ANTIGEN PROCESSING AND
PRESENTATION
LAURENCE C. EISENLOHR AND JAY L. ROTHSTEIN
Thomas Jefferson University
In the ongoing search for effective and reliable immune-based approaches to cancer therapy,
much of the work is focused on T lymphocytes as effectors. CD8+ T lymphocytes (TCD8+)
are of particular interest as they combine specificity and lethality at a level that no current
chemotherapeutic or radiation regimen can match. One can only marvel at the effectiveness
with which these cells are able to clear an acute respiratory tract infection, leaving the involved
tissues intact—the precise goal of cancer therapy. CD4+ T lymphocytes (TCD4+), relatively
specific, but generally less cytotoxic than TCD8+, can also mediate potent anti-tumor effects in
certain settings. While a great deal has been learned about how TCD4+ and TCD8+ responses
are induced and sustained, further exploration will be necessary if the full potential of these
populations is to be harnessed. One aspect worthy of closer inspection is that of antigen
processing and presentation—the various intracellular steps that prepare antigen for T cell
recognition. It is intuitive that greater understanding and controlled manipulation of these
events, which usher in the adaptive response, could have profound influence on the final
character of the anti-tumor immunity that is engendered.
1. INTRODUCTION
This chapter will review fundamental aspects of antigen processing and presentation
with special emphasis on how they pertain to tumor-specific immunity. Three points
must be made at the outset. First, there is no intent to evaluate the relative efficacy of
various therapeutic strategies that have been based on principles of antigen processing
and presentation. Only a handful of possible permutations have been tested at this
point and, in any event, outcomes will certainly be different depending upon the
experimental model or clinical situation. Second, there is minimal segregation of
findings in animal models (usually mouse) and humans. Most of the fundamental cell
biology is similar even though decades of experimentation and practical application
have made it clear that success in mouse models does not ensure success in patients.
Finally, the topic of tumor antigen processing and presentation is now sufficiently
large that a comprehensive review in a single chapter is not possible. While an attempt
has been made to cover a large amount of conceptual territory, space does not allow
for all of the relevant work to be mentioned here.
2. THE BASIS FOR T CELL RECOGNITION: FRAGMENTS OF ANTIGEN
DISPLAYED AT THE CELL SURFACE BY SPECIALIZED “PRESENTING”
MOLECULES
2.1. Peptide Binding
While B cells and their antibody products recognize antigens in their native forms,
T cells respond to pieces of antigens held at the cell surface by various “presenting
molecules” and generated by a variety of intracellular, and even extracellular
processes known collectively as antigen processing. Class I molecules are made up
of a heavy chain encoded within the major histocompatibility complex (MHC)
and a noncovalently associated light chain, β2-microglobulin. Class I heterodimers
bind peptides that are generally 8–11 amino acids in length and present them to
TCD8+ whose most appreciated response is killing of the peptide-presenting cell.
Class II molecules, comprised of α and β chains, both encoded within the MHC,
generally bind peptides 11–17 amino acids in length, and present them to TCD4+
which respond by elaborating factors that guide and potentiate both B cell and
TCD8+ responses.1 The variation in lengths of peptide bound by class I and class II
molecules is due to distinct structural differences in the peptide-binding grooves (1).
The binding grooves of class I molecules are closed at both ends, with the consequence
that a peptide must be a specific length in order to be bound. In contrast,
class II binding grooves are open at both ends so that quite large peptides have the
capability of binding. Despite this, relatively short peptides are usually isolated from
class II molecules, presumably due to the exposure of any extended portions to
intracellular and extracellular proteases. As might be surmised from several different
crystal structures (2), peptides that directly interact with the binding groove of
both class I and class II molecules are resistant to proteolysis, as are the presenting
molecules themselves (3–7). Many readers may know that a key feature of class I and
II molecules is their tremendous polymorphism, with hundreds of versions of each
encoded by many different loci within the MHC existent in the human population.
Greatest variation is in the residues that line the peptide-binding grooves, leading
to distinct peptide-binding specificities and, thus, differences among individuals in
the parts of any antigen that are responded to. This variation is a powerful strategy
for a population to counteract the rapid replication and mutation rates that many
microbes are capable of, but constitutes a major impediment for tissue transplantation
and immune-based cancer therapy since both applications may require individuallytailored
therapies. The basis for binding specificity is a series of pockets in the floor
of any peptide-binding groove into which side chains of the peptide extend. Some
of these pockets provide anchoring points that are quite stringent in terms of the
side chains that are acceptable, while others are much more permissive. Thus, only
specific segments within a protein, with appropriate amino acids properly spaced
apart, are able to bind any particular MHC molecule. Those side chains that do
not participate in binding to the groove are available for interaction with the T cell
receptor. As mentioned at the outset, recognition of peptides by T cell receptors can
be highly specific and sensitive. Single amino acid changes in a peptide, including
residues that do not directly contact the T cell receptor and even simple phosphorylation
of a peptide, can profoundly influence T cell recognition (8–10). In terms
of sensitivity, relatively few copies of a particular peptide are required for full T cell
activation—on the order of tens to hundreds (11–13). This can be derived from an
amount of antigen that cannot be detected using standard biochemical methods (14).
Both specificity and sensitivity are highly variable among different T cell clones (15),
being determined by both intrinsic factors, such as receptor sequence and density,
and extrinsic factors such as the balance of stimulatory and suppressive cytokines.
These factors will obviously vary dependent upon the tissue(s) where the antigen is
expressed.
From the standpoint of peptide presentation, targets of T cell-mediated tumor
immunotherapy can be divided into three broad categories: foreign,mutated self, and
nonmutated self epitopes. Examples of the first category (foreign) are epitopes from
the growing number of viruses that establish persistent infections and induce transformation,
such as the papillomaviruses and herpesviruses.Within the second group
are the proteins altered by point mutations, deletions or chromosomal translocation,
which are incidentally or coincidentally connected with transformation. All of these
can result in new peptide sequences that have the ability to bind to an MHC class
I or class II molecule and potentially elicit a response. An emphasis must be placed
on the words can and potentially. Such mutations do not guarantee the generation of
a neo-epitope that can bind to an MHC molecule and binding does not guarantee
T cell stimulation. At least with respect to peptide binding, some level of prediction
is possible. Algorithms, based upon known epitopes, have been developed for many
mouse and human MHC molecules, such that one can query an open reading frame
for the presence of segments with a high likelihood of binding (16, 17). Nonmutated
peptides could be of potential interest if they are: 1) derived from antigens, such as
carcinoembryonic antigen, that are expressed at low levels or not at all in the adult,
but highly expressed in the cancerous cell, 2) expressed by a differentiated (specialized)
cell type, such as the melanocyte, that is expendable, 3) expressed by a fraction
of a particular cell type, expendable or not, such immunoglobulins, the product of
B cell lymphomas, that can provide unique T cell epitopes from the hypervariable
regions (18, 19), or 4) altered by cellular processes that have gone awry as a result of
transformation. An example of this would be phosphorylation due to aberrant kinase
activity, as recently suggested by the formation of antigens within papillary thyroid
carcinomas (20).
