Degenerate and promiscuous recognition by CTL of peptides presented by the MHC class I A3-like superfamily: implications for vaccine development.

Recent data demonstrate that HLA class I alleles can be grouped into superfamilies based on similarities of their peptide-binding motifs. In this study, we have tested the immunogenicity and antigenicity of peptides capable of degenerate binding to multiple HLA class I molecules of the A3-like superfamily. The assay systems utilized included both primary in vitro cultures of lymphocytes from healthy donors, as well as in vitro restimulation of lymphocytes from HIV-infected individuals. Several of the peptides capable of binding more than one HLA A3-like class I molecule were also found to be immunogenic in the context of this same group of A3-like molecules (degenerate CTL recognition). Furthermore, some of the CTL lines thus generated demonstrated promiscuous recognition of the cognate epitope in the context of MHC molecules from more than one member of the superfamily. The fine Ag specificity of this phenomenon was further analyzed using two promiscuous CTL clones derived from A3 and A11 individuals, respectively, and specific for an epitope in the HIV-1 reverse transcriptase. By the use of single-amino acid-substitution analogues, it was demonstrated that the fine specificity of the TCR is largely maintained between MHC-matched and MHC-mismatched presentation of peptide within the A3-like superfamily. These results indicate that the similar peptide-binding specificities among different members of the A3-like superfamily can be reflected in a remarkable similarity in the peptide-MHC complex structures engaged by the TCR and responsible for T cell activation.

Degenerate cytotoxic T cell epitopes from P. falciparum restricted by multiple HLA-A and HLA-B supertype alleles.

We recently described human leukocyte antigen (HLA) A2, A3 and B7 supertypes, characterized by largely overlapping peptide-binding specificities and represented in a high percentage of different populations. Here, we identified 17 Plasmodium falciparum peptides capable of binding these supertypes and assessed antigenicity in both vaccinated and naturally exposed populations. Positive cytotoxic T lymphocyte recall and cytokine (interferon-gamma and tumor necrosis factor alpha) responses were detected for all peptides; all were recognized in the context of more than one HLA class I molecule; and at least 12 of the 17 were recognized in the context of all HLA alleles studied. These data validate the concept of HLA supertypes at the biological level, show that highly degenerate peptides are almost always recognized as epitopes, and demonstrate the feasibility of developing a universally effective vaccine by focusing on a limited number of peptide specificities.

Identification of A2-restricted hepatitis C virus-specific cytotoxic T lymphocyte epitopes from conserved regions of the viral genome.

We have focused on conserved regions of the hepatitis C Virus (HCV) genome to identify viral peptides that contain HLA class I binding motifs and bind with high affinity to the corresponding purified HLA molecules. Accordingly, we have identified 31 candidate epitopes in the HCV that have the potential to be recognized by either HLA-A1, A2.1-, A3, A11- or A24-restricted cytotoxic T lymphocytes (CTL). Twelve conserved peptides that bind HLA-A2.1 with high or intermediate affinity were tested for immunogenicity in vitro in human primary CTL cultures and in vivo by direct immunization of HLA-A2.1/Kb transgenic mice. Six HLA-A2.1-restricted CTL epitopes were immunogenic in both systems. At least three of these peptide epitopes were endogenously processed and presented for CTL recognition. Overall, these data illustrate the value of this approach for the development of virus-specific, peptide-based vaccines.

Definition of an HLA-A3-like supermotif demonstrates the overlapping peptide-binding repertoires of common HLA molecules.

An HLA-A3-like supertype (minimally comprised of products from the HLA class I alleles A3, A11, A31, A*3301, and A*6801) has been defined on the basis of (a) structural similarities in the antigen-binding groove, (b) shared main anchor peptide-binding motifs, (c) the identification of peptides cross-reacting with most or all of these molecules, and (d) the definition of an A3-like supermotif that efficiently predicts highly cross-reactive peptides. Detailed secondary anchor maps for A3, A11, A31, A*3301, and A*6801 are also described. The biologic relevance of the A3-like supertype is indicated by the fact that high frequencies of the A3-like supertype alleles are conserved in all major ethnic groups. Because A3-like supertype alleles are found in most major HLA evolutionary lineages, possibly a reflection of common ancestry, the A3-like supermotif might in fact represent a primeval human HLA class I peptide-binding specificity. It is also possible that these phenomena might be related to optimal exploitation of the peptide specificity by human TAP molecules. The grouping of HLA alleles into supertypes on the basis of their overlapping peptide-binding repertoires represents an alternative to serologic or phylogenetic classification.

Differences and similarities in the A2.1-restricted cytotoxic T cell repertoire in humans and human leukocyte antigen-transgenic mice.

HLA-A2.1-binding peptides (n = 38) were screened for immunogenicity with human peripheral blood mononuclear cells in cytotoxic T lymphocyte (CTL) induction experiments in vitro and with splenocytes from HLA-A2.1/Kb transgenic mice following immunization in vivo. These data were compiled and analyzed to determine the level of overlap between the A2.1-restricted CTL repertoire of A2.1/Kb-transgenic mice and A2.1+ humans. In both humans and mice, a major histocompatibility complex affinity threshold of approximately 500 nM appears to determine the capacity of a peptide to elicit a CTL response. Good concordance between the human data in vitro and mouse data in vivo was observed with 85% of the high-binding peptides, 58% of the intermediate binders, and 83% of the low/negative binders. Although some peptides immunogenic for mouse CTL but not for humans (and vice versa) could be identified, the data as a whole suggest an extensive overlap between T cell receptor repertoires of mouse and human CTL and support the use of HLA-transgenic mice for the identification of potential human CTL epitopes.

Human CTL epitopes encoded by human papillomavirus type 16 E6 and E7 identified through in vivo and in vitro immunogenicity studies of HLA-A*0201-binding peptides.

Human papillomavirus type 16 (HPV16) is strongly associated with cervical carcinogenesis. The HPV16 E6 and E7 oncoproteins are constitutively expressed in the majority of cervical tumor cells and are, therefore, attractive targets for CTL-mediated immunotherapy. In mice, the outgrowth of a lethal dose of HPV16-induced tumor cells has been prevented by vaccination with a CTL epitope encoded by HPV16 E7, indicating the feasibility of peptide immunization to obtain antitumor CTL responses. In the present study, the immunogenicity of 9 HLA-A*0201-binding peptides encoded by HPV16 E6 and E7 was analyzed in vivo in HLA-A*0201Kb transgenic mice and in vitro in CTL cultures induced from PBMC of HLA-A*0201+ healthy donors. Four peptides with a good binding affinity were immunogenic in HLA-A*0201Kb transgenic mice, and three of them were also highly immunogenic in CTL induction experiments with PBMC of HLA-A*0201+ healthy donors. Human CTL clones specific for these three peptides were capable of lysing the HPV16 E7-containing HLA-A*0201+ cervical carcinoma cell line CaSki. These E7-derived peptides (11-20, YMLDLQPETT; 82-90, LLMGTLGIV; 86-93, TLGIVCPI), therefore, are likely to represent naturally processed human CTL epitopes of HPV16. Additionally, these three HPV16-encoded peptides have the highest affinity of binding to the HLA-A*0201 molecule. In this study, peptides with a lower binding affinity were less immunogenic. Therefore, our data illustrate that the HLA-binding affinity of a peptide has a major impact on its immunogenicity. In conclusion, we have identified immunogenic peptides encoded by HPV16 E6 and E7 that could be used in vaccines for the prevention and treatment of cervical carcinoma.

In vitro induction of primary, antigen-specific CTL from human peripheral blood mononuclear cells stimulated with synthetic peptides.

A protocol for in vitro induction of primary, antigen-specific CTL from human peripheral blood mononuclear cells (PBMCs) was developed. Antigen presenting cells (APCs) consisted of Staphylococcus aureus Cowan-I (SAC-I) activated PBMCs treated with a citrate-phosphate buffer at pH 3 to release endogenous peptides bound to surface MHC. This treatment resulted in transient expression of empty class I molecules which could be subsequently stabilized with peptide and beta 2-microglobulin (beta 2m). SAC-I activated PBMCs from HLA-A2.1 normal donors loaded with HBV core 18-27 peptide following acid treatment were used to stimulate PBMCs depleted of CD4+ T cells, in the presence of recombinant interleukin-7 (rIL-7). After 12 days, cells were restimulated with autologous, peptide-pulsed, adherent cells and tested for CTL activity 7 days later. In 23 independent experiments from 13 different HLA-A2.1 donors, this protocol resulted in induction of primary CTL more than 90% of the time. As indicated by both the frequency and magnitude of the response against peptide-sensitized target cells, SAC-I activated PBMCs treated with acid were the most efficient stimulator APC. Thirteen per cent of the cultures generated were capable of lysing target cells transfected with the HBV core antigen and, in general, these CTL cultures exhibited high avidity for the HBV core peptide. This protocol is generally applicable to different antigens and class I alleles, and thus, may be utilized to screen large numbers of peptides to identify human CTL epitopes.

Specific binding of leukemia oncogene fusion protein peptides to HLA class I molecules.

Many human leukemias are characterized by chromosomal translocations yielding hybrid RNAs capable of encoding fusion chimeric proteins. The unique amino acid sequences found in these oncogenic fusion proteins represent true tumor-specific antigens that are potentially immunogenic. Although these leukemia-specific fusion proteins have an intracellular location, they might be recognized immunologically by T lymphocytes if peptides derived from the unique sequences are capable of presentation by the major histocompatibility complex (MHC) molecules on leukemic cells. The ability of a series of synthetic peptides corresponding to the junctional sequences of chronic myelogenous leukemia (CML)-derived bcr-abl and acute promyelocytic leukemia (APL)-derived PML-RAR alpha fusion proteins to bind to purified class I molecules was studied. A series of 152 peptides 8, 9, 10, and 11 amino acids in length, spanning the b3a2 and b2a2 breakpoints for CML and PML-RAR alpha A and B breakpoints for APL were analyzed for HLA A1, A2.1, A3.2, A11, A24, B7, B8, and B27 binding motifs. Twenty-one CML peptides and 4 APL peptides were predicted to be potential HLA class I binders. The peptides were tested for binding to appropriate purified HLA molecules in a competition radioimmunoassay. Four peptides derived from b3a2 CML breakpoint bound with high (< 50 nmol/L) or intermediate (< or = 500 nmol/L) affinity to HLA A3, A11, and B8. None of the CML b2a2 or PML-RAR alpha A or B junctional peptides showed affinity of this magnitude for the HLA class I molecules tested. This is the first evidence that tumor-specific breakpoint peptides can bind human MHC class I molecules and provides a rationale for developing a therapeutic vaccine strategy.

Patients with chronic hepatitis C have circulating cytotoxic T cells which recognize hepatitis C virus-encoded peptides binding to HLA-A2.1 molecules.

Antiviral cytotoxic T lymphocytes (CTL) may play a role in clearance of hepatitis C virus (HCV)-infected cells and thereby cause hepatocellular injury during acute and chronic HCV infection. The aim of this study was to identify HLA-A2.1-restricted HCV T-cell epitopes and to evaluate whether anti-HCV-specific CTL are present during chronic hepatitis C. Peripheral blood mononuclear cells from four HLA-A2-positive patients with chronic hepatitis C and from two individuals after recovery from HCV infection were tested against a panel of HCV-encoded peptides derived from different regions of the genome, including some peptides containing HLA-A2.1 binding motifs. HLA-A2-negative patients with chronic hepatitis C as well as healthy HLA-A2-positive (anti-HCV-negative) donors served as controls. Peripheral blood mononuclear cells stimulated repeatedly with several HCV-encoded peptides (three in core, one in NS4B, and one in NS5B) yielded cytolytic responses. All four HLA-A2-positive patients with active infection had CTL specific for at least one of the identified epitopes, whereas two patients who had recovered from HCV infection had almost no CTL responses. Monoclonal antibody blocking experiments performed for two epitopes demonstrated a class I- and HLA-A2-restricted CTL response. CTL epitopes could partially be predicted by HLA-A2 binding motifs and more reliably by quantitative HLA-A2.1 molecule binding assays. Most of the identified epitopes could also be produced via the endogenous pathway. Specific CTL against multiple, mostly highly conserved epitopes of HCV were detected during chronic HCV infection. This finding may be important for further investigations of the immunopathogenesis of HCV, the development of potential therapies against HCV on the basis of induction or enhancement of cellular immunity, and the design of vaccines.

Characterization of human CD4+, HIV-SF2 Nef-specific T-cell clones for antigen-processing and presentation requirements and for cytotoxic activity.

We have described previously the generation of seven HIV-SF2 Nef-specific, CD4+ T-cell clones, identification of epitopes within which are recognized by these clones, and the MHC alleles that restrict their responses. In this study, we have extended this characterization to include evaluation of antigen-processing and presentation requirements and cytotoxic activity. Clones were generated from five HIV-1 uninfected donors by in vitro stimulation of peripheral blood mononuclear cells with purified recombinant Nef1. In experiments with fixed cells, with the exception of two clones, recognition of Nef, but not Nef peptides, required processing. Also, at higher concentrations of antigen, the clones themselves were capable of presenting Nef peptides, but not soluble Nef. All clones had the ability to specifically lyse autologous, Epstein-Barr virus-transformed lines sensitized with Nef synthetic peptides, or, in some cases, soluble Nef. The cytotoxic activity mapped to the same epitopes identified for the proliferative response (a.a. 14-22, 47-53, 68-77, 70-77, 195-203 and 185-192) and was restricted by the same HLA class II molecules (DRw6, DQw7, DRw15(2), DR1 and DP5). Sensitization of the cytolytic clones with specific Nef peptides, but not soluble Nef, resulted in autolysis.

Induction of anti-tumor cytotoxic T lymphocytes in normal humans using primary cultures and synthetic peptide epitopes.

Cytotoxic T lymphocytes (CTLs) recognize peptide antigens associated with cell surface major histocompatibility complex (MHC) molecules. The identification of tumor cell-derived peptides capable of eliciting anti-tumor CTL responses would enable the design of antigen-specific immunotherapies. Our strategy to identify such potentially therapeutic peptides relies on selecting high-affinity MHC binders from known tumor-associated antigens. These peptides are subsequently tested for their ability to induce CTLs capable of killing tumor cells. With this strategy, we have identified a nine-residue epitope, derived from the product of the tumor-associated gene MAGE-3, which has the capacity to induce in vitro CTLs that kill melanoma and other tumor cell lines. These results show the primary in vitro induction of tumor-specific human CTLs and illustrate the feasibility of ex vivo antigen-specific approaches to the immunological therapy of cancer.

Epitopes recognized by HIV-SF2 nef-specific CD4+ T-cell clones generated from HIV-1-uninfected donors.

Human T-cell clones with specificity to the HIV-1 nef protein were generated by the in vitro stimulation of peripheral blood mononuclear cells (PBMCs) from HIV-1-seronegative donors with purified nef from the HIV-SF2 isolate produced in genetically engineered yeast. Here the characterization is described of a total of seven discrete clones derived from five different donors. Each clone was CD3+ CD4+ CD8- as determined by FACS analysis. The epitopes recognized by these clones were identified using synthetic overlapping peptides spanning the entire length of nef. Six discrete helper T-cell epitopes located in five distinct regions of nef were identified by this approach. Three of these epitopes are more than 80% conserved among all HIV-1 nef proteins for which sequence data are available. The remaining epitopes are in regions of nef that vary among isolates. Many of the epitopes recognized by our clones overlap T-cell epitopes identified by others examining T-cell responses to nef in HIV-1-infected patients and immunized animals. Using partially class II-matched EBV-transformed B-cell lines, we were able to identify five different HLA class II alleles which encode restricting elements for the in vitro nef-specific proliferative response of these clones (DR1, DRw15(2), DRw6, DQw7, DP5).

Modulation of macrophage Ia expression by lipopolysaccharide: stem cell requirements, accessory lymphocyte involvement, and IA-inducing factor production.

The mechanism of induction of murine macrophage Ia expression by lipopolysaccharide (LPS) was studied. Intraperitoneal injection of 1 microgram of LPS resulted in a 3- to 10-fold increase in the number of IA-positive peritoneal macrophages (flow cytometry and immunofluorescence and a 6-to 16-fold increase by radioimmunoassay. The isolated lipid A moiety of LPS was a potent inducer of macrophage Ia expression. Ia induction required a functional myelopoietic system as indicated by the finding that the response to LPS was eliminated in irradiated (900 rads) mice and reinstated by reconstitution with bone marrow cells. Comparison of LPS-induced Ia expression in normal and LPS-primed mice revealed a faster secondary response to LPS. The memory response could be adoptively transferred to normal mice with nonadherent spleen cells prepared 60 days after LPS injection. Spleen cells prepared 5 days after LPS injection caused Ia induction in LPS-nonresponder mice; such induction was not observed in irradiated (900 rads) recipients. The cell responsible for this phenomenon was identified as a Thy-1+, immunoglobulin-negative nonadherent cell. The biosynthesis and expression of Ia were not increased by direct exposure of macrophages to LPS in vitro. Small amounts of LPS inhibited Ia induction by gamma interferon. LPS showed positive regulatory effects on Ia expression by delaying the loss of Ia expression on cultured macrophages and by stimulating the production of Ia-inducing factors. Supernatants from cultured spleen cells stimulated with LPS in vitro contained antiviral and Ia-inducing activity that was acid labile, indicating that the active factor is gamma interferon. We conclude that induction of Ia expression by LPS in vivo is a bone-marrow-dependent, radiation-sensitive process which involves the stimulation of a gamma interferon-producing accessory lymphocyte and a delay in Ia turnover.

Induction of macrophage Ia expression by lipopolysaccharide and Listeria monocytogenes in congenitally athymic nude mice.

Experiments were performed to analyze the mechanism by which lipopolysaccharide (LPS) modulates the expression of Ia by murine peritoneal macrophages in vivo. We investigated the effect of LPS on Ia expression in T cell deficient mice by using the congenitally athymic nude mouse model. Injection (i.p) of LPS into athymic (nu/nu) mice resulted in a dramatic increase in the expression and biosynthesis of Ia by peritoneal macrophages 7 days after injection. The magnitude and kinetics of this induction were equivalent to increases observed after LPS injection of euthymic (nu/+) mice. Viable Listeria monocytogenes also increased Ia expression in athymic mice, but in contrast to the induction observed in euthymic mice at 3 and 7 days after injection, increased Ia expression was not seen until 7 days. Ia induction by either LPS or L. monocytogenes in athymic mice was not due to the presence or development of mature T cell function as defined by assays for T cell mitogenesis and interleukin 2 production. We conclude that increased macrophage Ia expression by LPS and L. monocytogenes in vivo can occur in the absence of mature functioning T cells.

The antigenic and mitogenic response of murine T and B lymphocytes to soluble proteins of Listeria monocytogenes.

Solubilized constituents from Listeria monocytogenes were fractionated by various techniques including isopycnic gradient centrifugation, molecular sieve chromatography, and preparative SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Fractionated material was tested in vitro for mitogenic and antigenic activity by quantitating the proliferation of splenic lymphocytes and the interleukin production by peritoneal T cells. Fractionation by isopycnic gradient centrifugation revealed both antigenic and mitogenic material fractionating with the protein at a density of 1.3 g/ml. This characteristic density, together with the reduction of activity with trypsin treatment, defined the material as protein. This material was termed soluble listerial proteins (SLP). Fractionation of SLP by molecular sieve chromatography using Sephacryl 200 (S-200) revealed predominant antigenic and mitogenic activity in proteins of greater than 100,000 m.w. In contrast, fractionation of SLP by preparative SDS-PAGE (nonreducing conditions) showed activity in groups of proteins with m.w. of less than 76,000. This difference (S-200 vs SDS-PAGE) may indicate an aggregation or subunit composition which is disrupted by SDS. When fractionated by SDS-PAGE, antigens which induced macrophage-dependent interleukin production by Listeria-immune T cells were observed over a broad range of molecular sizes. Major groups of antigenic proteins were observed at 57,000 to 76,000 m.w., approximately 40,000 and less than 25,000 m.w. Mitogenic activity (spleen cell proliferation) was associated with a more restricted group of proteins with major peaks at 57,000 and 40,000 m.w., with some weak activity in proteins less than 20,000 and greater than 64,000 m.w. Experiments involving T or B lymphocyte-depleted spleen cells and spleen cells from athymic mice revealed that the mitogenic response of splenic lymphocytes to SLP was predominantly B cell-mediated. Thus, we have defined groups of listerial proteins with potent antigenic activity with respect to T lymphocyte activation and as mitogenic activity for B cells.

Cyclosporine inhibits macrophage-mediated antigen presentation.

The influence of cyclosporine on antigen-specific, macrophage-dependent T cell activation was analyzed in vitro. Murine T cell activation by antigens derived from Listeria monocytogenes was monitored by the production of interleukin 2. Pretreatment (2 hr, 37 degrees C) of macrophages with cyclosporine resulted in a cell population with a markedly diminished capacity to support the activation of T lymphocytes. When cyclosporine-pretreated macrophages were added to cultures of untreated T cells and antigen, the dose of cyclosporine that produced 50% inhibition (ID50) was 1.5 micrograms/ml, and if antigen was present during the drug pretreatment, the ID50 was 0.6 micrograms/ml. Pretreatment of T cells also inhibited their subsequent activation by antigen and untreated macrophages, but a higher dose of cyclosporine was required to produce similar inhibition (ID50 = 4.4 micrograms/ml). Additional experiments focused on the mechanism of inhibition of antigen presentation when macrophages were pretreated with the drug. The addition of interleukin 1 or indomethacin to the cultures did not alter the inhibitory effect of cyclosporine. Under conditions that produced greater than 90% inhibition of antigen presentation, macrophage surface Ia expression was not altered, and the uptake and catabolism of radiolabeled antigen remained normal. Thus, cyclosporine had profound effects on antigen presentation that appear to be unrelated to decreases in interleukin 1 production, increases in prostaglandin production, decreases in Ia expression, or changes in antigen uptake and catabolism.