Degeneracy and Complexity in the Immune System
From Santa Fe Institute Events Wiki
The purpose of the workshop is to discuss the increasing realization (just a decade old) that lymphocyte recognition is degenerate. Even more recent is the further knowledge that degeneracy is a ubiquitous property of biological systems, appearing in gene networks, neural networks and other levels of biologic organization. As Edelman and Gally have pointed out, degeneracy is almost invariably accompanied by complexity. In contemporary immunological usage it describes the ability of a single receptor to recognize and react to a heterogeneous assortment of ligands, with each lymphocyte having its own pattern of degeneracy (i.e. its own set of ligands that stimulate or inhibit it). Such a degenerate system is quite adaptable and it can be argued that evolution itself depends heavily on extensive degeneracy in a set of complex systems. Most of the crucial receptor elements in the immune system, aside from lymphocyte antigen receptors, are also degenerate.
It is time to reevaluate and consider the role of degeneracy in the immune system as one of its most characteristic and distinctive attributes, rather than a peculiar and occasional oddity. The immune system, even more than the nervous system, functions to receive and process molecular information. It is the major thrust of the workshop on this topic to question how the system, which cannot rely on a unique specificity of its receptors at the molecular level can behave with such marvelous specificity at the operational level.
Allen, Paul firstname.lastname@example.org
Celada, Franco email@example.com, firstname.lastname@example.org
Cohen, Irun Irun.Cohen@Weizmann.ac.il
DeBoer, Rob R.J.deBoer@bio.uu.nl
Eisen, Herman email@example.com
Forrest, Stephanie firstname.lastname@example.org
Garcia, Chris email@example.com
Goldstein, Byron firstname.lastname@example.org
Greenspan, Ralph email@example.com
Hafler, David firstname.lastname@example.org
Hlavacek, Bill email@example.com
Hodgkin, Philip firstname.lastname@example.org
Huseby, Erik email@example.com
Krakauer, David firstname.lastname@example.org
Nemazee, David email@example.com
Parnes, Ohad firstname.lastname@example.org
Perelson, Alan email@example.com
Pinilla, Clemencia firstname.lastname@example.org
Sercarz, Eli email@example.com
Strong, Roland firstname.lastname@example.org
Wucherpfennig, Kai email@example.com
DRAFT WORKSHOP AGENDA
WEDNESDAY, JULY 5
Participants arrive in Santa Fe.
THURSDAY, JULY 6
Session I Chairperson: Irun Cohen
8:00AM Van transportation from Hotel Santa Fe
8:15AM Continental Breakfast
9:00 AM Alan Perelson Clonal Selection and the Need for Immune System Degeneracy: A Shape-Space View
9:45 AM Irun Cohen The Degeneracy Problem and the Immunological Homunculus
10:30-11:00 AM Coffee Break
11:00 AM Rob deBoer Specificity and Self Complexity in the Immune System
11:45 AM Ohad Parnes Title
12:30 – 1:45 PM Lunch Break
Session II Chairperson: Eli Sercarz
1:45 PM Ralph Greenspan Degeneracy and Flexibility in Gene Networks
2:30 PM David Krakauer An Informal Review of the Evolution of Degeneracy in Signaling Systems
3:00 – 3:30 PM Break
3:30 PM Clemencia Pinilla Positional Scanning Combinatorial libraries: What Do They Reveal About Degeneracy of T cell Specificity?
4:30 – 5:00 PM General Discussion – Chairs: Parnes and DeBoer
5:30 PM Dinner at the Coyote Café, Santa Fe
FRIDAY, JULY 7
Session III Chairperson: K.C. Garcia
8:00AM Van transportation from Hotel Santa Fe
8:15AM Continental Breakfast
9:00 AM Herman Eisen Variety of Peptide-MHC Complexes Recognized by a Single Antigen Specific T Cell Receptor (the 2C TCR)
9:45 AM Kai Wucherphennig TCR Topology and Crossreactivity
10:30 – 11:00 Coffee Break
11:00 AM K.C. Garcia Structural Insights into Degeneracy by T cell and Shared Cytokine Receptors: Do Recognition Codes Exist?
11:45 AM Roland Strong Non-standard Molecular Mechanisms Enabling Degenerate Recognition in Innate Immunoreceptors: NKG2D and Siderocalin
12:30 – 1:45 PM Lunch
Session IV Chairperson: Alan Perelson
1:45 PM Phil Hodgkin An antigen valence-based explanation for the evolution and organization of the humoral immune response
2:30 PM David Nemazee Antibody degeneracy and B cell tolerance
3:15 – 3:45 PM Break
3:45 PM Byron Goldstein Kinetic Proofreading, Serial Engagement and Combinatorial Complexity Give Rise to Degeneracy in Ligand Recognition
4:30 – 5:30 PM General Discussion – Chairs: Nemazee and Strong
5:30 PM Opera buffs depart for the Santa Fe Opera in their own cars (Preview buffet dinner begins at 6:30 PM, The Magic Flute at 9:00 PM)
SATURDAY, JULY 8
Session V Chairperson: David Hafler
8:30AM Van transportation from Hotel Santa Fe
8:45AM Continental Breakfast
9:30AM Paul Allen Degeneracy and Specificity in the T Cell Recognition of Foreign- and Allo-Antigens
10:15AM Erik Huseby How Thymic Selection Establishes Specificity at the TCR – MHC + Peptide Interface
11:00 – 11:30 AM Coffee Break
11:30 AM – 12:15 PM Franco Celada Split Crossreativity Against Viruses: Modeling Clonal Dynamics During Heterologous Infection
12:15 – 1:45 PM Lunch
Session VI Chairperson: Franco Celada
1:45 – 2:30 PM David Hafler Endogenous Self-Peptide/MHC Complexes Select for Cross-Reactive T Cell Receptors
3:30 – 3:15 PM Eli Sercarz An Explanation for the Rarity of Cross-Reactivity by Molecular Mimicry
3:15 – 3:45 PM Break
3:45 – 5:15 PM Wrap-up Discussion
WORKSHOP PRESENTATIONS - TITLES, SUMMARIES & DOWNLOADABLE PRESENTATIONS
THURSDAY, JULY 7
Title: Clonal Selection and the Need for Immune System Degeneracy: A Shape-Space View
I will present a set of theoretical arguments to show the need for degeneracy in immune recognition based on the idea that immune recognition receptors and their ligands have complementary shapes. In order for the immune system to have a complete or nearly complete repertoire employing a finite number of receptors requires receptor degeneracy. The theory, while simple, makes a number of quantitative predictions that may be of interest.
Title: The Degeneracy Problem and the Immunological Homunculus
Despite the inherent ligand degeneracy of the adaptive immune receptors, the immune system learns to apply to body maintenance and to body defense the appropriate inflammatory programs. This learning is accomplished with the aid of a congenital "learning set" natural autoimmunity positively selected in utero to certain self-molecules. The primordial immunological homunculus is this learning set.
Simple mathematical models predict that the specificity of lymphocytes is determined by the number of self antigens. We enumerate the number of self antigens for CD8+ T cells by bioinformatic analysis of the human genome, and confirm that this is in agreement with the specificity predicted by the models. We will review other data that T cells are quite specific.
The extensive interactivity among genes that is now being revealed suggests that there is considerable flexibility in the genome’s capacity for responding effectively to diverse conditions. In model gene networks affecting behavioral phenotypes in Drosophila, a high degree of flexibility has been observed and the interactions underlying the various states of the network have been analyzed. The phenotypic space of the network's functionality will be discussed in terms of degeneracy and complexity.
Title: Genetics and Language
T cell recognition studies using positional scanning combinatorial libraries have revealed both the degeneracy and specificity of T cell receptors. Positional scanning libraries are unbiased and systematically arranged collections of billions to trillions of peptide ligands. A number of T cell ligands for CD4+ and CD8+ clones of known and unknown specificity have been identified using these libraries. The recognition of mixtures composed of billions of compounds and its successful use for the identification of individual peptides illustrates the co-existence and balance of degeneracy and T cell specificity.
FRIDAY, JULY 7
CD8+ T cells that express the 2C TCR have been previously shown to recognize various peptides associated with several class I MHC proteins (Ld, Kb, Kbm3, a non-classical MHC-I). In H2b mice that express the 2C TCR and no other TCR, many of the 2C TCR+ cells are CD8+, others are CD4+, and still others are CD8-CD4- ("double negative or DN). The identity of the MHC required for the positive and negative selection of the CD8, CD4, and DN 2C cells reveals that the range of pepMHC recognized by this single TCR is much greater than previously realized.
Title: TCR Topology and Crossreactivity
We have determined the crystal structure of a human TCR isolated from a patient with relapsing-remitting MS bound to the HLA-DR2/MBP peptide complex. The structure showed a highly unusual binding topology with a shift of the TCR towards the N-terminus of the peptide. Only two TCR loops made contacts to peptide side chains, placing physical limits on the specificity of this TCR for peptide recognition. Several microbial peptides have been identified that only have obvious sequence similarity in the peptide segment contacted by these two TCR loops.
Title: Structural Insights into Degeneracy by T cell and Shared Cytokine Receptors: Do Recognition Codes Exist?
My talk will focus on recent results from my laboratory exploring the structural basis of cross-reactivity by two classes of degenerate receptors: T cell receptors and the shared cytokine receptors gp130 and common gamma chain. We will examine the structural evidence for and against the idea of a conserved recognition code underlying the ability of these receptors to recognize diverse ligands
The dominantly-activating, homodimeric immunoreceptor NKG2D is broadly expressed on a variety of cell types (NK, alphabeta and gammadelta T cells) and recognizes a variety of polymorphic MHC class I homologs, including MIC-A/B, ULBPs and Rae-1s. Rather than induced-fit or predominantly hydrophobic interfaces, NKG2D-ligand recognition, highly degenerate in specificity, is mediated by an alternate mechanism termed 'rigid adaptation'. Siderocalin is an anti-bacterial protein that sequesters iron as ferric siderophore complexes. Though the binding site is remarkably rigid, Siderocalin tolerates wide variation in the two distinct families of siderophores bound, broadening its utility.
Antibody recognition of antigen follows the general rule that the better the fit the longer the time of association and the higher the affinity. Therefore, for B cells there can be no absolute answer to the question - how many of the universe of epitopes can be bound by a single antibody? The answer can only be given relative to the degree of complementarity, or affinity, used to define binding. The question of interest we must ask, therefore, is what complementarity level between antibody and antigen offers protection against pathogens? I have attempted to explore this question by manipulating an artificial universe of antigenic and BCR surfaces for which probabilities of reactivity at various affinities can be calculated. By creating an artificial BCR repertoire and linking it to the need for B cells to proliferate, differentiate and accumulate antibody in fluids the importance of time is emphasized.
An appropriate way to look at the difficulties faced by the humoral immune response is to ask this question: "What is the quickest way to reach humoral protection against an invader?" The answer is markedly affected by antigen structure. Many highly multivalent antigens drawn from the artificial universe can be protected by 'natural' preexisting multivalent antibody. Multivalent antigens that are missed by natural antibody can be rapidly eliminated (ie. 3-5 days) by antibody production from a large fraction of low affinity cells. This manner of response patterns the T-independent B cell activation. The response to mono- or paucivalent antigens takes many days and therefore is the slowest outcome. Clearly memory is only required and advantageous for the latter group of antigens. This slow response that leads to memory is similar to T-dependent B cell responses. Thus we can suggest that 'time to protection' and antigen valence have both served as powerful selection forces for the evolution and coexistence of natural, T-independent and T-dependent antibody production mechanisms.
A large fraction of the preselected B cell antigen receptor repertoire is autoreactive and must be regulated by tolerance. Receptor editing is a mechanism that promotes tolerance with relatively little cell attrition. Evidence demonstrating the prevalence, affinity threshold and efficiency of editing will be presented, as will evidence for distinct thresholds for tolerance in different peripheral B cell compartments.
I review the role of kinetic proofreading, serial engagement and combinatorial complexity in signaling mediated by immune recognition receptors and discuss how these features of signaling contribute to the receptor's ability to trigger responses to more than one ligand.
SATURDAY, JULY 8
Title: Degeneracy and Specificity in the T Cell Recognition of Foreign- and Allo-Antigens
We generated and characterized a series of high affinity TCRs, and observed that a TCR can exhibit both local degeneracy and global specificity. Through a novel approach to study the recognition of peptides in alloreactivity, we found that alloreactive T cells are comprised of two populations: one highly peptide specific, and a second in which the MHC dominates the recognition.
Title: How Thymic Selection Establishes Specificity at the TCR – MHC + Peptide Interface
TCRs bind MHC + peptide complexes, with a range of affinities which are relatively low (1-200?M). This poses the question, how is specificity established at such a low affinity interface? I will present data showing that thymic development selects for T cells bearing TCRs for which a large number of side chains of the peptide amino acids are “hot spots” for binding. In addition, data will be presented indicating that a large portion of TCR specificity is maintained by the inability to tolerate interface disruptive amino acid side chains within or near the TCR/MHC + peptide binding site. These “hot spots of disruption” occur at positions on the MHC as well as on the peptide.
Title: Split Crossreativity Against Viruses: Modeling Clonal Dynamics During Heterologous Infection
Split crossreactivity is the case where a challenging antigen is recognized by T-cell memory but results foreign to the humoral response. This situation is common in successive viral infections in mice and has momentous consequences on the clonal make-up and stability of memory, as discovered by virologists of Worcester, MA. In collaboration with Liisa Selin these novel events have been simulated by our group of theoretical Immunologists in Genoa and New York using the agent based model IMMSIM. I will show some results and perspectives of these studies with regard to private specificity, memory repertoire formation and clonal attrition, dominance and reduction. I will also propose to make use of split crossreactivity in vaccination, and , eventually to find a less derogatory term to indicate degeneracy, the fundamental property that makes possible life on earth and the immune defence.
The paradox of autoreactivity to self-peptides in physiologic as opposed to pathologic immune responses is not well understood. Here, we directly examined the T cell response in a series of healthy subjects to endogenous self-peptides. CFSE labeled T cells were stimulated with antigen presenting cells loaded with endogenous self-antigen and the resulting CD4 populations entering into cell cycle (CFSElow) or non-proliferating CD4+ cells (CFSEhigh) were single cell sorted and screened against a panel of self-antigens and microbial memory antigens to interrogate their antigen reactivity. The precursor frequency of CD4 T cells entering cell cycle in response to self-peptide/MHC was calculated to be ~0.04%. Generation of CD4 cells with stimulation by endogenous self-peptides was dependent upon CD28 costimulation. Clones derived from the CFSElow population exhibited significantly more cross-reactivity to multiple antigens, which was class II MHC restricted, than clones derived from both the CFSEhigh population and CD4+ clones generated after foreign antigen stimulation. Sequencing the TCRβ chains indicated that CFSElow clones were oligoclonal, similar to what was observed in antigen driven immune responses. These data demonstrate that T cell clones generated with stimulation by endogenous self-peptide have more degenerate, cross-reactive T cell receptors that most likely recognize MHC determinants. We speculate that lack of cross-reactivity by these degenerate T cell receptors may be due to stearic hindrances by peptides loaded into MHC that block T cell receptor recognition of MHC.
In response to an infectious microbe, the dominant determinants that activate the microbe-specific repertoire are only occasionally able to also stimulate self-reactive cells. This larger microbe-specific repertoire competes successfully with the self-reactive clones that can cross-recognize the dominant determinant on the microbe. This relationship insures a response focused on the invasive microbe rather than on self-determinants.