Summary Sessile colonial invertebrates—animals such as sponges, corals, bryozoans, and ascidians—can distinguish between their own tissues and those of conspecifics upon contact [1]. This ability, called allorecognition, mediates spatial competition and can prevent stem cell parasitism by ensuring that colonies only fuse with self or close kin. In every taxon studied to date, allorecognition is controlled by one or more highly polymorphic genes [2–8]. However, in no case is it understood how the proteins encoded by these genes discriminate self from non-self. In the cnidarian Hydractinia symbiolongicarpus, allorecognition is controlled by at least two highly polymorphic allorecognition genes, Alr1 and Alr2 [3, 5, 9–12]. Sequence variation at each gene predicts allorecognition in laboratory strains such that colonies reject if they do not share a common allele at either locus, fuse temporarily if they share an allele at only one locus, or fuse permanently if they share an allele at both genes [5, 9]. Here, we show that the gene products of Alr1 and Alr2 (Alr1 and Alr2) are self-ligands with extraordinary specificity. Using an in vitro cell aggregation assay, we found that Alr1 and Alr2 bind to themselves homophilically across opposing cell membranes. For both proteins, each isoform bound only to itself or to an isoform of nearly identical sequence. These results provide a mechanistic explanation for the exquisite specificity of Hydractinia allorecognition. Our results also indicate that hydroids have evolved a molecular strategy of self-recognition that is unique among characterized allorecognition systems within and outside invertebrates.

Bony fish are among the first vertebrates to possess an innate and adaptive immune system. In these species, the kidney has a dual function: filtering solutes similar to mammals and acting as a lymphoid organ responsible for hematopoiesis and antigen processing. Recent studies have shown that the mammalian kidney has an extensive network of mononuclear phagocytes, whose function is not fully understood. Here, we employed two-photon intravital microscopy of fluorescent reporter mice to demonstrate that renal dendritic cells encase the microvasculature in the cortex, extend dendrites into the peritubular capillaries, and sample the blood for antigen. We utilized a mouse model of systemic bacterial infection as well as immune complexes to demonstrate antigen uptake by renal dendritic cells. As a consequence, renal dendritic cells mediated T-cell migration into the kidney in an antigen-dependent manner in the setting of bacterial infection. Thus, renal dendritic cells may be uniquely positioned to play an important role not only in surveillance of systemic infection but also in local infection and autoimmunity.

We have previously shown that CX3CR1+ renal DC sample antigen via intravascular processes in the context of systemic infection and immune complex disease. Here, we investigated the role of intravascular DC processes in antigen presentation and effector T cell migration to the kidney. To investigate T cell migration, CFP-OVA E. coli and CD8+ DsRed+ OT-I effector T cells, pretreated with Pertussis toxin (PTX), were i.v. injected into CX3CR1gfp/+ mice 24h before 2P intravital microscopy (2PIM) of the kidney was performed. OT-I T cells migrated into the kidney, with 2PIM showing prolonged (30 min) DC-T cell interactions. OT-I migration was significantly higher (2-fold) when OVA was present (CFP-OVA E. coli vs WT E.coli), demonstrating that migration of effector T cells was antigen-specific. To test if antigen presentation by DC was necessary to mediate migration of T cells, we used CD11c-YFP Kb−/− bone marrow chimeric recipients (hematopoietic cells lack H-2k(b), parenchymal cells express H-2k(b)). CFP-OVA E. coli injection into a CD11c YFP Kb−/− bone marrow chimeric mouse abrogated effector OT-I T cell migration into the renal tissue. Our data demonstrate that migration of effector T cells in the setting of systemic infection is antigen-specific. In addition, antigen presentation by hematopoietic cells, most likely DC, is necessary to mediate antigen-specific effector T cell migration. H-2k(b) expression on parenchymal and endothelial cells does not mediate antigen-specific effector T cell migration. Our findings show that renal DC have two roles: immune surveillance by sampling and presenting antigen via intravascular processes, and guiding effector T cells to the site of antigen.

The existence of sophisticated self/non-self systems is not limited to vertebrates. Several colonial marine invertebrates have evolved systems that distinguish between conspecifics via cell-cell contact. This phenomenon, known as allorecognition, occurs when colonies encounter each other as they grow across their substrate. Compatible colonies fuse or co-exist, while incompatible colonies reject and often aggressively compete for space. In all taxa studied to date, allorecognition phenotypes are determined by highly polymorphic loci, which ensure that colonies are only compatible with themselves or close kin. How allorecognition molecules achieve this specificity remains unknown. Allorecognition in Hydractinia is controlled by at least two genes, alr1 and alr2, which encode highly polymorphic transmembrane proteins similar to immunoglobulin superfamily molecules. Colonies with matching alleles at alr1 and alr2 fuse, while colonies with no matching alleles reject. Using in vitro assays with recombinant proteins, we demonstrate that alr1 alleles bind to themselves but not to other alr1 alleles. We suggest that, in vivo, compatibility between colonies is also determined by allele-specific homophilic binding of alr proteins. Since fusion is rare in nature and single populations contain hundreds of unique alr alleles, the Hydractinia allorecognition system appears to be based on a biophysical mechanism with unprecedented allelic diversity and specificity.