Evolution of Cytokine Receptor Signaling

The generation of mechanisms that enable efficient cell–cell communication has been critical in the evolution of the complex cell systems that characterize multicellular organisms, including immunity. One such mechanism is mediated by an array of secreted factors collectively termed cytokines (Cytos). These small polypeptides are produced in response to a variety of stimuli and act via specific Cyto receptors (CytoRs) expressed on the surface of target cells (1). The so-called class I and class II families of CytoRs mediate key aspects of immunity in addition to other development and homeostatic roles (2).

The class I and class II receptors lack intracellular kinase activity and instead depend on associated tyrosine kinases, particularly members of the JAK family (3, 4). Ligand binding leads to conformational changes in the CytoR complex that activate associated JAKs to initiate various downstream signaling pathways, importantly including the STAT family of transcription factors (4–6), which represents one of the seven major signaling pathways controlling bilaterian development (7). Once tyrosine phosphorylated, the STATs dimerize and translocate to the nucleus where they influence transcription to mediate key phenotypic changes in the cell (3). Signaling is subsequently extinguished in a number of ways, including by the action of tyrosine phosphatases, such as the Src homology 2 (SH2) domain–containing protein tyrosine phosphatases (SHPs), as well as induction of suppressor of cytokine signaling (SOCS) proteins that inhibit CytoR signaling by several mechanisms, providing a classical negative-feedback loop (5, 8, 9).

The archetypal CytoR–JAK–STAT signaling module is typified by that found in extant protostomes, such as insects, which consists of one CytoR that initiates signaling via a sole JAK that activates a single STAT to mediate its effects, with one SHP and several SOCS proteins contributing to the negative regulation of this signaling pathway (10). The various CytoR–JAK–STAT signaling components expanded significantly during vertebrate evolution, such that mammals possess >50 CytoR molecules serviced by four JAKs and seven STATs, along with two SHP and eight SOCS proteins (4). This review summarizes the evolution of the class I and class II CytoR signaling pathway, exploring the origins of individual components and their subsequent expansion and diversification to become a cornerstone of communication between cells, particularly those of the immune system, the evolution of which CytoRs have likely shaped in profound ways.

Each of the core components of the CytoR signaling paradigm consists of modular protein domains, many of which have a long evolutionary history (Fig. 1). The initial steps in the evolution of CytoR signaling involved the generation of the individual components of the pathway and their subsequent consolidation into a functional signaling system. The components arose largely by accretion of pre-existing domains across a broad evolutionary time frame. However, the generation of all components of the canonical CytoR–JAK–STAT module and their coalescence into a functional pathway only occurred in bilateria (Fig. 1) (11).

Individual components of the cytokine signaling pathway were assembled largely from pre-existing domains that were modified for purpose.

The ligands that activate class I and class II receptors are small polypeptides ∼5–25 kDa, which derive from a four helix–bundle structure that has been used in diverse proteins across evolution, from cytochromes to ferritin (12). The first definitive representatives are present in extant bilateria, characterized by the fruit-fly Unpaired proteins, which possess a so-called “long” chain conformation most similar to vertebrate leptin and IL-6 (13, 14).

Class I and class II receptors are cell surface protein complexes that consist of one to four receptor chains, at least one of which can transmit an intracellular signal to mediate the effects of the cytokine. These receptor chains possess an extracellular CytoR homology domain (CHD) consisting of two fibronectin (FBN) type III folds with a connecting sequence associated with cytokine binding. They are divided into two families based on structural differences within the CHD: class I receptor chains possess two pairs of disulfide-linked cysteines within the first FBN fold and a highly conserved WSXWS motif toward the C terminus of the second FBN fold (15–17), whereas class II receptor chains have one cysteine pair located in each FBN fold of their CHD (18). The CytoR chains can also possess additional domains, including extracellular Ig-like and FBN-like regions, as well as transmembrane and intracellular sequences essential for signal transduction (15, 19). Ig and FBN domains were generated independently early in evolution and have been widely used as components of many protein families (12). However, the hallmark CHD appears to have arisen later, being found in extant placozoa in a protein displaying homology to vertebrate CytoR-like factor (CRLF)3, which consists predominantly of a class I–related CHD (11). Only in bilaterians did a CHD become associated with Ig-like, FBN-like, and transmembrane domains, as well as intracellular sequences, to form an archetypal receptor that is structurally related to the vertebrate class I receptor chain gp130 (GP130) and exemplified by the fruit fly Dome protein (20) and the related, but signaling-incompetent, ET/Latran protein that acts as a negative regulator (21).

JAKs possess a unique four-domain architecture consisting of an N-terminal four-point-one, ezrin, radixin, moesin (FERM) domain followed by a variant SH2 domain that collectively mediate interactions with CytoR intracellular domains (6, 22). Proximal to this is a so-called “pseudokinase” domain with homology to tyrosine kinase (TK) domains but no catalytic activity, which plays a regulatory role. Finally, at the C terminus is a classical TK domain (23, 24). FERM, SH2, and TK domains arose separately in early eukaryotes but only became combined later. A JAK-like protein was identified in extant porifera; it consists of a FERM, SH2, and single TK domain (25), which shows high conservation with JAK TK domains (11). However, incorporation of the additional pseudokinase domain to form a canonical JAK appears only in bilateria, as observed, for example, in fruit fly Hopscotch (26), which is likely due to duplication and subsequent modification of sequences encoding the adjacent TK domain (11).

STAT proteins consist of conserved central coiled-coil, DNA binding, and SH2 domains, as well as the more variable N-terminal and C-terminal domains (23). This protein family has ancient roots, with the most divergent being the plant GRAS proteins, which possess an SH2-like domain and a putative DNA-binding domain (27), whereas present-day slime mold possess a STAT-like protein that includes a coiled-coil domain involved in transcriptional regulation added via domain accretion (28, 29). The N-terminal and C-terminal transactivation domains formed later (30–33), with a definitive archetypal STAT clearly identifiable in bilateria (11, 34), as exemplified in the present-day fruit fly Marelle protein (34), which is most homologous to higher vertebrate STAT5 and STAT6.

SHP proteins consist of tandem SH2 domains, followed by a protein tyrosine phosphatase (PTP) domain, both of which predate holozoa. However, a SHP-like protein with a single SH2 domain linked to a PTP domain is observed in present-day choanoflagellates (35), suggesting that this topology evolved in holozoa (10). The archetypal SHP topology was generated in early metazoans, with the dual SH2–PTP structure identified in extant porifera (36).

SOCS proteins consist of a variable N-terminal domain, followed by SH2 and SOCS box domains. Proteins containing SOCS box domains were identified in extant choanoflagellates (12), but only porifera possess proteins in which a SOCS box is associated with an SH2 domain and small N-terminal sequence to form an archetypal SOCS protein (12). Interestingly, this pathway component was subsequently replicated to yield three distinct lineages in bilateria (37), which are differentially used in the CytoR signaling paradigm (38, 39).

Exactly how the archetypal CytoR–JAK–STAT signaling pathway arose is a matter of speculation. We recently argued that assembly of the core CytoR–JAK–STAT module is consistent with the “retrograde” model of pathway evolution, with STATs, then JAKs, and then CytoRs becoming associated into a functional pathway as a result of a series of relatively minor changes, whereas pre-existing negative regulators were able to be recruited from other pathways in a “patchwork” manner (11). As a corollary, it is almost certain that many, if not all, of the individual components functioned in alternative roles prior to their recruitment into the CytoR–JAK–STAT pathway. Evidence supporting this assertion can be inferred from the involvement of several components in noncanonical signaling in extant species. For example, several roles were identified for STAT proteins that are independent of CytoRs and JAKs. These include the control of cell growth, differentiation, chemotaxis and immune responses in slime mold downstream of G-protein–coupled receptors (29), and the regulation of metabolic functions in vertebrates by unphosphorylated STATs (40, 41). Similarly, several SOCS members are principally involved in the regulation of growth factor receptors and other pathways (39).

The archetypal CytoR–JAK–STAT signaling pathway has a diverse range of functions, as detailed most extensively in fruit fly. In this organism it is involved in embryonic segmentation (42), eye development (43), stem cell proliferation in the testes (44) and intestine (45), as well as controlling growth and metabolism (46). Importantly, it also functions in innate immunity, with a role in the maintenance of multilineage progenitors in normal hematopoiesis, immune cell proliferation in response to immune challenge, the production of antimicrobial peptides and intestinal epithelium repair in response to gut bacteria, as well as antiviral responses (45, 47) that have been confirmed in other insects, including mosquito (48).

The consolidation of an archetypal CytoR–JAK–STAT pathway over a large evolutionary time frame was followed by a relatively rapid multiplication of pathway components (Fig. 2). As a result, the majority of the CytoR–JAK–STAT signaling components were already in place at the time of divergence of bony fish (including mammals) and cartilaginous fish (including sharks) around 420 million years ago, representing the “core” CytoR signaling network. Local duplications of individual components appear to have contributed to this process, but the two rounds of whole-genome duplication (WGD) that occurred during early vertebrate evolution (49) represent the major mediator of the expansion (20, 34, 37).

Early in chordate evolution, a precursor of the class II receptors was generated, presumably via duplication of the original class I–related CytoR and representing the archetypal IFNR involved in initiating and coordinating antiviral responses (50). These original class I and class II receptor chains further expanded via a combination of local duplications and WGDs to generate a set of receptor chains covering all major CytoR topologies by the divergence of bony and cartilaginous fish (20, 37, 51–54). For class I receptor chains, there are representatives for each of the five structural groups (55) and a single copy of the alternate sushi domain-based receptor chain used in mammalian IL-2R and IL-15R. For class II, it includes “short” and “long” chains belonging to the IFNR and IL-10R families. The cytokine ligands are notoriously difficult to definitively identify, but the available data suggest that they largely coevolved in parallel (56, 57).

Assuming conserved assembly of these receptor chains into functional CytoR complexes, a “core” set of CytoRs can be deduced that covers all receptor topologies, which includes representatives of the majority of class I and class II CytoRs (Fig. 3).

The core receptors include precursors for each of the mammalian homodimeric group 1 receptors (EPOR, TPOR, GHR, PRLR) (Fig. 3, Supplemental Fig. 1). They also include homodimeric group 2 CytoRs (G-CSFR, LEPR), heterodimeric complexes of different group 2 receptor chains (LIFR, IL-12R, IL-23R, IL-27R, IL-35R), and group 2/group 3 heterodimeric (IL-6R, IL-11R) and heterotrimeric (CNTFR) complexes, all of which use GP130 or a related receptor for signal transduction (Fig. 3, Supplemental Fig. 2). For heterodimeric group 4/group 5 receptors, the core members notably include several using the common IL-2Rγc signaling chain (IL-4R, IL-7R, IL-21R and IL-9?), one using the related IL-13Rα (IL-13R) complex, and two using the common IL-3Rβc chain (IL-5R and an additional complex that is the likely precursor for IL-3R and GM-CSFR, termed IL-3/GMR), as well as the heterotrimeric precursor of IL-2R and IL-15R (termed IL-2/15R) (Fig. 3, Supplemental Fig. 3). Other “core” CHD-containing molecules present include several that interact with cytokines extracellularly to form active heterodimers (IL-12p40, IL-27Rβ, CRLF1) or act as a so-called “decoy” (IL-13Rα2).

Core receptors comprise single type I, type II, and type III IFNRs, as well as representatives for each of the IL-10R family members (IL-10R, IL-20R, IL-22R, IL-26R) (Fig. 3, Supplemental Fig. 4). They also include the CHD-containing IL-22 decoy (IL-22BP) and tissue factor (TF), the latter of which is not involved in cytokine signaling.

The bulk of the diversity of the other CytoR signaling pathway components was also generated within the same evolutionary period. Precursors of all four mammalian JAKs (JAK1-3, TYK2) arose from the archetypal JAK sequence via the sequential rounds of WGD (37, 58). In contrast, the archetypal STAT was duplicated locally early in chordate evolution, but the action of the two WGDs resulted in the generation of six STATs from these in two distinct subfamilies (precursors to mammalian STAT1–4 and STAT5–6) (34, 37). The IRF9 component of the variant heteromeric STAT complex activated in response to IFNs was also generated during this time frame (59). Finally, the archetypal SHP was expanded to two members (precursors of higher vertebrate SHP1–2), whereas the three original SOCS genes expanded to create precursors to all eight SOCS types, most likely via the WGDs (10, 37).

Available evidence suggests that the functions mediated via the majority of these “core” CytoRs have been conserved. For example, analysis in zebrafish showed that GHR functions in controlling growth (60), LEPR functions in metabolic control (61), G-CSFR participates in myelopoiesis (62), IL-7R contributes to T cell development (63) and type II IFNR participates in antiviral responses (64), similar to their mammalian counterparts. This was shown to extend to their downstream pathways where investigated, such as EPOR/JAK2/STAT5 in erythropoiesis (65, 66) and IL-2Rγc/JAK3/STAT5 in T lymphopoiesis (67). Although there are some exceptions [e.g., PRLR, which functions in mammopoiesis, lactogenesis, and reproduction in mammals (68, 69) but functions in osmoregulation and retinal development (70, 71) in teleosts], it remains reasonable to deduce the likely functions of the core receptors (Fig. 3). This indicates a strong trend toward specialization of functions with regard to a particular cell system (innate immunity, hematopoiesis, neurogenesis, stem cell maintenance, and control of growth and metabolism) and, indeed, to particular subsets within these (e.g., EPOR in erythropoiesis, G-CSFR in myelopoiesis). It has also underpinned the development of new functions, most notably in adaptive immunity.

Having established a core group of CytoR–JAK–STAT pathways, additional diversification occurred during subsequent vertebrate evolution (Fig. 2, Supplemental Figs. 1–4). Prior to the divergence of ray-finned fish (including teleost fish, such as zebrafish) and lobe-finned fish (including tetrapods, such as humans), limited additional components evolved, with available data suggesting just the ligand-specific receptor chain for OSMR and the negative regulator SHP3, as well as a potential partial precursor of TSLPR, probably by local duplication. Along the tetrapod lineage, local duplication subsequently generated the distinct ligand-specific chains for IL-2R and IL-15R, as well as IL-3R and GM-CSFR, along with a bona fide ligand-specific TSLPR chain (55). There has been a concomitant increase in the cognate cytokines for these receptors (55, 56), whereas the diversity of type I and type III IFNs has also increased (51, 72). However, downstream components remain largely unaltered, with the exception of the duplication of a single STAT (STAT5) in mammals and loss of SHP3, which remains as a pseudogene (37). In contrast, within the teleost lineage there was an additional round of WGD ∼305–450 million years ago (73), which largely underpinned the duplication of many components, including several receptor chains (PRLR, GHR, LIFRα, IL-2Rγc, IFNγR, and the noncanonical TF) (20, 74), whereas local duplications created an additional IL-12Rβ2, IL-4Rα, and the novel somatolactin receptor, as well as alternate type I IFNRs. Conversely, there was a loss of ligand-specific chains that complex with IL-3Rβc and that form the type III IFNR, along with their ligands (20, 53). In addition, duplicates for one JAK (JAK2), two STATs (STAT1, STAT5), and several SOCSs of particular relevance to CytoR signaling (SOCS1, SOCS3, CISH) were also generated through the additional WGD (37). Although the analysis remains incomplete, it is also clear that numerous cytokines (including IL-11, G-CSFR, leptin, and multiple IFNs) were duplicated by a variety of mechanisms (50, 51, 56, 57, 72). How these components are used within functional CytoR signaling pathways remains to be determined.

The ongoing diversification of CytoR–JAK–STAT components observed in higher vertebrates has been relatively minor. However, it has likely underpinned increased sophistication in the tetrapod/mammalian immune system through the unique contributions to hematopoiesis and immunity mediated by TSLPR (75)—and potentially IL-9R (76)—as well as the distinct functions of IL-2R compared with IL-15R (77, 78) and of IL-3R compared with GM-CSFR (79). However, among immune-related CytoRs, the increased complexity appears similar in teleosts, suggesting comparable sophistication in this lineage, although this has not been confirmed experimentally. The diversification also appears to reflect some lineage-specific environmental adaptations. For example, the divergence in IFNRs likely reflects the different pathogen spectrum and dynamics between water-dwelling teleosts and largely land-dwelling tetrapods (80). The role for the somatolactin receptor in body color regulation (81) and PRLR in osmoregulation (70) within teleosts also falls into this category.

The evolutionary history of the CytoR–JAK–STAT pathway provides insights into the function(s) observed for individual CytoRs in specific lineages (Fig. 3). Thus, the archetypal pathway, which was based on a GP130-related CytoR, likely possessed diverse and pleiotropic functions, including in hematopoiesis, innate immunity, neurogenesis, stem cell maintenance, and the control of growth and metabolism, as seen in extant insects (42, 43, 45–47, 82). This probably formed the basis for the functional diversity, pleiotropy, and redundancy of the CytoRs found in higher vertebrates (83), such as the pleiotropic functions of IL-6R (84) and LIFR (85); the specialist hematopoietic functions of EPOR (65, 86), G-CSFR (20, 87, 88), and TPOR (89); the innate immunity roles of type I and III IFNs (90, 91); the stem cell functions of TPOR (92); the neural functions of CNTFR (93); and the roles of GHR, PRLR, and LEPR in growth, metabolism, and reproduction (68, 94, 95). The increase in components allowed one major additional functionality to emerge, with the IL-2R and IL-3R families and type II IFNs exerting their major effects in adaptive immunity (63, 67, 74, 96). In addition, downstream components originally played pleiotropic roles (97), a property that has been maintained particularly in JAK1, JAK2, STAT3, STAT5, and SOCS3, although other duplicates have developed more specific roles, such as JAK3, STAT4, and STAT6 in adaptive immunity and TYK2, STAT1, STAT2, and SOCS1 in antiviral immunity (5).

Interestingly, but perhaps not surprisingly, the evolution of CytoRs is paralleled in the downstream signaling molecules, such that expansion of particular CytoRs appears to be a strong driver of the increase in relevant downstream components (37). Thus, the expansion phase that generated all major receptor types also generated all types of JAKs, STATs, SHPs, and SOCSs required to provide functional specificity downstream. Moreover, during the diversification phase, the additional receptor complexes generated in the tetrapod/mammalian lineage (TSLPR, IL-2R, IL-15R, IL-3R, GM-CSFR) all use STAT5, which was also duplicated along this lineage. Similarly, many of those duplicates generated in teleosts use JAK2 (PRLR, GHR, type II IFNR), and most act via STAT5 (PRLR, GHR, IL-2Rγc) or STAT1 (type I and II IFNRs), with several regulated by SOCS1, SOCS3, and CISH, all of which have been duplicated in this lineage.

The evolution of CytoR signaling, including a complex array of cytokines, CytoR chains, and downstream signaling molecules, requires strong selective pressures. One major driver is likely to be the powerful selective advantage of a robust and multifaceted immune system in which many of the CytoRs exert their influence (37, 72, 98) (Fig. 3). This is consistent with the emergence of the CytoR–JAK–STAT signaling pathway being coincident with the evolution of primitive multilineage innate immune cells, the function of which is influenced by this pathway, as exemplified in present-day fruit fly (45). It is further supported by the major expansion of the CytoR–JAK–STAT pathway being concurrent with the development of a sophisticated adaptive immune system during development (72, 98). This is supported by the extensive retention of WGD-generated duplicates of CytoR signaling components compared with the average retention rate of just 3–4% (73) and is consistent with other studies indicating that WGDs played a crucial role in adaptive immunity by providing new genetic materials for a range of essential components (99). Additional evidence for immunity as a key driver can be gleaned from the significant divergence in the IFN-responsive CytoRs between teleosts and tetrapods (51), likely as a result of the powerful selective forces associated with combatting ever-changing pathogens, particularly viruses, which develop subversion mechanisms to counter the antiviral effects of IFNs in fish and mammals (53, 80, 100). This is borne out, for example, by other data showing that the majority of type I IFNs have been subjected to purifying selection (101). However, selective pressures outside of immunity also likely contributed to the evolution of the CytoR pathway, such as the need for functional specialization as organismal complexity increases to provide fine tuning of specific biological processes.

This review highlights the long and complex evolutionary history of the class I and class II CytoR signaling paradigm. This is reflected in the myriad roles played by numerous CytoR–JAK–STAT modules across diverse extant species, but especially the key conserved roles in immune and blood cell development and function. This history can be divided into three broad stages:

• Emergence: The generation of individual CytoR–JAK–STAT pathway components largely by accretion of pre-existing domains and their subsequent coalescence into a complete signaling module. This archetypal pathway had diverse and pleiotropic functions, including a role in innate immunity.

• Expansion: The rapid increase in components during early vertebrate evolution principally driven by two rounds of WGD that ultimately produced core representatives for each of the major CytoR groups, along with their cognate ligands, and the majority of JAK, STAT, and key negative-regulatory components. This was associated with increased specialization and, notably, the emergence of adaptive immunity.

• Diversification: Further lineage-specific increases in specific components, differentially mediated by local duplication in tetrapods and WGD in teleost fish, with duplication of downstream components linked with that of the relevant CytoR. This may be related to increased immune system sophistication, as well as environmental adaptations.

Collectively, this suggests that CytoR–JAK–STAT signaling has played an important role in shaping evolution at the organismal level, with the strong selective advantage of developing an increasingly more sophisticated immune system representing a likely major driver. Future work examining CytoR component expansion following each WGD during vertebrate evolution will be highly informative in providing additional mechanistic details. In addition, further analysis of salmonid fishes, in which another round of WGD occurred (102), with retention of duplicates for many CytoR chains (74, 96, 103), is also likely to uncover unique insights into CytoR signaling evolution.

The authors have no financial conflicts of interest.