The relationship between infections and autoimmunity is complex. Microbial and viral infections might act as environmental triggers inducing or propagating autoimmune and inflammatory processes, resulting in symptomatic presentation of a disease in genetically high risk individuals (Lerner, 2015; Arleevskaya et al., 2017). An autoimmune disease onset following an infectious agent exposure has been well-documented (Pordeus et al., 2008; Bogdanos et al., 2015; Sakkas and Bogdanos, 2016). At least, for CD, the following infections were suggested to be associated with the disease: viruses: enterovirus, Epstein-Barr virus (EBV), Cytomegalovirus (CMV), hepatitis C virus (HCV), hepatitis B virus (HBV), and rotavirus, microbes: Bacteroides species, Campylobacter jejuni, Pneumococcus, Mycobacterium tuberculosis, and Helicobacter pylori (Lerner, 2015). However, recent serological evidence suggests the opposite outcome, which is the protection against autoimmune conditions following bacterial/viral exposure (Christen and von Herrath, 2005). At least the suggestive protector agents for CD were CMV, EBV, Rubella, and Herpes simplex type 1 virus (HSV1) when compared to healthy people (Plot and Amital, 2009; Jansen et al., 2016). According to the “hygiene hypothesis,” the excessively sterile environment leads to the enhanced incidence of autoimmune disorders, asthma, and allergies, thus, associating surge of CD incidence with decreased infectious environment (Lerner et al., 2015b,e; Bloomfield et al., 2016).

Celiac disease is a life-long autoimmune disease (Lerner et al., 1996) mainly of the proximal intestine, affecting genetically predisposed individuals. Gluten, the storage protein of wheat, is the environmental inducer of the disease in addition to other structurally related molecules found in barley, rye, and oat (Lerner, 2014). Many environmental factors were suggested to induce or enhance the disease: multiple infections (Lerner and Reif, 2015), early infections (Myléus et al., 2012), early gastrointestinal infections (Beyerlein et al., 2017), lack of breast feeding (Lerner and Matthias, 2016c), time and amount of gluten consumption (Chmielewska et al., 2015), microbiome/dysbiome repertoire (Lerner et al., 2015a; Lerner and Matthias, 2017a,b), mode of delivery (Decker et al., 2011), early vaccination (Kemppainen et al., 2017) or early consumption of antibiotics (Canova et al., 2014) and geo-epidemiological influences (Lerner, 1994, 2015; Reif and Lerner, 2004b; Lerner and Matthias, 2015a). The abnormal immune response is directed, in particular, against tissue transglutaminase (tTG), representing the autoantigen, (Reif and Lerner, 2004a; Lerner et al., 2015c) and the two main autoantibodies, anti-endomysium and anti-tTG antibodies, are the most prevalent serological markers used to screen for the condition (Shamir et al., 2002; Lerner and Matthias, 2015d). Recently, the list of CD serological markers was expanded by two additional autoantibodies: anti-deamidated gliadin peptide and anti- tTG neo-epitope antibodies, found to be reliable for CD diagnosis (Rozenberg et al., 2012; Lerner and Blank, 2014; Lerner et al., 2015e). As yet, HLA-DQ2 and HLA-DQ8 are known predisposing genetic factors. The sequential events in disease progression were unraveled in the last years and gave rise to multiple future therapeutic strategies (Lerner, 2010). Notably, its epidemiological, incidental, and clinical presentation are changing continuously, and new clinical pictures are reported and expand the abundance of clinical variance of the disease (Lerner et al., 2015d). In fact, age of disease onset increases and the traditional enteric presentation is more and more replaced by extraintestinal manifestations. Skin (Lerner et al., 2015d), endocrine (Lerner and Matthias, 2016d; Lerner et al., 2017), hepatic (Anania et al., 2015), metabolic (Eliyah Livshits et al., 2017), skeletal (Lerner and Matthias, 2016a), rheumatic (Lerner and Matthias, 2015b), geriatric (Lerner and Matthias, 2015c), hematological (Branski et al., 1992), neurological (Zelnik et al., 2004; Lerner et al., 2012), gynecological and infertility (Mårild et al., 2012; Casella et al., 2016), oral and dental (Cantekin et al., 2015), hypercoagulability (Lerner and Blank, 2014), cardiac (Lerner et al., 2015d), and behavioral abnormalities (Zelnik et al., 2004) are often described. Those epidemiological and clinical changes can explain why the disease is diagnosed during the whole human life-span including in the elderly (Lerner and Matthias, 2015c). There is no doubt that in the last decades its incidence is constantly increasing, ranging between 1 and 3% nowadays (Lerner, 2014; Lerner et al., 2015b). The present review will concentrate, expand and update on the multiple faces of the inductive/protective roles that infectious agents might play in CD pathogenesis. This aspect is further interesting since pathogens are the major drivers of human selective genetic adaptation during evolution (Vatsiou et al., 2016), and the question of microbes that are bugging the celiac patient “are they friends or foes?” is the subject of the current review.

It should be clarified that, although the trigger role of microorganisms and viruses in the CD development was undoubtedly traced in numerous investigations, it substantially differs from other immune pathogenesis like rheumatoid arthritis (as a classic model of an autoimmune disease; Arleevskaya et al., 2016; Kemppainen et al., 2017).

The induction of rheumatoid arthritis most likely occurs under the influence of the burden of many trivial infections, influencing the patient's immune system due to the frequent and prolonged infectious episodes (Arleevskaya et al., 2014). Individuals at CD risk apparently do not have such features in their mucosal immunity, nor significant defects in systemic anti-infective protection, impacting infection susceptibility. Since all the CD studies are focused only on the disease link with various gastrointestinal infections, such association is different from what was shown in rheumatoid arthritis (Riddle et al., 2012).

The number of infectious agents related to CD is continuously increasing (Figure 1). Examples for viruses enterovirus, EBV, Cytomegalovirus (CMV), HCV, HBV, and rotavirus And for microbes Bacteroides species, C. jejuni, Pneumococcus, M. tuberculosis, and H. pylori (Lerner, 2015). However, links between CD and infections were more associative and less causative, thus, far from being elucidated. Moreover, the mutually exclusive hypotheses about the provocative and protective role of a particular microorganism/virus in CD pathogenesis were suggested and discussed in various publications.

It appears that the essential condition for the CD induction is a gastrointestinal infection (Kemppainen et al., 2017). Apparently, the other major condition is an early childhood—namely immature gastrointestinal tract, immature immune system, and gastrointestinal microbiome at the early phase of formation. The beginning of gut colonization by microorganisms set the stage for the cross talks between the epithelium, enteric lymphoid tissue, and microflora, all together establishing the intestinal barrier and as a consequence, a strictly dosed delivery of macromolecules into the internal environment and shaping of mucosal tolerance to food antigens and normal flora (Makarova et al., 2014).

So, in the vulnerable infant period any gastrointestinal infection, even a transitory one, is potentially able to disturb these processes of gut microbiome maturation and the establishment of local immunity and immune tolerance, including that against food and microbial antigens. Apparently, deleterious coincidence of these circumstances leads to an error in the negative selection of gluten-reactive lymphocyte clones.

Thus, the community of microorganisms, being extremely vulnerable during the ripening period, appears to be an inert system in adults. For example, in adults it needs no more than 30–60 days to restore gut microflora after the exposure to antibiotics (Spanhaak et al., 1998; Tannock et al., 2000). At the same time, a weekly clindamycin treatment of a newborn reduces Bacteroides diversity for the next 2 years (Jernberg et al., 2007). The same infections, capable of altering the fate of a sick infant, are merely a light ripples on the ocean surface for an adult microbiome.

In conclusion, there is an undoubted link between CD development and microorganisms, and this link looks to be rather specific. Gastrointestinal infections in predisposed infants with an immature gastrointestinal tract and immune system might shape gut microbiome in the immature and therefore labile circumstances. Such an unfortunate combination could trigger early CD development or becomes a ticking time bomb, represented by the structural features of the gut microbiome and persistence of gluten–reactive lymphocyte clones with latent basal cell proliferation without overt disruptive inflammatory activation.

It should be specified that Jansen and coauthors determined the antiviral antibody levels in the sera samples of 6-year old children (Jansen et al., 2016). CMV single infection and combined CMV, EBV, and/or herpes simplex virus type 1 infection antibodies were inversely associated with strongly tTg-IgA positivity. The authors suggested that the serological profile may indicate a protective effect of herpesvirus infections in the pathogenesis of celiac disease autoimmunity. It is accepted that by this age IgG and IgM production is close to that of the adults. However, the immune system is still immature and its reaction is not fully functional at this age even in healthy children, whereas the immune system of CD prone children, at least, the local immune defense, is somewhat delayed in the maturation. Thus, Jansen et al. suggested explanation should be taken with a grain of salt.

In Plot's publication, the age profile of the studied cohorts was not presented (Plot and Amital, 2009). There are some curious details in this publication which need to be interpreted. First, while the prevalence of the anti-EBV capsid antigen and anti-EBV nuclear antigen IgG in the CD patients was significantly lower than that in the controls, the prevalence of the anti-EBV early antigen IgG was comparable in the two groups and the prevalence of anti-EBV capsid antigen IgM, though unreliably, was more than twice higher in the CD group. In general, the anti-EBV capsid antigen IgMs are known to be produced during the acute phase (the first days—6 months from the onset of the disease) or during acute exacerbation of chronic EBV infection (http://www.tiensmed.ru/news/epstein-barr-bc1.html#nov1). In addition, if the anti-EBV early antigen IgG is not concurrently revealed, it may indicate incubation period or the very beginning of the infection (up to 1 week of symptoms). So, either in the CD patients this generally latent infection is apt to more frequent exacerbations, or, given the reduced frequency of the studied IgG antiviral antibodies, the existence of some features of the antiviral antibody production in CD can be assumed. Secondly, revealed by PCR, CMV DNA presence in the samples is not always accompanied by the presence of serum specific antibodies. This situation is typical in particular for the infants with immature immune system (Dong et al., 2004, 2005).

A number of publications on the HBV vaccine non-responsiveness in CD might attest for possible unique features of the antiviral antibody response formation (Noh et al., 2003; Park et al., 2007; Urganci and Kalyoncu, 2013), further criticizing the assumption that decrease antibodies activities in CD patients represent a protective effect.

As for the inverse correlation of the antiviral and anti-tTG antibodies demonstrated by Jansen et al. (2016), similar patterns are not uncommon in autoimmune diseases, and this fact by no means demonstrates a lesser exposure to the infections. For example, high levels of autoantibodies against double-stranded DNA, reflecting the activity and severity of systemic lupus erythematosus, are quite often combined with lower levels of antibodies to particular bacterial DNA (Pisetsky and Drayton, 1997). The inverse correlation of those indexes is usually explained by a distorted immune humoral response.

Additionally, it should be specified that, in general, in Herpesvirus infections and viral hepatitis the specific antibodies are not the principal players in the antiviral defense but are accepted to be the reliable serological markers for an infection (Grinde, 2013). It is advisable to note that caution should be used for the interpretation of the protective role of these infections in CD, relying only on data on the incidence of the corresponding antibodies.

As for Rubella infection, the specific antibodies are the major antiviral response players. However, the incidence of CD in children vaccinated with inactivated rubella virus as part of polio vaccine was close to that in the unvaccinated children (Myléus et al., 2012). Thus, a reduced incidence of the anti-rubella IgG antibodies demonstrated by Plot and co-workers can mean an equally probable lower exposure of CD patients to Rubella and the above-discussed features of the specific IgG antibody production. Moreover, since the antiviral immune response is always multi-componental, the disturbed antibody formation, although being a weak link of an antiviral defense, does not necessarily entail an increased susceptibility to Rubella.

The data which might testify for the direct and reverse links of CD to Herpes virus infection are summarized in Table 1. The analysis of the data demonstrates that the links of CD and herpes infections are multi-directional. On one hand there are some peculiarities, which can promote the viral infecting: (1) CD-associated DQA1^*0501/DQB1^*0201 genotype, which is also due to the imperfect response against Herpes viral infection; (2) the typical CD immature gastro—intestinal tract and the delayed process of microbiome maturation, which might be risk factors for the virus infecting; (3) the typical CD mucosal overexpression of epidermal growth factor receptors, by which Herpes viruses enter the cells; (4) increased expression of IL-33, suppressing local antiviral immunity in CD patients. On the other hand, the increased levels of several humoral factors with the antiviral activity; increased expression of some cytokines, which promote mucosa maturation and thus increase its' resistance to the viruses; as well as some potential features of CD microbiome that might indicate a backward link between CD and Herpes viral infection. The protective effect of the infection on atopic manifestations was demonstrated in the case of the early (infancy or early childhood) EBV exposure, while the later infection predisposes to the atopic disease (Nilsson et al., 2005, 2009). So, if to extrapolate the data on the links of herpes and atopic diseases to CD, it is likely that the early (infancy or early childhood) EBV exposure might play a protective role, while the later infection might trigger CD or have no impact on it at all.

The permanent interest in CD and Hp infection coexistence is quite natural, due to the gut-stomach axis (Lerner and Matthias, 2016b). The infectious inflammatory process directly in the gastrointestinal tract—CD epicenter, which might shape the local immune system and microbiota, might obviously play a role in CD pathogenesis. In addition, both CD and Hp infection in a number of cases are associated with the diffuse lymphocytic gastroenteropathy (Lynch et al., 1995; Broide et al., 2007; Pai, 2014). However, diffuse lymphocytic gastroenteropathy is far from being obligatory attributed only to both entities (Wu and Hamilton, 1999; Nielsen et al., 2014). Besides, lymphocytic gastritis and a subsequently villous atrophy are accepted to be a non-specific manifestation of many pathological conditions in the gastrointestinal tract, due to a wide variety of infectious, immunologic or any inflammatory stimuli raising intraepithelial lymphocyte numbers. Lymphocytic duodenitis and increased intraepithelial lymphocytosis are known to be associated with diseases that are completely different in their pathogenesis, such as autoimmune disorders like CD (Broide et al., 2007; Rostami et al., 2010), tropical sprue, food protein intolerance, Hp-induced duodenitis, peptic duodenitis, parasitic, and viral infections, intestinal lymphoma (Chang et al., 2005; Brown et al., 2006; Pallav et al., 2012; Rosinach et al., 2012; Shmidt et al., 2014) drugs' induced duodenitis (non-steroidal anti-inflammatory drug; Shmidt et al., 2014) and small-intestine bacterial overgrowth (Lappinga et al., 2010). The differential diagnosis of lymphocytic gastritis is not less restricted. CD and HP are not the only ones. Various non-HP infections, inflammatory conditions and several non-celiac autoimmune diseases were described (Broide et al., 2007; Polydorides, 2014).

The interest in the CD—HP infection link is fueled by the well-known data, indicating that childhood infection with HP could protect against the development of Crohn's disease, severe gastric-reflux disease, Barrett's esophagus and esophagus adenocarcinoma (Chen and Blaser, 2008). Yet, as for the protective role of Hp in CD, the available data are limited and quite contradictory. The prevalence of CD among Hp-positive adults was 0.05% compared with 0.09% among Hp-negative individuals (statistically non-significant) while the prevalence of Crohn's disease among Hp–positive patients was 0.07% compared with 0.24% among Hp-negative patients (Bartels et al., 2016). Based on these data, at least in adults, the protective effect of Hp on CD is minimal, if at all existing. However, given the fact that in childhood gastrointestinal infection appears to be a more important condition for CD triggering than in adults, this conclusion is not necessarily true in the case of early exposure to Hp (Kemppainen et al., 2017). Unfortunately, we failed to find the similar data on the CD frequency in children infected with Hp early in life.

The incidences of Hp and CD worldwide vary enormously (9–100%, 0.3–3.9%, respectively; Rostami-Nejad et al., 2016). Most of the studies aimed to determine the ratio of Hp infected individuals in CD and non-CD control groups. In our opinion, this study design gives less information about the inductive/protective effect of Hp infection on the CD development as on the susceptibility of the CD patients for the infection. The results of various studies on Hp both in children and adults are contradictory (Table 2). The spread in the ratios of both Hp infected CD patients and non-CD controls in these publications might be due to the national and age-related characteristics of the studied cohorts (Eusebi et al., 2014). As for the diametrically opposed regularities of HP incidence in CD and non-CD groups, a meticulously analysis of these publications, did not lead us to any reason for the conflicting results, but to the possible differences in the poorly described clinical characteristics of the controls. In these studies, the authors examined the collections of endoscopically obtained biopsies and sera allocating the cases into CD and non-CD (control) groups. It is obvious that the persons accepted as controls underwent the relevant examination because they had any gastroenterological problems associated or not associated with HP infection.

The analysis of the literature data which might testify the direct and reverse links of CD and Hp infection (Table 3) shows that the interaction of the two diseases represents an interweaving of differently directed processes. Perhaps the end result might depend on the balance of these processes, being deeply individual in each specific case. Important is the assumption that the direct or reverse CD/HP link may depend on the age at which the encounter with the bacterium occurred. At least, it is important for allergies–Hp links. Our attempt to test this hypothesis failed, because the literary data accumulated to the present moment are largely insufficient.

Endogenous and microbial enzymes are capable to generate intestinal enzymatic neo-antigens via PTMP. The modifications taking place in the intestinal lumen include peptides crosslinking, de/amination/deamidation by the transglutaminases, de/phosphorylation, a/deacetylation, de/tyrosination, and many other enzymatic modifications exist (Lerner et al., 2016). Related to the present topic, the human endogenous intestinal enzymes, tTG and its family member, the exogenous microbial transglutaminases, induce multiple neo-epitopes on the TG-gliadin cross-linked complex resulting in the formation of antibodies against the complexes in CD. CD is a classical disease where luminal PTMP is driving the disease. It seems logical that a microbial agent might modify non-self-peptide to self-one—reducing its immunogenicity. Additionally, in CD, some microbe strains might modify gluten in the lumen, thus preventing or aggravating the inflammatory cascade and the intestinal damage progression via PTMP (Caminero et al., 2016).

Given the extensive influence of the microbiota on human health, the gut-microbiome integrity is of prime importance for host health and survival. In this regards, our bodies' “second genome” cohabit with the human one to form a stable equilibrium for the two kingdom's long term survival. As opposed to long-term evolutionary events, newer genetic manipulations with microorganisms, plants, animals, or nutrients, applying new food technologies and/or microbial engineered delivery systems or novel mode of therapies are rapidly evolving.

Due to the close relationship and intimate cross-talks between the human and the gut's biospheres, consumption of the modified genetic cargo into the human intestinal ecosystem might occur. It was hypothesized that modern probiotic ingestion, genetically manipulated food consumption and genetically manipulated microorganism usage are potential genetic driving forces for changing the evolutionary equilibrium established during the last millions of years (Cho and Blaser, 2012). Horizontal gene exchange is the ability to transfer genetic material between contacting biological domains, including eukaryote (plants, animals, and man), prokaryote (microbes), and viruses (Aminov, 2011; Ruggiero et al., 2015). Despite not being investigated in CD, various virulent genes, the most studied one is the antibiotic resistance gene, were described to be laterally transferred. It is hypothesized that the opposite might occur. This infectious genetic cargo might include anti-inflammatory/pro-apoptotic/Treg or other immune-modulatory genes, attenuating or abrogating autoimmunity.

The tight junction protein, Zonulin, is involved in the regulation of the intestinal permeability between gut epithelial cells. Several clinical trials with Zonulin antagonist (Larazotide acetate, AT-1001, Alba Therapeutics, USA) demonstrated the promising therapeutic effect in CD (Lerner, 2010; Khaleghi et al., 2016). Larazotide acetate—an octa-peptide derived from a cholera toxin ZO, antagonizing zonulin via receptor blockade—is aimed to decrease the paracellular transport caused by gluten and thus to suppress the activation of the pathological immune cascade. In addition to the above mentioned cholera toxin derivative, many other factors produced by microorganisms can improve tight junction performance, modulating intestinal permeability. Salmonella enterica serovar, Escherichia coli, and C. jejuni modulated enteric epithelial barrier functions in chickens (Awad et al., 2012, 2014, 2015), and the probiotics Lactobacillus casei DN-114001 and E. coli strain Nissle 1917 decreased intestinal epithelium permeability in human intestinal originated cell lines (Parassol et al., 2005; Zyrek et al., 2007; Trebichavsky et al., 2010). Taken together, modulation of gut permeability by infectious agent, counteracting the breached tight junction integrity in CD, might represent a protective pathway.

Generally, molecular mimicry between foreign (infectious/environmental) and self-antigens is a well-described pathway of autoimmune disease induction. Recently, it was suggested that antigen mimicry between foreign and self-antigens might be due to the long-term regulation of inflammation (Pontes-de-Carvalho et al., 2013). In a cohort of African patients, infected with Schistosoma it was found that the parasite inhibited production of anti-nuclear antibodies (Mutapi et al., 2011). More recently, apparent effectiveness of rotavirus vaccination was found to prevent the onset of CD autoimmunity (Silvester and Leffler, 2016). This finding gives indirect evidence for the persistence of regulatory cells in the lack of stimulation of the immune system by pathogen-derived processes. The strength of the immune adjustment, however, may increase with the uninterrupted presence of the pathogen or its antigens. Thus, molecular mimicry can represent a protective mechanism of autoimmunity.

Inflammation, given rise by microbes, viruses and especially by parasites such as helminths, can shift the Th1 pathway to Th2 one, resulting in a more immunosuppressive state where regulatory T cells might be induced or be activated (Shor et al., 2013). Recent studies have supplied such an evidence for pathogen-specific regulatory cells in Leishmania major, Herpes simplex virus, and Friend retrovirus (murine leukemia virus) infections (Christen and von Herrath, 2005). During the acute phase of the infection this Th2 profile counter-regulates Th1-driven autoimmune pathologies. Along the chronic stage of infection, immune-regulatory networks arise, mainly led by regulatory T cells. These cells produce IL-10 and TGFβ, which has observative effect on Th1-related autoimmune diseases, such type 1 diabetes mellitus or CD. In fact, several helminths were tried in CD patients with encouraging results (Croese et al., 2015). Treatment with helminthes or helminthes ova ameliorated the clinical pictures of several autoimmune conditions in patients as well as in animal models (Smallwood et al., 2017). A major recent contribution to the field is the helminth phosphorylcholine proved to be an immunemodulatory molecule. Most recently, tuftsin-phosphorylcholine, a novel helminth-based compound was shown to reduce pro-inflammatory cytokine production and induced anti-inflammatory cytokine expression and Treg and Breg cell expansion in mouse models of rheumatoid arthritis, lupus nephritis, and colitis (Bashi et al., 2015b, 2016; Shor et al., 2015).

It is conceivable, that the ability of helminthic parasites to attenuate host immune responses into an anti-inflammatory/regulatory phenotype is attributed to the endogenous component that the parasites secrete and/or excrete interacting with immune effector cells to regulate their function (Lund et al., 2014; Selmi, 2016).

An additional mechanism was suggested for the helminth's immunomodulation of autoimmunity, in addition to the Th1 to Th2 shift. Accelerated T and B regulatory phenotypes, decreased levels of the inflammatory cytokines like IFNγ and Il-17 or vice versa, promoting IL-4, IL-10, and TGF-β release (Bashi et al., 2015a). Since CD is a Th1 profile disease, shifting the immune pathway to Th2 profile might reduce the intestinal damage (Lerner, 2010).

Inflammation can cause a substantial hyperactivation of auto-aggressive lymphocytes, leading to activation-induced cell death and attenuate the systemic load of aggressive T cells. It seems that repeated encounter with powerful antigenic stimuli leading to restriction of an immune response is well-established in viral infections, where the primary response undergoes a major restriction after antigen elimination. EBV, HBV, and CMV infections are some of the examples. Similarly, administration of mycobacterial products, such as bacilli Calmette-Guérin, prevented the onset and recurrence of type 1 diabetes mellitus in NOD mice by inducing apoptosis of autoreactive T cells (Christen and von Herrath, 2005). In view of the fact that viruses are inducers of immune cells apoptosis while sparing the Treg cells (Che et al., 2015) and apoptosis is enhanced in CD (Shalimar et al., 2013), it is suggested that viruses, by abrogating immune activation, might attenuate intestinal autoimmune progression in CD.

As suggested by Christen (Christen and von Herrath, 2005), an infectious inflammation elsewhere in the body might keep auto-aggressive cells from arriving into the sites of autoimmune destruction, that might be due to the abrogation of type 1 diabetes in NOD mice after LCMV infection. The authors suggested that this occurred because the “abrogative” virus grew predominantly in peripheral lymphoid organs and other sites rather than the pancreas or its islets themselves. Thus, the sites of severe inflammation might act as a filter for auto-aggressive T cells removing them from the circuit and depriving them from homing the pancreatic islets. Similar scenarios might operate where infection with B. coxsackievirus or Salmomella typhi murium protected against autoimmunity (Tracy et al., 2002; Raine et al., 2006).

Release of extracellular vesicles is a natural phenomenon of almost all cell types. They derive either from multivesicular bodies or from the cellular plasma membrane. Those vesicles contain a subset of cell derived proteins, lipids, including nucleic acids. Extracellular vesicles regulate immune responses against pathogens, as well as autoimmunity. It is suggested that these suppressive vesicles would prevent peripheral self-antigens and commonly encountered foreign antigens from causing chronic inflammation and autoimmunity. Following this lines, it is hypothesized that various infectious agents can induce those regulatory extracellular vesicles, counteracting autoimmune pathways, playing a protective anti-autoimmune role (Robbins and Morelli, 2014; Robbins et al., 2016).

Infections with helminths can prevent or attenuate auto-inflammatory/immune diseases. In addition to their Th1 to Th2 shift, most recently, Helminth secreted proteins were shown to prevent autoimmunity. The excretory/secretory products of Fasciola hepatica contain immune-modulatory molecules that arbitrate protection from autoimmune diabetes via the activation and provision of a regulatory immune environment (Lund et al., 2014). Such a mechanism was not studied in CD, but might explain the new potential therapeutic strategy to treat CD with Necator Americanus larvae (Croese et al., 2015; Giacomin et al., 2015).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.