Immunological Diversity Is a Cornerstone of Organismal Defense and Allorecognition across Metazoa

Immune systems operate at the forefront of the ongoing evolutionary arms race between hosts and microbes. The mechanisms that mediate host immune responses are subject to constant diversifying selection to keep pace with microbes that can evolve rapidly as a result of short generation times and high mutation rates. This evolutionary interplay has driven the emergence of many strategies to generate and maintain sufficient molecular diversity to detect a potentially infinite array of nonself-molecules (Fig. 1).

One of the most straightforward ways to generate diversity is to maintain a large repertoire of genes encoding immune receptors within the germline DNA. These receptors tend to recognize broadly conserved signatures of microbial life (e.g., LPS, flagellin, double-stranded RNA) (1). Duplications of such receptors, either individually or en bloc, may provide the raw material for subsequent functional diversification through mutation or genetic recombination. In this way, primary sequences vary, while the functional modularity and ability to interact with downstream adaptors remain conserved. This strategy ensures that self-reactive receptors are eliminated, whereas alleles conferring protective advantage are retained within the population, thereby providing immunological memory on an evolutionary scale. However, restricting diversity to the genome is slow; receptor diversity can increase only with new generations and is random, so reliance on this strategy is best suited to species that reproduce frequently and have many offspring. Many species have evolved ways to supplement these phylogenetically dispersed, genome-encoded pattern recognition receptors (PRRs) with sets of somatically diversifying receptors to avoid falling behind in the host–pathogen arms race. Initially, somatic diversification was thought to be restricted to jawed vertebrates. The RAG-mediated, V(D)J recombination of Ig and TCR genes remains the best understood mechanism for generating immune receptor diversity. However, as our knowledge of immune systems in other, phylogenetically diverse species has improved, it has become clear that somatic diversity can be generated in a myriad of ways, and that a vast array of molecules can be co-opted from other systems and used as building blocks for diversification.

We present a review of strategies for increasing immunological diversity by highlighting specific examples across a range of phylogenies, mechanisms, and functions. Given the relatively limited exploration of immune systems and the enormous breadth of animal, plant, and microbial life, we anticipate that these will be joined by additional means of generating immunological diversity in the future.

The gene families that encode PRRs have undergone many lineage-specific expansions (reviewed in Ref. 2). This strategy, which is particularly common among invertebrates (3–7) and plants (8, 9), is exemplified by the purple sea urchin (Strongylocentrotus purpuratus). Immune recognition in purple sea urchins is mediated by three types of PRRs: TLRs, NOD-like receptors (NLRs), and scavenger receptors containing scavenger receptor, cysteine-rich (SRCR) domains (6). Although functional work on these proteins remains limited, orthologs of immune receptors can be computationally identified in distantly related species using conserved domain architectures (e.g., TLRs are characterized by a series of leucine-rich repeats [LRRs], a transmembrane region, and a TIR domain that mediates signaling) (6, 10). For each of these receptor types, the S. purpuratus gene family consists of ∼10-fold more orthologs than homologous families in vertebrates or Drosophila. The ∼250 S. purpuratus TLRs can be classified into nine subfamilies based on sequence similarity (11). Genes within each subfamily exhibit distinct expression patterns across larval and adult tissues and in response to immune challenge. The highest expression levels are observed in tissues with immune functions (i.e., coelomocytes [circulating immune cells] and gut tissues), and transcript expression is not developmentally regulated, suggesting that, for the most part, these TLRs function within the immune response. Although the ligands for these receptors remain unknown, clusters of residues under positive evolutionary selection may form potential binding sites. Phylogenetic analyses reveal that the sea urchin TLRs genes are the product of a lineage-specific expansion and do not form clades with specific vertebrate TLRs (10). The diversity of the sea urchin TLRs is complemented by equally expanded gene families encoding NLRs and SRCRs. The S. purpuratus genome contains ∼200 NLR genes and ∼1100 SRCR domains. The NLR genes remain largely uncharacterized, but individual animals appear to express unique repertoires of SRCR transcripts in coelomocyte populations (12). Together, this expansive PRR repertoire likely provides the purple sea urchin with a broad capacity for immune recognition that is tightly linked with the more conserved aspects of the deuterostome immune response. Given the complex pattern of gene duplications/losses, long evolutionary times, and distinct selection pressures imposed on each species (e.g., pathogen encounters and environmental exposure), we suspect that even closely related species will have very different PRR repertoires. Analysis of genome sequences from several additional echinoderm species will allow this to be formally tested.

Vertebrate adaptive immune systems emerged in the context of pre-existing invertebrate innate immunity (Fig. 1). Although, in general, metazoan PRRs exhibit conserved protein domain structures and initiate conserved signaling pathways, lineage- and species-specific differences are apparent within the PRR repertoires even within the vertebrates. For example, exothermic vertebrates (cartilaginous fishes, bony fishes, and amphibians) generally have more complex TLR repertoires than endotherms. These include “fish-specific” and other novel TLRs, as well as the typical “mammalian” TLRs (reviewed in Ref. 13). Analysis of these sequences suggests that the ancestor of jawed vertebrates possessed a much larger TLR repertoire (∼19 genes) than most modern species (14). In addition, data suggest that the TLR repertoires within the extant jawed vertebrates are the result of lineage-specific diversification, differential gene loss, and gene duplication events. Relative to mammals, teleost fish genomes also exhibit considerable expansions within the gene families encoding NLR (15) and TRIM (16) proteins. In contrast, RLR repertoires are highly conserved across vertebrates (14).

Thus, although expansion of all PRR families occurs frequently in invertebrates, vertebrate PRR families appear to differ in their tolerance for expansion. We speculate that some vertebrate PRR families are evolutionarily constrained because they are tightly integrated with the adaptive arm of the immune response (e.g., TLRs), whereas PRR families with less influence on adaptive immunity are not similarly restricted (e.g., NLRs). Support for this hypothesis comes from the study of gadiform fishes, a group that includes Atlantic cod (Gadus morhua) and haddock (Melanogrammus aeglefinus). Analysis of ∼30 gadiform genome sequences revealed that this entire lineage lacks several molecules that were previously believed to be critical to vertebrate adaptive immunity (17, 18). This includes MHC class II α- and β-chains, invariant chain (CD74), and CD4. However, genomes of these fish species exhibit an expanded array of MHC class I genes and unique PRR repertoires, which may compensate for the absence of these hallmarks of adaptive immunity (18, 19). Based on the loss of several TLRs that recognize bacterial surface molecules (including TLR1/6, TLR2, and TLR5), as well as expansions of fish-specific NLRs and TLR families involved in nucleic acid recognition (TLR7, TLR8, TLR9, TLR22, and TLR25), the Atlantic cod immune system appears to rely heavily on intracellular nucleic acid detection to sense potential pathogens (19, 20).

Notably, for most nonmammalian PRRs, ligand specificities have not been experimentally confirmed. In most cases, orthology can be inferred based on conserved domain architectures, rather than the sometimes divergent primary sequences. Given that immune recognition receptors often evolve rapidly and the ligand binding domains are often subject to positive evolutionary selection (21), it cannot be assumed that PRRs share the recognition profiles of their mammalian orthologs. For example, although TLR4 orthologs are present in many bony fish and amphibians, these lineages are not particularly sensitive to Escherichia coli–derived LPS, which is the primary ligand for mammalian TLR4 (22). Instead, the TLR4-LPS signaling network appears to have evolved only after the divergence of amphibians (13, 14). Likewise, a recent study demonstrated that, despite broad structural similarities, LPS isolated from deep-sea, Gram-negative γ-proteobacteria were poor stimulators of the LPS-sensing PRRs from mice and humans (23). Thus, although much work remains in mapping the binding specificities of nonmammalian PRRs, this should provide a framework for understanding the frequently encountered pathogens in these species, as well as the selective pressures exerted on their immune systems.

In addition to their extensive PRR repertoires, purple sea urchins also express a diverse family of immune effectors known as SpTransformer (previously Sp185/333) (24). SpTransformer transcripts were originally identified in an expressed sequence tag library enriched in transcripts expressed by immune-stimulated coelomocytes. Aligning these sequences, which exhibited remarkable diversity (25, 26), required the insertion of large gaps, suggesting that this diversity might be generated by extensive alternative splicing. Surprisingly, however, analysis of the genes encoding SpTransformer transcripts revealed only two exons (27, 28). Instead, the observed transcript diversity is a consequence of a complex gene family in which the second exon is composed of shared sequence blocks known as “elements.” The S. purpuratus genome sequence contains only six SpTransformer genes, although other estimates suggest that this gene family may consist of up to 50 orthologs, and that copy number may vary among individuals (27). Additional evidence suggests that SpTransformer genes may also be subject to somatic diversification (29) and posttranscriptional modifications (30). The mechanisms that generate this diversity remain unknown, although the presence of a RAG1/2 cluster (31) and AID/APOBEC homologs with deaminase activity (32) within the S. purpuratus genome may provide some clues. SpTransformer proteins have an N-terminal glycine-rich region, a multimerization motif, and a histidine-rich region, yet lack predictable secondary structure. In the absence of ligands, the SpTransformer proteins are intrinsically disordered but, in the presence of microbes or derivative molecules (e.g., LPS), these proteins exhibit structural changes, transforming into ordered α helices or β strands (24). SpTransformer proteins are produced as membrane-bound and secreted forms, and they increase both phagocytic and bactericidal activity, possibly by interacting with the sea urchin complement system (6, 33). Importantly, sea urchin SpTransformer appears to be expressed in a clonal manner, with individual phagocytes producing only a single (identical or near-identical) transcript (34). Thus, assuming the membrane-displayed form of SpTransformer can trigger sufficient levels of intracellular signaling on interaction with Ag, through clustering, for example, then clonal selection could be operational in sea urchins.

Some arthropods employ an alternative strategy to somatically diversify a repertoire of proteins known as Down syndrome cell adhesion molecules (DSCAMs). Throughout Bilateria, orthologs of DSCAMs regulate axonal guidance during neuronal development (35). However, within insects and crustaceans, DSCAMs may also take on the additional role of immune protection (reviewed in Ref. 36). Dscam transcripts, which are expressed on hemocytes (circulating immune cells), are diversified through a process known as mutually exclusive alternative splicing. In Drosophila, each mature Dscam transcript consists of 20 exons. The dscam gene, however, contains a total of 115 exons, of which 95 are tandemly arrayed exon duplications (12 copies of exon 4; 48 of exon 6; 33 of exon 9, and 2 of exon 17) (35). As the transcript is processed, only one of each duplicated exon is retained in the mature transcript. This alternative splicing is stochastic, such that combinatorial diversity can result in >35,000 unique DSCAM proteins (35). DSCAM proteins consist of extracellular Ig and fibronectin domains (37) and are present in both transmembrane and soluble forms. The secreted form is generated via either proteolysis of the membrane-bound form or through alternative splicing that produces a transcript lacking the transmembrane domain (37). The latter resembles the alternate splicing that generates the cell-bound and soluble forms of IgM in jawed vertebrates. Soluble DSCAM constructs containing different combinations of exons exhibit differential binding specificities (e.g., see Ref. 38) and in many insect and crustacean species, the dscam transcript repertoire is highly dynamic in response to immune challenge (reviewed in Ref. 36). Finally, suppressing DSCAM production impairs pathogen clearance (37, 39). Together, the hypervariability of the dscam repertoire, presence of these transcripts in hemocytes, and that their expression patterns change in response to pathogen challenge indicate a role in the immune response. However, given the complex, and sometimes contradictory, nature of the data, the precise role of DSCAM in the immune response remains unclear (reviewed in Ref. 40). Further, given that splicing factors are generally not target specific, it is unknown how host cells would regulate the expression of specific DSCAM isoforms after immune challenge. Although the role of DSCAM proteins in pancrustacean immunity needs to be confirmed, this appears to be another example of a nonimmune protein being co-opted and diversified for use by immune cells.

Within the major vertebrate lineages, two parallel adaptive immune systems have evolved in which Ag receptors are somatically diversified. However, receptor structure and the mechanisms involved in repertoire diversification differ significantly between the jawed and jawless vertebrates.

The presence of adaptive immune mechanisms within jawless fishes (hagfish and lamprey) had long been suggested by experimental data showing the production of specific agglutinins after immunization with bacteria or foreign RBCs, as well as the rejection of second-set allogeneic skin grafts faster than first-set grafts (41, 42). However, genome and transcriptome searches failed to find genes with similarity to the jawed vertebrate immune receptors (i.e., MHC, TCR, or Ig), despite the presence of many other immune-relevant genes (43). The mystery was solved when Pancer and colleagues (44) identified cDNA clones containing tandem arrays of LRRs that were upregulated in lamprey lymphocyte-like cells after immune stimulation. These “variable lymphocyte receptor” (VLR) clones shared invariant 5′ and 3′ ends, but the sequence of individual LRR cassettes and the number of LRR cassettes present varied among clones. Subsequently, three VLR gene loci, VLRA, VLRB, and VLRC, were described (45, 46). Within the germline, VLR genes are incomplete, with arrays of LRR cassettes flanking each gene. As lymphocytes develop, these “donor” LRRs are incorporated into the incomplete VLR gene through a gene conversion-like process (44, 47). VLR gene rearrangement is mediated by two cytidine deaminases within the AID-APOBEC family (CDA1 and CDA2) (48). In fact, embryonic deletion of CDA2 results in the absence of mature VLRB genes (49). As with the jawed vertebrate Ig/TCR genes, successful rearrangement occurs at a single locus such that each lamprey lymphocyte expresses a single VLR transcript and undergoes clonal expansion in response to immune stimulation (44).

VLR genes are expressed by distinct lymphocyte subsets that can be distinguished by their immune gene expression profile (50). From a functional and transcriptional perspective, VLRB⁺ cells resemble the B cells of jawed vertebrates. VLRB⁺ cells dominate in the larval blood and kidneys and express orthologs of genes associated with B cells, including the transcription factor PAX5, the BCR-triggered tyrosine kinase SYK, and the adaptor protein BCAP (50). VLRB proteins are initially attached to the lymphocyte membrane by GPI linkage but, after Ag stimulation, are secreted as multimers composed of five disulfide-linked dimer pairs (functionally reminiscent of mammalian pentameric IgM). Further, lamprey C1q homologs can interact with secreted VLRB Abs to promote lysis of target cells, suggesting that a variant of the complement classical pathway exists in jawless fishes (51). Variability among VLRB molecules is concentrated within the concave surface that forms the Ag-binding pocket (52, 53); however, there is currently no evidence to support either somatic hypermutation or affinity maturation of the VLR repertoire.

In contrast, VLRA⁺ and VLRC⁺ cells parallel mammalian T cells (50). These cell lineages predominate in the larval thymoid (the lining of the lamprey pharynx and presumed thymus equivalent) and express homologs of genes present in gnathostome T cells (e.g., GATA2/3, NOTCH1, IL-17). Notably, expression of the γδ T cell–specific transcription factor SOX13 distinguishes VLRC⁺ from VLRA⁺ lymphocytes (54), suggesting that the two T cell lineages existed in the common vertebrate ancestor. Like TCRs, VLRA and VLRC are expressed exclusively as membrane-associated proteins. Direct binding of Ag to VLRA cells has not been demonstrated (50, 55), suggesting that VLRA (and possibly VLRC) may bind processed Ags. However, mechanisms for Ag processing/presentation have not been identified in either lampreys or hagfishes.

Many questions remain about the function of this “alternative” adaptive system in the jawless fishes, including: How is self-tolerance maintained? If (and where) does Ag-driven selection of the VLR repertoire occur? And finally, do jawless fishes exhibit immunological memory? However, it is apparent that the partitioning of humoral and cell-mediated adaptive immune functions, as well as the presence of three distinct lymphocyte lineages, occurred early in vertebrate evolution and before the emergence of the distinct Ag receptor systems.

Even within the TCR- and BCR-based adaptive immune system, there is surprising plasticity among jawed vertebrates in how these Ag receptors are deployed. For example, in addition to canonical TCRs, nonconventional TCR chains have been discovered in species as distantly related as cartilaginous fishes and nonplacental mammals. These nonconventional chains take three general forms. In the first form, the V domain is an Ig-TCR chimera, generated through trans-rearrangement of Ig/Ig-like V or V-D segments with the D-J or J segments of a canonical TCR chain (usually TCRδ). In the second form, the TCR chain is composed of two V domains: a membrane-distal Ig-like V domain linked to canonical TCR V and C domains. In cartilaginous fishes, the distal V domain is related to the novel, H chain–only isotype IgNAR, whereas the “supporting” domains are TCRδ-like. Both V domains in this “NAR-TCR” are the product of distinct VDJ rearrangement events (56). In the “TCRμ” chain of monotremes and marsupials, the membrane-distal V domain is IgM-like but also sits atop TCRδ-like V and C domains. The platypus TCRμ chain is doubly rearranging (as in the shark NAR-TCR), but in opossums, a nonrearranging, invariant TCRδ-like V domain is located between rearranging IgM V and TCRδ-like C domains (57, 58). Current data suggest that only the CDRs of the membrane-distal V domain interact with Ag; however, these chimeric molecules may bind soluble Ag to trigger signaling through the TCR machinery. The final noncanonical TCR form incorporates unique IgH-like V segments present within the TCRαδ locus that, so far, have been found in all jawed vertebrate lineages except teleost fishes and placental mammals (reviewed in Ref. 59). Interestingly, a few studies have now shown AID-mediated somatic mutation of TCR V regions, as well as Ig genes, in several species from across jawed vertebrate phylogeny (60–62). The role(s) these molecules play in immune protection has yet to be established. However, it is clear that the range of vertebrate Ig/TCR-based recognition molecules is much greater than would be anticipated from studies conducted solely in mice or humans.

Finally, the role of immunological diversity is not restricted to microbial recognition by Ag receptors. Immune systems also regulate allorecognition, that is, the capacity to differentiate self- from nonself-tissues. Allorecognition is central to two biological phenomena: mating, which ensures that gametes are fertilized by conspecifics and, in some cases, prevents self-mating (63, 64); and tissue fusion, which occurs when colonial invertebrates compete for space (65) or in the context of sexual parasitism among deep-sea anglerfishes (66). Evolutionary derivatives of allorecognition systems persist within the vertebrates as the well-characterized MHC loci, which plays important roles in both mate choice and defense (67, 68), as well as tissue compatibility (69). Allorecognition has been well studied in only a handful of organisms; we describe self/nonself-discrimination systems in a poriferan sponge, a hydroid cnidarian, a colonial tunicate, and vertebrate anglerfishes. The molecular mechanisms that mediate allorecognition rely on high levels of allelic polymorphism, RNA editing, and in some cases, loss of key factors in the immune system (70).

Sponges are remarkably amenable to fragmentation and fusion (reviewed in Refs. 71, 72) and have long served as model systems to understand invertebrate allorecognition. Data suggest that many sponges can discriminate self-tissues from those of another species and/or unrelated individuals. In fact, at certain developmental stages, some sponges can re-form after complete disaggregation to individual cells (73). The molecular mechanisms that mediate allorecognition likely vary among species in this ancient lineage. However, one gene family that has been implicated in allorecognition in the sponge Amphimedon queenslandica encodes proteoglycans known as aggregation factors (AFs). In this species, five genomically clustered AF genes encode secreted molecules consisting of tandemly repeated calx-β and von Willebrand domains (74). Notably, von Willebrand domains are also common in the proteins of the jawed vertebrate complement system. Analysis of 24 Poriferan genome sequences reveals that this family is restricted to Demosponges, and that gene copy number and sequence vary considerably among species. In A. queenslandica, AF transcripts are diversified in two ways. First, polymorphism is extremely high among AF alleles. This polymorphism is not random; nonsynonymous substitutions are significantly overrepresented. Second, comparison of gene and transcript sequences from individual animals reveals that the transcripts undergo significant RNA editing to increase sequence diversity. Although this system has been well characterized at the genomic and transcript levels, the mechanisms used by the AF proteins to mediate the fusion/rejection outcome remain unknown.

Hydractinia symbiolongicarpus is a colonial, sessile hydroid cnidarian that lives on hermit crab shells in intertidal marine environments. When two Hydractinia colonies on a substrate interact, they either fuse to form a single colony or undergo a rejection reaction in which nematocytes migrate to the interaction site and release cytotoxic compounds (75). In this way, allorecognition mechanisms ensure that animals avoid competing for space if they are sufficiently related, but also protect colonies from germline parasitism by circulating germ cells (76). This process is mediated by two proteins known as Allorecognition 1 (Alr1) and Alr2. These transmembrane proteins consist of two or three extracellular Ig domains. Homophilic binding between identical or nearly identical Alr1 or Alr2 alleles leads to colony fusion; in the absence of this interaction, a rejection reaction occurs (77). Analysis of gene sequences indicates that hundreds of highly polymorphic Alr1 and Alr2 alleles are present within natural populations of Hydractinia (78). Notably, in rare cases, rejection reactions occur among colonies with matching Alr1/Alr2 alleles, suggesting that additional uncharacterized proteins may also contribute to allorecognition in Hydractinia (70).

The capacity for discriminating between self and nonself is not limited to basal metazoans; one of the best studied model systems in allorecognition is the colonial tunicate Botryllus schlosseri (79). Work in this system also highlights the potential complexity of invertebrate allorecognition. Although the specific mechanisms that regulate fusion/rejection outcomes in Botryllus remain under debate, it is clear that these reactions are controlled by allelic variations in a series of linked genes: fuhc, fester, uncle fester, and Botryllus histocompatibility factor (BHF). By employing a forward genetic strategy to identify the locus controlling fusion or rejection, a highly polymorphic gene was identified. Known as fuhc, this gene encodes a transmembrane protein with two extracellular Ig domains (80). The fuhc alleles predicted allorecognition phenotypes in wild-type populations. A second polymorphic gene within the allorecognition locus, known as fester, is expressed as both secreted and transmembrane forms and is subject to extensive alternative splicing such that individuals express unique repertoires (81). The FuHC locus also contains a relatively conserved gene, known as uncle fester. Functional experiments indicate that uncle fester is required to initiate rejection reactions but is not involved in determining compatibility (82). The fester and uncle fester proteins both contain sushi domains (also known as complement control protein domains), which are common in vertebrate proteins that regulate complement activity and protect host cells from damage (83). Finally, analyses of recently available genome and transcriptome data from genetically defined B. schlosseri lines have identified a third polymorphic gene known as BHF (84). BHF is an intracellular protein that lacks recognizable domain structure. These genes, and others within the allorecognition locus, vary in their levels of polymorphisms, signatures of evolutionary selection, and the capacity to predict fusion/rejection outcomes (85). It remains unclear whether, or how, these proteins interact.

Finally, evolutionary pressures to enable tissue fusions between two individuals have resulted in the loss of nearly all adaptive immune mechanisms in some deep-sea ceratioid anglerfish species (Lophiiformes). The driving force behind these losses appears to be sexual selection; deep-sea anglerfishes are unusual among vertebrates in that they exhibit a form of sexual parasitism such that one or more tiny (6–10 mm) males attach, either temporarily or permanently, to the bodies of the much bigger female anglerfish (reviewed in Ref. 86). In some taxa, attachment is followed by full anatomical fusion, including connection of the two circulatory systems, to form self-fertilizing chimeras (87). How anglerfish achieve fusion without immunological rejection has long been a topic of speculation. However, recent analysis of the genomes of 10 anglerfish species revealed losses in the adaptive system of all species and gross deficiencies in the adaptive armament (including pseudogenization or deletion of aicda, rag1 and rag2, CD4, and CD8, plus multiple genes encoding TCR and BCR complex components) of those species that undergo permanent fusion of multiple males (66, 88). Thus, these anglerfishes have lost almost the entirety of their adaptive molecular repertoire and must rely exclusively on innate systems for immune protection. Whether the evolutionary advantage of sexual parasitism drove the observed changes in the adaptive immune system, or the loss of some components facilitated the evolution of sexual parasitism has yet to be resolved. However, given that anglerfish species that form transient attachments have an intermediate system, lacking only some adaptive immune components (66, 89), the latter currently appears most likely.

Immune systems are continuously shaped by evolutionary pressures from rapidly evolving microbes, driving substantial novelty in the mechanisms used to diversify the molecules involved in immunity. Just as hosts use molecular diversity to detect nonself in the form of microbial life, molecular diversity is also central to allorecognition. The outcomes of fertilization, tissue fusion, and sexual parasitism rely on complex systems that are subject to evolution. This Brief Review is not a comprehensive catalog of the approaches used to generate immunological diversity. Analyses of whole-transcriptome sequencing data suggest that RNA editing is much more frequent among invertebrates than previously believed (90–92); it would be surprising if this strategy was not used in the context of immune systems as well. In addition, prokaryotes rely on sophisticated mechanisms to protect themselves from phage infections (93, 94). We anticipate finding many more mechanisms for the creation of diversity as we continue to investigate immune protection in an ever-broadening phylogenetic range of organisms. It is already clear, however, that in immune systems, as with most things in life, diversity is the key to success and survival.

The authors have no financial conflicts of interest.

Institutional History

• Assistant Professor, Auburn University, Auburn, AL, 2019–current.

• Principal Research Biologist, Carnegie Mellon University, Pittsburgh, PA, 2018–2019.

• Assistant Research Professor, The George Washington University, Washington, DC, 2016–2018.

• Postdoctoral Fellow, University of Toronto, Toronto, ON, Canada, 2009–2016.

• Ph.D., Biological Sciences, The George Washington University, 2008.

• B.A., Biochemistry, Hood College, Frederick, MD, 2002.

Research Interests

• Comparative immunology

• Echinoderm immunity

• Evolution of immune systems

• Gene family evolution, gene regulatory networks

• Larval biology

I grew up in the monoculture of rural Pennsylvania. My parents were considered outcasts of the family for moving “across the river” – about 30 miles away. My childhood canon was filled with stories of my hardworking great-grandparents, who, in search of work during the Great Depression, left their child and traveled to California, opening a diner specializing in Pennsylvania-style potpie (a crustless delicacy). They did not care for California. Although my classmates were deeply religious in varying shades of Protestantism, my parents’ beliefs were rooted in the church of capitalism as they worked to establish a small machine shop, leaving my sister and me to fend for ourselves. Although we never felt “without,” looking back, I realize that our imaginary worlds and toys were largely sheets, sticks, and pieces of scrap metal scrounged from the shop. We learned to interpret the world by exploring and asking questions. The two of us made a fantastic team: my fearless curiosity would get us into trouble; her diplomacy and instincts would get us out.

In college, I studied biochemistry for what my practical family considered to be the most absurd of reasons: I loved it. To my family’s dismay, I then embarked on the even less apparently lucrative field of evolutionary biology and, eventually, academia.

I remain bewildered that this Pennsylvania girl has now traveled the world as a scientist. My family still opposes federal funding for basic science at every opportunity, despite the fact that my career now depends on it. The truth, though, is that the inadvertent lessons of my childhood remain my advice for aspiring scientists: study what you love, trust your instincts, and, most importantly, ask all the questions.

Katherine M. Buckley, Ph.D.

Assistant Professor, Auburn University

Institutional History

• Assistant Professor, University of Maryland School of Medicine, Baltimore, MD, 2016–current.

• Honorary Senior Lecturer, University of Aberdeen, UK, 2016–current.

• Senior Lecturer, University of Aberdeen, UK, 2015–2016.

• Lecturer, University of Aberdeen, UK, 2012–2015.

• Principal Research Scientist, Wyeth LLC/Pfizer Inc., 2008–2011.

• Research Associate, University of Maryland School of Medicine, 2007–2008.

• Postdoctoral Fellow, University of Maryland School of Medicine, 2002–2007.

• Scientist, Auvation Ltd., Aberdeen, UK, 2001–2002.

• Ph.D. Antibody Engineering, University of Aberdeen, UK, 2001.

• M.Sc. Antibody Engineering, University of Aberdeen, UK, 1998.

• B.Sc. (Hons) Genetics, University of Aberdeen, UK, 1996.

Research Interests

• Comparative immunology

• Cartilaginous fish immunity

• Evolution of immune molecules and networks

• Adaptive immunity

• Immunoglobulins

Growing up in Liverpool during the 1980s, the only scientists I “knew” were on the nature documentaries and science programs I watched with my dad. Although my parents were always very supportive and encouraged my curiosity, it did not occur to any of us that a career as a scientist was something a nerdy girl from a working-class household could achieve. Given my own experience, I am always mindful that there are many others who are also excited by some aspect of science but do not realize a career in STEM is a possibility for them or are unsure how to take their first steps along that path. Knowing this, I regularly participate in “careers in STEM” and “young women in STEM” Q&A sessions, aiming to demystify STEM career paths. My team also hosts pop-up science stalls regularly in public places, such as shopping malls, gaming conventions, or arts festivals, aiming to chat with as broad an audience as possible about both science and STEM careers. However, as a proud “First Gen,” I am also aware that recruiting a diverse population of scientists is only the first step. Those of us who have already established ourselves in STEM need to remain mindful that our students and mentees have extremely diverse life experiences and will advance an environment in which novel questions and ideas are encouraged. Most importantly, we must foster a more supportive environment for everyone if anyone is to thrive.

Helen Dooley, Ph.D.

Assistant Professor, University of Maryland School of Medicine