Bats and Human Health: Ebola, SARS, Rabies and Beyond

I

INTRODUCTION

1

BAT IMMUNOLOGY

1.1 INTRODUCTION TO THE IMMUNE SYSTEM OF BATS

A number of studies have explored the bat immune system in order to determine its components and their activity levels. Bats possess immunocompetent organs and cells similar to those in humans and mice, including the thymus, bone marrow, spleen, lymph nodes, neutrophils, T and B lymphocytes, monocyte/macrophages, eosinophils, basophils, and follicular dendritic cells. These leukocytes (white blood cells) are found in ratios similar to those in mice. They mount a delayed and somewhat smaller humoral and cell‐mediated immune response than mice (Paul & Chakravarty 1986; Sarkar & Chakravarty 1991; Schinnerl et al. 2011). Regulatory T lymphocytes which dampen the immune response appear to be responsible for the delay (Chakravarty & Paul 1987). Another notable difference between bats and terrestrial mammals is the loss of AIM2 and IFI16 genes which sense microbial DNA, perhaps reducing bat sensitivity to bacteria (Stockmaier et al. 2015).

1.1.1 White blood cell count and other serological parameters

White blood cell (WBC) numbers in Saccopteryx bilineata (greater sac‐winged bat) decrease with age within individuals. IgG antibody levels, however, are higher in older bats. Individuals of this bat species that have higher WBC counts or IgG concentrations had a lower chance to survive the next 6 months (Schneeberger et al. 2014). Energetically costly immunological responses are traded against other costly life activities, leading to a reduction in overall lifespan. Immune‐mediated generation of pro‐oxidants may contribute to this reduction. In the neotropical fruit bat, Carollia perspicillata, WBC numbers correlate with indicators of oxidative stress (Schneeberger et al. 2013a). Interestingly, infection with trypanosomes or nematodes does not correlate with higher WBC counts, IgG concentrations, or survival.

Among 26 species of neotropical bats, the total WBC counts were lower for insectivorous emballonurid, molossid, and vespertilionid bat species than for plant‐eating phyllostomid bats, with Ectophylla alba (Phyllostomidae), being less than half of that of all other bat species examined (range = 1714 ± 297/μl for Molossus bondae to 7339 ± 1503/μl for Trachops cirrhosis) (Schinnerl et al. 2011). The insectivorous diet, with its higher energy demands, may be at least partially responsible for the decreased WBC numbers. Many of the lymphocytes have an indented nucleus and cytoplasmatic granules, unlike humans, whose lymphocytes have round nuclei and are agranular. Bats, in general, also have higher than normal red blood cell count, hematocrit values, and hemoglobin concentrations than most mammals, perhaps due to the great energy expenditure and aerobic respiration activity and, therefore, oxygen levels, required for flight. Accordingly, total WBC count inversely correlates with hematocrit values. The highest hematocrit levels were found in M. bondae and Molossus sinaloae (Schinnerl et al. 2011). Additionally, polychromatophilic erythrocytes (young red blood cells) levels were high in these animals.

Among wild‐caught, healthy Indian flying foxes (Pteropus giganteus), the mean lymphocyte differential count is higher for juveniles than adults. Plasma biochemistry, however, is similar between males and females, juveniles and adults, and lactating and nonlactating females. Blood urea nitrogen and cholesterol concentrations are lower in P. giganteus than in other tested mammalian groups, but correspond with that seen in other Pteropus species. Alanine aminotransferase and AST levels, however, are higher than those reported for closely related Pteropus vampyrus (McLaughlin et al. 2007).

When Pallas’s mastiff bats (Molossus molossus) are administered lipopolysaccharide (LPS), an immune system agonist, in order to study their acute phase reactions, they lose body mass. Unlike other LPS‐stimulated mammals, however, they do not develop either leucocytosis or fever. During flight on a daily basis, bats’ internal body temperature rises to 40°C, mimicking fever. LPS also does not affect the subsequent energy‐conserving reduction in temperature, down to approximately 28°C, which occurs during torpor (O’Shea et al. 2014; Stockmaier et al. 2015).

1.1.2 Innate versus adaptive immunity

Active adaptive immune system activity consumes a great deal of energy that could be used for other essential activities, such as mating and reproduction, as well as longevity. Innate immunity tends to require lower energy expenditure than cell‐mediated or adaptive immunity, suggesting that bat species may differ from other mammals in the type and amount of innate versus adaptive immune responses, with an increased reliance upon the former (Schneeberger et al. 2013b). Innate immunity also is more rapid than adaptive immunity, perhaps allowing bats to clear viral infections earlier than occurs in humans (Baker & Zhou 2015).

The swelling induced by the phytohemagglutinin skin test is used to measure delayed‐type cellular activity of the adaptive immune response. In the Brazilian free‐tailed bat (Tadarida brasiliensis), this test revealed an early peak of lymphocyte influx, followed by a later peak in infiltrating neutrophils, as well as a high degree of intraspecies variation. Host roosting ecology, diet, life history, pathogen exposure, and age may contribute to this variation (Turmelle et al. 2010). Adaptive immune responses of bat species also vary with body mass.

Bactericidal activity of whole blood utilizes phagocytosis by neutrophils and complement‐mediated cytotoxicity of the innate immune response, both of which are important in defense against, and rapid responses to, infection. The subsequent onset of adaptive T cell‐mediated immunity is more important in clearance of bacterial infections than in preventing infection (Allen et al. 2009). In T. brasiliensis, bactericidal activity negatively correlated with shelter permanence. While significant immune activity varies among individuals, colony‐level effects also play a role in the extent of bactericidal activity. Females roosting at one cave had lower blood bactericidal activity than blood from females at three other sites, whether caves or bridges. It would be interesting to study whether the bactericidal levels are constant within a given roost or vary with time as the colony faces different bacterial or viral threats.

T cell‐mediated immunity is also associated with roost location, as females from two caves had higher responses than females roosting in two bridges. Animals roosting in caves also bear a higher ectoparasite presence, since females in the cave with the lowest blood bactericidal activity also carry a greater burden of mites. Both T cell‐mediated immunity and bactericidal activity show negative correlation on the individual level (Allen et al. 2009). T. brasiliensis maternity roosts form very large colonies, ranging from several thousand to several million individuals in caves and under highway bridges. This type of roosting ecology allows increased exposure to pathogens, with the resulting effects shaping immune defenses. Such a relationship between colonial living and immune responsiveness has also been reported in several avian species (Allen et al. 2009).

1.1.3 MicroRNA

Deep sequencing of the small RNA transcriptome of the black flying fox (Pteropus alecto) detected 399 microRNAs, of which more than 100 are unique among vertebrates. MicroRNAs are important negative regulators of eukaryotic gene expression. Clusters of rapidly evolving microRNAs appear to target genes regulating virus–host interaction in bats by dampening inflammatory responses, thus limiting immunopathology and possibly energy expenditure as well. Such genes include those active in antiviral immunity, DNA damage response, apoptosis, and autophagy. Understanding the roles of these microRNAs is important since P. alecto may be a natural reservoir of the human pathogens Hendra virus and Australian bat lyssavirus (Cowled et al. 2014). MicroRNAs have also been identified in the little brown bat (Myotis lucifugus), the big brown bat (Eptesicus fuscus), and the Jamaican flying fox (Artebius jamaicensis) (reviewed by Cowled et al. 2014).

1.2 VIRAL PATTERN‐RECOGNITION RECEPTORS AND THE BAT IMMUNE RESPONSE TO MICROBES

Molecular patterns used by the host to recognize viral infections are more limited than those used to recognize bacteria and commonly consist of nucleic acid recognition. Viral DNA and RNA are detected by several different classes of host pattern‐recognition receptors, such as retinoic acid inducible gene I (RIG‐I)‐like receptors (RLRs) in the cytoplasm, Toll‐like receptors (TLRs), NOD‐like receptors (NLRs), and the cyclic GMP‐AMP synthase (cGAS) and 20‐50‐oligoadenylate synthetase (OAS) nucleotidyltransferases.

TLRs 3, 7, 8 and 9 are found in endosomes and detect dsRNA after endocytosis. TLRs 3, 7, 8 recognize viral RNA, while TLR 9 recognizes viral, bacterial and protozoan DNA. TLRs’ ligand recognition properties vary among species. Bat TLRs 3, 7, 8, and 9, in general, evolved under similar functional constraints as other mammals and those of Desmoids rotundus display the classic genetic characteristics and three‐dimensional structure seen in other mammals (Escalera‐Zamudio et al. 2015). TLR 9 of bats, however, form a monophyletic clade positioned externally to all other eutherian mammals. Comparison of TLR among eight bat species revealed that TLR evolution in bats is order‐specific. This may reflect the need of different bat groups to adapt to a wide variety of ecological niches containing different pathogens profiles. While most bat‐specific mutations of the ligand‐binding site are unlikely to alter their function, some unique, nonconservative mutations are also present in the ligand‐binding sites of bat TLR 9 that might influence its ligand‐binding specificity. The adaptations found in the TLRs among bat groups and between bats and other mammalian TLRs may aid in resistance to infection by specific pathogens found in different environments (Escalera‐Zamudio et al. 2015).

TLRs 1, 2, 4, 5, 6, and 11 are expressed on the cell surface and recognize protein, lipid, and carbohydrate moieties in bacteria, protozoa, and fungi (Cowled et al. 2011). RIG‐I‐like receptors are cytoplasmic and detect viral RNA generated during their replication. The cytoplasmic cGAS recognizes short pieces of double‐stranded DNA and activates the Stimulator of Interferon Genes (STING) in the endoplasmic reticulum. This stimulates expression of type I IFN genes via TBK1‐IRF3 (TANK binding kinase 1/interferon response factor 3) signaling. It recognizes DNA viruses and bacterial DNA and as well as some RNA viruses. Three‐dimensional X‐ray crystal structures of cGAS and OAS1 show considerable similarity, despite the fact that OAS1 recognizes double‐stranded RNA and that the proteins have very different DNA sequences (Hancks et al. 2015). Binding of OAS and cGAS to double‐stranded RNA or double‐stranded DNA, respectively, produces nucleotide second messengers that activate RNase L (OAS) and STING (cGAS), initiating antiviral responses. Both of these genes are under positive selection and may undergo parallel evolution (Mozzi et al. 2015). Long stretches of unmodified dsRNA, while found in RNA and DNA viruses, are not produced by host cells. Host dsRNA sensors include protein kinase R (PKR), which suppresses viral protein synthesis, and RLR melanoma differentiation‐associated gene‐5 (MDA‐5), which induces interferon production. In addition to its antiviral activities, OASs may also play a role in antibacterial defense and cancer suppression (reviewed by Lohöfener et al. 2015). The RIG‐I like helicases retinoic acid‐inducible protein (RIG‐1) and MDA‐5 are important cytosolic pattern‐recognition receptors that detect viral RNA, with RIG‐I recognizing short dsRNA and MDA5 recognizing long dsRNA (Siu et al. 2014).

The TLR mRNAs in P. alecto and Rousettus leschenaultia have been cloned. Genome or transcriptome data also detect TRL in M. lucifugus and Artibeus jamaicensis (Schountz 2014). P. alecto TLR 1 to TLR 10 have a high degree of similarity to those of humans and other mammals. TLR 3, however, is highly expressed in bat liver, unlike the case in other mammals where it is primarily expressed in dendritic cells (Cowled et al. 2011). Cowled et al. (2012) also cloned the genes for RIG‐I, MDA‐5, and LGP2 in P. alecto and found that their primary structure and tissue expression patterns are similar to that found in humans. Bat databases also contain genes for the NLR members Ciita, Nod1, Nod2 (Schountz 2014).

1.3 INTRODUCTION TO THE INTERFERONS

Humans produce a number of type I IFNs: IFN‐α, with 13 subtypes, and IFN‐β, in addition to a single gene for IFN‐κ, IFN‐ε, and IFN‐ω (Kepler et al. 2010). Bat IFNs are only distantly related to those of humans and other mammals and those from Megachiroptera and Microchiroptera are separated into two genetic groups (He et al. 2014). Sixty‐one ORF for type I IFNs were found in the bats M. lucifugus and P. vampyrus. They are divided into several distinct subfamilies, including IFN‐α, IFN‐β, IFN‐κ, IFN‐ω, and IFN‐δ (Kepler et al. 2010). The single type II IFN is IFN‐γ (immune interferon), while the type III IFNs are composed of groups of IFN‐λ genes. The latter family includes four groups in humans, IFN‐λ1 (IL‐29), IFN‐λ2 (IL‐28A), IFN‐λ3 (IL‐28B), and IFN‐λ4. Of these, IFN‐λ1 and IFN‐λ3 genes have been also found in P. alecto (reviewed in Virtue et al. 2011a). Dobsonia viridis contains eight IFN‐α gene types (amino acid similarity 88.4–99.4%) plus one pseudogene. Phylogenetic studies which compare the type I IFNs of bats with those of other mammals show that these genes are under positive selection and diversity is due to duplication and gene conversion (He et al. 2010).

1.3.1 Regulation of interferon production

Interferon production relies upon a family of nine IFN‐response factors (IRFs) in humans, of which only IRF1, IRF3, IRF5 and IRF7 appear to be positive regulators of type I IFN transcription, with IRF3 and IRF7 promoting antiviral activity. IRF7 is the master regulator of type I IFN‐dependent, and perhaps also type III‐dependent, immune responses. It is constitutively expressed in plasmacytoid dendritic cells, cells of the innate immune response which specialize in IFN production, and at low levels in most other cell types. IRF7 is found in lymphatic tissues while nonimmune tissues express almost undetectable levels unless stimulated by type I IFN (reviewed by J. Zhou et al. 2014).

IFN induction in fibroblasts utilizes an intracellular pathway in which dsRNA or 5'‐triphosphorylated ssRNA of RNA viruses bind to one of two cellular RNA helicases, MDA‐5 or RIG‐1, respectively, to phosphorylate IRF3 via TBK‐1 or IKKε. Phosphorylated IRF3 forms a homodimer that translocates into the nucleus where it stimulates IFN‐β gene expression via the transcriptional coactivators p300 and CREB‐binding protein. In order to fully activate the IFN‐β promoter, IRF3 acts in concert with the transcription factors NF‐κB and AP‐1. NF‐κB is activated in part by PKR, a protein kinase that also recognizes dsRNA. This first‐wave of IFN production triggers expression of IRF7. IRF7 may be activated in the same way as IRF3, stimulating a positive‐feedback loop that stimulates production of IFN‐α in a second wave (reviewed by Thiel & Weber 2008).

The primary IFN producers of the lymphatic system are myeloid dendritic cells (mDC) and plasmacytoid dendritic cells (pDC). The mDC utilize the intracellular pathway as well as a second, endosomal TLR 3 pathway. Additionally, mDC, as well as monocytes, specifically produce IFN‐β, IFNλ1, and IFNλ2. In contrast, pDC use endosomal TLR7 and TLR8 to recognize ssRNA to produce all IFN types (reviewed by Thiel & Weber 2008; Lazear et al. 2015). All TLRs except TLR 3 activate IRF7 via the adaptor protein, MyD88 (myeloid differentiation primary response gene 88). MyD88 forms a complex with the kinases IRAK‐4 (interleukin 1 receptor associated kinase 4), IRAK‐1 and TRAF‐6 (TNF receptor‐associated factor), which binds directly to IRF7. This leads to TRAF‐6‐mediated ubiquitination and IRAK1 or IKK‐1(IκB kinase‐1)‐dependent phosphorylation and nuclear translocation of IRF7. IRF7 then binds promoter elements and induces IFN transcription. The human IFN‐β and the IFN‐α promoter regions have four or two to three positive regulatory domains, respectively, that are binding sites for IRFs (reviewed by J. Zhou et al. 2014).

TLR 3 and TLR 4 activate IRF7 via the adaptor molecule TRIF (TIR‐domain‐containing adapter‐inducing IFN‐β) which forms a complex with TBK1, IKK‐ε (inhibitor of nuclear factor‐κB kinase‐ε), and IRF7. The phosphorylated IRF7 forms a homodimer or a heterodimer with IRF3 prior to nuclear translocation and induction of type I or type III IFN (reviewed by J. Zhou et al. 2014). A large amount of constitutively expressed IRF7 is found in pDCs and the levels are further upregulated by a positive feedback loop to produce high levels of IFN‐α and IFN‐β (reviewed by Thiel & Weber 2008).

1.3.2 The JAK‐STAT pathway and interferon‐stimulated genes

IFN‐α/β bind to the type I IFN receptors present on almost all cells. Conformational changes in the intracellular region of the receptor activate the Janus kinase/signal transducer and activator of a transcription (JAK‐STAT) signaling pathway. The JAK family members JAK‐1 and TYK‐2 phosphorylate two STAT proteins (signal transducer and activator of transcription 1), STAT‐1 and STAT‐2. They form a heterodimer that recruits IRF‐9 to form the IFN stimulated gene factor 3 (ISGF‐3) complex that translocates to the nucleus where it binds and activates IFN‐stimulated response elements (ISRE) in promoter regions of IFN‐stimulated genes (ISG) (reviewed by Thiel & Weber 2008).

Some ISG have antiviral activities, including the GTPase Mx1 (orthomyxovirus‐resistant gene 1), PKR, and the 2'‐5' oligoadenylate synthetases (2‐5 OAS)/RNaseL system. Mx1 protects against infection with many RNA and some DNA viruses by binding and inactivating their ribonucleocapsid. PKR is a serine‐threonine kinase that phosphorylates the eukaryotic translation initiation factor eIF2, thus blocking translation of cellular and viral mRNAs. The 2‐5 OAS catalyzes synthesis of short 2'‐5' oligoadenylates that induce the latent endoribonuclease RNaseL to degrade viral and cellular RNAs. PKR and OAS/RNaseL eliminate virally infected cells by suicide resulting from reduced basal activity. They are constitutively expressed in an inactive form and are upregulated by type I and type III IFNs. Mx1 is not found in resting cells, but is induced by type I and type III IFNs (reviewed by Zhou et al. 2013). The promoter region of human PKR contains conserved KCS (kinase conserved sequence)‐ISRE promoter elements, permitting a high degree of PKR induction following IFN stimulation. Additionally, IRF‐1 activates PKR in the absence of IFN signaling in stimulated human cells. Human Mx1 and OAS1 also contain ISREs, but, unlike PKR, their induction is highly dependent on IFN signals.

Transcriptome analysis of stimulated immune cells from P. alecto detected a number of ISGs including Mx1, Mx2, OAS1, OAS2, OASL and PKR (Zhou et al. 2013). The functional domains and promoters of the bat P. alecto’s Mx1, PKR, and OAS1 are highly conserved with respect to those of other mammals, but P. alecto Oas1 has two ISRE in its promoter while the human Oas1 has only one. This may increase the inducibility of the bat gene by type I and type III IFNs. Bat OAS1 and Mx1 were induced in a highly IFN‐dependent manner after stimulation by IFN or dsRNA, but, as is the case in humans, PKR may be induced by an IFN‐independent mechanism.

Pteropine orthoreovirus NB (PRV1NB) (Nelson Bay virus) is a dsRNA reovirus of fruit bats while Sendai virus is a negative‐strand RNA paramyxovirus widely used to induce IFN. Bat Oas1 was most readily induced of these ISGs by IFN stimulation or Sendai, or to a lesser extent PRV1NB, infection. While Mx1 was inducible by either virus, Pkr was barely upregulated at all, nor was it induced by IFN stimulation, as occurs in humans. Bat Pkr is induced by the dsRNA analog poly (I:C), a viral‐associated molecular pattern which induces type I IFN, however. Due to its greater inducibility, OAS1 may therefore have the major antiviral role at least in this species or group of bats (Zhou et al. 2013). Sendai virus also induces a stronger IFN‐β and IFN‐κ2 response than PRV1NB. Both of these viruses may antagonize PKR responses in bats. Reoviruses have been shown to encode proteins that sequester dsRNA and reduce activation of Pkr and Oas1 by dsRNA. Oas2 is upregulated in vesicular stomatitis virus‐infected P. vampyrus immune cells to a greater extent than in poly (I:C)‐treated cells (Kepler et al. 2010).

Stimulation of Rousettus aegyptiacus primary kidney cells with human IFN‐α induces phosphorylation and nuclear translocation of STAT‐1. As is the case with human cells, infection with rabies virus inhibits nuclear translocation in IFN‐stimulated bat cells but not its phosphorylation. R. aegyptiacus STAT1 mRNA is highly expressed in the liver and to a low extent in muscle and spleen (Fuiji et al. 2010). RIG‐I, STAT1, and IFN‐β were also cloned and sequenced in R. sinicus and R. affinis horseshoe bats (Li et al. 2015). The Rhinolophus RIG‐I sequences have 87% nucleotide and 82% amino acid identity to that of humans and the most similarity to that of P. alecto (91% nucleotide and 86% amino acid identity). The Rhinolophus STAT‐1 sequence has 91% nucleotide and 95% amino acid identity to that of humans and the most similarity to that of R. aegyptiacus fruit bats (94% nucleotide and 97% amino acid identity). The Rhinolophus IFN‐β sequence has the greatest difference with other species, having only 74–76% nucleotide and 59–61% amino acid identity to that of humans and the greatest similarity with the P. vampyrus and R. aegyptiacus fruit bat (81–84% nucleotide and 69–74% amino acid identity) (Li et al. 2015). RIG‐I, STAT‐1, and IFN‐β are all highly expressed in bat spleen, lung, and intestines. Poly (I:C) stimulated a 30 000‐fold increase in interferon in bat cells and only a several hundred‐fold increase in mouse cells (Li et al. 2015). Taking these results together, RIG‐I and STAT‐1 from several species of bats have similar structures and functions to those of humans.

Other ISG in humans include the RNA‐specific adenosine deaminase acting on RNA 1 (ADAR 1), the product of ISG56, and ISG20. ADAR 1 deaminates adenosine on dsRNAs to inosine, leading to genomic mutation. ADAR 1 activation is also inhibited by reoviruses (Zhou et al. 2013). ISG56 binds the eukaryotic initiation factor 3e subunit of eIF3 to suppress viral RNA translation (reviewed by Thiel & Weber 2008), while IGS20 is a 3'‐5' exonuclease that degrades ssRNA.

Some viruses, including highly pathogenic members of the Flaviviridae, Filoviridae, Rhabdoviridae, Bunyaviridae, and Reoviridae, use acidic endosomal entry pathways to gain access to the host cell’s cytoplasm. The human immune system inhibits viral entry via an ISG, the IFN‐induced transmembrane protein 3 (IFITM3) (Benfield et al. 2015). IFITMs block cytoplasmic entry by blocking fusion of viral and host cell membranes by multimerization and increasing membrane rigidity. Mouse IFITM plays an important role in limiting influenza‐induced morbidity and mortality. In bat cells, poly (I:C) up‐regulates IFITM3 expression. When expressed in the A549 cell line, the M. myotis IFITM3 orthologue co‐localized with transferrin (found in early endosomes) and CD63 (present in late endosomes or multivesicular bodies). It blocked cytoplasmic entry of pseudotyped viruses expressing glycoproteins from rabies, Mokola virus, Lagos bat virus, and West Caucasian bat virus about 100‐fold. IFITM3 reduced viral yield mediated by hemagglutinin from multiple types of influenza virus by over 100‐fold as well. Virus production was increased by siRNA knockdown of IRITM3 (Benfield et al. 2015). In addition to bats, pigs were also shown to express protective IFITM3.

1.3.3 Type I interferons

Type I IFNs act in a direct antiviral capacity, but also inhibit cell proliferation, regulate apoptosis, and modulate adaptive immunity. They are produced by all nucleated mammalian cells and are upregulated early after infection, activating expression of >300 antiviral and immunomodulatory genes (Thiel & Weber 2008). Dendritic cells produce high levels of IFN‐α, while epithelial cells, fibroblasts, and neurons initially release IFN‐β and later switch to IFN‐α.

All known mammalian type I IFN genes are unusual in that they contain no introns (generally a trait of bacterial genes). The types and numbers of functional subtypes of type I IFNs vary between bats and other mammals as well among bat species. IFN‐ω has the greatest number of subtypes in bats, 12 intact members for M. lucifugus and 18 for P. vampyrus. While the IFN‐α family is large in humans, M. lucifugus has only pseudogenes, while P. vampyrus has 7 intact genes (Kepler et al. 2010). The IFN‐δ family consists of 5 intact genes in P. vampyrus and 11 genes in M. lucifugus. Pig placenta is the only other tissue found to contain a functional IFN‐δ gene, and it is involved in embryonic development in pigs, not in antiviral activity (Kepler et al. 2010). The genome of M. lucifugus additionally contains 1 complete IFN‐β, 2 IFN‐ε, and 2 IFN‐κ genes and P. vampyrus has 1 intact member of each of these genes (Kepler et al. 2010).

1.3.3.1 IFN‐α and IFN‐β

Characterization of IFN‐α and IFN‐β from Rousettus aegyptiacus revealed that they are most closely related to those found in swine (72% amino acid identity) (Omatsu et al. 2008). The IFN‐α ORF contains 562 base pairs and encodes a 187‐amino acid protein while the IFN‐β ORF is 558 base pairs and encodes a 186‐amino acid protein. Stimulation of Rousettus leschenaulti primary kidney cells and the Tb‐1 Lu bat lung cell lines with poly (I:C) leads to increased transcription of IFN‐β in the former, but not the latter, cells. IFN‐α gene expression occurs later, in response to the presence of IFN‐β. The production of IFN‐β is rapid and transient while that of IFN‐α is longer‐lasting (Omatsu et al. 2008). The difference in gene expression could be due to differences in tissue type or may result from the use of primary versus immortalized cell lines.

E. helvum cells react to viral stimulation by a high degree of induction of type I IFN mRNA, IFN protein secretion, and efficient ISG induction. When infected by O’nyong‐nyong virus, E. helvum strongly induces IFN genes, but this virus still evades the IFN system by a translational block (Biesold et al. 2011).

There is a high seroprevalence for Henda virus among Australian fruit bats despite an absence of illness, unlike the high degree of pathogenicity in humans. In humans, henipavirus protein P gene products interfere with IFN‐α and IFN‐β production via cellular MDA5 and STAT proteins (reviewed by Virtue et al. 2011b). Additionally, infection of human cells with either Hendra or Nipah viruses fails to induce IFN transcription. The same study found that when exogenous IFN was present in henipavirus‐infected human cells, ISG transcription was only partially blocked and that the exogenous IFN greatly reduced numbers of infected cells and syncytia. Thus, in humans, henipavirus immune evasion appears to be due to a large degree to failure to produce type I IFN (Virtue et al. 2011b).

Since the bat IFN response system is important for protection against adverse effects of other viruses that cause severe human illness, IFN responses may also be responsible for the persistent, nonclinical, Hendra infections of bats. Lung cell lines from T. brasiliensis and an interscapular tumor line from Myotis velifer incautus are resistant to henipavirus infection (Virtue et al. 2011a). However, Hendra or Nipah infection of lung, kidney, and fetal cell lines derived from P. alecto does not induce IFN‐α or IFN‐β expression and expression of IFN‐λ is reduced by 50%. IFN signaling is also antagonized in these cell lines since ISG54 and ISG56 transcription in response to exogenous IFN‐α was blocked by henipavirus infection. In these cell lines, therefore, henipavirus infection appears to be controlled by unidentified mechanisms and not by interferon responses (Virtue et al. 2011a). It is important to determine whether this is also the case in fetal and adult bat primary cell cultures. Interestingly, in humans, henipavirus infection of human cells inhibits IFN production but not the interferon signaling pathway (Virtue et al. 2011b).

Zho et al. (2014) have shown the P. alecto contains a single, functional, full‐length variant of IRF7 that has a wider tissue distribution than that of other mammals. In humans and mice, IRF7 expression is very low in cells other than pDC and cells which are active, while P. alecto IRF7 is present in comparable levels in immune‐related and nonrelated tissues, including brain, heart, kidney, liver, lung, small intestine, and testis. Stimulation of bat kidney cell lines induces peak levels of IRF7 at 9 h, 3 h later than peaks in bat type I and type III IFNs but similar to that of bat ISGs Mx1, OAS1 and PKR, consistent with IRF7 induction via a type I IFN feedback loop as is seen in other species (P. Zhou et al. 2014). Even though the MyD88 binding domain of bat IRF7 has little sequence conservation with that region of human IRF7, the differences do not affect IRF7 function either in IFN transactivation activity or activation by MyD88. Bat IRF7 activates both IFN‐α and IFN‐β promoters and bat MyD88 and IRF7 have similar binding capacity as those from humans. Deleting the MyD88‐binding region of bat IRF7 also decreases IFN activation. Additionally, using siRNA to knockdown IRF7 functions impaired IFN‐β induction in Sendai virus‐infected cells and enhanced Pulua virus replication (P. Zhou et al. 2014).

1.3.3.2 IFN‐κ and IFN‐ω

While the roles of the type I IFNs, IFN‐α and IFN‐β, are well‐known, the importance of IFN‐κ and IFN‐ω is less well characterized. He et al. (2014) found that these genes from brain cell lines of Eptesicus serotinus are conserved among most microchiropteran species. Both of their promoters contain transcription factor binding sites typical of mammals, including IRFs, ISREs, and NF‐κB. Since differences exist in the various IRFs and positions of IRF and NF‐kB binding sites, these genes from E. serotinus are likely to have different regulatory mechanisms (He et al. 2014). In vitro, IFN‐ω strongly activates IFN‐induced genes and IFN‐κ is a weaker activator. IFN‐ω also has the stronger anti‐lyssaviruses activity in an E. serotinus brain cell line, with anti‐EBLV‐1 activity greater than anti‐RABV activity, and the least activity is directed against EBLV‐2. This is relevant since E. serotinus is a major host of EBLV‐I in Europe (He et al. 2014). The situation is more complex, however, since there is a general silencing of IFN‐κ, IFN‐ω, and their induced genes during infection with bat‐associated lyssaviruses, perhaps permitting long‐term infection of bats by these viruses.

The IFN‐κ gene is found outside the type I IFN genetic locus, suggesting that this gene may undergo independent evolution in different groups of mammals. Indeed, phylogenetic analysis indicates that IFN‐κ sequences from Microchiroptera and Megachiroptera group separately from those of other mammals (He et al. 2014). While IFN‐ω and IFN‐κ sequences from E. serotinus grouped with those of other Microchiroptera, they are separate from Myotis IFNs (M. lucifugus, M. brandtii, and M. davidii). IFN‐κ from the Megachiroptera P. vampyrus clusters into a nonbat mammalian group (He et al. 2014).

1.3.4 Type II interferon

In humans, type II IFN (IFN‐γ; immune IFN) is mainly produced by activated T helper 1 cells and constitutively by natural killer (NK) cells. It acts in a paracrine or autocrine manner on macrophages, T cells, and NK cells. IFN II plays a role in the early innate as well as the adaptive immune responses responsible for long‐term control of viral infections (reviewed by Janardhana et al. 2012). It also stimulates antigen presentation by class I and class II major histocompatibility complex (MHC) molecules and effects cell proliferation and apoptosis via stimulation. Its primary function is not antiviral, although it does repress viral genes and up‐regulates host antiviral proteins, such as 2,5‐OAS, PKR, guanylate binding protein, and adenosine deaminase (reviewed by Janardhana et al. 2012).

IFN‐γ from the Hendra virus host, P. alecto, is conserved and functionally similar to that of other mammals. P. alecto IFN‐γ shares 99% amino acid identity with P. vampyrus and 70% with M. lucifugus, but only 44% similarity with the mouse homolog. The IFN‐γ genes Ifngr1 and Ifngr2 have been detected in A. jamaicensis as well. Features that are conserved with type II IFNs of other species include the proteins’ six α helical structure, essential regions in the C‐terminal, a high degree of hydrophobicity, and conserved potential N‐linked glycosylation sites (Janardhana et al. 2012). As is true of other species, mitogen‐stimulated P. alecto splenocytes secreted IFN‐γ, which inhibited viral growth in Semliki Forest virus‐infected P. alecto kidney cells and the microchiropteran T. brasiliensis lung cells. Hendra virus infection of P. alecto kidney cells was also inhibited (Janardhana et al. 2012).

1.3.5 Type III interferons

The human type III IFNs are the highly conserved IFN‐λ1, IFN‐λ2, and IFN‐λ3. They resemble IL‐10 structurally and use the IL‐10 receptor as a co‐receptor (Lazear et al. 2015). IFN‐λ receptors in human and rodents are primarily restricted to epithelial cells and differ from those of type I and type II IFNs (Donnelly & Kotenko 2010). While P. vampyrus has three IFN‐λ genes that are similar to those present in humans, the closely related P. alecto appears to only have two functional IFN‐λ genes. IFNλ expression is greater in P. alecto splenocytes infected with Tioman virus. Ifit1 recognizes 5′ triphosphate‐RNA from single‐stranded RNA viruses. IFNλ also inhibited replication of Pulau virus, a dsRNA bat orthoreovirus, and dramatically increased expression of Ifit1 and, to a lesser extent, Ddx58 in a P. alecto cell line. These immune molecules have similar antiviral activity to type I and type III IFNs from other mammals. IFNβ and IFNλ trigger expression of the P. alecto Mx1 gene, a GTPase that may target viral nucleoproteins, and Oas1, which activates RNaseL and degradation of viral RNA, but not Pkr (Schountz 2014), a GTPase that appears to target viral nucleoproteins.

Bat type III IFNs are differentially induced upon exposure to synthetic dsRNA. Type I and type III IFNs are produced early in infection, with type I induced as early as 30 min and type III IFNs at 1.5 h. Peak expression of both groups occurs at 6 h and decreases by 24 h. IFN‐λ2 response to poly (I:C) was approximately100‐fold greater than that of IFN‐λ1 and expression of IFN‐β was higher than either (Zhou et al. 2011b). IFN‐λ2 may cause as much as a 25‐fold induction of ISG56 and 4‐fold induction of RIG‐I. Type I and type III IFNs utilize different induction pathways, with type I IFN being activated by both endosomal and cytosolic pattern‐recognition receptors and type III IFN being activated predominantly by cytosolic molecules such as RIG‐I.

Tioman virus is a ssRNA virus belonging to the paramyxovirus family, which includes the henipaviruses, Hendra and Nipah, that infect P. alecto and P. vampyrus, respectively. The natural bat host of Tioman virus is the closely related Pteropus hypomelanus. Type III IFNs are upregulated by infection of bat cell lines by Tioman virus (reviewed in Virtue et al. 2011a). In humans and Pteropus genera of giant fruit bats, Tioman virus interacts very weakly with STAT2 (Caignard et al. 2013), fails to degrade STAT1 in human cells or prevent its nuclear translocation, and is unable to inhibit type I IFN signaling. Tioman virus does, however, bind to human STAT3 and MDA5 and interferes with IL‐6 signaling and IFN‐β promoter induction in human cells (Caignard et al. 2013). Interestingly, while Tioman virus does not upregulate splenic type I IFN production in P. alecto, it does induce a type III IFN response (Zhou et al. 2011b; Lazear et al. 2015). IFN‐λ2 is also able to protect P. alecto from Pulau virus.

Zhou et al. (2011a) cloned and characterized the genes for P. alecto IFNλR1 and IL10R2, which compose the type III IFN receptor complex. This complex is functional and has a wide tissue distribution in these bats. Expression of IFNλR1 is greatest in the spleen and small intestine. Epithelial and immune cells are responsive to IFN‐λ. Humans produce two splice variants of the IFNλR1 chain, a soluble and truncated transmembrane form. No such alternative splicing of IFNλR1 is present in P. alecto. The two splice variants found in humans may negatively regulate IFN‐λ and their absence in P. alecto may allow for greater IFN‐λ activity in at least some bat species.

IFN‐λ are believed to be more closely related to the IL‐10 cytokine than to type I IFNs, even though they serve as antiviral agents whose biological activities have some overlap with those of type I IFNs, including inducing similar subsets of ISGs. IFN‐λ are induced by a variety of viruses, including the human metapneumovirus; respiratory syncytial virus; SARS coronavirus; rotavirus; reovirus; and Sindbis, dengue, vesicular stomatitis, encephalomyocarditis, influenza, hepatitis B, hepatitis C, and Sendai viruses. They play a major role in preventing viral infection via hepatic, respiratory, gastrointestinal, and integumentary epithelia, as well as through the blood:brain barrier.

In response to infection by many viruses, IFN‐α amplifies IFN‐λ production and IFN‐λ also amplifies IFN‐α/β production by inducing IRF‐1 and IRF‐7 (reviewed by Lazear et al. 2015). Type III IFNs also suppress T helper 2 responses, increase IFN‐γ production, reduce numbers of T regulatory cells, increase degranulation by CD8+ T killer cells, and attack tumors (Donnelly & Kotenko 2010; Lazear et al. 2015).

Viral infection often coinduces type I and type III IFN production by similar pathways, although type III IFN responses are usually the weaker of the two. IFN‐λ1 and IFN‐β transcription are activated by both IRF3 and IRF7, while IFN‐λ2 and IFN‐α utilize primarily IRF7. The IFN‐λ1 enhanceosome, however, differs from that of IFN‐β, suggesting that they are differently regulated and, together, may bypass some of the viral evasive mechanisms, for additional host protection (Stoltz & Klingstrom 2010). When human epithelial cells were infected with Hantaan virus, IFN‐λ1 induction preceded that of MxA and IFN‐β, and IFN‐α was not produced. IFN‐λ1 and MxA were also produced in Hantaan virus‐infected Vero E6 cells, which do not produce type I IFNs, therefore this virus can induce IFN‐λ1 and ISGs without the need for either IFN‐α or IFN‐β. Activation of IFN‐λ1 requires replicating Hantaan virus since inactivated virus did not induce these genes (Stoltz & Klingstrom 2010).

1.3.6 Viral avoidance of the host IFN response

Most disease‐causing viruses at least partly block production of IFN‐α/β or their downstream mediators. Negative‐strand RNA hantaviruses do so by escaping recognition by RIG‐I and MDA5, disrupting TBK1‐TRAF3 complex formation, or preventing NF‐κB nuclear translocation. Host protective responses lead to the production of IFN‐λ1, however, which turns on Mx1 (Stoltz & Klingstrom 2010). SARS‐CoV and MERS‐CV block innate antiviral signaling by blocking type I IFN induction in several cell lines in vitro (Matthews et al. 2014). MERS‐CoV from humans and BtCoV‐HKU4 and BtCoV‐HKU5 from bats contain accessory proteins that inhibit IFN‐β induction in their hosts (Matthews et al. 2014). One of their accessory proteins, however, only weakly blocks the NF‐κB signaling pathway.

In order to avoid host IFN responses, some viruses block IFN transcription or ISGs. Henipavirus V protein blocks IFN production by sequestering STATs in a cytoplasmic complex that is unable to undergo nuclear translocation (Fujii et al. 2010). Upon stimulation R. aegyptiacus cells rely on nuclear translocation of phosphorylated STAT1, which bears 96% amino acid similarity to human STAT1. In a bat kidney cell line, rabies virus also inhibits nuclear localization of STAT1 rather than blocking its phosphorylation.

Mapuera virus is a paramyxovirus of the Rubulavirus genus that was originally isolated from an asymptomatic Sturnira lilium fruit bat in Brazil. Mapuera virus may or may not be pathogenic in humans and its host range is unknown. Mapuera virus V protein serves as a type I IFN antagonist by preventing nuclear translocation of STAT1 and STAT2 following IFN stimulation, without affecting their phosphorylation. Cytoplasmic sequestration blocks formation of the ISGF3 transcription factor complex in cells from diverse mammalian species, including those from bats, humans, monkeys, dogs, horses, and pigs, but not mice. Since some STAT1 is induced in the infected cells, it appears that at least some IFN is being produced. Mapuera virus V protein binds to mda‐5, but not rig‐1, and thus inhibits only IFN induction by the former pathway. Other paramyxoviruses have been shown to induce IFN via RIG‐I. The antagonism of the IFN pathway in bat and human cells suggests that another protective immune response may be used (Hagmaier et al. 2007).

1.4 ANTIBODIES AND B LYMPHOCYTES

Eutherian mammals produce five classes of antibodies: IgG with multiple subclasses, IgM, IgA, IgE, and IgD. Birds, by contrast, lack IgD, IgE, and IgA. Microchiropterians, however, transcribe all five antibody classes, indicating that the restriction of weight required for flight need not alter representation of different antibody classes. Megachiropterans do not produce IgD (Baker & Zhou 2015). An IgG isotype has been detected in C. perspicillata, E. fuscus has two IgG isotypes, while M. lucifugus has five (Bratsch et al. 2011). Fruit bats had been found to possess lower levels of antibodies that agglutinate, hemagglutinate, and fix complement upon antigenic stimulation than common laboratory animals. Additionally, antibody production is delayed in these bats (Iha et al. 2009).

Antibodies are composed of two identical heavy and two identical light chains, each containing variable (V) and constant (C) regions. The V regions are responsible for recognition of the antibodies’ targets (antigens) which initiates a cascade of events, eventually leading to the production and release of highly specific antibodies. The V region is divided into complementary determining regions (CDR) and framework (FW) regions. The three CDR are the regions of the antibody that actually bind antigen, while the FW regions provide structure. The specificity of individual antibodies and the presence of a vast number of microbial and nonmicrobial antigens necessitate a similarly great number of antibodies and a mechanism to allow production of such a large range of diverse antibodies. In contrast to most mammals, one of the primary mechanisms used by primates and rodents to generate antibody diversity is to rearrange regions of antibody genes that encode the variable, antigen‐binding component of the antibodies. The V region of antibodies is encoded by one each of multiple, distinct V, D, and J genes. Formation of antibodies involves genetic rearrangement, in which one of the V genes binds to one of the D genes and to one of the J genes to form large numbers of antibodies specific for different antigens.

The variable heavy chain repertoire (VH) is divided into families and three clans. An analysis of the expressed, rearranged antibody VH regions from P. alecto and the unarranged repertoire of P. vampyrus found that these bats use representative VH genes of families from all three studied VH clans (I, II, III). Most studied mammals, with the exception of primates and rodents, have few or no genes from at least one of the three clans (Baker et al. 2010). Pteropid bats also use the same sort of genetic rearrangements of their numerous VH genes and extensive number of D and J gene segments, a higher number than seen in humans. This permits a large number of possible diverse VDJ rearrangements vital for recognition of numerous antigens, including those of microbes. The two studied Pteropid bats, primates, and rodents are the only eutherian species known to have retained a high level of the VH diversity (Baker et al. 2010). At least some bats have over 250 germline VH3 genes, 5–15 times greater than that of primates and rodents (Bratsch et al. 2011). This should allow a high degree of antibody diversity via VDJ recombination. M. lucifugus has indeed been found to have a very high level of diversity of VDJ loci.

One of the key antigen‐binding regions of bat antibody variable heavy chain, CDR3, has fewer tyrosine and more arginine in comparison with other animals, perhaps forming antibodies with a greater degree of specificity with a weaker capacity to bind antigen. Bats also have some mutations in the FW3 areas which distinguish them from humans and mice (Bratsch et al. 2011). Bats also fail to produce neutralizing antibody upon infection with some viruses (Baker et al. 2010).

In addition to the VDH rearrangements discussed above, many mammals use somatic hypermutation to increase the amount of antibody diversity needed to respond to a large number of antigens. In this process, antibodies undergo a very high rate of mutation in the areas critical to binding antigen. While M. lucifugus bats possess a diverse VH gene repertoire which includes five of the seven human VH gene families, they have a very low mutation frequency, decreasing the role of somatic hypermutation in its generation of antibody diversity and indicating a greater reliance on VDJ rearrangements and junctional diversity (ability to rearrange V, D, and J gene segments at more than one site) to generate a highly diverse antibody repertoire (Bratsch et al. 2011; Schountz 2014).

B cell activating factor (BAFF) and aproliferation‐inducing ligand (APRIL) are members of the proinflammatory tumor necrosis factor (TNF) cytokine family that share two receptors. They are vital to B cell survival and activities, such as B lymphocyte proliferation, maturation, antibody secretion, isotype switching, T cell activation, and T‐independent antibody responses. Full‐length cDNA of BAFF and APRIL were cloned from the Vespertilio superans Thomas bat. They are encoded by 873 and 753 base pair ORFs that encode 290 and 250 amino acids, respectively. Both bat BAFF and APRIL express the typical TNF signature of a transmembrane domain, a putative furin protease cleavage site, and three cysteine residues. BAFF amino acid level identities between bat and dog, horse, human, and mouse are 80.82, 82.76, 77.59, and 55.28%, respectively. APRIL identity with dog, horse, humans, and cattle all exceed 80%. Cloned BAFF and APRIL are functional in that they promote survival and growth of mouse splenic B lymphocytes (You 2012a, 2012b). BAFF expression is high in the spleen and lower in the kidneys and intestine, similar to its localization in humans. APRIL expression is also highest in the spleen but may be found in other tissues, including bone osteoclasts and tumor cells (You 2012b).

Seasonal horizontal transmission of antibodies appears to occur between young bats and adult females. Seronegative bats typically seroconvert to many antigens, including microbial components, at 16–24 months of age, clustering temporally with late pregnancy of adult females (reviewed by Baker et al. 2010). Additionally, Pteropus and Myotis species show seasonal excretion peaks of henipavirus and coronaviruses associated with periods of pregnancy and lactation. Seroconversion in adult males, however, occurs in mid‐year (May 2010 and July 2011), close to the April–June mating period in E. helvum (Mutere 1968), when aggression among males increases as well as males having more intimate contact with females.

1.5 MACROPHAGES, DENDRITIC CELLS, AND PROINFLAMMATORY CYTOKINES

Mammalian macrophages are typically potent producers of type I IFNs as well as potent pro‐inflammatory cytokines, including TNF‐α and interleukin (IL)‐1 and IL‐6. These cytokines have antiviral activity, but are some of the leading causes of immunopathology. There are two types of dendritic cells with different origins and somewhat different roles. Rapid response to viruses or viral components is performed by pDC, which produce large amounts of type I IFN, have direct antiviral activity, and modulate natural killer cell and CD8 T killer cell activity. While mDC produce large amounts of type I IFN and other immunomodulatory cytokines, they are also antigen‐presenting cells which stimulate adaptive immune responses by T lymphocytes.

Monocyte/macrophages play a major role in filovirus pathogenicity in humans, triggering bystander apoptosis of lymphocytes and increased vascular permeability, leading to circulatory collapse. Macrophages also express a cell surface cytokine receptor that interacts with clotting factors VIIa and X to activate the coagulation cascades and the hemorrhagic manifestations of filovirus (reviewed by Basler 2012). Infection of human monocyte/macrophages by filoviruses in vivo and in vitro induces production of proinflammatory cytokines that attract additional cells to the site which, in turn, also become infected. Dendritic cells are also infected by ebolaviruses but do not produce inflammatory cytokines or initiate T helper cell responses. Interestingly, all of these cells produce little type I or type II IFN (reviewed by Basler). As discussed earlier in this chapter, bats appear to have lesser levels of adaptive immune responses than many other mammals and dampened production of proinflammatory cytokines. This may protect them against the damaging effects of viral infection that lead to life‐threatening disease in humans. It should be noted that a relatively small amount of studies has focused on adaptive immunity in bats or their production of proinflammatory or anti‐inflammatory cytokines.

1.6 T LYMPHOCYTES

T lymphocyte activity is vital for virus clearance in most viral infections, including coronavirus infections. This has been seen in a Middle Eastern respiratory syndrome coronavirus (MERS‐CoV) mouse model and appears to be important in human defense against this virus as well. Immunodominant epitopes which stimulate CD8 T‐cells are found in the MERS‐CoV S protein (Zhao et al. 2014). In humans, severe acute respiratory syndrome (SARS) survivors produce memory T cell responses against the products of the viral S, M, E, NP, and ORF3a genes as well (Oh et al. 2011). Six years after recovering from SARS, people still bore SARS‐specific memory CD4 T helper lymphocytes and CD8 T killer lymphocytes. Human T memory cells respond primarily to a dominant SARS‐CoV nucleocapsid protein by producing and releasing powerful inflammatory mediators, including IFN‐γ, TNF‐α, and macrophage inflammatory proteins 1α and 1β upon activation by antigen. The CD4+ memory cells produce the Th1 cytokines IFN‐γ, TNF‐α, and IL‐2 (Oh et al. 2011). The production of an excessive, detrimental, inflammatory response in humans and the absence of such a reported response in bats may at least partially explain the differences in the pathology of MERS and SARS, and perhaps other viral diseases, in bats and humans.

Recognition and activation of CD4 T helper cells requires interaction between antigens, the T cell receptor, MHC II, and CD4. The complete sequence of R. aegyptiacus CD4 cDNA reveals that bat CD4 has more homology to that of cats and dogs than to that of humans and mice. Bats’ CD4 Ig‐like C‐type 1 region contains an insertion of 18 amino acids. Bat CD4, like that of pig, cat, whale, and dog CD4, also lacks a cysteine, an amino acid which forms disulfide bonds and plays a major role in protein folding. Human, monkey, and mouse CD4 have this cysteine, indicating that human and bat CD4 differ in several key structural features (Omatsu et al. 2006).

Stressing the importance of CD4 and MHC II contributions to bat population health and fitness, there is a correlation between MHC II DRB alleles, hematophagous ectoparasite loads (ticks and bat flies), and the neotropical Noctilio albiventris bats’ reproductive state. Specific DRB alleles are associated with nonreproductive adult males and females, who also bear higher ectoparasite loads than reproductively active animals (Schad et al. 2012). The presence of ticks may affect immunity to co‐infection with other pathogens since antigen presentation by macrophages and T helper cell functions are reduced by compounds in tick saliva. Only one polyallelic DRB gene is found in N. albiventris, while two DRB gene copies are present in S. bilineata. Allelic variation in S. bilineata is believed to result primarily from intragenic recombination rather than intergenic recombination (Mayer & Brunner 2009; Schad et al. 2012).

The DRB gene, especially exon 2, is under positive selection, as evidenced by a greater than 2‐fold higher rate of nonsynonymous versus synonymous substitutions, particularly in the antigen‐binding sites (Mayer & Brunner 2009; Schad et al. 2012). DRB is believed to also alter individual bat body odor, as is the case in other mammals. Since bats are an extremely gregarious group of mammals with some colonies containing several million individuals, odor recognition is partially used as a means of recognition. DRB, therefore, may also be involved in recognition of family and mate selection (Schad et al. 2011) Male N. albiventris also have higher heterozygosity rate and genetic variability in the DRB gene than do females.

After recognizing antigen presented by MHC class II proteins, T helper lymphocytes produce and secrete cytokines. T lymphocyte‐derived cytokine production in bats is delayed in comparison with production by mice. Bat IL‐2, IL‐4, IL‐6, IL‐10, IL‐12 p40, and TNF‐α contain 152, 134, 207, 178, 329, and 232 amino acids, respectively. These genes are highly conserved in comparison with those from horses, dogs, cats, pigs, and cattle. Interestingly, all of these cytokines are encoded by a single exon (Iha et al. 2009).

1.7 OTHER PARAMETERS OF THE IMMUNE RESPONSE

Papenfuss et al. (2012) explored the transcriptome of P. alecto using stimulated spleen, white blood cells, and lymph nodes, in addition to unstimulated thymus and bone marrow. Approximately 18 600 genes were identified. Highly expressed genes were involved in routine cellular processes, such as cell growth and maintenance, enzymatic activity, metabolism, production of cellular components, and energy pathways. Approximately 500 genes, however, were associated with immune function and these composed 3.5% of the transcribed genes in this bat species. The largest proportion of immune genes was associated with T cell activation (79 genes). Other immune‐related genes include those involved with natural killer cell cytotoxicity (72), Toll‐like receptor cascades (70 genes), B cell activation (50), and antigen presentation (41). Transcriptome analysis also revealed the expression of genes such as pattern‐recognition receptors and some, but not all, natural killer cell receptors. Genes for NLRC5 and NLRP3 were also found to be transcribed. NLRC5 is believed to positively and negatively regulate bat antiviral immune responses, while NLRP3 is activated by danger signals, including viral and bacterial infections and environmental irritants. NLRP3 activates caspase‐1 in the inflammasome to cleave IL‐1β and IL‐18 into their mature, active forms (Papenfuss et al. 2012). As would be expected, the transcriptome also included IFN‐α and its receptor, as well as genes orthologous to the IFN stimulated genes Mx1, Mx2, OAS1, OAS2, OAS3, OAS‐like (OASL), PKR, RNaseL, and ISG15. Natural killer cell‐related molecules in the transcriptome include inhibitory CD94/NKG2A, CD24, CD16, and CD56. The MHC class I antigen‐loading and presentation pathway in the bat transcriptome include beta‐2 microglobulin, transporter associated with antigen processing 1, calnexin, and tapasin, CD1a, CD1b, CD1d, MR1, HFE, FcRn, and ULBPs. The bat MHC class II‐associated mRNAs present include homologs to class II invariant (CD74) chain, cathepsin S, the alpha chain homologs of DMA, DOA, DQA, and DRA, and the beta chain homologs of DMB, DOB, DQB, and DRB. Lymphocyte‐related molecules found in the transcriptome include α, β, δ, and γ chains of the T cell receptor, the TCRζ chain, CD3, CD4, CD8, and CD28, immunoglobulin variable and constant domains of the heavy and light chains, and B cell co‐receptors CD19, CD22, CD72, CD79a, and CD79b.

Transcriptional analysis of P. vampyrus bat kidney cells infected with the avian paramyxovirus, Newcastle disease virus, shows that 200–300 antiviral genes are highly upregulated, including genes for IFN‐β, RIG‐I, MDA5, ISG15, and IRF1. Infection with Hendra and Nipah viruses, by contrast, did not induce these innate immune response genes. Furthermore, the addition