Evolution and immunity

Abstract

This report describes a meeting organized by Ken Smith and Jim Kaufman, entitled Evolution and Immunity, which took place at the University of Cambridge on 24 September 2009 to honour the anniversaries of the birth of Darwin and the first publication of The Origin of Species. Ten internationally-known speakers described the effects of evolution on immunity, ranging in timescales from the deep-time evolution of adaptive immune systems in vertebrates and invertebrates to the evolution of pathogens and lymphocytes within a single individual. The final talk explored the application of phylogenetic analysis to non-biological systems.

Charles Darwin was born in 1809, and he published The Origin of Species in 1859,¹ when he was 50 years old. The year just past was thus a double anniversary, and the occasion for a 1-day meeting held at Peterhouse and the University of Cambridge on 24 September 2009. This meeting, entitled Evolution and Immunity, was the latest of the annual series now organized by Cambridge Immunology together with the British Society for Immunology.

Evolution of immunity happens at many timescales, from the deep time evolution of whole systems, to the evolution of variants of pathogens and individuals over the lifetime of a host species, to the changes in frequency of lymphocyte clones within an individual during an infection. It would have been impossible to do justice to even one of these vast subjects in a single day, so a few speakers were invited to touch on some of the highlights. The talks were held in the beautiful auditorium at Peterhouse, followed by a reception at the Fitzwilliam Museum during which the participants could enjoy the ongoing exhibition Endless Forms: Charles Darwin, Natural Science and the Visual Arts, followed by the conference dinner in the great hall of Christ’s College, where Darwin was a student.

It was long thought that adaptive immunity was a property of jawed vertebrates alone, so it has been a major revolution to find that some invertebrates have their own adaptive immune systems, based on molecules other than antibodies (Abs), T-cell receptors (TCRs) and major histocompatibility complex (MHC) molecules. The first glimpse of this reality was provided by work on snails,^2,3 which failed to shatter the old paradigm. Among other indications, the inability to find a multigene family of immunoglobulin variable (V) receptors in the genome sequence of the fruit fly Drosophila melanogaster confirmed the prevailing view that invertebrates relied only on innate immunity.

Therefore, it was particularly apt for Dietmar Schmucker (Harvard University, Boston, USA) to describe the properties of Down’s syndrome cell-associated molecule (DSCAM) in insects and crustaceans, in which a single gene generates thousands of variant isoforms by alternative splicing.^4–6 DSCAM in humans is expressed in the nervous system, and in fruit flies many variants are involved in development of the nervous system, including specifying axonal connections between neurons. However, insect and crustacean DSCAM variants are also expressed in cells involved in immunity, including haemocytes and the fat body, and are secreted into the haemolymph. Knockdown of fruit fly DSCAM in haemocytes decreases phagocytosis, and transient silencing in the mosquito Anopheles gambiae leads to decreased resistance to bacterial pathogens and malarial parasites. The DSCAM molecule contains eight immunoglobulin V domains arranged in an S-shape, and the variability clusters into two sites in the structure, one that specifies homophilic (and limited heterodimeric) pairing and one that may contribute to specific antigen binding. In the nervous system, homophilic interaction of DSCAM expressed on some neurons leads to repulsion.

At the base of the vertebrates are the jawless fish, lampreys and hagfish, which have yet another adaptive immune system,^7–9 described by Max Cooper (Emory University, Atlanta, USA). Lamprey lymphocytes could be identified by flow cytometry, and immunization with bacteria, foreign erythrocytes or plant mitogens led to the appearance of lymphoblasts. Large-scale sequencing of cDNAs from such lymphocytes resulted in no obvious Ab, TCR or MHC sequences, but many different leucine-rich repeats [LRRs, as are found in Toll-like receptors (TLRs)] were discovered and named variable lymphocyte receptors (VLRs). Analysis of genomic DNA revealed a few VLR genes in the haploid genome to which LRR sequences are added by a copy choice mechanism from LRR pseudogenes, resulting in upwards of 10¹⁴ different sequences with evidence for both allelic exclusion and clonal expression. The resulting proteins can be anchored in the cell membrane by glycolipid linkage, or may be secreted as pentamers of dimers. These pentamers have low affinity but high avidity for antigen, and X-ray structural analysis shows antigen bound ‘as in the palm of a cupped hand’.

Two kinds of VLR genes were found, VLR-A and VLR-B, for which monoclonal antibodies (mAbs) defined two kinds of lymphocytes with different distributions in lamprey tissues. Both kinds of lymphocytes respond to antigens; VLR-B cells bind antigen and secrete VLRs, while VLR-A cells do not but instead respond to a phytohaemagglutinin, a T-cell mitogen. Gene expression analysis showed many differences between these two kinds of lymphocytes, including cytidine deaminases, transcription factors, kinases, cell surface proteins, cytokines, chemokines and TLRs. Overall, the data suggest that the VLR-A lymphocytes may be like T cells and the VLR-B lymphocytes like B cells, leading immediately to the question of whether lymphocytes specialized into functionally different subsets independently in jawless fish and jawed vertebrates, or only once, after which different receptor molecules evolved (or persisted) in jawless fish and in the jawed vertebrates.

Coming to the adaptive immune system of jawed vertebrates, Jim Kaufman (University of Cambridge, UK) reviewed some of the data and speculations concerning the emergence of the key molecular components: Ab, TCR and MHC molecules.^10,11 It is generally accepted that the key event occurred at the base of the jawed vertebrates some 500 million years ago: the creation of the first split V gene from the insertion of a transposon into the immunoglobulin V exon of what is presumed to have been a membrane-bound receptor, which used recombination-activating genes (RAGs) to specify imprecise joining of the different parts of the split V gene. It has been argued that this first receptor was a TCR-like molecule, based on the fact that diversity is generated by recombinational joining only at complementarity-determining region 3 (CDR3), which is the major contact with peptide in MHC molecules. In this view, αβ T cells emerged first, followed by γδ T cells and then by B cells. The first MHC molecule was probably a class II-like homodimer, the gene for which duplicated and diverged to create a heterodimer, from which the first class I heavy chain and β₂-microglobulin genes could be generated by an inversion.

Many if not most non-mammalian vertebrates share the salient features first found in the chicken MHC, with MHC genes located close to polymorphic antigen-processing genes to allow co-evolution as stable haplotypes. The familiar organization of the human MHC probably arose from an inversion which put the class III region in between the class I and class II regions, leaving the tapasin, transporter associated with antigen presentation and inducible proteasome genes in the class II region. The organization of the primordial MHC allowed the ancestors of the MHC and antigen-processing genes to co-evolve from their original functions an antigen presentation pathway. Two rounds of genome-wide duplication at the base of the vertebrates led to MHC paralogous regions, including the CD1 region and arguably the regions containing natural killer (NK) receptor genes in humans. The discovery of NK receptor and CD1 genes in the chicken MHC was the first of many examples, including a TCR-like gene in the frog MHC, which suggest that the primordial MHC contained the genes that co-evolved into receptor–ligand pairs. In this view, the primordial MHC was the birthplace of the adaptive immune system, which has been breaking apart ever since.

A crucial property for an adaptive immune system with variable antigen-specific receptors is self-tolerance, part of which occurs in the thymus during development of T cells. Thomas Boehm (Max-Planck-Institute for Immunobiology, Freiburg, Germany) described the minimal gene networks which may have been necessary for the appearance of the T cells developing through the thymus.^12,13 Taking advantage of the transparency of zebrafish (Danio rerio) embryos and the ease of knock-downs, a chemokine system was identified for homing of thymic lymphoid precursors: zebrafish chemokine (C-C motif) ligand 25 (CCL25) (equivalent to human tyrosine kinase expressed in hepatocellular carcinoma) and zebrafish chemokine (C-X-C motif) ligand 12 (CXCL12) (equivalent to human CXCR12, or stromal cell derived factor 1). Both these genes were found throughout the jawed vertebrates, but only the equivalent of CCL25 was found in the jawless fish lamprey (Lampetra planeri), which lacks an obvious thymus, and neither was found in the protochordate amphioxis (Branchiostoma floridae), which lacks lymphocytes altogether. Using FoxN1 knock-out mice with adoptive transfer of transgenic cells, reconstitution of the thymic niche was shown to depend on Notch, delta-like ligand 4, kit ligand, CCL25 and CXCL12 under the control of the FoxN1 gene. Of these five genes, only the delta-like ligand gene is present in lamprey and amphioxis but still under control of the equivalent of FoxN1, suggesting that the stroma began to evolve the regulatory gene network for the formation of the thymus prior to the emergence of T cells.

On a shorter timescale, host species evolve adaptations to their pathogens, and different individuals may respond appropriately for protection, under-respond allowing infection, or over-respond leading to pathogenesis. Rick Maizels (Edinburgh University, UK) discussed the precarious balance between protective responses to multicellular parasites and pathogenic responses to harmless allergens or to self antigens.^14,15 Although at least 2·3 billion people (of approximately 6·9 billion) are infected with just five species of helminth worms, the majority escape severe pathology, with the parasites regulated by T helper type 2 (Th2) and regulatory T (Treg) cells. Many studies show an inverse correlation between the level of such parasites in a population and the prevalence of allergy, asthma and autoimmune diseases such as multiple sclerosis, a concept called the ‘hygiene hypothesis’. One example of how this can work is the reduction of lung allergy and autoimmunity in experimental allergic encephalitis (EAE) produced by infection of the mouse intestine with the nematode worm Heligmosomoides polygyrus, apparently through the parasite secreting a molecule(s) that expands Treg (FoxP3-positive) cells through binding the transforming growth factor (TGF)-β receptor. Overall, the evidence suggests that hosts evolved with strong selective pressure from parasites, leading to some individuals with an immune response that protects from parasites but leads to pathological responses in the absence of parasites.

This theme was continued by Ken Smith (University of Cambridge, UK) who described the importance of an antibody receptor called FcγRIIB in the balance between susceptibility to systemic lupus erythromatosis (SLE) and protection from malaria.^16,17 The FcγRIIB locus is one of the stronger susceptibility loci among the 40 found by genome-wide association (GWA) studies, which overall explain 25% of the genetic variation of SLE. Copy number variation (CNV) of the FcγRIIB gene is important, with knock-outs in mice leading to autoimmunity and with three or more copies leading to a reduction in SLE but also a reduction in the response to pathogens such as bacteria. Single nucleotide polymorphisms (SNPs) in FcγRIIB genes can also be important, with the I232T mutation in the transmembrane region abolishing function, apparently by blocking entry into lipid rafts in the cell membrane. The population distribution of T232 world-wide correlates both with susceptibility to SLE and with the geographic distribution of malaria, and the frequency of severe malaria in T232 homozygotes is half that of heterozygotes or I232 homozygotes. The balance between the inhibitory FcγRIIB molecule and the activatory FcγRIIIB molecule determines, at least in part, the immune response, with less inhibition apparently driven by the advantage of increased survival from malaria.

Like hosts, pathogens also evolve on several timescales, and Gordon Dougan (Sanger Institute, Hinxton, UK) described the evolution of some bacterial pathogens.^18–20 An emerging theme is the existence of groups of ‘promiscuous species’ infecting a variety of different host species using generalist strategies and differing in 1–2% of their DNA, which occasionally give rise to single species, clades or serovars adapted to a particular host with a specialist strategy and differing only in SNPs or indels. Examples include Salmonella, Pertussis, Yersinia, Mycobacteria and many others. For instance, Salmonella enteritidis infects the intestines of many hosts, resulting in inflammation and eventual clearance, but gave rise to Salmonella typhi, which infects only humans using a stealth approach which allows systemic infection, leading in some cases to eventual death. Complete genomic sequences of thousands of S. typhi isolates revealed only 2000 SNPs, a similar level of diversity as in humans, consonant with a jump to infecting humans about 30 000 years ago. New haplotypes can be formed by loss of genes, and there are examples of replacement of local haplotypes by ‘superfit’ strains, but now the major epidemics in South Africa are ‘non-typhoidal Salmonella’ (NTS) as a result of replacement by Salmonella typhimurium. Another kind of adaptation is illustrated by Clostridium difficile, different isolates of which have 2–4% difference in DNA sequence, apparently having crossed into human populations from several sources. In comparison, the ‘super-fit’ 027 ribotype has only 30 SNPs across the world. The emergence of this medical problem is presumably a result of the use of antibiotics, as C. difficile colonizes the intestine of humans after antibiotic treatment has changed the normal spectrum of microbiota, allowing such people to become highly infectious ‘super-shedders’.

Many pathogens can evolve over very short timescales, and Andrew McMichael (Oxford University, UK) discussed the evolution of human immunodeficiency virus (HIV) within individuals.^21–23 An HIV inoculum is generally very complex (10⁴–10⁶ viral sequences), but most individuals are initially infected by only a single virus (at least by the early peak of viraemia). The characteristics of this virus and the individual can determine the course of the infection, in particular the viral set point, the level of viraemia after the early peak. Many host genes contribute to resistance and susceptibility, and the virus evolves in response. As just one example of many, MHC class I genes present viral peptides to CD8-bearing cytotoxic T lymphocytes (CTLs) which can kill infected cells, with particular class I alleles [such as human leucocyte antigen (HLA)-B57] leading to low set points and greatly enhanced survival. As one example of the counter-response, the viral protein Nef down-regulates class I molecules. Whatever the class I molecule, the T-cell response destroys the infected cells, with selection eventually resulting in a viral sequence that has lost the relevant T-cell epitope. Viruses even escape highly protective MHC alleles like HLA-B57, but more slowly and with a clear fitness cost to the virus. The new viral sequence will in turn select for new CTLs while the old CTLs disappear, an evolutionary process inside a single individual that continues until the immune response is apparently exhausted.

Such evolution occurs constantly inside each individual of the jawed vertebrates, as clones of lymphocytes with different antigen-specific variable receptors wax and wane. Michael Neuberger (University of Cambridge, UK) described this process for the somatic evolution of B cells.^24–26. In mammals, the transition from high-avidity immunoglobulin M (IgM) to high-affinity IgG is accomplished by cycles of mutation and selection within germinal centres (GC) of lymph nodes. In the process, diversity is spread to the periphery of the CDRs, aided by the presence of somatic hypermutation (SHM) hotspots. The SHM process is triggered by a member of the apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like (APOBEC) family called activation-induced deaminase (AID), which is a B-cell-specific enzyme and thereby drives somatic mutation of Abs but not TCRs. AID is also the critical trigger of Ab diversity produced by gene conversion in birds, Ab class switching (first apparent in amphibians), somatic mutation (first apparent in jawed fish) and, quite probably, segmental recombination of LRR sequences in VLRs of jawless fish. The presence of such genes in jawless fish, cytidine deaminase (CDA) 1 in VLR-A cells and CDA2 in VLR-B cells, highlights the age of this gene family. Other members of the APOBEC family have appeared subsequently, but are involved in RNA editing and resistance to retroviruses, including HIV. The similarity of the structure of AID/APOBEC proteins to bacterial enzymes involved in nucleotide metabolism and in transfer RNA (tRNA) editing suggests a very ancient origin indeed.

Having touched in this meeting on some highlights of the importance of evolution to every level of immunity, Chris Howe (University of Cambridge, UK) reminded the audience that the only figure in The Origin of Species was a phylogenetic tree, a method that became a powerful workhorse of evolutionary analysis. This approach can be extended beyond the investigation of life to the analysis of manuscripts, language and even carpet designs.²⁷ We can only imagine what Darwin would have thought of the day’s presentations, but hopefully he would have appreciated how far we have come since his day and understood how far we have yet to go.