Nothing 'gainst Time's Scythe Can Make Defense...

See allHide authors and affiliations

Science  28 May 2004:
Vol. 304, Issue 5675, pp. 1262-1263
DOI: 10.1126/science.1098965

T cells recognize antigen with great precision. They use receptors that engage unique combinations of short antigenic fragments held in position by major histocompatibility complex (MHC) proteins. The two classes of MHC proteins are each specialized to serve distinct subsets of T cells. Class I MHC proteins are the target molecules recognized by CD8 killer T cells. They obtain peptide ligands for presentation to T cells through proteolysis of antigen in the cytosol of, for example, virally infected cells or tumor cells. Viral proteins present in virally infected cells are subjected to proteolysis in the cytosol, and the resulting peptides are presented at the cell surface in a complex with class I MHC proteins. This MHC-peptide combination indicates to the CD8 killer T cells that there is a bona fide target that must be eliminated. How such peptides contribute to the priming of naïve CD8 T cells to initiate this immune response is a complicated affair, as we now learn from Wolkers et al. (1) and Norbury et al. (2) on pages 1314 and 1318 of this issue. With far-reaching implications for vaccine design, their work ties T cell priming to the life-span of the protein segment that contributes the antigenic peptide. The long-lived, and presumably more abundant, protein segments carry the day: They seem to be the best T cell primers. This conclusion is also reached by Shen and Rock reporting in a recent issue of Proceedings of the National Academy of Sciences U.S.A. (3).

The mere presence of the appropriate MHC-peptide combination is sufficient to engage committed CD8 killer T cells to destroy virally infected cells or tumor cells. However, priming of a T cell response also requires that T cells perceive the correct costimulatory molecules on the surface of professional antigen-presenting cells (APCs), such as dendritic cells. Many viruses, because of their pronounced tropism, do not infect professional APCs, and this raises the question of how a T cell response against such viruses can be initiated at all. Most immunologists now agree that the key is cross-presentation (cross-priming). During cross-presentation, the capture of remnants of virus- infected cells by phagocytosis results in the transfer of viral antigens to professional APCs, which then load the antigen- derived peptides onto their class I MHC proteins (4). The pathways of peptide generation are manifold, as are the proposals for how antigenic proteins or their digestion products leave the phagosomal compartment and arrive in the cytosol (see the figure). Cross-presentation, in principle, could exploit many different sources of peptide.

Cross-presentation of cell remnants.

Cell remnants contain three potential pools of stable peptides for presentation to T cells: peptides still embedded within their polypeptides, peptides that form complexes with chaperones, and peptides that form complexes with class I MHC molecules. (1) Cell remnants are taken taken up by APCs into phagosomes during phagocytosis. (2) The phagosomes recruit ER components, including ER membranes and ER protein channels that may help to translocate protein substrates across the phagosomal membrane. (3) Lysosomes fuse with the phagosomes to form phagolysosomes. Cell remnants are broken down in the phagolysosomes with the subsequent liberation of stable peptide pools. (4) Some peptides are exported from phagolysosomes possibly via ER constituents recruited during phagosome formation. (5) Alternatively, proteasomes in the cytosol or associated with phagosomes break down antigenic proteins, liberating peptides. (6) These peptides then obey the standard rules for loading onto class I MHC molecules and are displayed on the surface of APCs. (7) Neither peptides bound to MHC products nor those bound to chaperones in the donor cell contribute to cross-presentation. Sec61, the pore through which newly synthesized polypeptides enter the ER.


The peptides produced by cytosolic proteolysis are quite short-lived (seconds rather than minutes) (5) unless they are stabilized by interaction with cytosolic proteins (such as heat shock proteins or chaperones) or by insertion into the peptide-binding groove of class I MHC proteins. The peptides ultimately recognized by T cells in the context of the appropriate MHC proteins are stabilized in two ways: Either they are still embedded in the polypeptides from which they must be liberated by proteolysis, or they are bound as peptides to chaperones or MHC proteins.

Cross-presentation itself is intimately linked to the acquisition (by phagocytosis) of the antigen in a particulate form (6). Cell remnants fit the bill, as do latex beads coated with antigen or even simple protein aggregates. Particulate antigens are taken up by phagocytosis into phagosomes, which then fuse with other endocytic compartments. Then, particulate antigens are deconstructed by the combined action of low pH, a reducing environment, and resident proteases. Peptides generated in these compartments may unite immediately with peptide-receptive class I MHC molecules. Alternatively, peptides, fragments of antigens, or even intact proteins are exported from the phagosome to the cytosol. In the cytosol, the usual rules apply: proteolysis and transport of peptides via the TAP peptide transporter protein into the class I MHC loading compartment. In a twist on this scheme, the phagosome itself may accomplish all of these tasks (7, 8). The phagosomal membrane arises through recruitment of endoplasmic reticulum (ER) membrane and associated ER channel proteins that may be essential for movement of protein substrates across the phagosomal membrane. The reported association of proteasome subunits with the phagosome and the presence of a functional TAP complex in phagosomal preparations could provide “one-stop shopping” for class I MHC proteins seeking to acquire peptide (see the figure). The conditions in a cross-presentation compartment may also allow disassembly of class I MHC-peptide combinations, either to generate peptide-receptive class I MHC molecules or to extract from donor MHC molecules their peptide cargo (both steps could assist in cross-presentation).

But which of these sources of peptide actually contribute to cross-presentation? Wolkers et al. (1), Norbury et al. (2), and Shen and Rock (3) show that relatively abundant, long-lived proteins are the predominant source of cross-presented antigen. There is no detectable contribution by antigenic peptides delivered to the ER in a signal sequence-dependent fashion, nor by the peptides processed and bound to donor class I MHC molecules or to heat shock proteins. Using an ingenious experimental design, Wolkers et al. (1) placed one peptide (MHC class I epitope) in a location where it was extremely short-lived (by insertion into the hydrophilic portion of the cleavable signal peptide) and a second epitope in a position where it was stable (by grafting it onto the carboxyl terminus of the stable protein). Cells that expressed such double-epitope constructs were recognized in vitro by CD8 cytotoxic T lymphocytes directed against either epitope (in a TAP-dependent manner). Upon immunization of mice with these engineered cells, the epitope placed in the stable position primed T cells far more effectively than the epitope placed in the short-lived position. Norbury et al. (2) showed that treatment of the cells donating antigen with proteasome inhibitors enhanced their ability to cross-present. This finding implies that proteasomal digestion products are not sources of antigen for cross-presentation. The investigators further excluded class I MHC-bound peptides by cleverly exploiting donor cells with the wrong class I MHC product—these cells still bind the relevant epitope added as a synthetic peptide. Such peptide-loaded cells did not transfer their peptide to the recipient's MHC proteins, as shown by a failure to activate the corresponding antigen-specific T cells in the recipient.

This work has far-reaching implications for the design of vaccines aimed at eliciting a CD8 T cell response. Such vaccines must include antigen in a form that is cross-presented effectively in vivo. We need to clarify how antigen is transferred from the phagosomal compartment to the cytosol, and how proteasomal substrates are delivered to the TAP peptide transporter. Ensuring delivery of the cross-presented antigen in sufficient quantity—and in the form of proteasome substrates rather than digestion products—will send vaccinologists scurrying back to the drawing board. There are other, less pragmatic implications as well: T cell development and induction of T cell tolerance both require presentation of MHC-peptide complexes. T cells that bear receptors of appropriate affinity for some combinations of a self peptide and MHC will be deleted in the course of development or will be rendered tolerant in the periphery. Cross-presentation makes an as yet unknown contribution to the T cell repertoire: both shaping it and rendering it tolerant. This contribution is difficult to assess because the assays for cross-presentation require a mismatch between donor and recipient MHC, or a defect in antigen presentation in the donor cell. Nonetheless, as Wolkers et al. suggest (1), there may be unexpected opportunities for tumor immunologists to raise beneficial T cell responses against epitopes that have failed to induce tolerance for lack of cross-presentation.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.

Navigate This Article