The amphibian immune system

Abstract

Amphibians are at the forefront of bridging the evolutionary gap between mammals and more ancient, jawed vertebrates. Currently, several diseases have targeted amphibians and understanding their immune system has importance beyond their use as a research model. The immune system of the African clawed frog, Xenopus laevis, and that of mammals is well conserved. We know that several features of the adaptive and innate immune system are very similar for both, including the existence of B cells, T cells and innate-like T cells. In particular, the study of the immune system at early stages of development is benefitted by studying X. laevis tadpoles. The tadpoles mainly rely on innate immune mechanisms including pre-set or innate-like T cells until after metamorphosis. In this review we lay out what is known about the innate and adaptive immune system of X. laevis including the lymphoid organs as well as how other amphibian immune systems are similar or different. Furthermore, we will describe how the amphibian immune system responds to some viral, bacterial and fungal insults.

This article is part of the theme issue ‘Amphibian immunity: stress, disease and ecoimmunology’.

1. Introduction

In the context of the major decline of amphibian populations worldwide, the immune system represents a major intersection that connects different endangered organisms with multiple factors such as climate change and environmental stressors that affect their ability to respond to infectious diseases. As such, filling the fundamental gap in knowledge on amphibian immunity is paramount for a full understanding of the issues causing amphibian disappearances as well as for any successful interventions. However, this task is complicated by multiple challenges including the remarkable diversity of amphibian biology and the limited tools and resources necessary to enable the immunological studies of non-model amphibian species. Nevertheless, important progress has been achieved in recent years thanks to genomic and transcriptomic technologies that can build upon fundamental studies from the previous decades. The goal of this review is to provide an overview of the basic amphibian immune system as it is related to other jawed vertebrates as well as to summarize recent developments with an attempt to identify main gaps and priorities.

As mentioned, amphibians are very heterogeneous in morphology, biology and physiology with animals adapted to a huge variety of environments from arid desert to frigid permafrost. Amphibians are classified into three orders: anurans (frogs, toads), urodeles (salamanders) and legless caecilians. The majority of what we know to date about the immune system in amphibians has been learned using model species such as the African clawed frog Xenopus laevis (reviewed in [1]), and to some extent the urodele Ambystoma mexicanum (axolotl) [2]. More recently, extensive immunological studies have been performed in the Chinese giant salamander [3]. Because the immune system of X. laevis remains to date the most extensively characterized, we have used this species as a guide in our discussion for comparison with mammals and for providing additional information when available for other amphibian species (table 1).

Table 1.

Outlines of the main features of the Xenopus immune system compared to other anuran and urodele species (differences between Xenopus and other amphibians are highlighted in bold).

	Xenopus species	Anuran species	Urodele species
lymphoid organs and immune cells	— thymus (T cell development) — spleen and liver (B cell development, myelopoiesis) — bone marrow rudimentary (no B cells) — no structured lymph nodes	— thymus (T cell development) — spleen and liver (B cell development, myelopoiesis) — bone marrow rudimentary — putative lymph nodes in some species	— thymus (T cell development) — spleen and liver (B cell development, myelopoiesis) — bone marrow rudimentary — no structured lymph nodes
cells	— lymphocytes (B & T) — granulocytes — natural killer — monocyte/macrophages — dendritic cells	— lymphocytes (B & T) — granulocytes — natural killer — monocyte/macrophages — unknown	— lymphocytes (B & T) — granulocytes — unknown — monocyte/macrophages — unknown
innate	— skin AMPs: potent against bacteria, Bd, ranavirus — 12 TLRs — complement system: classical, lectin, and alternative — expanded intron and intron-less IFN type I — IFN-II (γ) expressed by immune cells — type III and IV IFN — distinct CSF-1 and IL-34-promoted macrophages — macrophage involved in regeneration	— skin AMPs: potent against bacteria, Bd — TLRs (number unknown) — complement system: classical, lectin and alternative — type I IFN (not characterized) — likely similar — unknown — macrophages present but distinction unknown — macrophage likely involved in regeneration	— active against ranavirus — TLRs (number unknown) — complement system: classical, lectin and alternative — type I IFN (possible role in regeneration) — likely similar — unknown — macrophages distinction but distinction unknown — macrophage involved in regeneration
adaptive	— one polymorphic MHC-I per genome — expanded non-polymorphic nonclassical MHC-I — thymocyte differentiation (RAG, TCR, CTX, CD5) — thymic education/selection (AIRE) — MHC-I restricted CD8 T cells cytotoxicity — MHC-II restricted helper T cell function — MHC-like restricted iT cell — IgM + B cells — IgM, IgX and thymus-dependent IgY — ranavirus neutralizing antibodies	— up to 3 polymorphic MHC-I per genome — nonclassical MHC-I found in some species — not characterized — not characterized — unknown — not characterized — not characterized — IgM + B cells — IgM, IgY (IgX?) — anti-ranavirus IgY produced but activity not tested	— multiple polymorphic MHC-I per genome — nonclassical MHC-I found in some species — not characterized — not characterized — unknown — not characterized — not characterized — IgM + B cells — IgM, IgY (IgX in some species, IgP in others) — neutralizing antibodies against ranavirus produced
ontogeny	— distinct tadpole immune system from adult — larval thymocyte death at metamorphosis — adult T cell education from new progenitor — differential expression of MHC-I and MHC-II between tadpoles and adults — tadpole stage more susceptible to ranavirus — tadpole tolerigenic anti-mycobacteria immunity	— likely similar — unknown — unknown — unknown — tadpole stage generally more susceptible to ranavirus — unknown	— likely progressive — unknown — unknown — unknown — all stages susceptible to ranavirus — unknown

2. Lymphoid organs, tissues and cells

As in all jawed vertebrates, the development of T cells in X. laevis and all anuran amphibians occurs in the thymus, which is a primary lymphoid organ that derives early during embryogenesis from the third pharyngeal pouch [4]. The thymus provides an epithelial microenvironment specialized in the generation and selection of T cells. By contrast, the differentiation of other immune cells including B cells and leucocytes in amphibians typically takes place in the spleen and the liver, whereas the role of bone marrow, a primary lymphoid organ in mammals, is generally more rudimentary in amphibians [5,6]. However, some evidence suggests that committed progenitors of macrophages and granulocytes are present in X. laevis bone marrow. Also within bone marrow are mesenchymal pluripotent stem cells that can differentiate into osteocytes, chondrocytes and adipocytes [7–9]. While no structures resembling lymph nodes are found in Xenopus, lymph node-like organs (jugular, procoracoid and prepericardial bodies) have been described in Rana pipiens, which contain lymphocytes, plasma cells, granulocytes and monocytes and are situated in the throat and axillary regions of the frog [10]. In urodeles, the bone marrow is not the site of hematopoiesis, except in the slimy salamander, Plethedon glutinosus, where it has been shown that granulopoiesis and lymphopoiesis take place in the bone marrow [2,11]. In axolotls and salamanders, hematopoiesis takes place in the spleen and the liver [2,12]. As for the thymus, the localization and number of nodules varies among species. There are no known lymph nodes in salamanders [2]. In the giant salamander the thymus, spleen, liver and kidney appear to all serve as the sites of lymphopoiesis [3].

Besides, B and T lymphocytes, myeloid lineage cells such as granulocytes (basophils, eosinophils, neutrophils) and monocytes/macrophages are likely present across all amphibians. However, the status of dendritic cells, which in mammals are central cells connecting the innate and adaptive arm of the immune system, is only partially elucidated in X. laevis and unclear in other amphibians. In X. laevis, a cell type with dual functions of follicular dendritic cells (FDCs) specialized in activating B cells and conventional dendritic cells has been characterized [13]. In addition, gene orthologs of the fms-related tyrosine kinase 3 (Flt3) and its ligand (Flt3lg), which are important regulators of dendritic cell homeostasis, were identified [14]. The production of a recombinant tagged Flt3lg further supports the DC nature of the dual FDC/cDC subset and suggests the presence of other DCs or DC-like subsets distinct from macrophages. This distinction is of importance owing to the multiple crucial roles of monocyte/macrophages found in X. laevis. Indeed, macrophages are not only central innate immune cell effectors against pathogens but are also involved in X. laevis tail regeneration and can specifically activate T cell response [15–17]. In urodeles, while little is yet known about their role in host defences against pathogens, macrophages have been shown to have a role in limb and organ regeneration. In axolotls and salamanders, histological studies confirmed by single cell sequencing experiments demonstrate a large influx of macrophages during regeneration [2,18,19]. Studies in salamanders have also shown the involvement of macrophages in cardiac regeneration through fibroblast mediation [2,20].

3. Innate immunity

Although the distinction between innate and adaptive immunity has become blurred over recent years, innate immunity generally refers to any type of host responses elicited by receptor-encoded genes or gene families that do not undergo somatic modifications [21] The innate immune system can respond broadly and fast (e.g. within hours) upon infection or injury by producing and secreting various compounds (antimicrobial peptides, interferon) and activating cellular defence such as cell-mediated cytotoxicity by natural killer (NK) cells and phagocytosis by neutrophils and macrophages. The innate receptors typically recognize conserved molecular patterns derived from pathogens or cellular stress. Molecular products can be as diverse as protein fragments (peptides), nucleic acids (RNA or DNA), sugar or lipids. Contrary to common misconceptions, the recognition from innate receptors is specific and sensitive. For example, the LPS binding protein from the horseshoe crab (Limulus polyphemus) can detect as little as 0.01 ng endotoxin per ml; lectins can distinguish L from D sugar; and TLR3 can detect nucleoside modification of RNA [22–24]. In addition, innate receptors can be encoded by widely diversified gene families (e.g. NITR in bonyfish; FcR-and KIR-related receptors in Xenopus [25,26]. For example, a cluster of over 70 paired FcR-related genes and four KIR-like genes have been identified in the genome of X. tropicalis and X. laevis [27]. Virtually nothing is known about these genes, although the presence of immunoreceptor tyrosine-based inhibitory motif (ITIM) in the cytoplasmic tails of many of them suggests an involvement in immune modulation. It is also unknown whether these genes are as polymorphic in gene as their equivalent in mammals [28].

(a) . Innate or pattern recognition receptors

Pattern recognition receptors (PRRs) are innate receptors able to recognize specific molecular structures derived from pathogens or pathogen-associated molecular patterns (PAMPs), as well as apoptotic host cells and damaged cells or damage associated molecular patterns (DAMPs). The stimulation of these PRRs typically induces signals that activate antimicrobial and pro-inflammatory responses important to contain and combat infectious agents as well as to potentiate adaptive immune responses. There are different types of PRRs including Toll-like receptors (TLRs), nucleotide oligomerization domain (NOD)-like receptors (NLRs), retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs), C-type lectin receptors (CLRs) and absent in melanoma-2 (AIM2)-like receptors (ALRs) [29]. Studies in amphibians are still limited to a few PRRs.

Toll-like receptors are a prominent family of PRRs. The TLR signalling pathway in vertebrates is a central system recognizing and triggering host defence against pathogens. Different cell surface and intracellular receptors can bind and recognize PAMPs that can be nucleic acids, proteins from bacterial cell wall and flagellin [30]. In contrast to mammals, X. laevis and its sister species, X. tropicalis, exhibit a higher number of TLRs (16 genes) [31,32]. TLRs are expressed both at tadpole and adult stages in Xenopus, which suggests that they play an important role in host protection across development, although more functional studies are needed [31]. TLR5 and TLR12 have been studied to some extent in adult X. laevis and shown to be activated at the transcriptional level (e.g. increased gene expression detected by qPCR) by bacterial RNA and flagellin, respectively [33]. Interestingly, stimulation of TLR5 with flagellin, a bacterial component, triggers the reactivation of quiescent ranavirus FV3 infection residing in peritoneal macrophages both in vitro and in vivo [33]. This suggests some links between commensal or pathogenic bacteria and deadly disease outbreaks in amphibian populations with persistent asymptomatic ranavirus infection. TLR22 that recognizes bacterial RNA is present in fish and Xenopus but not in mammals [33]. In urodeles, a high number of TLRs have also been found (at least 21), but their function is not well characterized [34].

The inflammasome is another important innate immune sensor system that can detect pathogens and cellular stress and induce rapid inflammatory responses [35]. Most genes of the inflammasome pathway are present in the Xenopus genomes. In particular, the NLR gene family is expanded in Xenopus. These genes encode intracellular sensors of PAMPs. However, functional studies are still lacking in amphibians to our knowledge [36].

(b) . Antimicrobial peptides (AMPs)

This class of small, secreted peptides is an important element of innate immune defence that is relatively well characterized in amphibians, especially regarding resistance against pathogenic chitrid fungi. The skin of multiple amphibian species, both anurans and urodeles, have been shown to possess glands that produce and release a variety of peptides that effectively inhibit growth of bacteria and chytrids in vitro [37]. Skin peptides have also been shown to neutralize ranavirus [38]. In Xenopus, several of these AMPs have been characterized such as magainins [37]. Some of the AMPs identified have anti-HIV qualities and derivatives have been synthesized to potentiate their antimicrobial qualities [39,40]. In urodeles only a small number of AMPs have been characterized. Specifically, tylotoin has been identified in urodeles [2,41]. In mammals, tylotoin is associated with wound healing as well as TGF-β and IL-6 secretion; it remains unclear if the same association is true in urodeles [2,41]. Interestingly, the mucosome of the eastern hellbender (Cryptobranchus alleganiensis alleganiensis) containing AMPs has also neutralizing activity against the ranavirus FV3 [42].

(c) . Interferon complex

This is a complex system of cellular defences against pathogens that in most vertebrates including amphibians is composed of four main classes: type I, II, III and the newly identified IV interferon (IFN), with their respective receptors and effector molecules. Type I IFN response can be activated by any cell of an organism upon infection (especially viral infection) as a way to alert and heighten defences of neighbouring cells against the pathogens [35]. By contrast to mammals that have alpha and beta type I IFN, amphibians possess numerous genes (e.g. 42 genes in X. laevis and 51 genes in X. tropicalis) including both intron-spanning and intron-less genes that cannot be confidently classified as alpha or beta by phylogenetic study (reviewed in [43]). Xenopus type III IFNs also comprise an expanded set of fish-like intron-containing and mammalian-like intronless genes [44]. To date amphibians are the only class known to have both intron containing and intron-less IFN genes [45]. The existence of intron and intron-less IFN genes allows for an increased diversity through the ability to create genetic variants compared to mammals. It also coincides or perhaps dictates an increased diversity in IFN receptors, regulatory factors and stimulated genes [43]. The intron and intron-less IFN genes in amphibians could serve as a model to understand IFN gene diversification. In mammals, while the distinct type I IFNs are preferentially produced by different cell types and type III IFNs are produced in the context of mucosal tissues, type II IFN or IFNγ is typically produced and released by more specialized immune cell effectors, including CD4 helper T cells, CD8 T cells, innate-like T cells and NK cells. In amphibians, an IFNγ response has only been characterized in bulk tissues or cells. For example, IFNγ transcript levels increase in adult X. laevis spleen, kidneys and liver during FV3 and mycobacteria infection [46,47]. Type III IFN or IFNλ has also been shown to be critical in X. laevis for host response against the ranavirus FV3 and to signal through a heterodimeric receptor constituted by the IFNR1 and IL10R2 [48]. In addition, the IFN type I and type III response is distinct between tadpoles and adult frogs [49]. While X. laevis encodes for an IFN-IV gene, other Xenopus species only contain pseudogene sequences for IFN-IV [43]. The members of class II cytokine receptors, IFN-υR1 and IL-10R2, have been identified as the receptor complex of IFN-υ and are associated with IFN-υ stimulated gene expression and antiviral activity in zebrafish (Danio rerio) and X. laevis [50]. It is unclear whether in Xenopus IFN-IV signals through a receptor complex for activation [43]. The composition of IFN genes in urodeles remains poorly understood [43].

(d) . Complement

The complement system comprises a complex set of interacting proteins that can enhance immune responses against pathogens. It plays not only a crucial role in host defence, but also in homeostasis and tissue regeneration and thus serves as a bridge between the innate and the adaptive immune systems (reviewed in [51]). The different components of the complement system of amphibians are conserved across vertebrates. The structural conservation in the complement system has allowed the synergistic interactions between amphibian and mammalian complement proteins (e.g. presence of complement elements in the amphibian serum can be assayed using mammalian erythrocyte lysis assay) [52]. All three pathways of the complement system, classical, alternative and lectin, are conserved between amphibians and mammals [52,53]. The complement system may not play a major role during ranavirus infections [54]. The knowledge we have on the amphibian complement system has mainly been obtained with hemolytic assays in past decades, while more recent data are now coming from transcriptomic studies focusing on the common frog's response to Batrachochytrium dendrobatidis (Bd) and ranavirus exposure [54]. These results will be discussed in a later section. In urodeles, the complement system has been mainly studied in the context of regeneration due to its role in macrophage recruitment [55]. In axolotls, C3, a key complement component, is upregulated during limb regeneration. In newts, both C3 and C5 are upregulated during limb regeneration [56]. These observations are also based on transcriptomic studies. Recently, the complement system has been implicated in several developmental processes including patterning, organogenesis and neurogenesis [57,58]. As such, connection between immune function and ontogenesis is an area that merits more investigation.

4. Adaptive immunity

The adaptive arm of the immune system is composed of B cells and T cells that express a hugely diverse antigen receptor repertoire (BCR and TCR). These antigen receptor repertoires are generated during the differentiation of these lymphocytes by a somatic diversification system dependent on RAG 1 and 2 genes as well as additional diversification by terminal deoxynucleotidyl transferase [59].

(a) . T cells

In mammals, immature thymocytes that do not express surface CD4 or CD8 co-receptors (double negative) become double positive CD4+ and CD8+ cells after having rearranged their TCR and TCRα chains allowing them to express a functional receptor. This then allows them to undergo positive and negative selection by interacting with major histocompatibility complex class I or class II (MHC-I or MHC-II). During positive selection, the TCR is tested for antigen recognition to ensure it recognizes the MHC properly. During negative selection, the TCR is tested to make sure it does not react against self-antigens in the context of self MHC. After selection, mature T cells exit the thymus to circulate in the body and accumulate in lymphoid organs such as the spleen and lymph nodes, where they can be activated by antigen presenting cells (APCs). The expression of the MHC-I-antigen complex together with co-stimulatory molecules (e.g. B7) by APCs activates CD8 T cells to expand and become cytotoxic effectors able to recognize and kill infected cells. Likewise, antigen presented by MHC-II plus co-stimulation activates CD4 T cells into various T-cell effectors producing different cytokines and assisting other immune cells during immune responses.

In X. laevis, the thymus develops at two weeks of age (stage 47), with the cortex and the medulla becoming recognizable by stage 48 [60]; X. laevis express MHC-II on their thymic epithelial cells [61]. Additionally, the cortical thymocyte-specific antigen of Xenopus (CTX) is expressed in the surface of cortical thymocytes [60,62,63] CTX can also serve as a marker to differentiate immature (CTX⁺) from mature (CTX⁻) T cells. By stage 51, corresponding to about three weeks of age, the thymus development is completed and thymic epithelial cells in the medulla begin to include Hassall's corpuscles (an area of the medulla) [6]. In mammals, this area is associated with clearance of apoptotic thymocytes, regulatory T cell (Treg) induction and thymocyte maturations. In mammals, Tregs maintain central tolerance by preventing autoimmunity. Negative selection is achieved through sporadic expression by thymic epithelial cells of cognate tissue-specific antigens. Thymic epithelial cells in X. laevis, as in mammals, express the autoimmune regulator gene (AIRE), which is thought to stimulate the expression of a wide range of self-proteins that are not specific to the thymus [64]. T cells that strongly react to endogenous proteins are eradicated by negative selection, which has been elegantly demonstrated in X. laevis by thymus/lymphocyte embryonic chimeras' experiments [65].

During metamorphosis, the Xenopus thymus undergoes a reduction in size and drastic loss of most T cells [66,67]. After metamorphosis, a new wave of stem cells migrates into the thymus and new adult-type T cells are generated [68]. It is currently unknown how much thymic development varies across anuran amphibian species with different timing for metamorphosis. In the giant salamander, the thymus appears to follow a more progressive development without drastic cell death from larval to adult stages [69].

In X. laevis, the αβTCR repertoire potential is estimated to be as high as in mammals (10¹⁵ or more) [70]. As in mammals, X. laevis T cells exhibit αβTCRs that are restricted by classical MHC-I and MHC-II molecules [15,71,72]. The γδTCR repertoire has also been studied in X. tropicalis and shown to use antibody-like V genes as in mammals [73]. However, little is known about the function of γδT cells in amphibians. In addition to these T cells, X. laevis is to date the only non-mammalian species in which preset or innate-like (i)T cells have been characterized [70,74]. These iT cells express a very limited αβTCR repertoire and usually interact with non-polymorphic nonclassical MHC-I molecules (see next §4.2). Using reverse genetic loss-of-function by transgenesis and MHC tetramer technology, one iT cell subset expressing the invariant TCRα chain Vα6α1.43 was shown to be critical for the host response against the ranavirus FV3, whereas another iT cell subset characterized by the invariant TCR rearrangement Vα45α1.14 is important for resistance against mycobacteria [47,70].

In urodeles, while transcriptomics suggests the occurrence of typical T cell differentiation and function, there is a lack of more direct characterization. The CD3 protein complex associated with the TCR has been resolved to some degree in the axolotl [75]. Some experiments also suggest that T cells are involved in limb regeneration in axolotl [2]. Although not formally investigated, TCR rearrangement is likely to be RAG-dependent as in all jawed vertebrates [3]. Interestingly, in the giant salamander, RAG-1 and RAG-2 genes are consistently expressed outside the thymus in the kidney, liver and spleen, which raises the possibility of additional sites of somatic diversification in this species, although more studies are needed to substantiate this [3].

(b) . Major histocompatibility complex

The classical Major Histocompatibility Complex (MHC) is a set of genes that encode polymorphic cell-surface molecules involved in the presentation of antigenic peptides to T cells. MHC-I molecules present antigens to CD8 T cells and MHC-II molecules present antigens to CD4 T cells. X. laevis has only one class MHC-I locus per genome that, as in mammals, is highly polymorphic and that resides in the MHC locus [76]. A crystal structure of one X. laevis MHC-I molecule complexed to an antigenic peptide derived from the ranavirus FV3 has been solved [77]. Notably, tadpoles do not express high levels of classical MHC-I until the onset of metamorphosis when it becomes expressed on all splenocytes and erythrocytes [78–80]. No MHC-II is detected on thymocytes and T cells at the tadpole stage, whereas T cells become MHC-II positive after metamorphosis [81,82]. By contrast, thymic epithelial cells, B cells in several locations and other antigen presenting cells, such as macrophages and dendritic cells, express MHC-II in both tadpole and adult stages [81]. Even with the low levels of MHC-I surface expression, tadpoles have circulating CD8⁺ T cells and remain immunocompetent. Tadpole immunocompetency is likely due to their limited TCR repertoire and their repertoire of MHC-I-like molecules, which are non-polymorphic. While the CD4 gene is expressed in X. laevis tadpoles and adults, there is limited knowledge of CD4 T cells besides the detection of CD8-negative cells expressing the pan T cell marker CD5 in the thymus and spleen [63,83]. CD4 expressed on CD8-negative T cells also appears to be capable of binding IL-16 as in mammals [84]. Besides classical MHC-I and MHC-II, X. laevis, X. tropicalis and likely other Xenopus species possess an expanded family of more than 20 non-polymorphic nonclassical MHC-I (mhc1-uba) genes that cluster outside the MHC locus in the telomeric region of the same chromosome. Some of these nonclassical MHC-I gene products have been shown to regulate the development and function of iT cells [47,85] (see §5.1 and 5.3). However, the function of many other of these genes remains to be elucidated.

The MHC gene complex has been studied by transcriptomics and genomics in several other amphibians. In anurans, different numbers of expressed polymorphic MHC-I loci varying from two to three in different species have been reported, which contrasts with the single one in X. laevis [86]. It is generally thought that both the number of MHC genes and alleles should allow for an increased peptide-binding repertoire and thus, increase resistance against pathogens. In the case of Xenopus, the repertoire of the single polymorphic MHC-I locus may be complemented by the large number of nonclassical MHC-I genes mentioned above. However, the respective role or advantage of polygenic and allelic variations is not well understood. MHC-I and MHC-II allelic variation has been investigated in the context of chytrid fungus infection of the lowland leopard frog (Rana yavapaiensis) and the southern corroboree frogs (Pseudophryne corroboree), revealing associations with either resistance or susceptibility to this pathogen [87–89]. MHC-I immunogenetics has also been extensively explored across 30 species of salamander species [90]. The giant salamander also expresses MHC genes, which undergo increased transcription during the infection by a ranavirus, Chinese giant salamander iridovirus (GSIV), consistent with their involvement in antiviral immune responses [2].

(c) . B cells and antibodies

While in mammals B cells differentiate in the bone marrow, they typically differentiate in the spleen or liver in amphibians. B cells can produce and secrete antibodies that can directly neutralize pathogens or assist the complement system in opsonizing and lysing pathogens. In X. laevis, the first population of B cells develop around two weeks of age (stage 47) in the liver (reviewed in [91,92]). Similar to B cell development in mammals, there are three stages to B cell development in Xenopus, albeit the stages occur earlier in development in comparison. The three stages are pre-immune B cells with only a heavy chain IgM and presumably a surrogate light chain that remains to be identified, a mature B cell expressing a surface heavy and light chain, and a B cell that secretes immunoglobulins without Ig surface expression (reviewed in [6]). Each pre-immune B cell clone is different, and they can recognize a variety of antigens. In tadpoles, the pre-B progenitor stage does not have many cells; therefore, the later stages are also limited in their positively selected cells for recombination of the IgM heavy chain. IgM + B cells are first detected around stage 49; at this point light chain recombination occurs [92]. There is a conserved recombination mechanism and structure of the constant and variable domains of the heavy and light chains between mammals and amphibians [91,92]. The majority of splenic B cells express IgM receptor at their surface and produce IgM. A smaller portion of B cells produce IgY and IgX isotypes. Phylogenetic study of IgY suggests that it is an evolutionary predecessor to mammalian IgG and IgE [93]. IgX is considered to be a functional analogue of mammalian IgA [94]. IgX is likely linked to mucosal immune responses as suggested by studies with cholera toxin oral immunization [6]. IgX is also preferentially expressed in the foregut and skin, and its expression is not altered by thymectomy, which implies that its switch from IgM is not T-cell dependent. By contrast, the switch and differentiation of IgY producing B cells is thymus dependent and requires collaboration between T and B cells [92]. Furthermore, while affinity maturation occurring in germinal centres in mammals leads to the generation of antibodies of very high affinity for a given antigen (1000- to 10 000-fold higher), this is not the case for Xenopus. Instead, the antibody affinity in X. laevis increases only modestly (approx. 10 fold) despite hypermutations being generated at similar rates by a process dependent on the Activation Induced Deaminase (AID), as in mammals [95,96]. The absence of an effective selection of hypermutations and the lack of well-organized germinal centres in X. laevis likely explains the poor increase in antibody affinity. However, one should perhaps evaluate amphibian antibodies for their ability to protect the host against natural pathogens rather than just comparing them with mammals. In such case, it is noteworthy that X. laevis and the giant salamander both produce protective and effectively neutralizing anti-ranavirus antibodies [3,97,98].

Concerning other amphibians, B cell and antibody responses have been studied in the southern leopard frog, Lithobates sphenocephalus [99]. While IgM and IgY have been detected in the serum of several ranid and bufonid species by ELISA using X. laevis anti-IgM and IgY monoclonal antibodies, it is currently unclear how widespread the IgX homolog is outside of Xenopus [100]. However, in urodeles a putative IgX gene homolog has been reported in axolotl, whereas other urodeles species such as the Iberian ribbed newt Pleurodeles waltl possess another distinct isotype named IgP, which suggests that different families of urodeles can express different immunoglobulin isotypes [101]. Although B cell markers are not fully characterized in urodeles, the presence of homologs of the B cell receptor complex (CD79a) and immunoglobulin lambda-like polypeptide 5 (igII5) have been observed in scRNA-seq analysis [2]. There are marked differences between B cell populations in the blood stream and at the site where limb regeneration occurs. IgM and IgY have been characterized in salamanders [2]. In the giant salamander, the lasting protective vaccination with neutralizing antibodies obtained against the ranavirus GSIV has been characterized [3].

5. Pathogen host responses

(a) . Ranavirus

Ranaviruses are large double-strand icosahedral DNA viruses of the family Iridoviridae that have become major amphibian pathogens (reviewed in [102]). Ranavirus infection generally causes internal bleeding, edema and ulcer formation. We will focus here only on what has been learned about host immune responses to these viral pathogens.

Xenopus laevis adults can clear an FV3 infection in 2–3 weeks, but tadpoles usually cannot fully control the infection and most die within a month [103]. It is known that viral clearance is dependent on CD8 T cells and antibody responses [70]. Even without a fully developed adaptive immune system (classical MHC-I deficiency), tadpoles are immunocompetent and do mount an immune response against FV3 but are ultimately unable to fully clear the infection. Some data suggest that the susceptibility may lie in a gap in the innate immune response. In studies investigating gene expression post-FV3 infection, it was found that the expression of pro-inflammatory genes (TNF-α, IFN-γ and IL-1β) in tadpoles is delayed or less responsive in comparison to the adult frogs [33,70]. Additional studies in tadpoles also have revealed that the iT cell population with the TCR invariant Vα6α1.43 that is restricted by the nonclassical MHC-I molecule XNC10 plays an important role in the immune response against FV3 [104,105]. When XNC10 or iT Vα6α1.43 expression is silenced, the infected animals perish sooner [46,105]. FV3 targets innate immune cells which can contribute to the lack of response in tadpoles. Additionally, the morphology of adult and tadpole macrophages differ, which could be another contributing factor [105]. Notably, the two different ligands of the colony stimulating factor-1 receptor (CSF-1R), CSF-1 and interleukin-34 (IL-34) trigger the generation of two morphologically and functionally distinct macrophage subsets, the IL-34-derived one being more anti-viral than the CSF-1-derived one [17,106]. In addition, macrophages are involved in ranaviral persistence by harbouring quiescent virus in asymptomatic frogs that can be reactivated by inflammation or TLR5 stimulation into a more virulent and deadly infection [33,107]. The respective roles of CSF-1- and IL-34-dervied macrophages in persistence of asymptomatic ranavirus infections appear to be different [108]. Among other anuran species, persistence of ranavirus found in different asymptomatic or resistant populations in the wild is consistent with a similar quiescence mechanism involving macrophages and with a risk of reactivation leading to sudden outbreaks [109–111]. In addition, variation of MHC-II polymorphism associated with intensity of ranavirus infection in tadpoles of the wood frogs (Rana sylvatica) has been found [112] as well as selection of particular MHC-I haplotypes for wild common frog (Rana temporaria) populations exposed to ranavirus [113]. Notably, MHC heterozygosity and some groups of alleles were significantly associated with lower ranavirus infection intensity. However, it remains to be determined whether specific MHC alleles confer susceptibility or resistance to ranavirus.

The tiger salamander and the Chinese giant salamander are two other amphibian species that are highly susceptible to ranavirus pathogens. In the case of the Chinese giant salamander, the infection affects both juvenile and adult animals. By contrast to X. laevis, the infections accumulate in the spleen rather than in the kidney [3,46]. Unlike other amphibians such as the Hellbender salamander and other species of frogs, it remains unclear if giant salamanders have an added line of defence by producing skin antimicrobial peptides active against ranavirus [3]. The susceptibility of the Chinese giant salamander is likely caused by the recent infection of the population, suggesting that vaccination or priming could result in an effective protective measure [3]. In addition, an effective anti-ranavirus vaccine has been developed in the giant salamander [3]. The production of neutralizing antisera and B cell memory imply that these animals have a more functional adaptive immune system contrary to what was concluded from studies with other urodeles [114]. In axolotls, microarray data indicate that the expression of innate immune genes is upregulated over genes related to the adaptive immune system during ranavirus infection [3].

Overall, further studies are needed into the mode of activation of the innate immune system after ranavirus infection in tadpoles, which could lead to potentiating a more robust response. Investigating other gaps between the innate and adaptive system could also point to exact involvement of other pathways in clearance.

(b) . Chytrid fungi

Batrachochytrium dendrobatidis (Bd) and more recently discovered Batrachochytrium salamandrivorans (Bsal) are fungal pathogens responsible for major outbreaks decimating amphibians worldwide [115]. Bd kills the host by disrupting electrolyte transport across the epidermis of the skin leading to systemic metabolic dysregulation, electrolyte imbalance and subsequent cardiac arrest [116]. It was initially thought that amphibians were unable to detect and mount any immune responses against these pathogens. However, more careful investigation has revealed a more complex situation where Bd (even less is known for Bsal) is able to evade early immune detection, has the ability to inhibit T cell responses and as a result overcome host immune defences [117]. Several transcriptomics studies also suggest a disordered and dysfunctional response resulting from Bd infection [54,118].

Despite its resistance to Bd, X. laevis has been instrumental for a more functional characterization of anti-fungal immune responses [119]. Notably, IgY and IgM able to bind to Bd are secreted following immunization with Bd in the mucosal surfaces [120]. Importantly, studies in X. laevis have shown that Bd release products to inhibit T cell proliferation and induce apoptosis of T and some B cells [121]. Recently, a similar, albeit lower, lymphocyte inhibitory effect by Bsal has been reported [122].

As for other frog species, a decreased level of leucocytes has also been demonstrated in Litoria infrafrenata experimentally infected with Bd [123]. It has been reported that unlike X. laevis some frogs such as Rana mucosa do not produce detectable antibodies against Bd following immunization [124,125]. However, the immunization protocol in both studies was not adequate and efforts to prevent disease was not successful. Nevertheless, different immunogenetic studies have established a link between the occurrence of MHC-I and MHC-II alleles and either resistance or susceptibility to Bd infection [126,127]. Also, release of skin antimicrobial peptides has been shown to be a critical component of host resistance to Bd [38,128]. Still, very little is known about the role and antibody and of the complement system in host immune response to Bd infection [52].

Bd infections have led to a steep decline in amphibian populations in recent years, and there is still a major knowledge gap on host immune response to Bd that is critical to fully understand this infectious disease. Notably, little is still known of the early stage of host pathogen interaction at the site of infection (skin). It is unclear whether upon Bd infection infiltration of innate immune cell effectors such as neutrophils and macrophages, as well as localized release of cytokines (e.g. TNFα; IL-17) and chemokines (e.g. CCR3) occur. Investigation at the site of infection could benefit from the rapid progress in single cell transcriptomics. While Bd clearly can produce factors inhibiting T cell response, it is unclear when (at early, late or both stages of infection) these factors are released. The study of cellular innate responses to Bd in endangered species could take advantage of in vitro culture of skin explant or cell lines available or under development [129]. To initiate a T cell response that can be inhibited, APCs such as DCs and macrophages need to acquire Bd antigens to activate these T cells that in absence of draining lymph nodes can only occur in the spleen. The process of acquisition and presentation of Bd antigens by APCs to activate T cells remains to be characterized. Another important piece of information missing is detailed kinetics of host immune response against Bd from early to late stage of infection both at the infection sites (skin), the blood and the in the spleen. Studying the early stage of Bd infection is also critical because at later stages it becomes difficult to distinguish whether changes in immune function are related to Bd infection or to the overall pathology and physiological dysfunction of dying animals.

In urodeles, Bsal outbreaks are particularly devastating in East Asia and Europe, while to date Bsal has not been detected in the American continent (reviewed in [130,131]). Very little is known about host immune response against Bsal. One transcriptomic study in the Wenxian knobby newts (Tylototriton wenxianensis) indicates a weak expression response of immune genes suggesting that Bsal has an immune-dampening effect in this species [132]. In addition, co-infection for each pathogen between Bd and Bsal infection has been reported [131]. The consequences of co-infection for each pathogen and for the host are unclear. In one study, transcriptomics data from experimental infection of the eastern newt (Notophthalmus viridescens) suggest that co-infection Bd/Bsal compromises immune responses mainly against Bsal [133]. More investigation will be needed to determine the mechanism governing these co-infections.

(c) . Mycobacteria

Over 31 species of amphibians are known to be affected by mycobacterial infections [134]. Transmission occurs through aquatic contamination of the environment including the food. Mycobacteria cause granuloma formation and produce mycolactone, which causes apoptosis and necrosis, while inhibiting phagocytosis and proinflammatory cytokine production [134]. As for ranavirus, macrophages are key players during mycobacteria infection being both critical early innate immune cell effectors and targets for pathogens to evade host defences. Interestingly, as in the case of FV3, CSF-1- and IL-34-derived macrophages differ in their susceptibility to Mycobacterium marinum (Mm) with IL-34-derived macrophages being less susceptible to Mm and conferring resistance to infection in vivo [135].

Presumably because of developmental constraints, X. laevis tadpoles exhibit a more active disease tolerance toward mycobacteria that is dominated by iT cell responses and is minimally inflammatory (reviewed in [136]). By contrast, adult frogs elicit robust pro-inflammatory responses against mycobacteria, which are driven by conventional T cells. In X. laevis, impairment of the nonclassical MHC-I XNC4 or the TCR rearrangement iVα45-Jα1.14 results in a higher mortality rate of tadpoles infected with Mm. Since tadpoles mostly rely on iT cells, this indicates that the iT cell subset identified by the iVα45-Jα1.14 rearrangement and its interaction with XNC4 is critical for tadpole resistance to Mm. Adults clear mycobacterium infections better than tadpoles even with the impaired expression of iVα45-Jα1.14 [47]. Further studies characterizing antigen processing and cell surface expression of MHC-I-like molecules, as well as their engagement with the TCR of iT cells, will allow us to better understand the protective role of these iT cells against mycobacteria pathogens.

(d) . Other pathogens

Among other common amphibian pathogens, it is worthy to mention Aeromonas hydrophila, a Gram-negative bacillus, which is an opportunistic pathogen associated with red leg disease (a severe dermatosepticemia). Infection by this pathogen occurs when animals are stressed and their immune system is weakened [137]. More recently, Perkinsea, a dinoflagellate (protists), has been increasingly found to infect a wide diversity of amphibians throughout their developmental stages ultimately causing death to the host [138–140]. There are detection methods available, namely PCR [141]. However, given that this infectious disease has only recently been recognized, the mode of infection, its effects on the immune system, and any potential treatments remain unknown.

6. Concluding remarks

As mentioned in the introduction, because of the wide diversity of amphibian development, biology and ecology, expanding studies to various non-model species are likely to provide novel insights into the evolution and adaptation of the immune system. However, despite the recent progress in genomics and transcriptomics, immunological studies in non-model species are still underdeveloped. In addition to bulk transcriptomics, single-cell transcriptomics, as well as reverse genetics using CRISPR/Cas9 genome editing, can be adapted to non-model species, and thus provide new opportunity for immune investigations. We hope that the extensive knowledge gathered in Xenopus, axolotl and more recently the giant salamander that we have summarized in this review and in table 1 can help guide more extensive studies in a wide variety of amphibian species, including multi-organ (i.e. spleen, liver, infection sites) and longitudinal (kinetics) characterization of immune gene expression responses. Also, considering the unique development of anurans with distinct immune systems in tadpoles and adult frogs, more attention to immune processes during metamorphosis would be important even in Xenopus [142]. In the context of ongoing amphibian decline, an important tool for investigating host defences against viral, chytrid and other pathogens in different amphibian species, including endangered ones, is the availability, as well as the generation and characterization of new cell lines as discussed in a recent comprehensive review [129].

Data accessibility

This article has no additional data.

Authors' contributions

V.L.R.: conceptualization, writing—original draft, writing—review and editing; J.R.: conceptualization, funding acquisition, project administration, resources, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

Funding of the authors for this work: National Institute of Allergy and Infectious Diseases at the National Institutes of Health (grant no. R24-AI059830), and V.L.R. was supported by the Pathogenesis Training Grant (grant no. T32-AI118689).