Development of innate immunity in chicken embryos and newly hatched chicks: a disease control perspective

ABSTRACT

Newly hatched chickens are confronted by a wide array of pathogenic microbes because their adaptive immune defences have limited capabilities to control these pathogens. In such circumstances, and within this age group, innate responses provide a degree of protection. Moreover, as the adaptive immune system is relatively naïve to foreign antigens, synergy with innate defences is critical. This review presents knowledge on the ontogeny of innate immunity in chickens pre-hatch and early post-hatch and provides insights into possible interventions to modulate innate responses early in the life of the bird. As in other vertebrate species, the chicken innate immune system which include cellular mediators, cytokine and chemokine repertoires and molecules involved in antigen detection, develop early in life. Comparison of innate immune systems in newly hatched chickens and mature birds has revealed differences in magnitude and quality, but responses in younger chickens can be boosted using innate immune system modulators. Functional expression of pattern recognition receptors and several defence molecules by innate immune system cells of embryos and newly hatched chicks suggests that innate responses can be modulated at this stage of development to combat pathogens. Improved understanding of innate immune system ontogeny and functionality in chickens is critical for the implementation of sound and safe interventions to provide long-term protection against pathogens. Next-generation tools for studying genetic and epigenetic regulation of genes, functional metagenomics and gene knockouts can be used in the future to explore and dissect the contributions of signalling pathways of innate immunity and to devise more efficacious disease control strategies.

KEYWORDS:

• Chicken embryos
• newly hatched chickens
• innate defences
• pathogens
• PRRs
• ontogeny

Introduction

The immune system of neonates of all vertebrates is characterized by the progressive development of innate and adaptive defences, and by inadequate immunological memory, predisposing them to microbial infections (Wynn & Levy, 2010). During this period, development of innate defences is mainly stimulated by interaction with commensal microbiota (Levy, 2007; Jiao et al., 2009; Thaiss et al., 2016). Innate responses consist of cellular defences and soluble mediators, and operate as constitutive barriers or as inducible biochemical responses (Medzhitov, 2001). Classical innate responses of host cells are initiated after the recognition of microbial molecular motifs by pattern recognition receptors (PRRs) that are constitutively expressed on host immune cells (Akira et al., 2001; Akira, 2009). The engagement of PRRs, such as Toll-like receptors (TLRs), with pathogen-associated molecular patterns (PAMPs) triggers intracellular signalling pathways leading to the induction of effector molecules, including chemokines, cytokines and host defence peptides (HDPs) (Alexopoulou et al., 2001; Schoggins et al., 2011), which, in turn, coordinate specific cell- and antibody-mediated immune responses contributing to pathogen clearance (Palm & Medzhitov, 2009; Dar et al., 2015).

In vitro and in vivo studies in mammals have revealed marked differences in the magnitude and quality of innate responses of neonates compared to those of adults. The underlying immunological basis for age-dependent variation has been widely investigated in mammals with the discovery of several PRRs expanding the understanding of the mechanisms involved in the ontogeny of innate responses (Levy, 2007; Wynn & Levy, 2010). In chickens, however, there is a paucity of information on age-dependent development and initiation of innate responses. Since extensive differentiation and proliferation of innate immune system cells occurs during embryonic development and in the first few days post-hatch, constitutive and inducible levels of innate responses are expected to differ in embryos and chicks compared to mature chickens. One study indicated that, although the constitutive expression of global innate response genes was higher in one-week-old chicks, induced expression of these genes post infection was higher in four-week-old chickens (Reemers et al., 2010a). This suggests an increase in the magnitude of inducible immune competence with increasing age.

Exploring and understanding the relevant molecular mechanisms that cause qualitative and quantitative differences in innate responsiveness in embryos and newly hatched chicks compared to adult chickens of different genetic backgrounds may allow for targeted immunomodulation in ovo, or a few days post-hatch, subsequently leading to enhanced resistance to infections. Therefore, this review summarizes current knowledge on the ontogeny of the innate immune system as well as cellular and molecular mechanisms involved in initiation and control of innate responses during the embryonic stage and early post-hatch period.

Ontogeny of innate immune system cells and their functional characterization

The major components of the innate immune system in vertebrates are mucosal barriers, soluble serum proteins, and cells, which include phagocytic cells, γδ T cells and natural killer (NK) cells (Delves & Roitt, 2000). Phagocytic cells such as granulocytes and mononuclear phagocytes are derived from myeloid lineages, while γδ T cells and NK cells have a lymphoid lineage origin (Melchers, 2010). These cells arise from single pluripotent haematopoietic stem cells (HSCs) from the yolk sac, indicating an extra-embryonic origin of haematopoiesis (Melchers, 2010). In chicken embryos, the presence of a large number of CD2+ cells in the peripheral blood as well as CD45+ cells in the yolk sac of 2-day-old embryos prior to liver and thymus development implies yolk sac origin of immune cells (Zhou et al., 2008; Dóra et al., 2017). Furthermore, the detection of pan-leukocyte antigen (CD45) and αIIb integrin subunit (CD41) on the cells of aortic clusters (and para-aortic foci) suggests the embryo proper as the origin of definitive HSCs (Melchers, 2010). In all species with defined bone marrow, HSCs populate the developing bone marrow and give rise to the cell subsets residing in lymphoid tissues. In general, ontogenic pathways that determine haematopoietic cell development and differentiation into lineage specific cells are conserved in all vertebrates (De Kleer et al., 2014). Below, we describe the ontogeny and functional properties of chicken phagocytic cells (heterophils, monocytes/macrophages, and dendritic cells (DCs)), γδ T cells and NK cells, as well as including immunological data from other vertebrates, to provide a broader understanding of the innate immune system in avian species (Table 1).

Table 1. Ontogeny of innate immune system cells in chickens.

Tissues	Times lines
	Before hatch (Embryonation day, ED)					Post-hatch (Weeks)
	ED1-4	ED5-8	ED9-12	ED13-16	ED17-21	Week 1	Weeks 2–3	Weeks 4–6
Yolk sac	CD45+ cells
Blood	CD2+ cells					Heterophils, Monocytes	Heterophils
Bursa of Fabricius		CD45+ cells, DCs						Various types of DCs
Thymus					Thymic DCs
Spleen			CD45+ cells, Macrophages	NK cells, γδ T cells		γδ T cells		γδ T cells, various types of DCs
Caecal tonsil			CD45+ cells, γδ T cells
Intestine				γδ T cells	NK cells	Heterophils
Skin		CD45+ cells, Langerhans cells

Notes: First appearance of various cell types in tissues of pre- or post-hatch chickens is indicated in each column. In cases where there are reports of an increase in cell numbers during development in each tissue, cell types are repeated for that tissue. For example, γδ T cells appear in the spleen by ED13-16 and their numbers increase by week 1 and weeks 4-6 post-hatch. The data in this table are obtained from studies that used healthy broiler or specific-pathogen-free chickens.

Avian heterophils

Avian heterophils, similar to mammalian neutrophils, provide the first line of defence against bacterial pathogens by releasing microbicidal agents such as reactive oxygen species, proteolytic enzymes and microbicidal peptides from their cytoplasmic granules (Genovese et al., 2013). Avian heterophils can kill microbes through the extrusion of heterophil extracellular traps similar to neutrophil extracellular traps formed by mammalian neutrophils (Pieper et al., 2017).

Heterophils are the predominant cells in the blood and gastrointestinal tract of very young chickens compared to older birds (Bar-Shira & Friedman, 2006). Their functional competencies during early life have been investigated in several genotypes of birds. The phagocytic ability, degranulation and oxidative burst were found to be reduced in newly hatched chicks compared to mature birds (Wells et al., 1998; Genovese, 2000). This functional deficiency, which can be observed until 21 days of age, was correlated with susceptibility to Salmonella infections (Genovese, 2000). Similarly, in human neonates, neutrophils have reduced phagocytic ability and are less responsive to chemotactic molecules (Türkmen et al., 2000; Stålhammar et al., 2017). However, most of the impaired functions of neutrophils during the neonatal period in humans are believed to be restored to adult levels around the first year of life (Muniz-Junqueira et al., 2003).

Innate responses by heterophils can be triggered when PAMPs are engaged with TLRs (Kogut et al., 2005). In more detailed profiling of heterophil-dependent innate responses in chickens, a striking difference was observed when heterophils from genetically distinct parental lines of broilers were treated with TLR ligands (Kogut et al., 2006). Heterophils from day-old chicks from both lines respond to TLR ligand stimulation, but heterophils from the line which is genetically resistant to Salmonella enterica serovar Enteritidis when compared to a susceptible line have significantly higher oxidative burst and degranulation and expression of IL-1β, IL-6 and CXCLi2 (Kogut et al., 2006). Moreover, higher innate responses involving Myeloid Differentiation Primary Response (MyD88)-dependent pathways were observed when heterophils from resistant lines were treated with Salmonella enterica serovar Enteritidis (Kogut et al., 2012). These observations allowed for the generation of a progeny chicken with a particular cytokine/chemokine phenotype (Swaggerty et al., 2008). Although heterophils from newly hatched chicks show impaired phagocytosis, they can express the same spectrum of TLRs described in mature birds. Age-related differences in TLR expression and function in neutrophils have been reported in other species. In human neonates, neutrophils express lower levels of TLR4, but express similar levels of TLR2 compared to neutrophils of adults, but in neonates, signalling through MyD88 is defective upon ligation of TLR2 and TLR4 with their ligands (Melvan et al., 2010). In newborn foals, although TLR9 expression was reported to be lower in their neutrophils compared to neutrophils from 2-week-old and 2-month-old foals, age-related differences in cytokine expression induced by CpG-oligodeoxynucleotides (CpG-ODN) were not observed, thus suggesting that the TLR9 signalling pathway in foal neutrophils may not be defective and could be activated by CpG-ODN (Liu et al., 2009; Bordin et al., 2012).

TLRs not only influence cytokine or chemokine gene expression by engaging with the nuclear factor (NF)-κB pathway, but also promote cellular degranulation or oxidative burst, or both (Kogut et al., 2008). In addition, heterophils from one-day-old chicks activated via TREM-A1 (triggering receptor of the Ig superfamily expressed on myeloid cells) have greater phagocytic capability against Salmonella enterica serovar Enteritidis (Kogut et al., 2012). This explains the involvement of TLR4 for enhanced phagocytosis since TREM-1 regulates TLR signalling and activation (Radsak et al., 2004). Heterophils isolated from 2- and 10-day-old chicks that were in ovo inoculated with CpG-ODN, through the amniotic route at embryonation day (ED) 18 or 19, showed increased degranulation when stimulated by killed or live Salmonella enterica serovar Enteritidis (MacKinnon et al., 2009). This indicates that the innate responses of heterophils in newly hatched chicks could possibly be boosted to adult levels by using appropriate TLR stimuli as in other neonatal vertebrates (Cohen et al., 2014).

Mononuclear phagocytes

Phagocytes carry out specific effector functions during infections and play roles in inflammation, tissue growth and remodelling. In CSF1R-transgenic chickens, also called MacReporter birds, in vivo lineage-tracing studies revealed that mononuclear phagocytic cells originate from myeloid precursors (Balic et al., 2014). The Fms-intronic regulatory element, a critical control element in the first intron of the CSF1R locus in mammals, which is also conserved in birds, allows for the macrophage-specific expression of reporter genes (Sasmono et al., 2003). Using this method, specific CSF1R-reporter gene expression was found in the chicken macrophage lineage in the yolk sac and throughout embryonic stages and in all adult tissues indicating earliest stages of haematopoiesis in the yolk sac (Balic et al., 2014; Pridans et al., 2014). Moreover, tissue resident macrophages in adult chickens and mammals may be derived directly from yolk sac macrophages that seed embryonic tissues before hatching (Balic et al., 2014; De Kleer et al., 2014). This has been shown in one recent work whereby CD45 myeloid progenitors of yolk sac origin cells colonize all organ primordia to differentiate into multiple cell types, including tissue resident macrophages (Dóra et al., 2017).

Phenotypically, monocyte subsets in mice and humans (Italiani & Boraschi, 2014) and, to some extent, in chickens (He et al., 2009) are characterized by a combination of cell surface scavenger, chemokine, cytokine and Fcγ receptors (He et al., 2009; Italiani & Boraschi, 2014). These receptors regulate monocyte functions and facilitate their interaction with other innate cells. In humans, members of the Ig superfamily such as CD300 family receptors (lipid-binding receptors expressed mainly on myeloid cells) are involved in phagocytosis and chemotaxis and show age-related differences in their expression levels (Gasiorowski et al., 2013). Monocyte subsets from human neonates have a lower surface expression of CD300c, an activating receptor, compared to adult monocytes. This may, to some extent, explain the impaired chemotactic activity and altered phagocytosis by neonatal monocytes compared to adult monocytes (Arenson et al., 1979). The expression of CD300a, which is an inhibitory receptor, by adult and neonate monocytes remained at similar levels (Zenarruzabeitia et al., 2016). The chicken CD300 locus encodes activating (CD300L-X1) and inhibitory (CD300L-B1) receptors. The functions of these receptors are predicted to resemble those of mammals, but data confirming their functions in newly hatched chicks or mature birds are not available (Sperling et al., 2015).

Functionally, chicken embryonic macrophages identified at ED 12-16 in the spleen and liver are able to recognize and phagocytose microbial antigens (Jeurissen & Janse, 1989). Their numbers and phagocytic activity also increase over embryonic development implying that, like in mammals, avian embryonic macrophages have a major function in tissue growth and remodelling. Moreover, macrophages obtained from day-old chicks can be activated by several microbial stimuli indicating their functionality at hatch (Qureshi et al., 2000). Monocytes and macrophages in chickens express most TLRs, except for TLR5, and monocytes isolated from newly hatched chickens and expanded in vitro respond to Polyinosinic:polycytidylic acid (poly(I:C)) and CpG-ODNs, either alone or in combination, with the combination inducing a wider spectrum of pro-inflammatory and regulatory cytokines (He et al., 2006, 2012; Arsenault et al., 2013). In humans, cytokine expression levels profiled by RNA-seq from lipopolysaccharide-stimulated monocytes revealed a lack of major differences among neonates, young and older adults, although older adults exhibited relatively elevated basal inflammatory gene expression (Lissner et al., 2015). In other vertebrate species, age-dependent differences were observed in monocyte/macrophage responses to TLR-stimulants. For example, in vitro stimulation of ovine alveolar macrophages with TLR3 and TLR4 ligands has revealed marked differences in cytokine and chemokine (interleukin-1β, L-6, and IL-8) gene expression profiles between neonates and adults (Fach et al., 2010).

Dendritic cells

Dendritic cells (DCs) form the pillar of immunity and link innate and adaptive immune responses. In mice and humans, the ontogeny of DCs has been extensively explored and the heterogeneous nature of these cells is well recognized (Shortman & Liu, 2002; Merad et al., 2013; Georgina et al., 2018). However, DCs are not well characterized in avian species, but they may develop through diverse ontogenic pathways to give rise to phenotypically heterogeneous subtypes homing to specific tissues (Nagy et al., 2016).

The immunological functions of numerous DC subsets described in mice and humans are regulated by several genes and surface receptors (Georgina et al., 2018). Avian DCs can be defined by the expression of a combination of markers such as putative CD11C, DEC205, CD83 and MHC class II surface molecules (Hansell et al., 2007; Wu et al., 2010; Staines et al., 2013; Vu Manh et al., 2014). DEC205 and CD83 are also used as markers for DCs characterization in humans and mice (Swiggard et al., 1995; Lechmann et al., 2002). However, the state of cell activation, maturity or tissue of origin will determine the level of expression of most of these molecules (Hansell et al., 2007; Wu et al., 2010; Staines et al., 2013). DEC205 and CD83 are expressed mainly on DCs; activated lymphocytes and macrophages also express these surface glycoproteins to some extents (Hansell et al., 2007; Staines et al., 2013). In chickens, DEC205 detects C-type lectins (Staines et al., 2013) and in healthy chickens, DEC205 and CD83 expressing cells are abundant in areas of antigen presentation or retention in the spleen, bursa of Fabricius (BF), thymus and caecal tonsils indicating that these markers can detect mainly DCs (Staines et al., 2013). In chickens, DCs have been generated by differentiating bone marrow derived cells (Wu et al., 2010). Generally, these cells do not seem to show the heterogeneity in DC phenotype and function. Whether these bone marrow granulocyte-macrophage colony-stimulating factor-derived DCs represent the conventional DC (cDCs) lineage in chickens is not known. In other species such as mice, bone-marrow derived DCs require granulocyte-macrophage colony-stimulating factor (GM-CSF) for in vitro differentiation, whereas plasmacytoid DC populations require IL-3 for in vitro differentiation (Lutz, 2004; Diao et al., 2006; Georgina et al., 2018). In chickens, the existence of bona fide cDCs (also called myeloid DCs in mice) in the spleen has recently been confirmed based on gene expression profiles (Quéré et al., 2013; Vu Manh et al., 2014). These cells are found to express most genes described for mammalian cDCs (Robbins et al., 2008).

Phenotypic and functional differences among chicken DC subsets have yet to be evaluated. In vitro expanded DCs generated from bone marrow cells of chicken embryos (ED12-18) and mature birds were shown to express several classes of PRRs, and respond to stimulation with TLR ligands and viral infections (Wu et al., 2010; Liang et al., 2015). They express costimulatory molecules, cytokines and chemokines following TLR ligand stimulation or viral infections (Wu et al., 2010; Liang et al., 2015). Qualitative and quantitative age-specific changes in the expression of TLRs by human DCs and their responses to TLR stimulation have been reported (Sun, 2003; De Wit et al., 2004). Human neonatal plasmacytoid DCs show defects in type I interferon (IFN) production when stimulated with TLR7 or TLR9 agonists despite expression of adult levels of both receptors; in this context, innate responses may not be substantially increased with additional TLR engagement (De Wit et al., 2004). Application of exogenous cytokines in conjunction with rationally selected TLR ligands, combined for synergy, augments the functions of neonatal DCs to a comparable level to those in adults (Goriely et al., 2001; Krumbiegel et al., 2007).

Although studies comparing DCs from newly hatched chickens with those from adult birds are lacking, we speculate that DCs from newly hatched chickens and mature birds may likely respond differently to pathogens or pathogen-derived molecular motifs. Therefore, in chickens, comprehensive transcriptomic and functional analyses are required for further characterization of DCs isolated from lymphoid tissues of newly hatched chickens and mature birds. Generally, the low frequency of DCs in lymphoid organs can be a major problem for obtaining an adequate amount of DCs for characterization. In the future, single-cell level analysis methods, validated for characterizing immune cell populations in other vertebrate species, can be used to study cells present at lower frequencies in lymphoid tissues of chickens (Villani et al., 2017).

In addition to conventional DCs, efforts have been made to characterize the ontogeny and functions of different types of DCs residing in avian primary and secondary lymphoid organs. Dóra et al. (2017) showed that CD45+major histocompatibility (MHC) class II+ ramified cells arising from the yolk sac expand in the para-aortic region before colonizing the developing bursa of Fabricius at ED10. These cells in the bursa may give rise to precursors of bursal DCs (Nagy et al., 2004; Dóra et al., 2017) and their number increase in the bursa from hatch until 4 to 6 weeks of age (Nagy & Oláh, 2009). Furthermore, CD45+MHC II+ ramified cells from the yolk sac colonize the developing caecal mesenchyme and splenic mesenchyme at around ED10 (Dóra et al., 2017). Clusters of thymic DCs that express CD83 and an antigen detected by monoclonal antibody 74.3, which potentially targets chicken DCs (Jeurissen et al., 1992), start to appear at ED18, and after hatching are restricted to the thymic medulla and the cortico-medullary border (Nagy et al., 2016). Although the origin of avian follicular dendritic cells (FDCs) is unclear, a common haematopoietic origin, including ellipsoid-associated cells as precursors, has been suggested for FDCs (Jeurissen et al., 1992; Gallego et al., 1997). Avian FDCs reside in the germinal centres (GCs) of secondary lymphoid organs, similar to mammalian FDCs (Jeurissen et al., 1998). However, these cells in mammals have a stromal origin and develop from vascular mural cells (Krautler et al., 2012). Due to their specific localization in GCs, avian and mammalian FDCs may have a conserved function of inducing high-affinity antibody production, B cell activation and B cell memory formation (Jeurissen et al., 1998; Kranich & Krautler, 2016). In chickens, FDCs and interdigitating dendritic cells (IDCs) have been found to increase in numbers in caecal tonsils following infection with protozoa (Del Cacho et al., 2009). Langerhans cells, a highly migrating DC type, present in epidermal tissues of chickens, may originate directly from the yolk sac (Igyártó et al., 2006) as in mammalian tissue resident macrophages (Gomez Perdiguero et al., 2015). Dóra et al. (2017) revealed the presence of CD45+MHC II+ cells, starting from ED8 in the epidermis, that possibly represent precursors of Langerhans cells. ATPase/CD45/vimentin expression by Langerhans cells is similar for newly hatched chickens and adult birds (Igyártó et al., 2006).

γδ T cells

Chicken γδ T cells represent up to 50% of T lymphocytes (Kubota et al., 1999), compared to humans and mice where γδ T cells constitute only 2-10% of peripheral lymphocytes (Haas et al., 1993). Moreover, chickens have high numbers of γδ T cells in the intestinal/caecal epithelium (Bucy et al., 1988; Dunon et al., 1993; Pieper et al., 2008). An integrin-like protein expressed by γδ T cells may allow selective homing and retention of these cells in the intestinal epithelium (Haury et al., 1993). γδ T cells are the first T cells to appear during ontogeny in the thymus. Their primary role could be immunosurveillance or immunoprotection, since functional responses of conventional αβ T cells (which require clonal expansion through positive and negative selection), are severely impaired in human neonates (Gibbons et al., 2009). In mice, and calves, γδ T cells recognize non-peptide antigens and produce IFN-γ and cytotoxic peptides such as perforin (Price et al., 2007; Alvarez et al., 2009). It is not clear whether chicken γδ T cells behave in the same fashion, but a recent study indicated the capability of chicken γδ T cells to possess cytotoxic effect against target cells with chicken IL-2 and IL-12 enhancing this effect (Fenzl et al., 2017). Evidence supports that γδ T cells in chickens express trans-membrane (or T cell) immunoglobulin and mucin (TIM)-4 receptor (Hu et al., 2016). TIM4 expressing macrophages in mammals (Kobayashi et al., 2007) and chickens (Hu et al., 2019) are known for phagocytizing apoptotic cells. γδ T cells likely use this receptor for immunosurveillance mechanisms.

In chickens, T cell progenitors from para-aortic foci colonize the epithelium of the developing thymus. This first wave of colonization occurs at ED6. The second and the third waves occur at ED12 and ED18, respectively; migration occurs post-hatch from the bone marrow (Coltey et al., 1989). Thymic γδ T cells increase to 30% of all thymocytes between ED12-15 (Sanchez-Garcia & McCormack, 1996). They appear in the embryonic spleen and intestinal epithelium from ED15 (Cooper et al., 1991) indicating intestinal epithelia, similar to the thymic epithelium, may support T cell development. Furthermore, γδ T cells are exported to these tissues immediately post-hatch and between 6-8 days post-hatch (Kasahara et al., 1993; Sowder et al., 1988). As the bird ages, the percentage of TCR-γδ+ T cells co-expressing CD8α and chIL-7Rα in the blood increases from 1% to 10% (van Haarlem et al., 2009).

Activated avian γδ T cells up-regulate the expression of CD28, CD25, MHC class II antigens and CD5 expression (Koskela et al., 1998). While the functions of γδ T cells in chickens in vivo remain to be elucidated, their numbers, as well as expression of IFN-γ, increase in the caecum following mucosal infection by Salmonella (Pieper et al., 2011). We have recently reported that infection of day-old chickens with Marek's disease virus (MDV), an oncogenic herpesvirus of chickens, increases the number of γδ T cells in spleen and lungs (Laursen et al., 2018). During the early cytolytic phase of MDV infection, in which the virus causes significant B cell depletion in the bursa of Fabricius, γδ T cells expressed large amounts of IFN-γ and IL-13. γδ T cells harvested at day 4 post infection of day-old chicks also had higher TLR3 expression, implying a role for TLR3 in MDV recognition (Laursen et al., 2018). It is, therefore, conceivable that TLR ligand ligation could possibly enhance the cytotoxicity of γδ T cells against MDV transformed cells.

Murine and human γδ T cells express TLR2, 3, 4, 7, 8 and 9 (Wesch et al., 2011). γδ T cells in birds express all TLRs, except for type 1 TLR2 (Iqbal et al., 2005; Pieper et al., 2008). In other vertebrates, these cells respond to stimulation with PAMPs and mount their effector functions through cytokine production (Hedges et al., 2005). Although chicken γδ T cell responses to TLR stimulation are not yet documented, based on the TLR expression in these cells it is reasonable to suggest that chicken γδ T cells can recognize PAMPs and mount an effective innate response.

Natural killer (NK) cells

NK cells target virus-infected and neoplastic cells and are the primary early source of IFN-γ (Moretta et al., 2006). In chicken embryos, NK cells are present in higher densities in embryonic spleen and intestine (Göbel, 2000). Moreover, the intraepithelial lymphocyte compartment in younger chickens has an abundant NK cell population (Göbel, 2000; Göbel et al., 2001). NK cells are generated in the bone marrow and migrate to colonize embryonic spleen and intestinal epithelium. NK cells characterized as CD8+ cells lacking both B cell and T cell receptors, appear at or before ED14 in the embryonic spleen, prior to T cell migration from the thymus to the peripheral lymphoid tissues. By ED19, these cells are found in abundance in the intraepithelial lymphocyte compartment and also in the embryonic spleen (Göbel et al., 1994, 2001).

NK cell subsets in mice and humans are clustered based on the expression of CD56 and CD16 on their surface. Avian NK cells also express CD56 molecules (Neulen & Göbel, 2012). The extent of expression of a complex array of surface molecules determines the specific localization and function of NK cells (Cooper et al., 2001). In mammals, during ligand or target cell recognition, the interaction between activating and inhibitory receptors on NK cells and the integration of signals transmitted by these receptors determine the extent of NK cell responses (Lanier, 2005). In chickens, the activating chicken Ig-like receptor (CHIR)-A and inhibitory CHIR-B domains, encoded by the leukocyte receptor complex, regulate NK cell functions; the balance between the two receptors determines the level of NK cell activation (Viertlboeck et al., 2005). Compared to mammals, chickens have few C-type lectins encoded by the NK gene complex (Rogers et al., 2008). The inhibitory and activating receptors of avian NK cells, in conjunction with MHC class I molecule expression, may target virus-infected or transformed cells and play a role in immune surveillance as indicated for other vertebrate species (Saunders et al., 2015). It has been reported that in chickens, CHIR-A interacts with influenza A viruses in a sialic acid-dependent manner (Jansen et al., 2016).

Differences in the expression of surface receptors on neonatal and adult NK cells influence NK cell function (Cooper et al., 2001). NK cells from mammalian neonates display higher levels of inhibitory receptors (CD94/NKG2A) on their surface, which likely reduce their cytotoxic and lytic capabilities compared to adult NK cells (Guilmot et al., 2011). A lower level of expression of CD54 by human neonatal NK cells hinders their adherence to virus-infected or transformed cells for lytic activity (Lin & Yan, 2000). Chicken CD94/NKGD2 has the features described for human CD94 and NKGD2, but their functionality remains to be evaluated (Chiang et al., 2007).

NK cells, like classical effector T cells, make use of perforin and granzyme B to mediate lytic functions. The expression of these proteins differs by age, with human neonatal NK cells having higher expression than adult NK cells (Solana et al., 2012). Defects in exocytic trafficking of these proteins across immunological synapses between effector and target cells may reduce their cytotoxic effects and has been noted in neonates and extreme age groups. In chickens, much work remains to be conducted to functionally characterize NK cell derived lytic proteins. Limited data support an age-related, increased expression of NK-lysin-like protein and IFN-inducible transmembrane protein being lower at ED16 but higher in one-week-old chickens (Cui et al., 2004). Higher expression levels of granzyme by day 4 and 7 post infection during the early cytolytic phase of MDV infection suggest a role of NK cell derived proteins against MDV (Sarson et al., 2006). Functional assessment of avian NK cells against virus-infected and haemopoietic tumour cells has shown cytotoxic capabilities to differ among different strains of chickens, and cytotoxic potential was found to be not fully developed until they reached 6 weeks of age (Lillehoj & Chai, 1988). In other studies, however, NK cells harvested from chicken embryos at ED14 were found to be functionally similar to those isolated from the spleens of 4-week-old chickens (Jansen et al., 2010). In this case, NK cells stimulated with phorbol myristate acetate/ionomycin were found to express higher levels of CD107a (lysosomal-associated membrane protein-1). It is noteworthy here that the expression of CD107a has been shown to correlate with cytokine secretion and NK cell-mediated lysis of target cells in mammals (Alter et al., 2004).

TLR signalling in NK cells synergizes with chemokine- or cytokine-mediated signals to activate these cells. Human NK cells constitutively express nine TLRs (TLR1-9) and can be activated by viral- and bacterial-derived PAMPs (Chalifour, 2004; Sivori et al., 2004). Following stimulation with double stranded RNA(dsRNA), NK cells can induce DC maturation for effective priming of T cells but also can kill immature DCs (Moretta, 2002). Cord blood-derived NK cells are reported to have lower levels of TLR3 expression and lower responses to poly(I:C) stimulation compared to adult NK cells. TLR ligation in neonatal NK cells by combined TLR ligands may potentiate their innate immune functions; NK cells have higher expression of activation markers such as CD69 and CD25 after stimulation by or CpG-ODN (Sivori et al., 2004). To this end, avian NK cells, like conventional CD8 and CD4 T cells, can express most TLRs, and associated molecules, however, TLR functions in these cells remain to be investigated.

Collectively, a comparative analysis of avian immune system cellular development and function reveals striking conservation of the major features defined for mammalian immune systems. As outlined above, there are also characteristic differences in the ontogenic development of cells of the avian innate immune system compared to its mammalian counterparts.

Ontogeny of genes and molecules associated with innate immunity

Pattern recognition receptors

PRRs are germ-line encoded and evolutionarily conserved innate receptors, which are localized on cell surfaces, within intracellular vesicles or in the cytoplasm (Juul-Madsen et al., 2014). TLRs are the best characterized families of PRRs in vertebrates. In chickens, 10 TLRs have been discovered (Kogut et al., 2005; Wu & Kaiser, 2011). Similarities and differences in the basic biology and signalling pathways of TLRs in chickens and mammalian species have been well described previously (Brownlie et al., 2009; Keestra et al., 2013). TLR1, TLR2 (TLR2A and TLR2B), TLR4 and TLR6 recognize lipids/lipopeptides, whereas TLR3, TLR7, and TLR21 recognize nucleic acids (Brownlie & Allan, 2011). While TLR5 recognizes flagellin (derived from bacterial flagella), TLR15 recognizes fungal and bacterial proteases (Keestra et al., 2013). Chicken TLR21 is identified as an endosomal receptor for microbial DNA or unmethylated synthetic CpG-ODN, and serves as a functional orthologue of the mammalian TLR9 (Brownlie et al., 2009; Keestra et al., 2010). Chicken TLR15 has no molecular orthologue in mammals (Higgs et al., 2006).

The retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs) and Nod-like receptors (NLRs) are cytoplasmic PRRs present in mammals and some birds (Benko et al., 2010; Ciraci et al., 2010; Wilkins & Gale, 2010). RIG-I and melanoma differentiation-associated gene 5 (MDA5) recognize viral-derived and synthetic such as poly(I:C) leading to the activation of the IFN pathway (Barbera et al., 2010; Karpala et al., 2011). While MDA5 selectively recognizes long poly(I:C) molecules (>1 kbp), RIG-I preferably recognizes short poly(I:C) sequences (<1 kbp) (Kato et al., 2008). Additionally, RIG-I recognizes short, uncapped 5′-triphosphate (5′ppp) RNA (Hornung et al., 2006). Chickens, however, lack RIG-I, which is ubiquitously present in ducks (Barbera et al., 2010). Although NLRs are not well studied in chickens, existing evidence suggests the induction of NLRC5 in HD11 cells in response to Salmonella endotoxin (Ciraci et al., 2010).

The expression of some TLRs, during early and late embryonic stages, suggests their importance in organ development as well as in mediating the recognition of pathogenic and commensal microbial encounters immediately post-hatch. Most TLRs, including TLR2A, TLR3, TLR4, TLR5, TLR7, TLR15 and TLR21 are expressed as early as ED3 in the whole chicken embryo preparation (Kannaki et al., 2015). Primitive macrophages in mice that appear in the yolk sac play a role in organ development by inducing apoptosis, and have been shown to express PRRs for undertaking most cellular activities related to apoptosis (Melchers, 2010). Similarly, avian macrophages are identified in the yolk sac, and those from relatively mature embryos are functionally competent, indicating their role in embryonic development (Jeurissen & Janse, 1989). The CD45+ cells residing in the para-aortic region of chicken embryos have been shown to express TIM-1 and TIM-4 (Syrjänen et al., 2014). Functionally, by binding with phosphatidylserine, TIM4 particularly mediates the recognition and uptake of apoptotic cells and hence plays roles in tissue development (Kobayashi et al., 2007). Recently, TIM4-expressing macrophages have been detected in whole mounted chicken embryos at the earliest stages of embryonation (Hu et al., 2019). These cells also express other cell surface receptors such as TLR4. The expression of some TLRs on murine and human haematopoietic progenitor cells is known to stimulate innate immune system replenishment (Nagai et al., 2006). In chickens, higher expression of TLRs in some organs of late embryonic stages chicken embryos may indicate functional competency for induction of effector molecules following ligand recognition. For example, at ED12, TLR4 is highly expressed in liver and brain of chicken embryos and, by ED18, TLR21 is expressed in the liver and intestine (Kannaki et al., 2015). Expression of the three viral RNA-sensing PRRs, MDA5, TLR3 and TLR7, is relatively higher in the post-hatch period compared to pre-hatch, and expression of these receptors is higher in lung tissues than in spleen (Karpala et al., 2012). TLR7 expression gradually increases post-hatch, and by 2 weeks of age, its expression reaches adult levels. The early regulation of PRRs in post-hatch life in the intestine and lung might be related to the establishment and homeostasis of the gut and respiratory microbiota in the immediate post-hatch life. Therefore, based on PRR expression dynamics, the acquisition of immunological competence in late stage chicken embryos could partly be explained by a higher expression of most TLRs.

Cytokines

Cytokines regulate the survival, proliferation, differentiation and migration of HSCs, lineage-committed and mature innate immune system cells (Melchers, 2010). During chicken embryonic development, the survival and proliferation of primitive macrophages are regulated by macrophage colony stimulating factor (CSF1) by interacting with CSF1R and some transcription factors (Garceau et al., 2010). By providing key signalling molecules and milieu, cytokines and cytokine responses shape the developing adaptive immunity of hatchlings (Kaiser & Staeheli, 2014).

Cytokines with adaptive and regulatory roles

Cytokine genes are developmentally regulated with certain cytokine networks dominating particular times of the ontogenic period (Table 2). Basal expression levels of chicken IFN-γ, IL-4, IL-10 and IL-18 have been identified in the spleen at ED12 with levels of expression increasing at later embryonic stages and continuing up to day 7 post-hatch (Abdul-Careem et al., 2007a). While IFN-γ markedly increased during the first week post-hatch in the spleen, the relative levels of IFN-γ increased by week 2 post-hatch in the intestine implying tissue-specific ontogenic development and regulation of cytokines (Lammers et al., 2010). Cytokine gene expression profiling over an extended period revealed significantly higher expression of IL-1, IL-10, IL-12p40 and IFN-γ in the intestine of 14- to 42-day-old chickens compared to 3-day-old and 49- to 70-day-old chickens. Most cytokine mRNA expression levels decreased to basal levels between 49 and 70 days post-hatch (Lammers et al., 2010). Within the intestinal context, Schokker et al., (2009) using microarray analysis showed age-dependent sequential cytokine profiles in jejunal tissues of 1- to 21-days old chickens. During early post-hatch life, cytokine profiles are dominated by cytokines initiating the migration of immune cells. Genes associated with cellular differentiation and specialization, and those promoting cell maturation and immune regulation, dominate the later stages (Schokker et al., 2009). These lines of evidence indicate that the ontogenic expression of cytokine genes in chickens is both age-dependent and highly regulated.

Table 2. Ontogeny of innate response genes and receptors in chickens.

Tissue	Times lines
	Before hatch (Embryonation days, ED)					Post-hatch (Weeks)
	ED1-4	ED5-8	ED9-12	ED13-16	ED17-21	Week 1	Weeks 2–3	Weeks 4–6
Bursa of Fabricius						CATH1-3
Spleen			IL-4,10,18, IFNγ	IFNα, β IFNλ	IFNγ	CATH2-3, MDA5 TLR3, 7	IFNαR1, αR2, IL-10R2, IFNλR1, IFNλ	IFNα, β, IFNλ
Intestine/Caecum			CATH1-3, Most AvBD		CATH2, AvBD9 TLR21	IFNγ	AvBD1,2,4,6
Lung			Some AvBD		IFNα, β	MDA5, TLR3,7	IFNαR1, αR2, IL-10R2, IFNλR1	TLR7, IFNλ
Whole embryo	All TLRs

Notes: First evidence for expression of innate response or receptor genes in tissues of pre- or post-hatch chickens is indicated in each column. In cases where there are reports of an increase in expression during development in each tissue, gene names are repeated for that tissue. For example, expression of IFNα, β can be recorded as early as ED13-16 but their expression increases over time and there is evidence for enhanced expression of these molecules by weeks 4-6. The data in this table are obtained from studies that used healthy broiler or specific-pathogen-free chickens, with no clinical infections or treatment with TLR ligands.

Antiviral cytokines

The ontogeny of antiviral response genes such as type I and III IFNs and signalling networks were studied by Karpala et al., (2012) (Table 2). In the lungs, IFN-α and IFN-β expression peaked at ED18 and at hatch, with their expression declining in the post-hatch period. Expression of IFN-λ (a type III IFN) continued to increase post-hatch and higher expression occurred at week 2 post-hatch in the lungs. In the spleen, expression of type I and III IFNs increased gradually from ED14 with significantly higher expression at week 6 post-hatch (Karpala et al., 2012). Although the cell types which produce IFNs were not examined in the study, the organ-specific changes may be linked to differences in the constituent cells. In vitro, chicken embryonic cells can be stimulated to express IFNs as early as ED15, which supports the functional importance of this system starting from early life and in young birds (Sekellick et al., 1990). Cells prepared from young embryos (as early as ED5) required more time to develop the IFN system when cultured in vitro than did cells from older embryos (ED10) (Sekellick et al., 1990).

Although the expression of type I and III IFNs genes appears early during embryogenesis, the respective receptors mediating IFN responses appear late during development (Karpala et al., 2012). IFN-αR1 and IFN-αR2, as the receptor complex for type I IFN signalling, are poorly expressed in the lungs of embryonic stage chickens, while significant expression is detected in most post-hatch periods in both the lungs and spleen (Karpala et al., 2012). IL-10R2 and IFN-λR1, which mediate IFN-λ signalling (Kotenko et al., 2003) increase in the first two weeks post-hatch (Karpala et al., 2012). Even though there is some information available on the ontogenic development of IFNs in chickens, there is a paucity of information about the ontogenic development of interferon stimulated genes, which are key antiviral effector molecules initiated by IFNs.

Cell survival cytokines

IL-7 is a central regulator of lymphopoiesis and acts primarily on cells of the lymphoid lineage. In chickens, the expression of IL-7Rα is detected as early as ED7 in pooled primary and secondary lymphoid organs, even though expression levels are lower than at ED20 (van Haarlem et al., 2009). Early expression of IL-7Rα may be correlated with cell differentiation and survival. Furthermore, age-related expression of IL-7Rα in peripheral blood mononuclear cells and splenocytes of 1-, 3-, and 6-week old chicks observed (van Haarlem et al., 2009). While evaluating gut-associated lymphoid tissue development during the first 2 weeks of life in chickens, an increase in transcripts of IL-2 is observed in the small intestine of 4-day-old chicks, which may be related to the initiation of gut colonization by commensal microbiota (Bar-Shira & Friedman, 2006; Bar-Shira et al., 2003).

Chemokines

Chemokines are a large family of secreted proteins facilitating the migration of innate and adaptive immune system cells in a gradient dependent manner towards the source of chemokines. Chemokines are classified by the positions of sequentially conserved cysteine residues into CC, CXC, C, and CX3C subgroups (Zlotnik & Yoshie, 2000). The first two cysteines are adjacent in CC chemokines, whereas, in CXC chemokines, one amino acid residue separates the first two cysteines (Miller & Mayo, 2017). These two chemokines have the largest subfamilies in mammalian species. Conserved cysteine residues in CC and CXC chemokines pair up to form disulphide bridges, which determine the binding of these chemokines with their respective cell surface G-Protein Coupled Receptors (GPCRs) (reviewed in Griffith et al., 2014). While C type chemokines lack the first and third of their cysteines, CX3C chemokines have three amino acids between the first two cysteine residues (Zlotnik & Yoshie, 2000). Some of the CXC chemokine families have proinflammatory properties, in that their release can be induced during an immune response at the site of infection. Most chemokines have homeostatic roles and regulate cell recruitment, and migration during tissue development or maintenance (Raz & Mahabaleshwar, 2009). All four chemokine subgroups described for mammals are also conserved in chickens (Kaiser et al., 2005). Examples of CC chemokines that have been commonly described in chickens include monocyte chemoattractant protein, macrophage inflammatory proteins and RANTES (CCL5), and CXCLi1 and CXCLi2 chemokines are among the few CXC chemokines studied (Hughes et al., 2001).

The start of the ontogeny of chemokines in chickens is expected in the earliest stages of embryonation and may continue during the post-hatch period. CXCL14, which is responsible for connective tissue development appears in the earliest stages in chicken embryos (Torres & García‐Andrés, 2010). The stromal cell derived factor-1, which belongs to the CXC family of chemokines appears as early as the primitive streak stage of chicken embryos (Read et al., 2005). Infection of chicken embryos with pathogens or stimulation with TLR agonists enhances the expression of some of the chemokines implying the basal expression of these chemokines in embryonic tissues (Patel et al., 2008). Although chickens have fewer members of chemokine families compared to mammals, chemokines in chickens show polyfunctionality (Kaiser & Staeheli, 2014). As in mammals, chemokines in chickens may play roles in cellular development and migration as well as in gut maturation. A transient increase in the level of chemokines in the caecum of chickens in the immediate post-hatch life is probably induced by the colonizing gut microbiome for initiating a physiological inflammation that enhances gut immune system maturation (Crhanova et al., 2011). In chickens, B cells have pre-bursal, bursal and post-bursal phases, in contrast to the solely bone marrow development of B cells in mammals (Ratcliffe & Haertle, 2014). In mammalian species, such as mice, CXCL12 and its corresponding receptor CXCR4 are essential for early stages of B cell development in the bone marrow (Tokoyoda et al., 2004). Therefore, chickens may have an extended range of homeostatic B cell chemo-attractants to direct B cell trafficking (Kaiser & Staeheli, 2014). For example, a higher expression of stromal cell derived factor-1 in the bursa of Fabricius of 4- to 6-week-old birds (Read et al., 2005) can be related to its role as pre-B cell growth and stimulating factor (Nagasawa et al., 1996). Two of the chicken CXC chemokines, CXCLi1 and CXCLi2, are known to induce peripheral blood mononuclear cell chemotaxis (Poh et al., 2008; Kim et al., 2017). In chickens, CXCLi1 is a particularly strong chemoattractant to heterophils, whereas CXCLi2 is more efficient at inducing the migration of monocytes (Poh et al., 2008). Although profiling of the ontogeny of chemokines in chickens is limited in scope, one may anticipate a higher expression of chemokines in the blood and intestine of newly hatched chickens for coping with the higher number of heterophils at this stage of growth.

HDPs

HDPs are cationic peptides involved in innate immunity including immunomodulation. Their direct antimicrobial effects are related to pore formation on anionic bacterial surfaces (Ganz, 2003). As constituents of the neutrophil extracellular traps (Brinkmann, 2004) and nanonets in the intestine (Chu et al., 2012), HDPs can trap bacteria before intestinal translocation. HDPs can influence gut microbiota composition by influencing the expression of PRRs by intestinal epithelial cells (Tang et al., 2016; Cuperus et al., 2018). In addition to their effects on innate immunity, HDPs can boost adaptive immunity through the regulation of chemotaxis, maturation and activation of lymphocytes and DCs (Davidson et al., 2004). In chickens, 14 avian β-defensins (designated as AvBD1-14) and four cathelicidins (CATH1-3 and CATH-B1) and their antimicrobial effects against a range of bacterial pathogens have been described (Cuperus et al., 2013; Yacoub et al., 2015, 2016).

HDPs show constitutive and regulated expression in chicken embryonic tissues (Table 2) suggesting a role in innate immunity, particularly at early ages when other components of the immune system have not fully developed. As such, the amount of constitutive expression of AvBD in intestinal epithelial cells of chicken embryos has been reported to be higher in Salmonella enterica serovar Enteritidis resistant lines compared to susceptible lines (Derache et al., 2009). In the intestine, AvBD9 shows stable expression during embryonic development, but a higher expression was observed in advanced embryos (ED20), which correlates with the appearance of enormous AvBD9-producing entero-endocrine cells (Cuperus et al., 2016). Moreover, the relatively higher intestinal expression of all cathelicidins and several avian β-defensins at ED12 resembles that in other distant mucosal tissues (lungs) and lymphoid tissues (Meade et al., 2009). At ED18, CATH-2 is expressed in the small intestine and developing lymphoid tissues, which can be correlated with the influx of heterophils (Meade et al., 2009). At the cellular level, heterophils are the prime source of HDPs, while monocytes, macrophages and epithelial cells lining mucosal surfaces also produce substantial amounts of HDPs (Brogden et al., 2003).

In the post-hatch period, as the gut microbiota composition goes through various phases in diversity and richness, HDP expression profiles may vary. Ramasamy et al., (2012) showed constitutive, but different patterns of expression of avian β-defensins in different parts of the intestine during early post-hatch life. While AvBD4 and AvBD6 expression were higher in the duodenum, AvBD3 was higher in the caecum (Ramasamy et al., 2012). The segment-specific expression may be related to the richness of microbiota composition and differences in the composition of cells (Choi et al., 2014). AvBD1 and AvBD2 expression, originating mainly from myeloid cells, show a reduction in the intestine in the first week of life, but the expression is restored to higher levels in the following weeks due to the influx of heterophils and other innate cells to the intestine (Crhanova et al., 2011).

Expression of CATH-B1 (mainly of epithelial origin) in the lungs and caecum can increase several-folds in the first week of life, while expression of CATH1-3, which is of myeloid cell origin, declines after day 4 in the bursa of Fabricius (Achanta et al., 2012). As CATH-2 protein is exclusively detected in heterophils, with a large amount in mature heterophils, reduction in the number of these cells in the bursa of Fabricius due to extensive B cell proliferation may account for low CATH-2 expression. However, compared to the higher number of immature heterophils in the bursa of Fabricius at day 28 post-hatch, spleen and caecum contain more mature heterophils (Cuperus et al., 2016), which may explain, at least in part, the higher expression of CATH-2 in these tissues (Achanta et al., 2012).

Early life responses to pathogens

Bacterial infections

Understanding innate responses of embryonic stage chickens and hatchlings will benefit the development of disease control strategies against early life infectious diseases. Innate response studies, particularly in newly hatched chickens, have mainly used Salmonella spp as infection models, and the results indicate that early control of Salmonella infection in younger chickens is dependent on pro-inflammatory and T helper (Th1)-type cytokines (Wigley, 2013). Genetics of the host and the invasiveness of Salmonella serovars determine the kinetics and magnitude of observed innate responses (van Hemert et al., 2006; Berndt et al., 2007). These innate response variations, both in terms of kinetics and magnitude, are determined by differences in the levels of constitutive expression of innate immune genes prior to infection. In this case, rapid onset and higher levels of induction of proinflammatory cytokines and chemokines are observed for Salmonella enterica serovar Enteritidis resistant broiler chickens bearing a higher basal expression of these molecules (Swaggerty et al., 2004).

Newly hatched chickens appear to mount potent IFN-γ and pro-inflammatory responses with an accompanying influx of heterophils, macrophages and γδ T cells in the caecal mucosa within a few hours of oral infections with non-host specific Salmonella spp. (Berndt et al., 2006). Although young chickens are susceptible to Salmonella infection, CATH-2 pro-peptide packed heterophil granules in the lamina propria of the jejunum of chickens orally infected with Salmonella enterica serovar Enteritidis provide some levels of protection and the released HDPs may facilitate wound healing (van Dijk et al., 2009). Furthermore, heterophils protect against internal organ colonization by Salmonella in chickens (Kogut et al., 1994), as observed in mice (Fierer, 2001). The acute responses against enteropathogenic Salmonella are further characterized by the expression of C-C motif ligand 4, iNOS, IL-12 and CXCLi2 in the caecum as well as in caecal tonsils, which collectively results in influx of inflammatory cells (Berndt et al., 2007; Setta et al., 2012).

Viral infections

The innate responses of newly hatched chickens to viral pathogens involve the secretion of cytokines from the resident or recruited immune cells as well as from epithelial cells lining the mucosal surfaces. Experimentally, intratracheal infection of young chickens with infectious bronchitis virus triggers upregulation of pro-inflammatory cytokines and IFN-β in the respiratory tract (Kameka et al., 2014). An aerosol challenge with cell-free MDV significantly increases the expression of IL-1β, CXCLi2, iNOS, TLR3 and TLR7 in the lungs of 5-day-old chickens (Abdul-Careem et al., 2007b). Feng et al. (2013) showed differences in RIG-I-like receptors and downstream antiviral response pathways to be influenced by the genetic background of newly hatched chickens infected with MDV (Feng et al., 2013). The innate responses of chicken embryos to influenza virus infection using ex vivo and in vivo experiments resemble responses to infections with other respiratory pathogens such as Newcastle disease virus. These infection models induced the expression of potent antiviral factors including IFNs, interferon stimulated genes, proinflammatory cytokines, chemokines and collectins (Reemers et al., 2009a; Reemers et al., 2009b, Reemers et al., 2010a; Reemers et al., 2010b; Petersen et al., 2012; Reuter et al., 2014). The changes in gene expression evaluated in the lungs after infection of ED18 chicken embryos via the allantoic route, and via the oculo-nasal route of infections of 3-weeks-old chickens with lentogenic LaSota strain revealed the expression of at least one of the genes with innate antiviral effects, or genes involved in chemotaxis (Deist et al., 2017; Schilling et al., 2018). Newly hatched chickens with different genetic backgrounds also respond to infectious bursal disease virus (IBDV) infection by the production of antiviral molecules (Ruby et al., 2006). A more specific gene expression profile indicated that chicks of the resistant line had an earlier onset and higher expression of pro-inflammatory cytokines in the bursa of Fabricius, with up-regulation of IFNγ, IFNα, chemokines and iNOS (Ruby et al., 2006). Even using late-stage chicken embryos (ED18) as a host, Khatri and Sharma (2009) demonstrated the upregulation of IFNγ, IL-6 and CXCLi2 in splenocytes of IBDV-infected embryos pre-hatch (Khatri & Sharma, 2009).

Genetic control of innate immunity

The innate immune competency of a host to early life pathogens is not only determined by the age of the host, but also by the genetic basis of PRRs (Wells et al., 2003). For example, single nucleotide polymorphisms (SNPs) residing in the ligand-binding domain of PRRs alter the topology of the receptor and ligand-receptor interaction properties (Shinkai et al., 2006; Skevaki & Pararas, 2015). SNPs present in gene regulatory elements, such as in the promoter region of PRRs, have similar effects (Chang et al., 2015). These genetic variants further influence the initial signalling events and contribute to a differential recognition of pathogens or pathogen molecular mimicry (Wells et al., 2003). In such a scenario, the onset and magnitude of core and inducible innate effector mechanisms differ when hosts with such genetic backgrounds are infected with pathogens (Arbour et al., 2000). Two examples in chickens that determine susceptibility/resistance phenotypes to Salmonella infections are SNPs in the ligand-binding domain of TLR4 (Leveque et al., 2003) and at the promoter region of NLRC5 (Chang et al., 2015). The nuclear factor κB pathway has been found to be altered in the spleens of chickens with SNPs in NLRC5 (Chang et al., 2015). SNPs in innate effector genes can impact the quality and quantity of innate responses. In chickens, two opposing views are published regarding the role of SNPs in determining the antiviral properties of myxovirus resistance (Mx)1 gene based on allelic differences at amino acid position 631, located in the C-terminal GTPase effector domain (Ko et al., 2002; Benfield et al., 2010; Li et al., 2012). Such allelic differences resulted in differential expression of Mx1 in different tissues (Yin et al., 2010). However, whether there is a direct link between resistance/susceptibility to multiple pathogens and SNPs in PRRs or effector molecules is yet to be determined by comprehensive genome-wide association studies. Moreover, alterations in SNPs in PRRs, and their effects on developmental pathways, require investigation as PRRs are also involved in organ development, innate immune system replenishment and the establishment and homeostasis of the gut and respiratory microbiota (Nagai et al., 2006).

Epigenetic control of innate immunity

Although the scope of this review paper is limited to genetic control of innate immunity, it is worthwhile to briefly address the role of epigenetics in innate immunity of chickens. Epigenetic modifications are not encoded by changes in the nucleotide sequence of the DNA, but the activity of the gene is altered, which results in changes in the expression patterns of the gene (Youngson & Whitelaw, 2008). The effects of epigenetic regulation on immunity are well characterized in mice compared to what is known in chickens (Saeed et al., 2014). The epigenetic state of cells is believed to result in variation in innate responses among chickens and can determine the resistance/susceptibility patterns to pathogens (Gou et al., 2012). DNA methylation and histone modifications are important epigenetic modifications involving the promoter regions of pathogen sensors or signalling pathways leading to either activation or suppression of a series of immune responses (Gou et al., 2012; Hennessy & McKernan, 2016). While profiling innate responses in Salmonella enterica serovar Enteritidis susceptible and resistant birds, one study identified higher DNA methylation levels in the promoter region of TLR4 and TLR21 and in an exonic CpG island of TLR2-1 in the susceptible chickens (Gou et al., 2012). Moreover, the susceptible line had reduced TLR4, TLR2-1 and TLR21 transcript levels in leukocytes (Gou et al., 2012). A study by Tian et al. (2013) also indicated a reduction in DNA methylation levels in thymus cells of chickens belonging to an MDV-resistant line after infection with MDV.

Studies comparing the epigenome of innate immune system cells originating from embryos versus newly hatched chickens, and how that affects innate immune cell programming, are lacking. It is useful to explore the role of epigenetic modification in determining developmental stages of these cells, developmental expression of innate immune response genes and the functional competency of innate immune system cells.

TLR ligands as alternative prophylactic agents

Functional expression of TLRs and several other defence molecules by innate immune system cells of embryos and young chickens suggests that innate responses can be modulated at this particular stage of development. In this regard, TLR ligands are explored as prophylactics in the context of bacterial, viral and protozoal infections in chickens (Dalloul et al., 2004; Gomis et al., 2004). At ED18 or 19, chicken embryos respond to CpG-ODN applied by amniotic and chorioallantoic injections, and the hatched chicks can be protected early on from Escherichia coli and Salmonella enterica serovar Typhimurium infections (Gomis et al., 2004; Taghavi et al., 2008). Additionally, early post-hatch parenteral injection of CpG-ODN, flagellin, and loxoribine (guanosine analogue) detected by TLR21, TLR5 and TLR7, respectively, also provide similar protective effects against extracellular and intracellular bacteria (Genovese et al., 2007; Swaggerty et al., 2012). In both embryonic and parenteral mode of TLR ligand application, protected birds had higher survival rates, lower clinical scores and reduced bacterial burden in target organs (blood or caecum). However, in ovo administered CpG-ODN conferred more extended protection compared to intramuscularly administered CpG-ODN, and birds orally challenged 2 weeks post in ovo CpG-ODN administration still resisted Salmonella enterica serovar Enteritidis infection (MacKinnon et al., 2009). In an infectious laryngotracheitis virus infection model, in ovo administered lipoteichoic acid reduced virus replication in the respiratory tracts of newly hatched birds (Thapa et al., 2015). Similarly, CpG-ODN administered in ovo inhibit infectious bronchitis virus replication at the prime target site (the lungs) and other distant target tissues (Dar et al., 2009, 2014). Moreover, avian influenza virus shedding through the oropharyngeal and cloacal routes was dramatically reduced in two-weeks-old chickens treated with lipopolysaccharide and CpG-ODN by the intramuscular or mucosal application (Barjesteh et al., 2015). We and others have shown recently that delivering TLR ligands with nanoparticles (NPs) in vivo and in ovo improves the immune stimulating and protective capacity when tested against bacterial and viral pathogens (Alkie et al., 2017; Bavananthasivam et al., 2018a, Bavananthasivam et al., 2018b; Gunawardana et al., 2015; Taha-Abdelaziz et al., 2017; Taha-Abdelaziz et al., 2018). Delivering TLR ligands such as CpG-ODN encapsulated in polymeric NPs orally in broiler chickens also modulates gut microbiota (Taha-Abdelaziz et al., 2018b).

The results of these studies revealed that the nature of the ligands, method of delivery and routes of administration determine the innate responses measured. Chicken embryos receiving CpG-ODN by the chorioallantoic route at ED18 had increased expression of IFN-γ, IL-6, IL-1β and CXCLi2 in the spleen at 6 h post treatment, while IFN-γ and IL-6 expression levels remained elevated up to 72 h post treatment (Dar et al., 2009). Significantly increased and prolonged expression of IL-6 has been reported in one-week-old chicks treated in ovo via the amniotic route with bacterial DNA, which possibly contains unmethylated CpG motifs (Jenkins et al., 2009). Although the parenteral application of CpG-ODN into day-old chicks resulted in a higher magnitude of expression of cytokines and chemokines in the spleen as early as 3 h post-treatment, all these responses were short-lived (Patel et al., 2008). In ovo applied TLR ligands in late stage chicken embryos induce antiviral response genes in most embryonic organs examined by triggering antiviral response genes and proinflammatory cytokines. A recent study also showed the induction of antiviral response genes in the chorioallantoic membrane when TLR ligands were applied in ovo at ED10-14 and the induced genes reduced AIV replication in chicken embryos (Barjesteh et al., 2015).

The chemokines and cytokines induced by TLR ligands recognized by endosomal TLRs, and by those that engage cell surface-localized TLRs, in newly hatched chickens following in ovo delivery may further recruit innate cells to target tissues for modulating innate responses. Recently, CpG-ODN administration in ovo has been found to enrich innate cells and T cell subset compartments in spleen and lungs of embryos and chicks (Gunawardana et al., 2019). The overall complex patterns of innate responses including oxidative bursts and degranulation, secreted proinflammatory cytokines, IFNs, ISGs, and HDPs are crucial for defence against multiple pathogens.

Innate immunity and in ovo vaccination: perspectives

In ovo vaccination is a successful method of mass vaccination in the poultry industry for inducing early protection against pathogens (Sharma & Burmester, 1982; Negash et al., 2004). At the present time, several viral, bacterial, protozoal and vectored live vaccines have been administered to late-stage chicken embryos for induction of early immunity (Bublot et al., 2007; Abdul-Cader et al., 2018). Marek’s disease vaccination in ovo (with herpesvirus of turkey, or attenuated MDV strains SB-1 or CVI988) has been shown to accelerate immune system maturation and increase the proportions of T cell subsets in the spleen (Gimeno et al., 2015). Additionally, depending on the strain used, Marek’s disease vaccination in ovo can induce the expression of type I IFNs and TLRs in lymphoid tissues (Gimeno et al., 2018). Furthermore, in ovo vaccination of chicken embryos (ED18) with herpesvirus of turkeys resulted in higher NK cell activity than in chicks vaccinated at hatch (Sharma et al., 1984). The exact role of NK cells in early protection against MDV before the onset of antigen specific B and T cell immunity suggests that early in ovo vaccination increases the functional abilities of NK cells. In this MDV challenge model, activated NK cells, by releasing cytolytic proteins including granzyme, provided immediate protective innate responses in newly hatched chickens. Vaccination in ovo against IBDV leads to increased numbers of macrophages in the bursal tissues of vaccinated chicks. Moreover, FDCs loaded with IBDV antigen are detectable several days post in ovo vaccination in the bursal follicles (Jeurissen et al., 1998). The definitive roles of these cells for in ovo vaccination have not been well studied. However, in ovo administration of TLR ligands results in the recruitment of innate cells to the lungs, which subsequently enhance antigen uptake and processing (Thapa et al., 2015).

Vaccine antigens fused with cytokine genes (for example, fusion of granulocyte-macrophage colony-stimulating factor with spike glycoprotein of infectious bronchitis virus), delivered by adenoviruses as backbone vectors in ovo, resulted in rapid and long lasting antibody responses (Zeshan et al., 2011). In this antigen-cytokine fused vaccine formulation, expressed cytokines could shape the adaptive immunity by enhancing the proliferation and maturation of antigen presenting cells, as well as the expression of their antigen presenting and co-stimulatory molecules. Although several experimental trials have been conducted in chickens to administer antigens into embryos using virus vectors (Bublot et al., 2007), no mechanistic studies have been published evaluating the effects of the vector backbone on innate immune responses of embryos.

In ovo vaccination has been demonstrated to be a successful mode of Eimeria vaccine delivery in chickens (Weber et al., 2004). Inovocox EM1 is a commercially available live oocyst-based vaccine recommended for in ovo vaccination of 18- to -19-day-old embryonated chicken eggs (Sokale et al., 2017). As Eimeria parasites have a long and complex biological life cycle, understanding host-pathogen interactions is crucial for further developing efficacious vaccines mediating both cell- and antibody-mediated immune responses (Min et al., 2013). These responses depend on effective triggering and activation of the components of the innate immune system. Studies indicate the capability of orally-delivered Eimeria parasites to induce sets of PRRs and downstream effectors in the intestines within hours of infection (Zhang et al., 2012). In an in ovo vaccination experiment against Eimeria, recombinant profilin acts as an immunogen and an inducer of innate responses; and increased the level of transcripts for IL-1β, IL-15, and IFN-γ in the intestinal tissues (Lee et al., 2010). This may further amplify anti-Eimeria adaptive immune responses by modulating innate immune system cells.

Vaccine adjuvants, including TLR ligands, can be covalently linked to immunogens, or both components can be encapsulated or adsorbed on the surface of biodegradable polymeric NPs (Negash et al., 2013; Alkie et al., 2018). This implies that the potency of these vaccine formulations could be enhanced by several folds of magnitude due to directed vaccine antigen delivery, or the adjuvant-vaccine system could modify innate responses. More work is needed to clarify the mechanisms involved. A recent study indicated that birds vaccinated in ovo with inactivated inclusion body hepatitis virus and polyphosphazene (PCEP), (a water-soluble biodegradable polymer) and avian beta defensin 2 as vaccine adjuvants, induced significantly higher antibody production compared to a hexon protein, an adenovirus capsid protein-based vaccine formulation (Sarfraz et al., 2017). By three weeks of age, higher B cell and macrophage/monocyte counts were seen in the blood of vaccinated chickens. Moreover, the expression of IFN-α, IFN-γ and IL-12(p40) and IL-6 was substantially higher by 2 weeks of age due to polyphosphazene inclusion in ovo, suggesting that polyphosphazene can be involved in the activation and differentiation of B cells and T cells, and activation of macrophages and NK cells (Sarfraz et al., 2017).

In ovo-administered TLR ligands or vaccines can reach the intestine and lungs of developing embryos through ingestion or inhalation of the amniotic fluid before hatching. However, the induction sites of adaptive immune responses, such as the bronchus-associated lymphoid tissue (BALT) and the gut-associated lymphoid tissue, are underdeveloped in embryonic stages or early post-hatch but continue their development in the course of time after hatch (Fagerland & Arp, 1993). In chickens, BALT starts to appear in the lungs in the first week post-hatch, but not in the pre-hatch periods and immediately after hatching. However, mature BALT with distinct B- and T-cell areas and germinal centres appears in the lungs after 4 weeks post-hatch (Sutton et al., 2018). The study also suggests the presence of microfold-cells (M-cells) in the follicle associated epithelium overlaying the BALT, that may be involved in antigen sampling comparable to M-cells residing in Peyer’s patches of mice (Mabbott et al., 2013). In this context, innate immune stimulating molecules administered in ovo, which induce potent pro-inflammatory mediators that in turn affect the number and composition of immune cells in the BALT, and gut-associated lymphoid tissue (Thapa et al., 2015; Bavananthasivam et al., 2017; Abdul-cader et al., 2019; Gunawardana et al., 2019) may hasten the development of these lymphoid organs. This may have implications for the induction of both early and long-lasting adaptive immune responses.

Conclusion

Economic losses due to infectious diseases in chickens are significant and pose a threat to the poultry industry. The fast-growing poultry industry worldwide requires immediate research-based disease control options for enhancing poultry health and industry sustainability and productivity. In this context, a thorough understanding of the development and functioning of the avian immune system would be necessary. This review summarizes innate immunity from ontogeny and functional perspectives. There are certain differences in the ways that birds of different age groups and genetic backgrounds orchestrate innate responses from the point of microbial recognition, downstream signalling and regulatory molecules, to effector mechanisms. As optimal vaccine responses ultimately rely on innate immunity, understanding age-related innate responses is helpful for designing new vaccines and vaccine adjuvants. Although there is extensive literature regarding innate immunity in chickens, immune modulators that have demonstrated greater efficacy under experimental conditions have not yet been applied in the poultry industry. TLR ligands administered in ovo (at ED18-20), either in soluble form or encapsulated in biodegradable polymeric NPs, have shown efficacy in enhancing innate immunity. For instance, the ability of TLR ligands to enhance innate responses in embryos and at post-hatch suggest a role for these ligands as vaccine adjuvants, as stand-alone antimicrobial compounds to combat emerging pathogens, and as aids in reducing the use of antimicrobials for disease control. Despite the effectiveness of these ligands in their soluble form, there is a need to develop these compounds into commercially practical products with novel approaches, such as incorporation into biodegradable polymeric NPs for application in the poultry industry.