Thomas Boehm, Jorge Carneiro, Antonio Coutinho (organisateur), Maria Curotto-Lafaille, Jocelyne Demengeot, Sidonia Fagarasan, Leszek Ignatowicz, Ludger Klein, Juan Jose Lafaille, Nicole Le Douarin (organisateur), Diane Mathis, Helia Neves, Hiroko Ohki-Hamazaki, Alexander Y. Rudensky, Shimon Sakaguchi, Josselyne Salaun, Harry W. Schroeder, Harald Von Boehmer, Martin Weigert
Compte rendu
The adaptive immune system of the jawed vertebrates emerged about 500 million years ago involving, as a major novelty, the somatic generation of diversity of the antigen receptors by V(D)J recombination. This process, which likely results from a successful co-option of the mutagenic transposition machinery of an ancestral transposon, confers the organisms an unprecedented capacity to generate random molecular specificities that are employed in pathogen recognition and memory. Using this system, vertebrates can keep up with the evolutionary pace of infectious microorganisms, but, as a drawback, they become liable to autoimmune diseases, since very often the randomly generated antigen-receptors recognize the organism’s own components. Thus, together with generation of diversity, vertebrates evolved robust mechanisms of tolerance by which they avoid pathologic autoimmunity. The symposium gathered researchers from around the world to discuss and advance the understanding of tolerance as an essential developmental process of the jawed vertebrate organisms.
Nicole Le Douarin overviewed her groundbreaking work on tolerance induction using xenogeneic and allogeneic embryonic chimeras. Substituting chicken embryonic tissues by the same tissue from quail embryo resulted in chimeric birds that developed and hatched normally, but the grafted tissues would eventually be rejected upon the maturation of the immune system. An exception to this rule was the grafting of embryonic tissues that gave rise to thymic epithelium. Not only this quail thymic epithelium was not rejected after hatching, but the lymphocytes that matured therein became tolerant to other quail tissues. Similar tolerance induction was achieved in mice where grafts of allogeneic embryonic pharingeal pouches grafted onto athymic mice led to atopic thymuses where an immune competent T cell repertoire matured to be tolerant to both donor and recipient tissues. Furthermore, tolerant T cells generated in allogeneic thymic epithelium were shown to transfer tolerance to tissue grafts from the thymic epithelium donor, when adoptively transferred to otherwise immunocompetent, responding animals.
Figure 1 – Chick embryos grafted with quail limb buds develop wings but a few weeks after hatching the tissues of quail origin are infiltrated and rejected by thymic derived T-cells. The rejection of the xenogenic tissues is not observed if concomitantly with limb buds the chick embryo is grafted with quail tissues that will give rise to thymic epithelia where chicken T-cells develop and become tolerant to any quail tissue.
Le Douarin’s work demonstrated the key role of thymic epithelium in establishing tolerance to non-thymic tissues, and together with other seminal studies, namely by Shimon Sakaguchi, Antonio Coutinho, and Juan Laffaille, were instrumental in demonstrating the role of regulatory T cells in tolerance.
The thymus is the organ where T cells, including regulatory T cells, are generated and quality-controlled after random rearrangement of their TCR. The symposium proceeded with a thorough discussion on the development, evolution, and function of this organ. In embryo, the thymus originates, together with the parathyroid, from a common primordium that develops from pharyngeal pouches. Helia Neves addressed the specification of thymic epithelium by Foxn1 on localized territory within the pouches. She showed that somatopleural mesenchyme is necessary for the expression of Foxn1 in the pouches, through a mechanism that requires Bmp signalling, and is modulated by Noggin and Sonic hedgehog. According to her results, the thymic epithelium and the parathyroid are specified by a “French flag”-type mechanism based on inverse gradients of Bmp and Noggin, in turn triggered by Sonic hedgehog. The territory where Bmp expression predominates anticipates the thymus, while the territory dominated by Noggin correlates with future parathyroid location. This model was offered as an interpretation for the fact that in sharks all pouches except the first can form thymic tissues but in birds and mammals the origin of the thymus is restricted to fewer pouches, based on the fine tuning of these Bmp/Noggin gradients.
The evolution of the adaptive immune defense of the jawed vertebrates involved the co-option of a transposon of the transib family to generate diversity of the antigen receptors. The successful co-option of this transposon was accompanied by several morphological novelties, including the thymus, where a specific environment is offered for T cell differentiation and quality control of the somatically recombined T-cell receptors. How did the vertebrate thymus evolve to become the organ that quality controls T cells? Was it a true novelty or the result of rewiring of an ancient genetic network ? Thomas Boehm reported the reconstruction of the genetic changes in chordates and vertebrates by comparative genomics of several thymus specific genes with roles in the specification of thymic epithelium, the homing of lymphocyte precursors to the thymus, and the specification of these progenitors to the T cell lineage. Ancestral gene networks, involving Foxn1-like and Delta-like genes, are present and define specific territories of the pharyngeal epithelium of jawless fish, but there is no colonization of this tissue by lymphoid cells. In cartilagineous fish these ancestral networks are extended to control and turn on chemokine genes in prospective thymic epithelium, while chemokine receptor networks co-evolved in the lymphoid lineage. These complementary modifications allow lymphocyte precursors to home to the thymus using chemokine signaling, where they commit to the T cell lineage based on notch signaling. Considering that the transib transposon is present in many non-vertebrate genomes in an inactive form and that the MHC-peptide system might have evolved by modifying peptide carriers of the vomeronasal system (that perform interindividual discrimination in fish) to intraindividual discrimination, it was concluded that in a proto-jawed vertebrate everything was ready to accommodate and benefit from the transib mutagenic potential to the novel function of immune defense.
The lymphoid precursors that colonize the thymus differentiate to mature cells and are positively and negatively selected by interactions with thymic epithelial cells (TECs) and resident or migratory dendritic cells (DC). Interactions with these stromal cells are critically dependent on the recognition, by the newly formed TCR, of the MHC-peptide complexes they present. The establishment of tolerance requires the expression of peptides derived from body antigens by thymic epithelia, which is promoted by the AIRE transcription factor. Animals or humans deficient in the AIRE gene have a disturbed T-cell repertoire and develop a multiorgan autoimmune syndrome. Diane Mathis addressed the molecular biology of AIRE and its role in tolerance induction. She showed that AIRE reads the histone code to upregulate the transcripts of thousands of genes including so-called peripheral tissue antigens (i.e. genes that are otherwise expressed in unique tissues) through an intriguing mechanism that does not change the levels of unspliced pre-RNAs. Then, based on a doxycycline-regulated transgene that allows to control temporally the TEC expression of AIRE, she showed that it is necessary and sufficient to express AIRE in the perinatal period to prevent the multiorgan autoimmunity. AIRE can be shut down a few weeks after birth and remain off for long periods, without major deleterious effects.
Ludger Klein explored further the specialization of TECs to the function of selecting the CD4 T cell repertoire and inducing tolerance. He showed that TECs exhibit a high rate of constitutive starvation-independent autophagy, and that abrogation of autophagy specifically in TECs by target disruption of Atg5 (autophagy-related gene 5), a component necessary for autophagosome formation, leads to significant disturbances in the selection of particular MHC-II-restricted T-cell specificities. The alteration in the repertoire ultimately results in fatal colitis and multi-organ inflammation.
In the thymus, concomitantly with positive and negative selection of the T cell repertoire, there is the differentiation of regulatory T cells from precursors upon recognition of agonist MHC-II-peptide ligands on stromal cells, mostly TECs. Jocelyne Demengeot presented novel data indicating that TEC do crosspresentation of extra-thymic antigens, using as read-out the capacity of TEC to drive regulatory T cell differentiation in TCR-transgenics. She reported that antigen specific regulatory T cell differentiation was observed in an animal model, in which the nominal antigen was exclusively expressed by MHC class II deficient hematopoietic cells, while MHC itself was expressed only by TEC. This extends the spectra of mechanism used by TEC for antigen-presentation, but leads to the intriguing possibility that the antigens that are not encoded in the genome of the organism, may also drive the generation of specific regulatory T cells.
Shimon Sakaguchi discussed the role of TCR signaling strength in repertoire selection and tolerance using SKG mice. These mice carry a mutation in the SH2 domain of Zap-70, a hub in the TCR signaling pathway. Sakaguchi presented several lines of evidence indicating that there is a shift of repertoire selection in these mice, due to attenuated TCR signaling. Mice carrying the SKG mutation in the Balb/c background develop spontaneous arthritis dependent on IL-17 producing cells. Depletion of regulatory T cells in these mutants led to spectra of autoimmune diseases different from the one obtained by depletion of regulatory T cells in wild type mice. In mice models that spontaneously develop autoimmune gastritis or diabetes the introduction of the SKG mutations prevents the disease. Both conventional and regulatory T cells are affected by the SKG mutation, and mutant regulatory T cells are less suppressive than wild type in vitro.
Juan Lafaille asked how much does the intra-thymic development of regulatory T cells depend on the TCR specificity? Multiple TCR-transgenic mouse lines were constructed to express TCR encoded by genes cloned from peripheral Foxp3+ T cells of normal adult animals. The surprising observation in these TCR-transgenic mice was that the numbers of regulatory Foxp3+ T cells generated in the thymus was systematically very small, often below the detection levels of normal flow cytometry. A more thorough and quantitative examination of the relationship between the number of precursors and the output of Foxp3+ T cells indicated that the output increased with the number of precursors until it leveled off at a plateau. These results, together with similar results reported by Alexander Rudensky, were interpreted as evidence of the limited capacity of the TCR-dependent intrathymic niches that can drive the upregulation of Foxp3+ by thymocytes.
Development in the thymus is not the only stage in which T cells can differentiate from precursors cells. Regulatory Foxp3+ T cells can also be generated extra-thymically under appropriate conditions including TCR stimulation and TGFbeta signaling. Harald von Boehmer discussed the conditions for intra- and extra-thymic induction of tolerance, paying special attention to corresponding lineages of regulatory T cells. He discussed the role of demethylation of a specific conserved element in the Foxp3 locus in the stabilisation Foxp3 expression in thymic derived, but not in peripherally induced, regulatory T cells.
Leszek Ignatowicz addressed the selection of regulatory and conventional T cells, by comparing their respective repertoires in terms of global properties such average autoreactivity, diversity, and common usage of variable region sequences. As a model system, he utilized genetically manipulated mice that can only generate a very limited T cell repertoire, and where positive selection was mediated by a single peptide covalently bound to the MHC molecule. Hybridomas derived from regulatory and conventional T cells from these animals did not respond to the selecting antigen when stimulated in vitro, but they responded to unrelated foreign antigens. Sequencing hundreds of variable regions from these hybridomas indicated that the structural repertoires of both regulatory and conventional T cells were comparable. Based on the findings it was concluded that, at least in these animals with limited diversity, the regulatory T cell repertoire is not especially biased towards the organism’s own antigens.
Jorge Carneiro addressed conventional and regulatory T cell population dynamics in the periphery. Based on mathematical models, he derived quantitative predictions about the repertoires of these two populations, drawing a formal analogy between developmental patterning by positional information and the patterning of T cell repertoires in TCR-specificity space. He then went on to critically examine the statistical inference of repertoire properties such as diversity and overlap between two repertoires based on sequencing of TCR variants. After dismissing conventional methods that provide inaccurate and unreliable estimates of diversity, he proposed instead the use of more rigorous and more informative Poisson-abundance models (PAM) based on which one can additionally infer clonal size distributions.
Regulatory Foxp3+ T cells prevent autoimmunity as well as the immunopathology triggered by exuberant immune responses. Removal of these T cells, by disruption of the Foxp3 gene, leads to a cytokine storm and fatal multi-organ autoimmunity. Alexander Rudensky reported two examples, in which the selective ablation of transcription factors characteristic of a given class of T helper cells in regulatory Foxp3+ T cells, abrogates the capacity of the latter to suppress the corresponding class of responses in vivo. Animals in which the Irf4 gene, a transcription factor involved in Th2 responses, is disrupted in cells expressing Fox3, developed a selective deregulation of Th2 responses, characterized by massive germinal centers and hyper IgE. Similarly, mice in which Stat3, a transcription factor required from Th17 cell differentiation, was selectively depleted in regulatory Foxp3+ T cells, succumb to Th17-dependent intestinal inflammation. Thus, regulatory T cells need to express transcription factors characteristic of their T cell targets in order to mingle with them and inhibit their responses.
While T-cells develop and are quality controlled in the thymus, B cells, the sister population of lymphocytes, develop and are selected in the bone marrow of mice and humans. In an attempt to dissect the selective forces shaping the repertoire of B cells, Harry Schroeder examined how structural properties of the sequences encoding the antibody variable regions changed along different stages of differentiation, to conclude that there are biases and regularities in the CDR3H loop in terms of length, aminoacid usage, and average hydrophobicity. Furthermore, there are preferential biases in D segments and reading frames. Do these biases in repertoire generation have any evolutionary and physiological meaning? Genetically forcing the use of the alternative reading frames that are normally disfavored by selection in the bone marrow leads to significant changes in the antibody responses to haptens and viruses. From these observations he concluded that it can be hazardous to have some naturally disfavored sequences, and therefore that the biases in repertoire could be under selective pressure.
Martin Weigert addressed the mechanism of receptor-editing according to which B-lymphocytes with high-affinity autoreactive antigen-receptors escape deletion when they turn-on further rearrangement of their antigen receptor genes. He focused in particular in antibodies that bind DNA examining the molecular details of the interaction. Antibody binding to the highly negative charged DNA molecule is mainly based on arginine residues, and addition of arginines can increase affinity in a stepwise manner. However, the arginine side chain can be stabilized by hydrogen bonding to aspartate in the same antibody, thus preventing DNA binding. Weigert discussed the functional implications of antibodies containing multiple arginines and aspartates that are zipped together by hydrogen bonding. The affinity and even the specificity of these antibodies can evolve by somatic mutation in a dramatically discontinuous way, since disruption of any of these pair can radically change its potential binding partners.
Mucosa are the main interfaces between the body and the external world and, accordingly, they are immunologically special places. The gut represents one of such interfaces and a number of specific molecules and cell types are specifically localized there, being involved in maintaining the intestinal microflora under control. IgA antibodies are key players in this relationship with the microflora, as discussed by Sidonia Fagarasan. Mice deficient in AID, a molecule required for the differentiation of B cells into IgA producing B cells, display a marked shift in intestinal flora composition with an over-representation of segmented filamentous bacteria, which drives an enlargement of Peyer patches. Based on adoptive transfers of selective CD4 T cell subsets from the spleen and lymph nodes, Fagarasan concluded that the formation of germinal centers, where differentiation to IgA producing cells takes place, depends on former CD4 Foxp3+ T cells that lost Foxp3 expression and acquired a normal follicular helper cell phenotype. Other CD4 subsets, such as Foxp3- or RORgt+ Th17 cells are not able to provide such T cell help in Peyer patches.
The airways and lungs represent the other major interface with the external milieux where specialized immune regulatory mechanisms are called into action. Maria Coroto-Lafaille addressed the mechanisms that regulate allergic inflammation induced via the respiratory route. Using different transgenic animal models that differ in their ability to generate regulatory T cells, due to competency or deficiency in the Foxp3 gene, she provided evidence for the existence of complementary mechanisms for control of chronic allergic inflammation, that are either dependent or independent of Foxp3+ regulatory T cells. Foxp3+ regulatory T cells are required to control inflammation and to inhibit IL-4 production and mucosal lymphoid neogenesis, while Foxp3-independent process are sufficient to prevent eosinophilia.
Hiroko Ohki used avian embryonic chimeras to investigate the extent of tolerance induction and sexual differentiation in male/female chimeras. Avian sex determination is based on W and Z chromosome system, in which males are ZZ and females ZW. In this system, estrogen encoded in W chromosome leads to de-masculinization of the otherwise default male. Substituting the brain primordium in a recipient by that of a donor embryo leads to chimeric animals where several tissues, namely the brain, comb, beak, skin, and cartilage on face are of donor origin (fig.2a). When the donor is female and the recipient male, the female tissues are rejected upon sexual maturation (fig. 2b). In contrast, male to female grafts survive and mature to adulthood. This allows to ask: does the chimera with a male brain in a female body display a behavior characteristic of the wild type male or female chicken? Studying multiple behavioral traits that distinguish males from females, namely stepping, pecking, and safety posture, Ohki concluded that chimeric female recipients of a male brain behave mostly as a normal hen, despite some morphological changes in the grafted tissues (fig.2c) and slight delay in sexual maturity and egg laying rates. A similar conclusion was reached by studying female brain into male chimeras under heavy immunosuppressive treatments.
Figure 2- Tolerance induction in female/male brain embryonic chimeras. a) grafting of male (female) brain primordia in female (male) embryos generates chimeric animals where the brain and other local tissues are of different origin than the remaining tissues including the thymus. b) In female (WZ) to male (WW) chimeras the tissues of female origin are rejected upon sexual maturation, illustrated by the loss of comb, beak and skin coloration. c) In male (WW) to female (WZ) chimeras, the tissues of male origin are tolerated, as illustrated by the comb, which nevertheless does not reach the normal size in the male animal.
Antonio Coutinho, in the closing lecture of the symposium, discussed diseases of self-nonself discrimination, paying special attention to human primary immunodeficiencies that are very often associated with autoimmune diseases. Analysing the spectra of autoimmune manifestations such diseases as such as IPEX (Foxp3 deficiency), APECED (AIRE deficiency), ALPS (Fas signalling deficiency) and others, he concluded on the existence of recurrent patterns that demand explanations.
Overall, the symposium was very stimulating and thought provoking. While examining the mechanisms of tolerance in such depth and breadth, it was unavoidable to realize how the concept of tolerance evolved during the last two decades, from the simple notion of deleting autoreactive lymphocytes to a biologically fascinating regulatory process established in development but operating for the entire life span of the organism. As Harald von Boehmer pointed out, it is hard to find solid evidence for instances of tolerance induction that can be exclusively explained by deletional mechanisms, where a role for regulatory T cells is clearly ruled out. Many of the key researchers in driving this major shift in Immunology were present in this symposium, and shared the expectation that regulatory T cells will make their way into the clinical setting successfully. This optimism notwithstanding, it was also evident that there is still a long way to go before we reach a full understanding of tolerance and the fault-lines that lead to autoimmunity and allergy. More knowledge has to be acquired on the origin and plasticity of conventional and regulatory T cells, as well as on their cross-regulation and homeostasis. A significant number of “apparently established facts” in the field of regulatory T cells are likely to be revised, as some of the experiments demonstrating the plasticity of conventional and regulatory T cells need to be reexamined under the light of the fact that Foxp3+ T cells might be present in very low frequencies, normally undetected by conventional flow cytometry approaches. Major advances in the near future are expected from the characterization of the recurrent patterns in the manifestations of autoimmune diseases in mice and humans, and the constraints on T cell repertoire selection processes that underpin them.