The immune system of chicken and its response to H9N2 avian influenza virus

Abstract Influenza A virus is a negative-sense single-stranded RNA virus that belongs to Orthomyxoviridae family. Based on the antigenic characteristics of hemagglutinin (HA) and neuraminidase (NA) influenza viruses are classified into multiple subtypes. H9N2 belongs to the low pathogenic Avian Influenza Viruses (AIVs) and is one of the widely spread viruses in poultry, which can pose a threat to humans by directly infecting or providing internal genes for various zoonotic avian influenza strains. It has the potential to directly or indirectly participate in becoming an AIV that causes a human pandemic. When the virus enters a host, the innate immune system is activated first by pattern recognition receptors. The cytokines produced at the site of infection recruit innate immune cells and antigen-presenting cells and those cells subsequently transmit antigenic signals to adaptive immune cells (i.e. B cells and T cells), to trigger specific humoral and cellular immune responses. As a result, humoral and cellular immunity can clear virus and infected cells via antibody-mediated neutralization and cytotoxicity, respectively. Understanding how chicken immune systems respond to H9N2 is a top priority for effectively controlling the virus’s spread and designing vaccines. In this review, we comprehensively discuss the role of the chicken immune system in defending against H9N2, and clarify the current limitations in understanding chicken immune responses to H9N2 virus, thereby providing potential directions for future research as research on the chicken respiratory mucosal immune system has been stagnant for more than 20 years especially on how the mucosal immune system in chicken responds to avian influenza.


Introduction
Influenza A virus is a negative-sense single-stranded RNA virus that belongs to Orthomyxoviridae family. Based on the antigenic characteristics of hemagglutinin (HA) and neuraminidase (NA), influenza A viruses comprise 18 HA (1-18) and 11 NA (1-11). While the H17N10 and H18N11 subtypes are found exclusively in bats, the natural hosts of all other HA and NA subtype viruses are wild waterfowl and sea birds . Avian influenza viruses (AIVs) can be roughly divided into highly pathogenic AIVs (HPAIVs) and low pathogenic AIVs (LPAIVs) based on their pathogenicity in chickens and molecular markers of the HA protein.
The HA protein of HPAIVs contains multiple cleavage sites that can be cleaved by an endogenous cellular furin-like protease, leading to severe systemic infection and high mortality in chickens. By comparison, the HA protein of LPAIVs contains mono-, di-, or occasionally tri-basic cleavage sites, which can only be cleaved by extracellular trypsin-like proteases.
H9N2 belongs to the LPAIV and was first identified in turkeys in the U.S.A. state of Wisconsin in 1966 (Homme and Easterday, 1970). In the decades since, H9N2 has spread widely in different parts of the world, including Africa, Asia, Middle East, and Europe (Alexander, 2007). Originally, the virus did not initially infect chickens, only turkeys and occasionally quails (Wu et al., 2009). Due to its segmented genome and error-prone RNA polymerase, AIVs are prone to gene mutation and gene reassortment, processes that often allow them to overcome limitations in the range of potential hosts (Webster et al., 1992). Gene reassortment poses a greater impact for avian influenza. Amid the continuous evolution of H9N2, its host range has also expanded from mostly turkeys to also include chickens (Fioretti et al., 1999), pig (Cong et al., 2007), ferret (Ku et al., 2014), dog (Sun et al., 2013) and even man (Guo et al., 1999). H9N2 actively participates in gene reassortment and provides complete internal genes for other novel influenza viruses, including H5N1 (Guan et al., 1999), N7H9 (Lam et al., 2013), H10N8  andH5N6 (Shen et al., 2016). Most of those viruses are zoonotic and make humans highly susceptible to related diseases and death. At present, the primary AIV subtypes in commercial chickens and ducks have transformed from H5N6 and H7N9 to H9N2, and nearly all H9 AIV strains prefer human-type receptors (Bi et al., 2020). Compared with viruses of wild bird lineage, the H9N2 viruses of poultry lineage have been shown to preferentially recognize α-2, 6-linked sialic acids, and replicate in higher titers in mammalian cells and mice (Guo et al., 2021). Study has shown that H9N2 can infect civet cats and Asian badgers -causing respiratory symptoms in the latter -and may be at play in cases of human-to-wildlife virus transmission (He et al., 2022). At present, the prevalence of the H9N2 virus in humans is also not promising. Although there are fewer reported cases of human infection and illness or death caused by H9N2, from 2014 to 2016 the H9N2 AIV's seroprevalence rate among healthy poultry workers reached 11.2% in many provinces of China, and with a rising trend (Li et al., 2017;Quan et al., 2019). Other countries have also reported high seroprevalence rates of H9N2 among poultry workers, including Nigerian (Okoye et al., 2013), Romanian (Coman et al., 2013), Indian (Pawar et al., 2012), Cambodian (Blair et al., 2013), Vietnamese (Uyeki et al., 2012), Iranian (Alizadeh et al., 2009), andPakistanian (Ahad et al., 2013). H9N2 is considered as the AIV with the greatest potential to cause a cross-species pandemic (RahimiRad et al., 2016).
Chickens, as the largest host population for H9N2 AIV, play a critical role in the transmission and evolution of the H9N2 subtype. As vertebrates, avian species have many similarities with mammals in their immune systems and immune responses to invading pathogens. The avian immune system also exhibits an impressive diversity, including the bursa of Fabricius, a unique B cell-generating organ, and the Harderian gland, a local immune secondary lymphoid organ (Bang and Bang, 1968;Swayne and Kapczynski, 2008), as well as a lack of lymph nodes and the pattern recognition receptor RIG-1. Despite numerous differences in the immune systems of chickens compared with mammals, chickens' immune genes are generally consistent with those of mammals based on genome sequencing. Like mammals, chickens' innate immunity also drives adaptive immunity, which is essential for resisting pathogens and clearing pathogens and providing immune memory. Against that background, this review article primarily focuses on summarizing the immune response mechanism of chickens to H9N2 AIV infection, including their innate immunity and adaptive immunity, and provides information for exploring H9N2's pathogenesis and immune protection mechanisms.

Innate immune response to H9N2 AIV infection in chicken
The innate immune system is considered to be the first line of defense, which can combat pathogens as soon as they invade the body. The innate immune system includes (1) physical and chemical barriers; (2) innate immune cells, including dendritic cells, natural killer (NK) cells, and macrophages; (3) pattern recognition receptors (PRRs) expressed by innate immune cells; and (4) complementary proteins and cytokines. Compared with adaptive immunity, innate immunity is immediately involved in resisting pathogens and the first to appear in the body.

Pattern recognition receptors (PRRs)
PRRs are the sentinels of innate immunity, primarily expressed on the surface or in the cytoplasm of monocytes, macrophages, dendritic cells (DCs), and non-immune cells (e.g. fibroblast cells, endothelial cells, and mucosal epithelial cells), that recognize the pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns of pathogens (DAMPs) (Takeuchi and Akira, 2009). There are four major families of PRRs in chickens: Toll-like receptors (TLRs) (Nang et al., 2011), retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs) (Liniger et al., 2012), nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) (Chothe et al., 2020;He et al., 2021), and, most recently discovered, cyclic-GMP-AMP synthase (cGAS) . Among differences in innate immune signal pathways between chickens and mammals, chicken TLR21 is a functional homologous surrogate for mammalian TLR9 and may exhibit a relatively broad recognition of species (Brownlie and Allan, 2011;Dalpke et al., 2006). Beyond that, the important receptor molecule, RIG-I, in the RLR pathway is absent in chickens (Zou et al., 2009). Instead, RNA ligands are recognized by other cellular sensors such as melanoma differentiation-associated protein 5 (MDA5) and TLRs. Liniger et al. were the first to demonstrate that chMDA5 is involved in chIFNβ's induction against infection with H5N1 AIV (Liniger et al., 2012), as later confirmed for H9N2 (Hayashi et al., 2014). In addition, like RIG-I in mammals, the NS1 protein of AIV has been shown to inhibit chMDA5 signaling (Liniger et al., 2012). By knocking out MAD5 and TLR3, MDA5 has been identified as the primary sensor for sensing RNA ligands in chicken DF1 cells, with TLR3 being the secondary sensor (Lee et al., 2020). Moreover, MDA5 can complement the lack of RIG-I by interacting with the stimulator of IFN genes (STING) to form the MDA5-STING-IFN-β pathway, which confers a strong antiviral state against RNA virus (i.e. H9N2), while STING activates both NF-κB and IFN regulatory factor 7 (IRF-7) transcription pathways to induce type I IFN and IFN-stimulated genes (ISGs) (Figure 1) . The location-specific upregulation of TLR3, TLR7, TLR21, and MDA5 in oviducts has been observed in oviducts of laying hens infected with H9N2 . Furthermore, the expression of TLR3 and NLRC5 gene was found to be upregulated in DCs derived from chicken bone marrow infected with H9N2 virus in RT-Quantitative Real-time PCR (qPCR) analysis . NLRC5 can be activated by various LPAIVs and HPAIVs in primary chicken lung cells and chicken macrophage cell lines (HD11). However, the siRNA-mediated knockdown of NLRC5 has been shown to inhibit the replication of influenza virus, which suggests that NLRC5 acts as a negative feedback regulator during AIV infection in chickens (Chothe et al., 2020). In addition, the Asp-Glu-Ala-Asp (DEAD)-box polypeptide 3 X-linked (DDX3X) may also participate in the RNA virus-mediated IFN-β signaling pathway under the stimulation of H9N2 (Niu et al., 2019).

Dendritic cells (DC)
DCs play a crucial role in initiating adaptive immunity and act as the bridge between innate immunity and adaptive immunity. DCs can process antigens and present antigenic to T cells in the form of the antigenic peptide-majority histocompatibility complex (MHC) molecular complex and the co-stimulatory signal. They also have the unique ability to activate naive T cells. Immature DCs have a powerful phagocytic capacity. When stimulated by inflammatory responses and pathogens, they migrate along the vessel wall toward T cells while expressing high levels of chemokine receptors such as CCR7 Wu et al., 2011). In turn, DCs stimulate the activation and proliferation of T cells by secreting proinflammatory cytokines such as IL-12, IL-18, and TGF-β (Gutcher and Becher, 2007), and by cooperating with the first and second signals of antigen presentation. During recognition phagocytosis of pathogens, and antigen presentation, immature DCs gradually transform into mature DCs with reduced phagocytic capacity and the increased expression of MHC molecules, costimulatory molecules, and the chemokine receptor CCR7. DCs play an irreplaceable role in transmitting immune signals to T cells. Compared with DCs in mice and humans, chicken DCs have seldom been investigated. It was not until 2009 that Wu et al. cultured chicken DCs in vitro for the first time and initially proved their function, which can mature into a T helper type 1 Figure 1. innate immune response against H9n2 aiV infection in chicken. H9n2 virus is recognized by melanoma differentiation-associated gene 5 (MDa5) and asp--glu--ala--asp (DeaD)-box polypeptide 3 X-linked (DDX3X) which trigger downstream pathways through the stimulator of iFn genes (i.e. StinG) and the mitochondrial antiviral-signaling protein (MaVS), respectively. although MaVS also activates StinG after receiving the upstream signal, what remains uncertain is whether MDa5 directly activates StinG. in any case, StinG activates the transcription factors iFn regulatory factor 7 (irF7) and nuclear factor kappa B (nF-κB) by orchestrating the assembly of tanK-binding kinase 1 (tBK1). even so, the interaction between tBK1 and irF7 also remains unclear. Once activated, irF7 and nF-κB translocate into the nucleus and are phosphorylated, which further stimulates the transcription of type 1 interferons and proinflammatory factors. in turn, type i iFns stimulate the production of the iFn-stimulated gene factor (iSG) by both autocrine and paracrine signaling through cognate type i iFn receptor recognition.
Few studies have investigated the role of DCs in chicken immune systems during H9N2 AIV infection. They mainly focused on proteomics and transcriptomics or how to improve H9N2 AI vaccine efficacy by targeting the stimulation of DCs. Even so, a gene ontology (GO) enrichment analysis of global gene expression revealed that 130 and 120 GO terms were up-and downregulated, respectively in chicken DC cells after H9N2 AIV infection. These terms were primarily enriched in cellular components, molecular functions, and biological processes . Other authors have also reported transcriptomic and microRNA data in chicken DCs infected with the H9N2 virus Yang et al., 2020). When DC stimulators such as CpG DNA, DC-binding peptide (DCpep), and antibody to the CD83 receptor were used as potential adjuvant for H9N2 AI vaccine, CpG DNA was found to effectively activate TLR21 in chicken bone marrow-derived DCs in vitro, increase the expression of IRF7 and tumor necrosis factor (TNF) receptor-associated factor 3 (TRAF3), and further stimulate the expression of IL-1, IL-6, and IL-10 (Lin et al., 2014). In another work, virus-like particles expressing the haemagglutinin-neuraminidase of Newcastle disease virus and the HA of H9N2 were decorated with DCpep. They activated chicken DCs in vivo and promoted sIgA secretion and splenic T cell differentiation, and better inhibited the viral shedding of H9N2 when administered intranasally .
CD 83 is a marker of early DC activation and plays an important role in B cell antibody production against influenza A viruses (Akauliya et al., 2020;Lee et al., 2012). When HA antigen was fused to single-chain fragment variable (scFv) antibodies specific to the CD83 receptor, it improved the expression level of IFN-γ, IL6, IL1β, IL4, and CxCLi2 mRNA in chicken splenocytes in vitro, as well as inhibited viral shedding after H9N2 challenge in chickens . Another study showed that scFv antibodies specific to chicken CD11c receptors, when fused to H9 HA protein to form trimers, can improve the immunogenicity of HA . Moreover, H9N2 has been able to replicate in the monocyte-derived dendritic cells (MoDCs) of chickens and produce progeny viruses. In that study, compared with H5N1, relatively low levels of IFN-α, IFN-β, and IFN-γ mRNA expression were induced after H9N2 AIV infection of MoDCs. The same trend was observed for TLRs 3, 5, 15, and 21 and death rate of MoDCs. However, H5N1 caused more severe damage. Those findings suggest that MoDCs play an important role in immune deregulation in chickens during H5N1 and H9N2 AIV infection (Kalaiyarasu et al., 2016).

Macrophage/monocyte cells
As important members of the innate immune response, macrophages are generally considered to be the first cells to encounter pathogens (Schat et al., 2014). They can perform various functions such as clearing pathogens, modulating innate immune responses, activating adaptive immunity, and maintaining tissue homeostasis (Taylor et al., 2005). Previously, it was thought that macrophages differentiated from monocytes in blood vessels after infiltrating tissue through the vascular endothelium. However, a more recent study has shown that macrophages in tissues are not terminally differentiated; on the contrary, the differentiation phenotype of macrophages depends on the type of tissue and is influenced by the tissue environment that regulates their gene expression (Taylor et al., 2005). In mammals, macrophages are susceptible to infection with certain influenza viruses and express viral proteins, but cannot produce progeny viruses. This phenomenon has also been observed in chickens (Mock et al., 1987;Van Campen et al., 1989). Nevertheless, some studies suggest that certain influenza viruses can replicate productively in the macrophages of different mammals (Cline et al., 2017). These studies indicate that the ability of influenza viruses to replicate productively in macrophages is strain-specific.
Macrophages are generally divided into two major phenotypes-M1 and M2-polarized by IL-4 and/or IL-13 and IFN-γ, respectively. The macrophages activated by IL-4 and/or IL-13 are involved in regulating the Th2 immune type, which is responsible for controlling parasitic infection, promoting tissue remodeling, and modulating immunity (Murray et al., 2014). The macrophages activated by IFN-γ are involved in regulating the Th1 immune type, which is characterized by the high production of reactive nitric oxide (NO) and oxygen intermediates, high expression levels of proinflammatory factors, and strong microbicidal and tumoricidal activity (Sica and Mantovani, 2012). In 1989, Jeurissen et al. reported that macrophages appear in chicken embryos on embryonic incubation day (EID) 12 in the liver and EID 16 in the spleen (Jeurissen and Janse, 1989). The secretion of NO and reactive oxygen is considered to be an important mechanism for the clearing of pathogens by activated macrophages in mammals (Babior, 1992;Babior, 2004), which has also been demonstrated in chickens (Palmquist et al., 2006;Qureshi, 2003). Interestingly, IL-4 can stimulate NO production in chicken HD-11 cells, which has not been observed in mammals (Paoliello-Paschoalato et al., 2005;Xu et al., 1999), and this effect is stronger than IFN-induced NO production. However, when macrophages were stimulated by microbial agonists form bacteria and viruses, IL-4 showed a strong inhibition of NO secretion (He et al., 2011). The mechanism of IL-4's bidirectional regulation of NO production by macrophages could elucidate the function of chicken innate immunity in response to disease invasion, which occurs at stage of innate immunity regulation by macrophages. Research has shown that IL-4 expression is downregulated after HTC macrophage cell line infection with H9N2, which plays a key role in regulating Th2-type immune responses (Xing et al., 2008). However, other studies have reported contrasting results, with IFN-γ and IL-4 significantly upregulated in DH11 cells infected with H9N2 at 12 h and 24 h, respectively (Chu et al., 2020). In mammals, alveolar macrophages exhibited polarization toward the M1 phenotype 4 h after H9N2 AIV infection, with the upregulation of M1-associated marker genes STAT1, TNF-α, MCP1, INOS, IL-6, and IL-12. After 8 h, however, the macrophages tended to polarize toward the M2 phenotype as indicated by the upregulation of M2-associated marker genes IL-10. After 24 h, they showed an immunosuppressive phenotype characterized by the downregulation of both M1-and M2-associated genes STAT1, MCP1, IL-12, and MGL1 (Zhao et al., 2014). The expression levels of cytokines in macrophages after H9N2 AIV infection are unclear and divergent and the mechanism of how macrophages regulate T cells through cytokines in chickens remains unknown.
Influenza viruses can activate anti-apoptotic P13k-Akt signals at the early and middle (5-9 h) post infection (p.i) stages of infection and thereby preventing cells from apoptosis and ensure viral replication to produce progeny viruses in mammalian cells. However, at late stages of infection, both p53-dependent and alternative p53-independent apoptotic and/or necrotic signals were activated to release the progeny virus from the cells (Zhirnov and Klenk, 2007). Compared with H6N2, H9N2 induced severe apoptosis at 18 h after infecting macrophages (Xing et al., 2008). The authors also demonstrated that the NS1 protein of H9N2 could inhibit Fas/Fasl-mediated apoptosis at 10 h and 24 h after infecting HTC macrophage cell line, thereby improving the infectivity of chicken macrophages (Xing et al., 2009).

Natural killer (NK) cells
NK cells are considered to play an important immunomodulatory role as effector cells in innate immunity. NK cells can kill target cells by MHC-independent and antibody-dependent cytotoxicity (ADCC), and produce various cytokines, including interferon γ, TNF-α, and granulocyte-macrophage colony-stimulating factor (Straub et al., 2013). Perforin and granzymes are effector granules that are mobilized and secreted by activated NK cells to exert cytotoxic effects and induce the apoptosis of target cells (Kägi et al., 1994;Voskoboinik et al., 2006). In mammals and chickens, accumulating evidence suggests that NK cells play an important role in pulmonary immune responses against respiratory disease Rogers et al., 2008;Vervelde et al., 2013). In chicken embryos, NK cells first appear at 13-16 d in the spleen, where they express CD8 molecules and lack TCR (Alkie et al., 2019;Göbel et al., 1994). In adult chicken, classical-like NK cells mainly exist in the intestine, whereas other organs or tissues mainly have NK T cells. Unlike mammalian NK cells which account for about 20% of blood lymphocytes, NK cells in chickens appear in very low proportions in the blood, spleen, and cecal tonsils, namely from only 0.5 to 1.0% (Rogers et al., 2008). CD3 − CD56 + CD16 + and CD3 − NKp46 + are often used to define NK cells in man and mice, respectively (Cong and Wei, 2019). But there are currently no commercial mature antibodies available to identify NK cells in chicken. As an alternative, some researchers have developed and used antibody 28-4 with CD3 to identify NK cells (Abdolmaleki et al., 2018;Meijerink et al., 2021).
NK cells are involved in the early immune response against viral infection. In humans, the NKp46 and NKp44 receptors of NK cells can directly recognize the HA protein of influenza viruses, meaning that the NK cells themselves can directly lyse cells infected with such viruses - Ho et al., 2008;Mandelboim et al., 2001). In mouse models, NK cells play a dual role in responding to influenza virus infection, with deleterious effects on the host at high doses and the opposite at low doses (Stein-Streilein and Guffee, 1986;Zhou et al., 2013). In chicken, research has revealed that the ratio of CD3 − CD8α + cells in both the lung and peripheral blood mononuclear cell (PBMC) was upregulated 3 d after H9N2 AIV infection, as were lung CD107 + cells, which suggests that lung NK cells were indeed activated. NK cells thus play an active immune role in defending against H9N2 AIV infection .

Adaptive immune response to H9N2 in chicken
Although innate immune response can resist the invasion of influenza viruses, adaptive immunity is ultimately required for virus clearance, a complex process that involves multiple immune cell systems, including DCs, B cells, T cells, macrophages, and NK cells (Figure 1), as well as cytokines.

CD4 and CD8 T cells
T cells, which mature in the thymus, mediate adaptive cellular immune responses and play an important auxiliary role in thymus-dependent antigeninduced humoral immunity. Thus, T cells play a central role in adaptive immunity. T cells require three signals to work together to make a productive response and to avoid death and/or immune tolerance (Curtsinger and Mescher, 2010). The first signal is the recognition of the antigen by naïve T cells through the specific binding between the surface TCR and the MHC molecule of APCs. The second signal is the interaction of multiple pairs of costimulatory molecules, mainly CD28, on the surface of T cells and APCs. The third signal is inflammatory cytokines secreted by APCs and critically determining the direction of T cell differentiation. The generation of Th1 cells is mediated by the synergy of IL-12 and IL-18 (Figure 2), whereas TGF-β secretion can polarize naive cells toward a regulatory phenotype with Regulatory T cells or an auto aggressive phenotype with Th17. In the latter case, the production and maintenance of Th17 cells also requires the cytokines IL-6 and IL-23. T cells secrete IL-4 in an autocrine manner and act on themselves to polarize Th2 cells, a process induced by the interaction between APCs and T cells (Gutcher and Becher, 2007). Th1 cells mainly respond to the infection of intracellular pathogens, including intracellular parasitic bacteria and viruses, by secreting IFN-γ, IL-2, and TNF-α. Meanwhile, Th2 cells mainly respond to the infection of extracellular bacteria, fungi, and parasites, by secreting IL-4, IL-6, and IL-10 (Szabo et al., 2003).
In chicken, T cells first appear in the thymus of embryonic chicks at 10-12 d of development (Seto, 1981). Study has shown that although the T cells of 1-d-old chicks are phenotypically mature, they have no immune function and cannot achieve the immunity of adult chicken until they are one week old (Lowenthal et al., 1994). In chicken, IL-12 and IL-18 produced by APCs also are important cytokines in the differentiation of CD4 T cells to the Th1 type (Degen et al., 2004;Guo et al., 2013;Schneider et al., 2000) and regulate IFN-γ production (Szabo et al., 2003). Although humoral immunity mediated by antibodies can prevent AIVs infection, they are strain-specific. T cells, especially CD8 T cells, can provide a broad range of cellular immunity to different subtypes of AIVs (McMichael et al., 1983;Sridhar et al., 2013). In previous research, T lymphocytes from chickens infected with H9N2 virus protected naive chickens from a lethal H5N1 virus challenge (Seo and Webster, 2001). In addition, lymphocytes from chickens infected with H9N2 have been shown to induce cross-reactivivity of cell-mediated immunity against the H7N2 isolate (Kapczynski et al., 2011). Such evidence suggests that H9N2-induced cellular immunity can provide cross-protection against strains of other subtypes of AIVs.
The change trends of CD4 and CD8 T cells in the PBMCs of H9N2-infected chickens have varied in different studies. Qiang et al. found that the proportion of CD4 T cells in PBMCs after H9N2 AIV infection was lower than that in the uninfected group, and there was a significant difference at 14 and 21 d post-infection (dpi). Meanwhile, the proportion of CD8 T cells in PBMCs was significantly lower only at 14 dpi, and the difference was not significant at any of the other time point (Qiang and Youxiang, 2011). However, other studies have reported no significant differences in the proportion of CD4 T cells in PBMCs after H9N2 AIV infection but observed a significant increase in the proportion of CD8 T cells at 3 and 5 dpi (Huang et al., 2012;Kwon et al., 2008). Moreover, Dai et al. found that although the trend of CD8 was also upregulated at 5 dpi, the proportion of CD4 cells was significantly upregulated after 5 dpi with H9N2 (Dai et al., 2021). Another study, however, reported no change in the proportion of CD4 and CD8 cells (Teng et al., 2006) similarly to our unpublished study. Although those differences may derive from differences in chicken breeds and MHC genotyping, we believe that H9N2 AIV infection in chickens may not Figure 2. the Possible adaptive immune response against H9n2 aiV infection in chicken. the mechanisms of cytokines and signaling molecules in adaptive immunity remain largely elusive in chickens, so relevant studies of mice and humans are cited to furnish explanations. after the H9n2 virus infects chickens, it is recognized by antigen-presenting cells (aPCs), among which DCs are the most important. DCs present extracellular and intracellular antigens to naive CD4 and CD8 t cells after capturing and processing antigens, respectively. CD4 t cells subsequently differentiate into th1, th2, and th17 cells. th2 cells are induced by autocrine il-4 from naive CD4 t cells. th1 cells are induced by il-12 and il-18 secreted by DCs. th1 cells promote the activation of cytotoxic t lymphocyte (Ctl) cells, which induce cytotoxicity by secreting il-2 and iFn-γ, and, then granzyme and perforin are important effector molecules of Ctl cells. Meanwhile, th2 cells secrete il-4 and il-5, which promote the production of virus-specific antibodies by B cells. cause significant changes in the proportion of CD4 and CD8 T cell in PBMCs, because the H9N2 virus is a LPAIV and cannot cause obvious clinical symptoms.

Regulatory T (Treg) cells
Treg cells are a subset of CD4 T cells that regulate the immune system negatively by providing feedback mechanisms. They protect hosts from excessive immune responses during infection and maintain self-tolerance and mucosal tolerance (Workman et al., 2009). Treg cells were first described as CD4 + CD25 + in mice (Sakaguchi et al., 1995), and later FOXP3 was identified as a critical regulator of Treg cells (Schubert et al., 2001), which is necessary for their feedback function. Mammalian Treg cells do not produce IL-2 (Shevach, 2009) and inhibit the proliferation of T cells by competing with them to bind to IL-2 via CD25, a high-affinity IL-2 receptor. This suppressive effect of Treg cells can be reversed by adding IL-2 in vitro (Scheffold et al., 2005;Selvaraj, 2013). These mechanisms have also been demonstrated in chicken (Shanmugasundaram and Selvaraj, 2011). In vivo, the immunosuppressive response of Treg cells can be reduced by adding anti-CD25 antibodies (Shanmugasundaram and Selvaraj, 2012;Shanmugasundaram and Selvaraj, 2013). Treg cells also express CTLA-4 which can competitively bind to and lower the levels of CD80/ CD86 co-stimulatory molecules on APCs, thus inhibiting T cell activation. Moreover, Treg cells also produce various immune-suppressive cytokines, such as IL-10, TGF-β, and IL-35 (Tanaka and Sakaguchi, 2017). However, excessive activation of Treg cells can severely inhibit the immune function of T cells, B cells, NK cells, DCs, and macrophages by cytokines and cell-to-cell contact mechanisms (Sakaguchi et al., 2008), leading to chronic disease and poor defense against pathogens . Treg cells not only affect immune response to self-antigens, but also participate in the immune response to viruses in almost every case studied (Belkaid and Tarbell, 2009;Rouse et al., 2006;Rudensky and Campbell, 2006).
It was previously thought that chicken did not have the FOXP3 gene and used CD4 + CD25 + as Treg cells (Selvaraj, 2013;Shack et al., 2008). However, in 2013 researchers found FOXP3 in Pseudopodoces humilis using sequencing technology and bioinformatics (Qu et al., 2013). More recently, Burkhardt et al. successfully cloned chicken FOXP3 despite the highly GC-rich content of the chicken FOXP3 gene that makes it difficult to amplify. They found that FOXP3 is expressed on CD4 + CD25 + in peripheral lymphoid organs such as the spleen and cecal tonsils, but not on CD4 + CD25 + in the thymus, which correlates with the origin of chicken FOXP3 + T cells. They also identified a CD4 − CD25 + FOXP3 high subset of thymic lymphocytes, which suggests a distinct FOXP3 population in chicken compared with mammals (Burkhardt et al., 2022). Another study showed that CD25 + , CD4 + CD25 + , and CD8 + CD25 + cells in the spleen of chicken were significantly upregulated at 2 dpi with H9N2 but then decreased and stabilized (Teng et al., 2006). Other studies have shown that only CD8 + CD25 + cells were significantly upregulated in the PBMCs of H9N2-infected chickens at 3 dpi and then decreased (Huang et al., 2012). Similar results have also been found in mice. Influenza A virus induced a strong FOXP3 + CD4 + T cell response that was highly suppressive regarding CD4 + and CD8 + T cell effects in vitro. These Treg cells proliferated rapidly in response to influenza virus antigens (Betts et al., 2012). However, the effects of Treg-mediated suppression on infection in vivo remain uncertain. On one hand, Treg activity helps to control excessive immune responses induced by pathogens that could damage tissue (Workman et al., 2009). On the other hand, several studies have shown that Treg cells delay the clearance of virus from hosts and prolong the infection in mammals (Cabrera et al., 2004;Kinter et al., 2004;Stoop et al., 2005;Suvas et al., 2003). Moreover, this persistent infection-mediated immune response affects subsequent infections of repeated pathogens (Belkaid et al., 2002).

Antibody response and B cells
The humoral immune response is an important mechanism that enables hosts to maintain the homeostasis of the extracellular humoral microenvironment. In this process, antibodies produced by B cells are the mainly means of humoral immunity against infectious agents (Hua and Hou, 2013). There are two types of B cell-mediated immune responses: thymus-dependent antigens (TD-Ag) and thymus-independent antigens TI-Ag. Similar to T cells, B cell activation by TD-Ag needs three signals: an antigen stimulation signal, co-stimulatory molecules on the surface of Th cells and B cells, and cytokines, including IL-21 (Depoil et al., 2009;Kuchen et al., 2007). As APCs, B cells transmit antigens and costimulatory signals to T cells. Then T cells send a second signal to B cells by co-stimulatory molecules after receiving the antigen signals (Parker, 1993).
Researchers have demonstrated that the activation of chicken B cells is similar to that of mammals, in that both require an MHC class II-restricted interaction between the helper T cell and the responding B cell (Vainio et al., 1984). Chicken B cells can also form GCs, which have been shown to induce the proliferation of B cells (Yasuda et al., 1998) and immunoglobulin class switching (Arakawa et al., 1996;Yasuda et al., 2003), but evidence of whether they induce antibody affinity maturation is lacking (Ratcliffe and Härtle, 2022). Antibodies are an important line of defense against pathogen invasion, as well as an important indicator for evaluating the immune effect of vaccines (Nutt et al., 2015). Studies have revealed that natural IgM antibody, generated by innate-like B cells called 'B-1 cells' , was present in mice not infected with influenza virus and played a role in early infection (Baumgarth et al., 2000;Choi et al., 2012;Savage et al., 2017). Subsequently, a part of B cells co-activated by influenza Ag and Th cells differentiate into plasma cells to secrete multiple antibodies (Lam and Baumgarth, 2019). Meanwhile, another part undergoes a germinal GC reaction that generates a broad repertoire of antibodies and B memory cells in preparation for the next exposure to the virus (Lam and Baumgarth, 2019). This interaction between T and B cells is critical for the development of a robust immune response. The antibodies produced by B cells can recognize and neutralize specific antigens, thereby preventing or limiting the spread of infection. Overall, the humoral immune response is a complex process involving multiple signaling pathways and cell types that work together to protect the body against infectious agents.
The bursa of Fabricius is unique in birds and generates primary B lymphocytes. It is populated by B cell precursors from about EID 8 to EID 14 (Houssaint et al., 1976;Le Douarin et al., 1975). Although it was initially suggested that H9N2 could inhibit humoral immunity in chickens leading to poor production of neutralizing antibodies (Xing et al., 2008), this may not be reliable because those chicken came from commercial egg farms infected with an unknown virus. In contrast, studies have clearly shown that H9N2 can effectively induce HI antibodies after infecting chickens (Dai et al., 2021). Other studies have predicted the B cell epitopes of H9N2 based on the development of peptide-based vaccines and revealed that the discontinuous epitopes of HA and HA are in amino acid residue positions 136-521 and 107-469, respectively (Ramamurthy et al., 2019). They also found potentially and highly immunogenic linear and conformational B-cell epitopes, which may support the development of a vaccine against AIVs for both humans and poultry.

Mucosal immune response
The respiratory mucosa is the first target of influenza virus attack, constituting the first line of defense against infection and effectively preventing the virus from entering (Ross and Herzberg, 2016). The protective immunity against respiratory virus infection involves a complex interaction between systemic and mucosal responses (Broadbent et al., 2015). The mucosal immune system in the upper respiratory tract plays a vital role in controlling influenza virus infection. It produces antiviral and immunomodulatory factors, including natural immunoglobulins, antimicrobial peptides, lectins, complement molecules and mucins, which respond first to influenza virus infection (Malik and Zhou, 2020). Then adaptive immune responses are activated to exert antiviral effects, among which antibodies play an important role, especially IgG and IgA. IgG antibodies mainly reduce viral pneumonia rather than prevent upper respiratory tract infections via the mucosa (Ito et al., 2003). Compared with IgG, IgA plays a more important role in preventing influenza virus infection in the upper respiratory tract (Suzuki et al., 2015) and clearing from infection with epithelial viruses (Brandtzaeg, 2007). In addition, influenza virus-induced respiratory mucosal IgA shows stronger cross-reactivity to different subtypes of virus strains than serum IgG (Asahi-Ozaki et al., 2004;Sano et al., 2018;Tamura et al., 1991). The blood-gas barrier of bird lungs is much thinner than that of mammals of equal weight, while the breathing surface is larger. These characteristics are considered to be factors that make pathogens more likely to cause pathological damage to bird lungs (Reese et al., 2006). The peripheral lymphatic system involved in the mucosal immune response of birds respiratory tract mainly includes: Harderian gland (HG), conjunctiva-associated lymphoid tissue (CALT), head-associated lymphoid tissue (HALT), and bronchus-associated lymphoid tissue (BALT) (Rumińska et al., 2008). BALT is a normal lymphoid structure in poultry that compensates for the lack of lymph node system . Fully developed BALT has been found in birds at 6 weeks of age and older, with almost every bronchus opening surrounded by lymphoid tissue . The Ig secreted by HG can enter the upper respiratory tract, providing localized protection (Dohms et al., 1981). Surgical removal of the HGs indicates that the main source of IgA in tears is HGs (Baba et al., 1990). The level of virus-specific IgA is correlated with the level of protection against ocular challenge (Gelb et al., 1998;Toro and Fernandez, 1994). Harderian glands play a role in the development of local immunity against various pathogens (Burns, 1979;Davelaar and Kouwenhoven, 1980;Survashe and Altken, 1977). Unfortunately, in-depth research on the chicken respiratory mucosal immune system was mainly concentrated in the last century, and has been stagnant for more than 20 years, and there is no recent research on how the mucosal immune system responds to avian influenza.

Conclusions
Knowledge about the avian immune system has contributed greatly to our understanding of modern basic immunology, especially that of chicken. Studies on the immune system of chicken have pioneered an explanation for the two major systems of adaptive immunity: humoral immunity and cellular immunity (Davison, 2022). In an accidental experiment, Glick B discovered that the bursa of Fabricius is an essential organ for mediating antibody immune response (Glick, 1987). In the years that followed, mainstream immunologists began to take an interest in chicken lymphoid organs and concluded that the thymus controls cell-mediated immunity . In 1965, Cooper et al. proposed an initial description of the bursa of Fabricius and thymic lymphatic system of chicken (Cooper et al., 1965). Since then, the central concept of immunology, involving T-cell-mediated cellular immunity and B-cell-mediated humoral immunity, has been established. To commemorate the discovery of this special immune organ structure of birds and its contribution to immunology, the cells that mediate humoral immunity are named 'bursa-derived lymphocytes'that is B lymphocytes (Davison, 2022).
However, compared with studies on mammals, research on the immune system of chicken has been limited and lagging behind in recent decades. This is largely due to the following challenges: non-standardized methods of assessing avian cellular immunity; a lack of antibodies for chicken, including the unstable reproducibility of intracellular IFN-γ cytokine staining; and the lack of antibodies against markers for T cell activation and migration (Hao et al., 2021). Nevertheless, chicken have a population numbering in the tens of billions worldwide, they provide a vast pool of hosts for AIVs, which provides countless opportunities for avian influenza to multiply and mutate and, ultimately honing its ability to cross the host barrier from chickens to mammals and even humans. Among these viruses, the H9N2 virus is a major source of gene segments for a variety of influenza viruses that infect humans. Given these trends, it is crucial to control the prevalence of H9N2 in chicken, which requires a profound understanding of the immune system of chicken and its role in defending against H9N2.
In this review, we have discussed how innate and adaptive immune responses cooperate to control viral infection following H9N2 AIV infection in chicken. As is known, a host needs multiple immune responses to work together against H9N2. Although a single immune response can be pivotal, it is not enough to completely eliminate the virus. In past studies, more attention was paid to the importance of antibodies in protecting against H9N2 subtype AIVs. As a consequence, current understanding of the role of cellular immunity and innate immune responses in preventing H9N2 in chickens remains limited. Therefore, future research needs to explore the immune response of chickens to H9N2 from multiple perspectives and provide a theoretical basis for controlling its spread in chickens, or an improved strategy for constructing vaccines against it.

Disclosure statement
The authors report there are no competing interests to declare.