APLs and Oranges: Induction of T Cell Anergy by Altered Peptide Ligands Free

A long-lived and still-prevalent view of TCR signaling is that it is digital: an encounter with an appropriate peptide/MHC ligand either leads to a full T cell response or none at all. Failure might arise from insufficient TCR engagement, leading to the unraveling of a nascent signaling cascade, or the action of inhibitory pathways that thwart TCR activation. However, evidence that TCR stimulation is in fact more versatile has been in the literature for decades.

Pivotal studies in the early 1990s found evidence that some amino acid substitutions in the peptide component of the peptide/MHC complex could produce ligands that occupy the TCR without inducing a functional response, that is, ligands that behaved as TCR antagonists (1). Other altered peptide ligands (APLs) were defined that could induce some but not all responses by mature T cells (e.g., cytokine production but not proliferation) (2). Pharmacologically, such ligands could be classified as partial agonists, capable of limited induction of the TCR signaling cascade, which terminates before full “commitment” of the T cell to the complete panoply of functional responses. This and similar work led to the conclusion that there is a hierarchy of T cell effector functions, the induction of which depends on the quantitative or qualitative “degree” of TCR stimulation.

One of the most surprising and intriguing cases of such partial activation was defined in the current Pillars of Immunology article, published by Allen and colleagues (3) in 1993. Previous studies discovered that encounter with some APLs drives induction of the nonfunctional state called anergy in T cells. Anergy had been characterized as the consequence of TCR activation in the absence of suitable costimulation, that is, cells receiving “signal 1” but not “signal 2” (4, 5), with far-reaching implications for ways in which costimulatory (and coinhibitory) molecules regulate T cell function. In their short report (immunologists of a certain age may recall, nostalgically, a time when groundbreaking papers could consist of less than a dozen simple graphs, with no supplementary material), Sloan-Lancaster et al. (3) showed that nonresponsiveness could also be driven by encounter with subtly altered peptide/MHC ligands, despite the availability of costimulatory cues. Allen and colleagues studied mouse CD4⁺ T cell clones reactive to peptides derived from an allele of hemoglobin β-chain presented by MHC class II molecules. Some conservative substitutions in the peptide produced ligands that could stimulate early features of T cell activation (such as elevated expression of CD25 and LFA-1) but were unable to provoke proliferation or cytokine production. Stimulation by such APLs induced a long-lasting state of nonresponsiveness to a subsequent encounter with APCs bearing the normal hemoglobin β-chain peptide. APL-induced anergy could be prevented by blocking the NFAT pathway with cyclosporine A, in keeping with the idea that APL encounter drives partial activation of the TCR signaling cascade, leading to anergy.

Subsequent studies have shown that partial agonist and antagonist APLs typically interact with their matching TCRs with lower affinity than the original peptide/MHC ligand (6, 7). However, more recently, it has become apparent that there is a range in the biophysical “quality” of the TCR–peptide/MHC bond that dictates functionality, a feature that has been exploited for enhanced tumor immunotherapy (8). An unresolved issue is whether APLs truly induce a qualitatively distinct TCR signal from strong agonists: certain studies argue that at least some APL-like responses (including the induction of anergy and promotion of thymocyte maturation) can be mimicked by strong agonists presented in a narrow, limiting dose range (9, 10). However, the ability of APLs to provoke TCR partial agonism or antagonism across a wide concentration range demonstrates that there are qualitatively distinct thresholds for TCR stimulation by different peptide/MHC ligands, which might determine in vivo responses. Indeed, mutations that produce APLs in viruses can, in some situations, act dominantly to skew the response to unmutated pathogens (11).

It is important to appreciate that the response to APLs is conditioned by the differentiation state of the T cell, however. For example, although TCR antagonists do not provoke overt activation of mature T cells, they can drive maturation of thymocytes with the same TCRs, in the process of thymic positive selection (12), and at least some self-peptide/MHC ligands can also act as positively selecting TCR antagonists (13). Interactions with self-peptide/MHC ligands during homeostasis can also affect the response of mature T cells. Indeed, Allen and colleagues (14) would later show that naive T cells from two TCR transgenic strains that recognize an identical foreign Ag with the same affinity exhibit distinct in vivo responses due to their different levels of “tonic” TCR signals. Hence, the outcome of encounter with TCR agonists and antagonists may be dictated by a T cell’s developmental and homeostatic experiences.

The work from Sloan-Lancaster et al. (3) constituted an important step in the realization that TCR engagement can direct varied responses in T cells, dictated by the nature of the TCR–peptide/MHC interaction. How T cells integrate these interactions, including recognition of low-affinity self-peptide/MHC ligands, determines whether T cells respond, survive, or even develop in the first place (15).