T cells recognize foreign antigens as peptide fragments presented in the groove of major histocompatibility complex (MHC) class Iencoded molecules expressed on the cell surface. Cytotoxic T cells, the CD8+ subset of T lymphocytes, recognize and kill cells expressing foreign peptides derived from infection with viruses or intracellular bacteria, or from self-proteins that are overexpressed or mutated in tumors. An elaborate system has evolved to allow peptides generated by proteasomes in the cell cytosol to cross the endoplasmic reticulum membrane and to load onto newly synthesized MHC class I molecules. The stable peptide/MHC complex traffics to the cell surface, thus exposing even intracellular antigens to immune surveillance by the cytotoxic T cell arm of our immune defenses. The CD4+ subset of T lymphocytes recognizes forms of antigen that have been taken up and processed in phagosomes and presented in the groove of MHC class II molecules. On antigen recognition these cells can activate other inflammatory cells, such as macrophages or dendritic cells, and help the response of B lymphocytes and CD8+ T lymphocytes to antigen.
Following an acute infection with pathogenic viruses or bacteria, T cells specific for antigens encoded by the pathogen may go through 15 or more cell divisions within 6 days to produce enormous numbers of effector cells at the peak of the response. When antigen is cleared, these large numbers of effector cells are useless and more than 90 percent of them die during a phase of contraction, leaving behind a population of memory cells that can maintain their numbers for the lifetime of the animal. Memory T cells, which can respond faster to antigen rechallenge than naive T cells, are dispersed in all tissues to deal rapidly with reentry of a pathogen. What controls the contraction phase of the response? We asked whether the extent of effector cell death following pathogen clearance is regulated by competition for survival factors such as interleukin-7 (IL-7) or IL-15, or competition simply for "space" in the lymphoid organs. A system was designed in which large numbers of effector cells did not die when the infection had been cleared because they lacked a pro-apoptotic molecule. We used these competitors to study the effect this had on the rate and degree of contraction of other effector cells in the same animal. Remarkably, the added competition had no effect on the contraction phase, implying that competition for survival factors does not play a role. Rather, the decision to die during contraction or to live and go on to differentiate to a memory cell seems to be programmed during the expansion phase.
How do CD4+ T cells help the response of CD8+ T cells? It was thought that helper T cells and cytotoxic T cells were brought together by recognizing their antigen on the same antigen-presenting cell, and that help was provided by cell-cell contact or the short-range provision of cytokines. More recently, this model has been replaced by one that posits that the antigen-presenting cell (most likely a dendritic cell) is the essential go-between for CD4-mediated assistance in the cytotoxic T cell response. In this model, CD4+ T cells, recognizing antigen, send signals to the presenting cell, thereby activating or "licensing" it to be competent to stimulate the CD8 response. Without CD4 activation, the presenting cell would be unable to stimulate CD8+ T cells. This idea fit nicely with the knowledge that cytotoxic T cell responses to weak antigens, such as tumor cells, are CD4-dependent, while responses to viruses or bacteria, which can themselves activate dendritic cells, are CD4-independent. It was a surprise then when we showed that even though the primary CD8 response to the intracellular bacterial pathogen Listeria monocytogenes, which has a plethora of immune system alarm signals, is CD4-independent, the secondary or memory CD8 response wanes in the absence of CD4+ T cells. Memory appears to be CD4-dependent.
Our recent evidence suggests that IL-2, perhaps produced by CD4 cells, may be a key factor in the early programming of CD8+ T cell effector and memory differentiation. IL-2 was originally described as T cell growth factor, which is essential to keep T cells growing in culture. In vivo, however, CD8+ T cells that lack the high-affinity receptor for IL-2 respond to pathogen challenge robustly, going through many cell divisions and differentiating into memory cells. However, these "unhelped" effector cells are less effective killers and the memory cells die on rechallenge with the pathogen. Thus, it seems that IL-2 signals are not required to promote the growth of CD8+ T cells responding to an infection, but they are required to push their complete differentiation to functional effector and memory cells.
To avoid autoimmunity, the immune system must maintain self-tolerance. This begins in the thymus when T cells with dangerously high affinity for self-antigen are deleted in a process referred to as central tolerance. How peripheral self-antigens, which are not supposed to be present in the thymus, could be represented in this negative-selection screen has been a mystery. Amazingly, many tissue-specific antigens are ectopically, or promiscuously, expressed in the thymus to mediate central tolerance. Epithelial cells in the medulla of the thymus express pancreatic, adrenal, and ocular antigens, and this ectopic expression leads to the deletion of newly arising T cells that could damage these peripheral targets. Humans and mice that lack this ectopic expression of tissue-specific antigen in the thymus succumb to multiorgan inflammation and autoimmune disease.
In a system in which the insulin promoter is used to drive expression of a model antigen in the islet of Langherhans as well as ectopically in the thymus medulla, we have studied the efficiency and cellular requirements for the negative selection of CD4+ and CD8+ T cells in the thymus. The results show that the most effective presentation of antigen for deletion in the thymus medulla is performed not by the cells making antigen, but by professional antigen-presenting cells, probably dendritic cells, that acquire antigen from the medullary epithelial cells and "cross-present" it to developing T cells. How the professional antigen presenters do this—whether by phagocytosis of dying medullary cells or via uptake of membrane blebs from healthy cells—is not known. T cells bearing receptors with high affinity for the self-antigen are efficiently deleted. We find, however, that below a certain threshold of affinity, self-reactive T cells escape the thymus and populate the peripheral lymphoid organs. Under normal conditions, these cells are harmless. But when they are activated following infection by a pathogen that introduces a cross-reactive antigen, or higher levels of the self-antigen, autoimmunity can result. As effector cells, these weakly autoreactive T cells can cause autoimmune tissue damage, such as destruction of pancreatic islets resulting in diabetes.
Part of this work is supported by a grant from the National Institutes of Health.
As of August 17, 2010