Considerations in the design of vaccines that induce CD8 T cell mediated immunity.

The protective capacity of many currently used vaccines is based on induction of neutralizing antibodies. Many pathogens, however, have adapted themselves in different ways to escape antibody-based immune protection. In particular, for those infections against which conventional neutralizing antibody-based vaccinations appear challenging, CD8 T-cells are considered to be promising candidates for vaccine targeting. The design of vaccines that induce robust and long-lasting protective CD8 T-cell responses however imposes new challenges, as many factors such as kinetics and efficiency of antigen-processing and presentation by antigen presenting cells, T-cell repertoire and cytokine environment during T cell priming contribute to the specificity and functionality of CD8 T-cell responses. In the following, we review the most prominent aspects that underlie CD8 T-cell induction and discuss how this knowledge may help to improve the design of efficient CD8 T-cell inducing vaccines.

The components of the proteasome system and their role in MHC class I antigen processing.

By generating peptides from intracellular antigens which are then presented to T cells, the ubiquitin/26S proteasome system plays a central role in the cellular immune response. The proteolytic properties of the proteasome are adapted to the requirements of the immune system by proteasome components whose synthesis is under the control of interferon-gamma. Among these are three subunits with catalytic sites that are incorporated into the enzyme complex during its de novo synthesis. Thus, the proteasome assembly pathway and the formation of immunoproteasomes play a critical regulatory role in the regulation of the proteasome's catalytic properties. In addition, interferon-gamma also induces the synthesis of the proteasome activator PA28 which, as part of the so-called hybrid proteasome, exerts a more selective function in antigen presentation. Consequently, the combination of a number of regulatory events tunes the proteasome system to gain maximal efficiency in the generation of peptides with regard to their quality and quantity.

Efficient identification of novel HLA-A(*)0201-presented cytotoxic T lymphocyte epitopes in the widely expressed tumor antigen PRAME by proteasome-mediated digestion analysis.

We report the efficient identification of four human histocompatibility leukocyte antigen (HLA)-A(*)0201-presented cytotoxic T lymphocyte (CTL) epitopes in the tumor-associated antigen PRAME using an improved "reverse immunology" strategy. Next to motif-based HLA-A(*)0201 binding prediction and actual binding and stability assays, analysis of in vitro proteasome-mediated digestions of polypeptides encompassing candidate epitopes was incorporated in the epitope prediction procedure. Proteasome cleavage pattern analysis, in particular determination of correct COOH-terminal cleavage of the putative epitope, allows a far more accurate and selective prediction of CTL epitopes. Only 4 of 19 high affinity HLA-A(*)0201 binding peptides (21%) were found to be efficiently generated by the proteasome in vitro. This approach avoids laborious CTL response inductions against high affinity binding peptides that are not processed and limits the number of peptides to be assayed for binding. CTL clones induced against the four identified epitopes (VLDGLDVLL, PRA(100-108); SLYSFPEPEA, PRA(142-151); ALYVDSLFFL, PRA(300-309); and SLLQHLIGL, PRA(425-433)) lysed melanoma, renal cell carcinoma, lung carcinoma, and mammary carcinoma cell lines expressing PRAME and HLA-A(*)0201. This indicates that these epitopes are expressed on cancer cells of diverse histologic origin, making them attractive targets for immunotherapy of cancer.

The role of the ubiquitin-proteasome pathway in MHC class I antigen processing: implications for vaccine design.

Proteasomes are multisubunit enzyme complexes that reside in the cytoplasm and nucleus of eukaryotic cells. By selective protein degradation, proteasomes regulate many cellular processes including MHC class I antigen processing. Three constitutively expressed catalytic subunits are responsible for proteasome mediated proteolysis. These subunits are exchanged for three homologous subunits, the immunosubunits, in IFNgamma-exposed cells and in cells with specialized antigen presenting function. Both constitutive and immunoproteasomes degrade endogenous proteins into small peptide fragments that can bind to MHC class I molecules for presentation on the cell surface to cytotoxic T lymphocytes. However, immunoproteasomes seem to fulfill this function more efficiently. IFNgamma further induces the expression of a proteasome activator, PA28, which can also enhance antigenic peptide production by proteasomes. In this review, we will introduce the ubiquitin-proteasome system and summarize recent findings regarding the role of the IFNgamma-inducible proteasome subunits and proteasome regulators in antigen processing. We review the different ways by which tumors and viruses have been found to target the proteasome system to avoid MHC class I presentation of their antigens, and discuss recent progressions in the development of computer assisted approaches to predict CTL epitopes within larger protein sequences, based on proteasome cleavage specificity. The availability of such programs as well as a general insight into the proteasome mediated steps in MHC class I antigen processing provides us with a rational basis for the design of new antiviral and anticancer T cell vaccines.

Differential influence on cytotoxic T lymphocyte epitope presentation by controlled expression of either proteasome immunosubunits or PA28.

The proteasome is the principal provider of major histocompatibility complex (MHC) class I-presented peptides. Interferon (IFN)-gamma induces expression of three catalytically active proteasome subunits (LMP2, LMP7, and MECL-1) and the proteasome-associated activator PA28. These molecules are thought to optimize the generation of MHC class I-presented peptides. However, known information on their contribution in vivo is very limited. Here, we examined the antigen processing of two murine leukemia virus-encoded cytotoxic T lymphocyte (CTL) epitopes in murine cell lines equipped with a tetracycline-controlled, IFN-gamma-independent expression system. We thus were able to segregate the role of the immunosubunits from the role of PA28. The presence of either immunosubunits or PA28 did not alter the presentation of a subdominant murine leukemia virus (MuLV)-derived CTL epitope. However, the presentation of the immunodominant MuLV-derived epitope was markedly enhanced upon induction of each of these two sets of genes. Thus, the IFN-gamma-inducible proteasome subunits and PA28 can independently enhance antigen presentation of some CTL epitopes. Our data show that tetracycline-regulated expression of PA28 increases CTL epitope generation without affecting the 20S proteasome composition or half-life. The differential effect of these IFN-gamma-inducible proteins on MHC class I processing may have a decisive influence on the quality of the CTL immune response.

MHC class I antigen processing of an adenovirus CTL epitope is linked to the levels of immunoproteasomes in infected cells.

Proteasomes are the major source for the generation of peptides bound by MHC class I molecules. To study the functional relevance of the IFN-gamma-inducible proteasome subunits low molecular mass protein 2 (LMP2), LMP7, and mouse embryonal cell (MEC) ligand 1 in Ag processing and concomitantly that of immunoproteasomes, we established the tetracycline-regulated mouse cell line MEC217, allowing the titrable formation of immunoproteasomes. Infection of MEC217 cells with Adenovirus type 5 (Ad5) and analysis of Ag presentation with Ad5-specific CTL showed that cells containing immunoproteasomes processed the viral early 1B protein (E1B)-derived epitope E1B192-200 with increased efficiency, thus allowing a faster detection of viral entry in induced cells. Importantly, optimal CTL activation was already achieved at submaximal immunosubunit expression. In contrast, digestion of E1B-polypeptide with purified proteasomes in vitro yielded E1B192-200 at quantities that were proportional to the relative contents of immunosubunits. Our data provide evidence that the IFN-gamma-inducible proteasome subunits, when present at relatively low levels as at initial stages of infection, already increase the efficiency of antigenic peptide generation and thereby enhance MHC class I Ag processing in infected cells.

Abrogation of CTL epitope processing by single amino acid substitution flanking the C-terminal proteasome cleavage site.

CTL directed against the Moloney murine leukemia virus (MuLV) epitope SSWDFITV recognize Moloney MuLV-induced tumor cells, but do not recognize cells transformed by the closely related Friend MuLV. The potential Friend MuLV epitope has strong sequence homology with Moloney MuLV and only differs in one amino acid within the CTL epitope and one amino acid just outside the epitope. We now show that failure to recognize Friend MuLV-transformed tumor cells is based on a defect in proteasome-mediated processing of the Friend epitope which is due to a single amino acid substitution (N-->D) immediately flanking the C-terminal anchor residue of the epitope. Proteasome-mediated digestion analysis of a synthetic 26-mer peptide derived from the Friend sequence shows that cleavage takes place predominantly C-terminal of D, instead of V as is the case for the Moloney MuLV sequence. Therefore, the C terminus of the epitope is not properly generated. Epitope-containing peptide fragments extended with an additional C-terminal D are not efficiently translocated by TAP and do not show significant binding affinity to MHC class I-Kb molecules. Thus, a potential CTL epitope present in the Friend virus sequence is not properly processed and presented because of a natural flanking aspartic acid that obliterates the correct C-terminal cleavage site. This constitutes a novel way to subvert proteasome-mediated generation of proper antigenic peptide fragments.

Efficient generation of a hepatitis B virus cytotoxic T lymphocyte epitope requires the structural features of immunoproteasomes.

Interferon (IFN)-gamma-induced cells express the proteasome subunits low molecular weight protein (LMP)2, LMP7, and MECL-1 (multicatalytic endopeptidase complex-like 1), leading to the formation of immunoproteasomes. Although these subunits are thought to optimize MHC class I antigen processing, the extent of their role and the mechanistic aspects involved remain unclear. Herein, we study the proteolytic generation of an human histocompatibility leukocyte antigen (HLA)-Aw68-restricted hepatitis B virus core antigen (HBcAg) cytotoxic T lymphocyte (CTL) epitope that is recognized by peripheral blood lymphocytes from patients with acute self-limited but not chronic hepatitis B virus (HBV). Immunological data suggest that IFN-gamma-induced rather than uninduced HeLa cells process and present the HBV CTL epitope upon infection with HBcAg-expressing vaccinia viruses. Analyses of 20S proteasome digests of synthetic polypeptides covering the antigenic HBcAg peptide demonstrate that only immunoproteasomes efficiently perform the cleavages needed for the liberation of this HBV CTL epitope. Although the concerted presence of the three immunosubunits appears essential, we find that both catalytically active LMP7 and inactive LMP7 T1A support CTL epitope generation. We conclude that LMP7 influences the structural features of 20S proteasomes, thereby enhancing the activity of the LMP2 and MECL-1 catalytic sites, which provide cleavage specificity. Thus, LMP7 incorporation is of greater functional importance for the generation of an HBV CTL epitope than cleavage specificity.

The proteasome inhibitor PI31 competes with PA28 for binding to 20S proteasomes.

PI31 is a previously described inhibitor of 20S proteasomes. Using recombinant PI31 we have analyzed its effect on proteasomal hydrolyzing activity of short fluorogenic substrates and of a synthetic 40-mer polypeptide. In addition, we investigated its influence on the activation of 20S proteasome by the proteasome activator PA28. PI31 inhibits polypeptide degradation already at concentrations which only partially inhibit fluorogenic substrate turnover and immunosubunits do not influence the PI31 binding affinity. Furthermore our data demonstrate that PI31 is a potent competitor of PA28-mediated activation.

Mutational analysis of subunit i beta2 (MECL-1) demonstrates conservation of cleavage specificity between yeast and mammalian proteasomes.

Proteasomes are the major protein-degrading complexes in the cytosol and regulate many cellular processes. To examine the functional importance of the MC14/MECL-1 proteasome active site subunits, cell lines expressing a catalytically inactive form of MECL-1 were established. Whereas mutant MECL-1 was readily incorporated into cytosolic proteasomes, replacing the constitutive MC14 subunit, removal of the prosequence was incomplete indicating that its processing required autocatalytic cleavage. Functional analyses showed that the absence of the MC14/MECL-1 active sites abrogated proteasomal trypsin-like activity, but did not affect other catalytic activities. Our data demonstrate a conservation of cleavage specificity between mammalian and yeast proteasomes.

The Listeria monocytogenes-secreted p60 protein is an N-end rule substrate in the cytosol of infected cells. Implications for major histocompatibility complex class I antigen processing of bacterial proteins.

Cytosolic antigen degradation is an initial step in the generation of major histocompatibility complex (MHC) class I-associated cytolytic T lymphocyte epitopes. Intracellular Listeria monocytogenes secretes p60, a murein hydrolase, into the host cell cytosol, where it is degraded by proteasomes. Roughly 3% of degraded p60 gives rise to p60 217-225, a nonamer peptide that is bound by H-2Kd MHC class I molecules. Herein, we introduce targeted deletions throughout the p60 gene to identify potential proteolytic signals within p60. Degradation of mutant forms of p60 was investigated in macrophages infected with recombinant L. monocytogenes. We found that deletions within the amino-terminal two-thirds of p60 enhanced cytosolic degradation. In contrast, truncation of the C terminus resulted in modest stabilization of p60 in the host cell cytosol. Because a protein's N-terminal amino acid can determine its rate of degradation, we mutagenized this residue in p60 into known stabilizing and destabilizing residues. Valine substitution dramatically stabilized cytosolic p60 molecules, while substitution with aspartic acid resulted in rapid degradation. The number of p60 217-225 epitopes isolated from infected cells directly correlated with the rates of p60 degradation. Our data, therefore, indicate that the N-terminal amino acid and multiple internal regions of p60 influence its stability in the cytosol of infected cells. Antigen degradation and epitope generation are linked, and different degradation signals can channel bacterial proteins into the MHC class I antigen processing pathway.

MHC class I antigen processing of Listeria monocytogenes proteins: implications for dominant and subdominant CTL responses.

Listeria monocytogenes (L. monocytogenes) secretes proteins associated with its virulence into the cytosol of infected cells. These secreted proteins are degraded by host cell proteasomes and processed into peptides that are bound by MHC class I molecules in the endoplasmic reticulum. We have found that the MHC class I antigen-processing pathway is very efficient at generating the epitopes that are presented to cytolytic T lymphocytes (CTL). Depending on which antigen is investigated, from 3 to 30% of degraded antigens are processed into nonamer peptides that are bound by MHC class I molecules. Surprisingly, neither the efficiency of epitope generation nor the absolute number of epitopes per infected cell determines the magnitude of the in vivo CTL response. One of the least prevalent epitopes, derived from an antigen that is virtually undetectable in infected cells, primes the immunodominant CTL response in L. monocytogenes-infected mice. Our studies suggest that immunodominant and subdominant T-cell responses cannot be predicted by the prevalence of antigens or epitopes alone, and that additional factors, yet to be determined, are involved.

Enhanced intracellular dissociation of major histocompatibility complex class I-associated peptides: a mechanism for optimizing the spectrum of cell surface-presented cytotoxic T lymphocyte epitopes.

Association of antigenic peptides with newly synthesized major histocompatibility complex (MHC) class I molecules occurs in the endoplasmic reticulum and is a critical early step for the initiation of cytotoxic T lymphocyte (CTL)-mediated immune defenses. Pathogen-derived peptides compete with a plethora of endogenous peptides for MHC class I grooves. We find that two H2-K(d)-restricted peptides, which derive from the Listeria monocytogenes p60 antigen, accumulate in infected cells with different kinetics. Although competition assays suggest that both epitopes are bound with equivalent affinity, they dissociate from MHC class I molecules at markedly different rates. p60 217-225 forms complexes with H2-K(d) with a half-life >6 h, while p60 449-457 dissociates from H2-K(d) with a half-life of approximately 1 h. We find that p60 449-457-H2-K(d) complexes retained intracellularly with brefeldin A have a half-life of 30 min, and thus are less stable than surface complexes. While peptide dissociation from retained MHC class I molecules is enhanced, retained H2-K(d) molecules maintain a remarkable capacity to bind new T cell epitopes. We find that intracellular H2-K(d) molecules can bind new CTL epitopes for up to 3 h after their synthesis. Our studies provide a glimpse of peptide interaction with MHC class I molecules in the endoplasmic reticulum/proximal Golgi complex of intact, infected cells. We propose that the increased intracellular lability of peptide-MHC class I complexes may function to optimize the spectrum of peptides presented to T lymphocytes during cellular infection.

A single residue exchange within a viral CTL epitope alters proteasome-mediated degradation resulting in lack of antigen presentation.

CTL epitope (KSPWFTTL) encoded by AKV/MCF type of murine leukemia virus (MuLV) differs from the sequence in Friend/Moloney/Rauscher (FMR) type in one residue (RSPWFTTL). CTL experiments indicated defective processing of the FMR peptide in tumor cells. Proteasome-mediated digestion of AKV/MCF-type 26-mer peptides resulted in the early generation and higher levels of epitope-containing fragments than digestion of FMR-type peptides, explained by prominent cleavage next to R in the FMR sequence. The fragments were identified as 10- and 11-mer peptides and were efficiently translocated by TAP. The naturally presented AKV/MCF peptide is the 8-mer, indicating ER peptide trimming. In conclusion, a single residue exchange can cause CTL epitope destruction by specific proteasomal cleavage.

CTL epitope generation is tightly linked to cellular proteolysis of a Listeria monocytogenes antigen.

Listeria monocytogenes is a pathogenic intracellular bacterium that secretes proteins into the cytosol of host cells. A major secreted protein, p60, is processed by the host cell into the nonamer peptides p60 217-225 and p60 449-457, which are presented to CTL by H-2Kd MHC class I molecules. Herein, we use two membrane permeable peptide aldehyde protease inhibitors, LLnL and Z-LLF, to inhibit cytosolic proteolysis in L. monocytogenes-infected cells. These inhibitors, which have been shown to inhibit proteasomes, completely abrogate cytosolic p60 degradation. The effect of LLnL and Z-LLF on p60 epitope generation was determined by acid-eluting, HPLC-purifying, and quantifying p60 217-225 and p60 449-457 from infected cells. We show a direct linkage between p60 degradation and epitope generation. However, the two inhibitors have quantitatively different effects on the generation of the two epitopes. Our findings implicate proteasomes in the earliest stages of Ag degradation and suggest that different CTL epitopes can be generated by distinct proteolytic processes.

Two Listeria monocytogenes CTL epitopes are processed from the same antigen with different efficiencies.

Listeria monocytogenes is an intracellular bacterium that elicits MHC class I-restricted CTL in infected mice. A major CTL specificity is the nonamer peptide p60 217-225, which is derived from the bacterial murein hydrolase p60 and presented by the H-2Kd MHC class I molecule. In this report, we identify a second H-2Kd presented peptide, encompassing residues 449-457 of p60, that is detected by L. monocytogenes-specific CTL. Both p60-derived CTL epitopes are good competitors for H-2Kd binding and TAP (transporter associated with Ag processing) transport. CTL clone WP11.12 lyses L. monocytogenes infected cells and recognizes naturally processed p60 449-457 acid eluted from L. monocytogenes-infected macrophages. Although both epitopes derive from the same Ag and bind the same allelic form of MHC class I, quantitative analysis reveals that the amount of p60 449-457 in infected cells is approximately 10-fold greater than the amount of p60 217-225. Shuffling p60 217-225 into position 449-457 decreases its processing efficiency, indicating that the large number of p60 449-457 epitopes cannot be entirely attributed to epitope-flanking sequences. Our findings indicate that CTL epitopes can be processed from Ags with markedly different kinetics and efficiencies. Intrinsic qualities of an epitope and its location within a protein influence the efficiency of Ag processing.

Listeriolysin is processed efficiently into an MHC class I-associated epitope in Listeria monocytogenes-infected cells.

Listeria monocytogenes is an intracellular pathogen that enters the cytoplasm of infected cells by secreting listeriolysin (LLO), a protein that destroys the phagosomal membrane. In infected mice, LLO is a major Ag detected by protective, MHC class I-restricted CTLs. Although the role of LLO in pathogenesis and host immunity is well established, its rate of intracellular synthesis has yet to be determined. Herein we show that cytosolic L. monocytogenes secrete LLO at a relatively low rate of approximately one molecule per bacterium per minute. Under extracellular labeling conditions, the rate of LLO secretion is approximately 50-fold higher. Intracellular LLO synthesis suffices, however, for the accumulation of 600 to 1000 H-2Kd-associated LLO 91-99 epitopes per cell. We calculate that between four and 11 LLO molecules are degraded for each LLO 91-99 epitope bound by H-2Kd. Our findings indicate that the antigenicity of LLO, with respect to MHC class I-restricted CTLs, cannot be attributed to high levels of intracellular secretion. Rather, LLO is a dominant Ag because it is rapidly degraded and very efficiently processed into an MHC class I-associated epitope.

Major differences in transporter associated with antigen presentation (TAP)-dependent translocation of MHC class I-presentable peptides and the effect of flanking sequences.

The MHC-encoded transporter associated with Ag presentation (TAP) translocates peptides from the cytosol to the ER lumen, where association with MHC class I molecules occurs. The MHC class I/peptide complex is subsequently transported to the cell surface for presentation to CD8+T cells. We studied TAP-dependent translocation of defined MHC class I presentable murine peptides by competition for translocation of a radiolabeled model peptide, to address whether efficient peptide presentation by MHC class I molecules is preceded by equal efficient peptide translocation by TAP. Surprisingly, we observed that four immunodominant viral peptides of 16 peptides tested were very inefficiently transported by TAP. Inefficient translocation could be overcome by substitution of a proline residue present at position 3 in the peptides. Furthermore, addition of natural flanking amino acids directly surrounding a poorly transported peptide could considerably improve translocation by TAP. Our data suggest that some peptides are efficiently transported by TAP in their optimal size for MHC class I binding, whereas other peptides are transported as larger peptide fragments that need further trimming in the ER for MHC class I binding.

Identification of an H-2 Kb-presented Moloney murine leukemia virus cytotoxic T-lymphocyte epitope that displays enhanced recognition in H-2 Db mutant bm13 mice.

Upon infection with the Moloney murine sarcoma virus-murine leukemia virus (MuLV) complex, H-2b C57BL/6 (B6) mice respond with a class I Db-restricted cytotoxic T-lymphocyte (CTL) response, which protects against virus-induced tumorigenesis. In the B6-derived Db mutant B6.CH-2bm13 (bm13) strain, part of the class I Db antigen-presenting groove is shaped by a class I Kb-encoded sequence. Like B6 mice, bm13 mice reject Moloney virus-induced tumors, but the protective CTL response is Kb restricted. In this study we show enhanced levels of Moloney MuLV-specific CTLp with a restriction for Kb in bm13 mice. Through the use of CTL clones from Moloney virus-immunized bm13 mice, the class I Kb-presented CTL epitope was identified. The epitope is located in the Moloney virus gp70 envelope protein region (Moloney envelope, amino acids 189 to 196 [Mol env (189-196)]), SSWDFITV and has the Kb allele-specific binding motif. The Dbm13 molecule does not present the env(189 to 196) epitope to Kb-restricted bm13 CTL. In B6 mice, Mol env(189-196)-specific CTL could be induced by peptide vaccination. B6 mice thus have CTL precursors specific for this epitope but at considerably lower levels than do bm13 mice. We hypothesize that additional positive selection of Kb-restricted CTL on the Dbm13 molecule in bm13 mice explains this difference in precursor frequencies. We examined related strains of MuLV for the presence of Mol env(189-196) sequence equivalents. Rauscher, Friend, and AKV MuLV-encoded Mol env(189-196) epitope equivalents were properly recognized in cytotoxicity assays, both as synthetic and as endogenously expressed (Rauscher MuLV) peptides. In contrast, the mink cell focus-forming virus MuLV-encoded epitope equivalent, lacking a Kb anchor residue, was not presented for CTL recognition and hence can be excluded as an important CTL epitope for mink cell focus-forming viruses.