Human leukocyte antigen (HLA) class II genes are found in the three most polymorphic loci, and are associated with more diseases than any other region of the genome [1], [2]. The primary function of HLA class II proteins is to present immunogenic peptides to CD4+ T cells and each has a series of binding pockets that include distinct residues in polymorphic positions, which serve to create allele-specific peptide binding preferences [3] (Figure 1A, 1B, and 1C). These preferences define the set of peptides that are available for T cell selection in the thymus and activation of CD4+ T cells by antigen presenting cells (APCs) in the periphery [4]. As such, the unique motifs of HLA class II molecules play a fundamental role in shaping the CD4+ T cell-mediated immune responses in each individual. Indeed, experimental evidence reveals HLA-specific differences in T cell receptor distribution, indicating that HLA genotype shapes the T cell repertoire [5]. For T cell-mediated autoimmune diseases in particular, an overwhelming proportion of the genetic risk can be assigned to susceptible HLA class II haplotypes [6]. Therefore, it can be surmised that HLA class II proteins associated with susceptibility to autoimmune disease contribute to this susceptibility by facilitating the selection of a potentially autoreactive T cell repertoire.

The immune system, whose primary aim is to combat infectious diseases for the benefit of the host, with the secondary aim of preventing malignancies via immune surveillance, is self-referential [7]. This means that tolerance to self is acquired first in the thymus for T lymphocytes (and in the bone marrow and elsewhere for B lymphocytes) from gestation through the neonatal period, and on into to adulthood [8]. After birth, the periphery acquires an important role in tolerance maintenance, complementing that of the thymus. Thus, T and B cells recognizing components of self with high affinity should be eliminated through deletion, while those recognizing such components with intermediate affinity are controlled through regulation (known in earlier eras as suppression, Figure 2). The remaining surviving thymocytes should possess T cell receptor (TCR) molecules that recognize self-peptide-MHCI or self-peptide-MHCII combinations with low affinity and are selected, respectively, into mature CD8+ and CD4+ T cells. A sizeable fraction of the thymocytes formed prior to selection cannot recognize any pMHCI/II complex with appreciable affinity. These consequently die within the thymus by neglect. The selected and combinatorially-formed TCR repertoires (out of a truly astronomic scale of variety, ca. 1018 different TCRαβ sequences) should react with any component of an infectious or noxious agent that may enter the host. As a consequence of successful negative selection, their respective affinities for “foreign” non-self-MHCI/II combinations should be higher than for combinations containing the self-peptides against which the cognate TCRs were originally selected. Hence, the potency of immune responses toward foreign antigens (Figure 3).

The degeneracy of the thymic selection process can be seen as follows (Figure 3): assuming, 15 possibilities in any of the TCR-potential contact residues (to allow for steric hindrance between bulky residues and other structural constraints), and 3 possibilities for each of the anchoring positions, yields (15)8 × (3)5 = 6.228 × 1011 different peptide combinations. The number of possible self-peptide combinations (with 23,000 protein-coding genes and an average of 10 strongly-, 10 intermediate-, and 20 weakly-binding epitopes per coded polypeptide, an upper-limit average) we end up with 9.2 × 105 self-epitopes / HLA II molecule, a number almost 6 orders of magnitude lower than the conservatively-estimated number that can be theoretically accommodated by the given HLA II molecule. Post-translational modifications (PTMs) can increase the number of self-epitopes according to their frequency of occurrence; however, according to current data, the number of self-epitopes will still be several orders of magnitude lower than the total number of TCRαβ combinations [8]. The thymic selection process will further reduce this degeneracy, but it will, nevertheless be substantial great [10], [11].

The discriminatory power of TCRs, even though their affinity for pMHC is in the μM range, is exquisite [12]. The pattern-recognition receptors (Toll-like receptors) that recognize inherent regular structural features of bacteria, viruses, and other parasites form a credible first line of defense to such agents [13], but in a conventional response this first wave of innate responders is supplemented and then supplanted by the adaptive responses of T and B cells. Autoimmunity may arise when CD4+ T lymphocytes bearing TCRs that are generally weakly reactive to self-peptide-HLAII complexes (pHLAII) are activated. In turn, such activated CD4+ T effector cells inappropriately help B cells to produce relevant autoantibodies, and cytotoxic CD8+ T cells to attack vulnerable tissues, resulting in full-fledged autoimmune disease [14].

Published work indicates that epitope sequences that arise as a consequence of PTMs comprise an important facet of the self-reactive immune responses that underlie autoimmunity [15], [16]. The phenomenon of neo-epitope presentation and recognition is well established in celiac disease (CD) and rheumatoid arthritis (RA), for which the recognition of deaminated and citrullinated (deiminated) peptides, respectively, have established importance in disease pathogenesis [17]. However, there is an increasing appreciation that modified neo-epitopes play a role in other T cell-mediated autoimmune diseases such as type 1 diabetes (T1D). This brief review will cover the structural features of autoimmune-associated HLA class II proteins, which converge to create peptide binding motifs that accommodate post-translationally modified neo-epitopes. Such neo-epitopes can be recognized with higher affinity than their wild type counterparts and are likely to be less subject to the tolerance mechanisms that limit autoimmunity. The biological functions of PTMs while well known in many instances, remain a mystery in many others. Over 45 years ago, a review found some 140 different covalent modifications of the 20 natural amino acids in proteins [18]. The transient phosphorylation of select Ser/Thr or Tyr residues in signal transducing or receptor tyrosyl-kinases are some of the first well-known examples. By contrast, while transglutamination as a means of forming isopeptide bonds or converting glutamine to glutamate residues in peptides/proteins is well established, and the respective enzyme (especially isoform 2, tissue transglutaminase 2; tTG2) has been implicated in many biological processes, yet the precise in vivo purpose of the reaction remains elusive [19]. This enzyme is targeted by autoantibodies in CD [20]. Likewise, arginyl deimination is catalyzed by several isozymes, with some having suspected developmental and tissue-specific roles [21]. Peptidyl arginine deiminase IV (PAD4) is found primarily in monocytes and neutrophils [22], [23], but expression has also been reported in the neonatal mouse thymus [24]. Detailed studies of PAD4-specific thymocytes have generated fascinating insights about pMHC-II/TCR complex orientations and dwell times that lead to regulatory T cell (Treg) formation or selection of conventional T cells versus negative selection [25], [26]. From the point of view of protection from horror autotoxicus [27], a balanced occurrence of PTMs within the thymus, lymph nodes, peripheral non-lymphoid tissues, and body tissues (within which the modification first occurs), would be required in order to have self-tolerance in place. However, it remains to be established whether the PAD4 enzyme is sufficiently active to carry out the deimination of argynyl residues within the thymus (thereby contributing to central tolerance) or within peripheral lymphoid tissues (thereby contributing to peripheral tolerance) toward to-be-modified self-proteins. Likewise, tissue transglutaminase seems to increase in activity during thymic involution, yet its participation in PTM of tissue-specific antigens has not been fully explored [28].