing a Th2 immune response.
Although the critical involvement of DCs in Th2 development is clear, the precise signals that lead to priming of a Th2-inducing DC are incompletely characterized. In recent years, the importance of epithelial-derived cytokines such as interleukin-25 (IL-25), IL-33 and thymic stromal lymphopoietin (TSLP) in allergy and parasitic infection has become appreciated. 7 These cytokines can act directly on DCs, skewing resultant responses to Th2, and also directly activate type 2 innate lymphoid cells (ILC2s), inducing a rapid innate type 2 response.
ILC2s are innate lymphocytes, lacking antigen-specific receptors, which produce large amounts of type 2 cytokines IL-5, IL-13 and IL-9, as well as proresolving factors amphiregulin 8 and IL-10. 9 Activated ILC2s also express class II MHC, can present peptide antigen and can supply IL-4R signals in type 2 response initiation. 10 Thus, type 2 innate epithelial cell cytokines, DC and ILC2s are involved in the earliest responses to allergens and helminth parasites, in initiation of the Th2 immune response, and in amplification of allergic and antiparasite immunity. However, antigen specificity and the control of ongoing immune responses are critically dependent on adaptive immunity.

| HELMINTH INFEC TI ON S , DAMAG E AND T YPE IMMUNE RE S P ON S E S
While type 2 immune responses have clear pathological roles in allergy, they are generally beneficial in helminth infections. Increased susceptibility to a range of intestinal and tissue-dwelling parasites can be seen in mice lacking essential elements of the type 2 response pathway. 11 Likewise, in human populations, single nucleotide polymorphisms (SNPs) in type 2 response elements such as IL-13 and STAT-6, and immunoregulatory elements IL-10 and TGFβ correlate with both decreased susceptibility to allergy and increased susceptibility to parasitic infection. [12][13][14] Many helminth species remain in the host for a prolonged time, and type 2 responses may be more beneficial for the survival and integrity of the host than the more inflammatory Th1/Th17 alternative. Indeed, asymptomatic infections with, for example filarial worms are associated with type 2 responses, 15 whereas in individuals suffering from helminth infections linked to pathology and clinical symptoms, Th1 or Th17 cell responses are often found. [16][17][18] In many parasitic infections, sterile immunity is not common: most individuals living in endemic areas are constantly reinfected, even after drugmediated clearance of parasites. 19 As a consequence, host immune responses in endemic areas are often characterized as a "modified Th2" response that results in control (but not clearance) of parasite load, low-level parasite transmission and minimal host pathology: an acceptable host/parasite compromise.
Tissue damage caused by helminth infections is also a powerful stimulus for type 2 responses, which in turn lead to a rapid type 2 response-mediated healing phenotype. 20 For example, in the lung, type 2 immune responses are important in healing damage caused by migrating Nippostrongylus brasiliensis larvae, while recruiting eosinophils that damage larvae, hamper their fecundity and fertility upon arrival in the gut, leading to an early expulsion. 11 However, type 2 responses may also cause pathology due to aberrant healing, such as in the case of fibrotic granulomas formed around schistosome eggs. These granulomas cause mild to more severe pathology, linked to local fibrotic tissue, liver and splenomegaly, and an increased risk to develop cancer in the liver or the bladder, depending on the species. 21 Thus, depending on context and infecting species, parasite products can induce epithelial cell proliferation, encourage healing, control fibrosis 22 and cause transformation and cancerous growth. 23 In the absence of helminth infection, type 2 responses are often perceived to be only involved in pathological allergic responses, ultimately leading to decreased lung function and airway hyperactivity in asthma and rhinitis, 24 pruritus (itching) and damage to the skin barrier in atopic dermatitis, 25 and itching, pain and/or swelling of the mouth, pharynx and oesophagus, diarrhoea and abdominal pain in food allergy. 26 However, type 2 responses in the absence of helminth infections can also have beneficial roles: circulating IgE specific to venom toxins, which can cause dangerous anaphylaxis on exposure, can also be protective with release of mast cell proteases that degrade venom toxins and counteract the venom's detrimental effects. 27 During pregnancy, the type 2 cytokine milieu in the womb protects the "non-self" foetus from abortion (which conversely is linked to increased Th1/17 responses). 28 Finally, perinatal type 2 responses in the lung are required for establishment of lung homoeostasis and development of anti-inflammatory type 2 macrophages. 29 Therefore, just as there is no such thing as "weeds" in a garden (just plants in the wrong place) perhaps there is no such thing as a "bad" immune response, just inappropriate in its context. How parasites modulate these useful and/or pathological responses, and what happens when the balance is perturbed, will be covered in the next sections.

| TALE S OF WORMS IN MEN
In the 1970s, the "hygiene hypothesis" was proposed as an explanation for the steep and alarming rise in the prevalence of childhood allergies and asthma among urban, Westernized societies. The hygiene hypothesis links changes in housing, sanitation and health care to increased allergic disease and proposes that this is in part due to reduced endemic infections. The prevalence of parasitic infections in particular has been drastically reduced in Westernized societies over the last century and is therefore proposed to be an important contributing factor in this hypothesis. Multiple epidemiological studies have been used to support this hypothesis by indicating that in helminth-endemic rural areas relatively few people have allergic symptoms. 30,31 The fact that some local African languages contain no words to describe allergic symptoms could support this hypothesis, indicating that allergic diseases have never been a problem among these populations. 32 However, an examination of the many population studies of the past 30 years either by meta-analyses 33,34 or in some excellent systematic reviews 35,36 shows that the direction of the effect by helminth infections is far from consistent. For example, while hookworm infections were associated with a protective effect against asthma, other helminths like Trichuris trichiura, Enterobius vermicularis and Strongyloides stercoralis did not show any effect, and conversely Ascaris lumbricoides infection increased the risk of developing asthma and wheeze. 33 Greater consensus was observed regarding protection to atopic sensitization and allergic skin reactivity, although the outcome varied with the allergen studied. [34][35][36] Similarly, studies using anthelmintic treatment of helminthendemic populations show mixed results (for a full overview, see Wammes, 2014 35 )-some show an increased frequency of allergen skin prick test (SPT) after worm clearance, 37 while others did not observe any differences between treatment and control groups within one or 2 year time frames. 38 Part of the inconsistent findings and dissimilarities in conclusion in the epidemiological and interventional studies may be explained by variations in factors such as the age of the population studied, age of helminth exposure (and consequent early-life immune imprinting) and the infectious burden, endemic parasite species and chronicity of infection, or differences in study parameters such as clinical symptoms in asthma, rhinitis or eczema or methods used to measure allergen sensitization (SPT versus allergen-specific IgE). 35,39 Epidemiological studies applying antihelminthic treatment during pregnancy provide an interesting approach to evaluate the relationship between helminths, early immune priming and allergy: these show an increased risk of early-life eczema in babies of treated mothers 37,40 ; however, a 9-year follow-up showed that this effect was not maintained to later life. 41 In contrast to human studies, experiments in mice showed more consistent findings in the prevention of allergic airway inflammation by a wide range of helminth species: N. brasiliensis, 42 Heligmosomoides polygyrus, 43 Litomosoides sigmodontis, 44 Schistosoma mansoni, 45 Trichinella spiralis 46 and Schistosoma japonicum. 47,48 Interestingly, transmaternal protection against allergic airway inflammation by helminth infection in mice implied that this was dependent on the phase of the infection during pregnancy: offspring from schistosome-infected females were protected if they had mated during the initial Th1 phase, or the chronic immuneregulatory phases of schistosome infection, but conversely disease was exacerbated if mating occurred during the high Th2 phase of infection (coinciding with egg deposition). 49 This may further complicate assessment of human population studies, as protection against allergy may depend on far more complicated interactions than simple presence of infection, but also prenatal stimuli and phase of infection.

| HELMINTH INTER AC TI ON S WITH OTHER INFEC TI ON S AND THE MI CROB I OTA
Severe respiratory syncytial virus (RSV) and/or rhinovirus (RV) infection in early life gives a sevenfold increased risk of developing asthma, 50 while asthma exacerbations are associated with concurrent respiratory viral infection in up to 80% of cases. 51 Furthermore, in experimental RV infections of asthmatic volunteers, IL-33 and other type 2 cytokines were released into the airways, correlating with severity of asthma exacerbation. 52 Parasite products can suppress IL-33 release 53,54 ; thus, they could directly suppress viral proallergic responses, or, via suppression of IL-33 release, lead to increased antiviral immune responses. 55 Interestingly, the interactions between helminths and viruses have recently received a great deal of attention. Helminths not only change the response to other infectious agents within the host but can also affect the balance of commensal organisms with which they share an environment. The host genome and the total diversity of the microbiota (the "microbiome") are important in mediating or reflecting health and disease in the intestine, and in other barrier sites such as the lung and skin. Changes in the gut and lung microbiomes are seen in allergic diseases such as asthma, 60 reflecting the immune axis between these mucosal sites or transfer of bacterial populations through processes such as inhalation of airborne bacteria, bacterial migration along mucosal surfaces and microaspiration of gastric contents. 61 Although many studies of the microbiome focus on faecal contents, it is important to note that intestinal helminths infect specific niches within the intestine: that is Trichuris spp in the large intestines, and human hookworms, H. polygyrus and N. brasilienesis in the small intestines. In humans living in helminth-endemic areas and during experimental helminth infection, helminth infection is associated with increased diversity and abundance of the microbiome. [62][63][64][65] Although it has not been demonstrated whether these differences are related to differences in lifestyle and hygiene or causally linked to current helminth infection, the fact that changes in the microbiota are partially abrogated on anthelmintic treatment supports the latter hypothesis. 63 In several mouse models, intestinal helminth infections induced a decreased prevalence of commensals associated with inflammation 66,67 and increase in commensals associated with immune regulation. 63,68 This is proposed to be an active process, mediated by secreted antimicrobials from the parasite (such as host defence peptides), 69 or mediated by (type 2) immune responses directed against or modulated by the parasite. 63,70 Consequently, changes in the microbiota (due to, eg changes in diet) and microbial metabolite levels (such as short-chain fatty acids) mediate changes in allergic responsiveness. 71 Interestingly, a recent study also found altered fatty acid production by the microbiota of H. polygyrus-infected mice and their additional role in protection against allergic airway inflammation. 72

| RE WORMING THE WE S T
The growing support for the idea that helminth infections suppress inflammatory responses led to the proposal of using helminth infections as therapeutic agents in these diseases. In the first clinical trials of "helminth therapy," patients with inflammatory bowel disease (Crohn's disease or ulcerative colitis) were treated with eggs (ova) from the porcine intestinal parasite Trichuris suis (TSO), leading to significant reduction in symptom scores in a series of small trials. 73 These initial studies formed the basis of clinical trials treating patients with IBD or multiple sclerosis in the United States and Europe with TSO. However, to date, these trials have shown disappointing F I G U R E 1 Helminth infections are associated with both promoting and reducing allergic symptoms. Helminths interact strongly with the host immune system, but the type of response is heavily influenced by the chronicity of infection, the species involved and/or worm burden, ultimately tipping the balance towards more detrimental and type 2 response or beneficial and regulatory responses. Subsequently, this balance is further influenced by cofactors such as host genetics, socioeconomic status, coinfections and the composition and diversity of the microbiome leading to the development of clinical symptoms and allergies or tolerance in the host. Image is adapted from Servier Medical ART response rates and no significant reduction compared with placebo controls. [83][84][85] Of most relevance to this review, studies using hookworm infection to treat asthma or TSO to treat allergic rhinitis patients have also been undertaken. However, no change in clinical measurements was seen in either study, 86 and although type 2 specific responses were detected against the hookworm, no regulatory immune responses were found. 87 Thereby, the promise of helminth therapy has so far not translated to a practicable treatment for human disease. Reasons for this may well include difference between prevention and cure (ie parasitic infection may need to precede allergic sensitization, and effect could be in utero 49 ), the difference between infection and administration of parasite products (most of the therapeutic effects in mouse models were based on the application of helminth products rather than a full infection), the single parasite infective dose given (which is generally determined by that which causes no notable side effects, but may therefore be too low and too little for functional suppression of pathology) or disease endotypes which are responsive or refractory to these treatments, precluding statistical significance of effects when the disease population is taken as a whole. 88 Critically, however, the mechanism of action of helminth immunomodulation is not well understood, and whether this is shared between all helminth infections, or more likely unique to each parasitic species, is presently unknown ( Figure 1).

F I G U R E 2 Immune responses during helminth infections.
Depending on their life cycle, various helminth species will pass or reside in (the proximity of) the lung and the gut. Consequently, damage will occur, leading to the release of alarmin cytokines IL-33, IL-25 and thymic stromal lymphopoietin (TSLP) by epithelial cells and tuft cells (gut). These cytokines will act on innate lymphoid cells (ILC)2 and on dendritic cells (DCs), which will migrate to the draining lymph node and skew naïve T cells towards polarized Th2 cells, producing the cytokines IL-4, IL-5 and IL-13, in a similar fashion as ILC2s. These cytokines are central to the type 2 immune response and drive the isotype switch to IgE immunoglobulins, act on eosinophils, mast cells and drive the development of alternatively activated macrophages. All these elements are instrumental in worm expulsion but can also promote tissue damage, anaphylaxis and allergic responses towards bystander antigens. These responses are balanced by various cells from the regulatory network: for example regulatory T and B cells, regulatory macrophages and tolerogenic DCs. These regulatory cells can act on other cell types directly or through the production of anti-inflammatory cytokines IL-10 and TGFβ, as well as by the induction of anti-inflammatory IgG4, leading to immune tolerance and damage control, but at the same time prevent worm expulsion, promoting chronic helminth infections. Image is adapted from Servier Medical ART In the following sections, we will focus on mechanisms by which type 2 immune responses are suppressed or induced in helminth infections, and how this could affect allergic responses (Figure 2 for a schematic overview). A deeper understanding of the interaction between helminths and their host will help to translate these mechanisms into a better therapeutic approach. The immunoglobulin isotype IgG4 is often associated with a tolerized allergic response following allergen-specific immunotherapy, and its production is also increased in many helminthinfected individuals. 99 Thus, in this context, its main function appears to be a blocking one, and possibly instrumental in preventing IgE-mediated inflammation. High levels of anti-Ascaris IgG4 have been negatively associated with allergen SPT positivity, 102 while in a S. mansoniendemic area-although higher levels of both IgE and IgG4 were found in infected individuals-a higher ratio of IgE to IgG4 predicted clinical allergic symptoms, 103 just as in allergen-specific immunotherapy. 104 As many helminth products are homologous to common allergens, IgG4 responses raised against helminth products may also bind and block IgE epitopes on allergens, reducing responses to allergens and directly reducing SPT responses. [105][106][107] Mechanistic research into the role of IgE and/or IgG4 is hampered by the lack of good experimental animal models, as IgG4 does not exist in mice. Analysis of splenic CD1d hi B cells from schistosome-infected mice showed increased Tlr7 expression, and TLR-7 ligation increased the IL-10 production in splenic CD1d hi B cells from naïve animals. 124 Adoptive transfer of TLR-7 stimulated splenic CD1d hi B cells reduced allergic airway inflammation through the recruitment of regulatory T cells. Further mechanistic insight was recently supplied by the finding that the S. mansoni-derived molecule IPSE/alpha-1 could drive Breg differentiation in vitro. 125 In addition, and separately to "con- FcεRI. 149 In the absence of DCs, type 2 responses in allergy models are profoundly abrogated, 150 but other myeloid cell populations may also be important to support Th2 cell development in either allergy models or helminth infection, like monocyte-derived dendritic cells. 140,141 Macrophages differentiate into alternatively activated or M2 macrophages in response to IL-4 and IL-13. They can be distinguished from classically activated or M1 macrophages by the upregulation of markers such as RELMα, Ym1 and arginase and are associated with wound healing and parasite killing. 11,151 M2 macrophages are also anti-inflammatory, producing IL-10 and TGFβ 152 and arginase, which restrict T-cell function through amino acid starvation 153 and suppress liver fibrosis in S. mansoni infection. 154 Furthermore, retinoid acid production by M2 macrophages during S. mansoni infections promotes the development of Treg cells at the sites of inflammation. 155 In allergic asthma, macrophages are considered to play a key role in inflammatory responses associated with lung injury, fibrosis and goblet cell hyperplasia 156 161,166,167,[176][177][178] Further studies are needed to clarify whether these molecules can also be used in allergic and asthmatic patients in a therapeutic setting without the disadvantages of the infection itself.

| Innate lymphoid and epithelial cells
The importance of early, innate, epithelial cell-derived cytokines in type 2 response initiation has only recently begun to be understood.
Recently, it was shown that ILC2s can also be activated to produce IL-10, providing a immunoregulatory pathway (similar to that seen in T cells) which could be amenable to parasite immunomodulation. 9 Proximal epithelial cell responses therefore represent an ideal target for intervention, as blocking these cytokines could blunt protein in HES which directly binds IL-33, abrogating its release. 54 Recently, it was shown that cholinergic neurons activate [182][183][184] and adrenergic neurons inhibit 185 ILC2 responses, while type 2 cytokines in turn activate neurons, 186 forming a new field of neuroimmune interactions for future studies, and potential helminth modulation.

| CON CLUDING REMARK S
Helminth parasites have coevolved with their mammalian hosts for millions of years, and in doing so have developed an intimate relationship. On the one hand, this relationship can be seen as antagonistic, with the parasite attempting to subvert immune responses to its own ends, while the host immune system attempts to damage or kill the parasite. Alternatively, it can be seen as a mutualistic interaction, with parasite immunomodulation expected by the host immune system, and in fact required for healthy immune development, and avoidance of immune-mediated disease. It is likely that both forms of interaction occur, depending on context, parasite and environment. "Gene-environment" interactions in allergy and helminth infections are further complicated by "infection-