Abstract
A conformationally biased decapeptide agonist of human C5a anaphylatoxin (YSFKPMPLaR) was used as a molecular adjuvant in stimulating an Ag-specific CTL response against murine P815S target cells expressing an Ld-restricted CTL epitope of the hepatitis B surface Ag (HBsAg). Groups of BALB/c mice (H-2d) were immunized with aqueous solutions of the HBsAg CTL epitopes (IPQSLDSWWTSL and IPQSLDSWWTSLRR); the C5a agonist (YSFKPMPLaR); the C5a agonist and HBsAg CTL epitopes admixed (IPQSLDSWWTSL and IPQSLDSWWTSLRR + YSFKPMPLaR); the C5a-active, HBsAg CTL epitope-C5a agonist constructs (IPQSLDSWWTSLYSFKPMPLaR, IPQSLDSWWTSLRRYSFKPMPLaR, and IPQSLDSWWTSLRVRRYSFPMPLaR); a C5a-inactive, reverse-moiety construct (YSFKPMPLaRRRIPQSLDSWWTSL); and a C5a-attenuated, carboxyl-terminal-blocked construct (IPQSLDSWWTSLRRYSFKPMPLaRG). Ag-specific CD8+ CTL responses were observed after the secondary boost in the absence of any added adjuvant only in mice that were immunized with C5a-active contructs, IPQSLDSWWTSLRRYSFKPMPLaR and IPQSLDSWWTSLRVRRYSFKPMPLaR. These two C5a-active immunogens contained potential subtilisin-sensitive linker sequences between the HBsAg CTL epitope and the C5a agonist; i.e., a double-Arg (RR) and a furin protease sensitive sequence (RVRR). The introduction of these potentially cleavable sequences may be a method of increasing the likelihood of liberating the CTL epitope from the C5a agonist by intracellular proteases, thereby facilitating entry of the epitope into Ag-processing pathways via an exogenous route.
The generation of CTLs is crucial to the development of an effective immune response to intracellular pathogens. This is particularly true for viral infections, where CTLs are induced by intracellular viral Ags, which are processed as peptides that associate with MHC class I determinants for their presentation on the cell surface. Immunization with exogenous viral protein or peptide Ags rarely induces MHC class I-restricted CTL responses unless the Ags are formulated with lipidic adjuvants, which facilitate Ag passage through cell membranes and into intracellular Ag-processing pathways. Such lipophilic adjuvants, like the majority of experimental adjuvants in use today, are nonspecific in their immunoenhancing effects. This is also true for alum, the only adjuvant presently approved for human use in the United States and whose activity appears restricted to promoting Th2-type Ab responses. Consequently, there is considerable interest in the development of adjuvants capable of inducing robust immune responses, but with a degree of specificity that focuses their immunopotentiating activities on the cell(s) involved in generating specific types of humoral and/or cellular immune responses. Ideally, this immunologic specificity should be the result of a single molecular component that potentiates an essential step (or steps) in Ag targeting, processing, presentation, and cellular activation, a so-called molecular adjuvant (1).
The complement system (C) plays a central role in mediating humoral and cellular responses between the innate and acquired arms of the immune system. As such, the various components of C represent potential candidates for use as molecular adjuvants for inducing immune responses via more specific pathways of activation. The C components C3b and C3d, for example, have been used as molecular adjuvants when covalently attached to a protein Ag (1, 2, 3). The resulting, enhanced Ag-specific immune responses were attributable to the ability of these C components to facilitate specific steps in Ag processing and presentation.
We recently reported on the use of a conformationally biased decapeptide agonist of the C-derived human anaphylatoxin C5a (YSFKPMPLaR) (uppercase letters designate the l stereoisomeric form of the amino acids and lowercase letters designate the d stereoisomeric form) as an effective molecular adjuvant for inducing Ag-specific Ab responses to human mucin type 1 (MUC1)3 glycoprotein (4) and the human μ- and κ-opioid receptors (4, 5). These Ag-specific responses were generated in mice and rabbits immunized with C5a-active constructs in which the B cell peptide epitope was covalently attached to the N terminus of the C5a agonist. YSFKPMPLaR was used as a molecular adjuvant in these studies because it expresses conformational features shown to be important for selective C5a receptor (C5aR) binding and activation. For example, YSFKPMPLaR is a significantly more potent and selective agonist for C5aRs expressed on resident tissue macrophages in the induction of smooth muscle contraction of human umbilical artery than it is in inducing enzyme release from C5aR-bearing neutrophils (6, 7, 8, 9). Also, the conformational features in YSFKPMPLaR responsible for these response-selective activities render it resistant to the cleavage of the biologically important C-terminal Arg by serum carboxypeptidases (10).
Results from the aforementioned studies demonstrated that C5a agonist activity was essential to the induction of the Ab responses and suggested that the adjuvant effects of the epitope-YSFKPMPLaR constructs resulted from the ability of the C5a agonist moiety to simultaneously deliver both the epitope and stimulatory signals to C5aR-bearing APCs (4, 11, 12). Upon C5aR binding, the C5a agonist moiety, like natural C5a, likely induces the synthesis and release of cytokines necessary for eliciting T cell help. Following C5aR binding and activation, the ligand-C5aR complex was predicted to internalize like the natural C5a/C5aR complex (13), enabling the entry of the attached peptide epitope into intracellular class II Ag-processing pathways and subsequent association with MHC determinants on the surface of the APC.
The results obtained with the MUC1 (4) and opioid receptor B cell epitopes (4, 5) are consistent with the mechanism described above and support the use of YSFKPMPLaR as a molecular adjuvant in inducing Ag-specific Ab responses. However, given the need for prophylactic and therapeutic vaccines that stimulate cellular responses against intracellular pathogens and tumors, we sought to determine whether YSFKPMPLaR was capable of inducing Ag-specific CTL responses against poorly immunogenic peptides. In this report, we present the results of a study in which YSFKPMPLaR was used as a molecular adjuvant for inducing Ag-specific CTL responses against a defined CTL peptide epitope from the hepatitis B surface Ag (HBsAg). The HBsAg CTL epitope was covalently attached to either the N terminus of YSFKPMPLaR (C5a-active constructs) or to its C terminus (C5a-inactive constructs). Mice were immunized with these C5a-active and C5a-inactive constructs in the absence of any added adjuvant to evaluate the ability of YSFKPMPLaR to provide the necessary signals and/or targeting required for the induction of an Ag-specific CTL response. Immunizations were also performed with C5a-active constructs containing protease-sensitive linker sequences between the HBsAg CTL epitope and the C5a agonist. The results of this study are discussed in terms of a possible mechanism by which YSFKPMPLaR induces Ag-specific CTL responses and the importance of a protease-sensitive sequence between the epitope and the C5a agonist.
Materials and Methods
Peptide synthesis
Peptides were synthesized by standard solid phase methodologies on an Applied Biosystems (Foster City, CA) model 430A synthesizer. Syntheses were performed on a 0.25-mmol scale and employed the 9-fluorenylmethyloxycarbonyl method of repetitive residue linkages. Peptide purification was accomplished with analytical and preparative HPLC on columns packed with C18-bonded silica. All peptides were characterized by amino acid compositional analysis and mass spectrometry. The details of these methods of synthesis, purification, and characterization have been described previously (6).
Pharmacologic assays
C5a-like agonistic activity was assessed by the ability of the peptides to induce smooth muscle contraction of human umbilical artery and myeloperoxidase (MPO) release from human polymorphonuclear leukocytes (PMNs) according to previously published methods (6, 7, 8). Full concentration-response curves were generated for individual peptides and natural C5a in each assay and the EC50 values were calculated. pD2 transforms [−log EC50 (M)] were calculated for each concentration-response curve and reported as the mean ± SE. Peptide binding affinity to the C5aR was evaluated on intact human PMNs by a competition assay using [125I]-C5a according to previously described methods (7, 8). Statistical analysis of the values obtained from pharmacologic assays was performed using one-way ANOVA.
Animals
Female BALB/c (H-2d) mice 10–12 wk old were purchased from The Jackson Laboratory (Bar Harbor, ME). On arrival, mice were housed in isolator cages, provided autoclaved food and water ad libitum, and quarantined for 7 days. On release from quarantine, the mice were entered into the study following written protocols on file with the Animal Care and Use Committee and in compliance with the Animal Welfare Regulations (9CFR).
Immunization protocols
Groups of six to eight BALB/c mice were given bilateral s.c. injections in the inguinal area with (100 μl each side) of a PBS solution containing 50–100 μg of the peptides. Booster injections were given s.c. in the inguinal region at 21-day intervals. Sera for Ab analysis were obtained by retro-orbital bleeds of mice under CO2 narcosis.
Spleen cell cultures and in vitro stimulation
Three weeks after secondary (2°) immunization and 2 weeks after tertiary (3°) immunization, two to three mice from each experimental group were sacrificed by cervical dislocation and their spleens removed aseptically. Pooled splenic single-cell suspensions were prepared in RPMI 1640 medium (Sigma, St. Louis, MO) containing 10% FBS (HyClone, Salt Lake City, UT) and the following supplements: 10 mM HEPES buffer, 50 mM 2-ME, 2 mM l-glutamine, 50 μg/ml gentamicin sulfate, 100 U/ml penicillin, and 50 μg/ml streptomycin (all supplements from Sigma). This supplemented complete medium is designated RP10-SC. For culture, 75 × 106 pooled spleen cells in 5 ml of RP10-SC were pipetted into a 25 cm2 T-flask (Corning, Corning, NY). Next, 5 ml of RP10-SC containing 150 nM of synthetic, Ld MHC class I-restricted peptide, IPQSLDSWWTSL, were added to the flask. The flasks were incubated undisturbed in an upright position at 37°C in 5% CO2. After 4 days of incubation, the cells were recovered from the T-flasks, washed once by centrifugation in fresh RP10-SC, resuspended in 5 ml RP10-SC, counted, and adjusted to 5 × 106 viable cells/ml.
Target cell lines
The specific cell target used for measuring CTL activity was P815S, a transfectant cell line of P815 (H-2d) expressing the HBsAg (14). P815S was grown in RP10-SC medium containing 400 μg/ml geneticin disulfate (G418; Sigma). The parental H-2d mastocytoma cell line, P815 (TIB64; American Type Culture Collection, Manassas, VA) grown in RP10-SC medium, was used as a nonspecific target for CTL assays to measure percent nonspecific lysis. In all the experiments shown, this value was <5% at E:T ratios of 50:1.
Target cell labeling
The target cells, either P815S or P815, were washed two times in RP10-SC. For labeling, 5 × 106 target cells were mixed with 250 μCi 51Cr-labeled sodium chromate [400–1200 Ci (14.8–44.4 TBq)]/g; NEN DuPont, Boston, MA) in a 1.0-ml volume in a 50-ml conical tube and incubated in a 37°C water bath for 90 min. After incubation, the labeled cells were washed three times by centrifugation using 15-ml volumes of fresh RP10-SC and allowed to stand at room temperature for 30 min. The pelleted cells were resuspended in RP10-SC at 1 × 105 cells/ml.
Cytotoxicity assay
The recovered splenic effector cells at 5 × 106 cells/ml were serially diluted in triplicate in wells of round-bottom 96-well plates (25850; Corning) in a total volume of 100 μl/well and using RP10-SC as the diluent. Next, 100 μl volumes of 51Cr-labeled targets, P815S or P815, at 1 × 105 cells/ml were added to the wells. Maximum release (MR) wells contained 100 μl of target cells and 100 μl of 2% (v/v) Tween 20, while spontaneous release (SR) wells contained labeled cells in medium alone. E:T ratios of 50:1, 25:1, 12.5:1, and 6.25:1 were routinely employed. The plates were centrifuged at 400 × g and incubated at 37°C in 5% CO2 for 4 h. After incubation, the supernatant fractions in the wells were collected using a Skatron Supernatant Collection System (Skatron Instruments, Sterlin, VA). The amount of 51Cr radioactivity in the supernatant fractions was measured using a Wallac 1470 Wizard gamma counter (Turku, Finland). Percent specific lysis was calculated as [(experimental release − SR/(MR − SR)] × 100. The SR was always <10% of the MR. All assays were performed in triplicate.
Results
Peptide design
Peptide immunogens were designed to evaluate the requirement for C5a agonist activity in the induction of Ag-specific CTL responses. C5a-active constructs were generated by the covalent attachment of the HBsAg CTL epitope to the N terminus of the C5a agonist (IPQSLDSWWTSLYSFKPMPLaR, IPQSLDSWWTSLRRYSFKPMPLaR, and IPQSLDSWWTSLRVRRYSFKPMPLaR). This positioning of the CTL epitope relative to the C5a agonist leaves the biologically important conformational features expressed in the C-terminal region of YSFKPMPLaR free to interact with C5aRs expressed on the cells involved in Ag uptake and processing. The latter two peptides were designed to evaluate if predicted protease-sensitive linker sequences placed between the HBsAg CTL epitope and the C5a agonist might facilitate intracellular release of the epitope into Ag-presenting pathways and thereby enhance the response. The linkers consisted of a dibasic double-Arg (RR) sequence, which is susceptible to cleavage by proteases of the subtilisin family (15) and trypsin-like proteases. The other was a sequence sensitive to the ubiquitous intracellular subtilisin-like protease furin, RVRR (16). This latter sequence is found at the junction of the A and B fragments of diphtheria toxin and it is believed that furin plays a prominent role in the intracellular proteolytic activation of diphtheria toxin and several other bacterial toxins as well as in the processing of proproteins and prohormones that contain the consensus sequence RX(K/R)R (16, 17). C5a-inactive constructs were generated by blocking the functionally important carboxyl group on the C-terminal Arg of YSFKPMPLaR with either the HBsAg CTL epitope (YSFKPMPLaRRRIPQSLDSWWTSL) or with a Gly residue (IPQSLDSWWTSLRRYSFKPMPLaRG) (4).
Pharmacologic activities
All peptides were evaluated for C5a agonist activities in assays that measured peptide-mediated contraction of smooth muscle in human umbilical artery (Table I⇓), the release of MPO from human PMNs (Table II⇓), and binding to C5aRs expressed on the surface of human PMNs (Table III⇓). Constructs in which the HBsAg CTL epitope was attached to the N terminus of the C5a agonist (IPQSLDSWWTSLYSFKPMPLaR, IPQSLDSWWTSLRRYSFKPMPLaR,and IPQSLDSWWTSLRVRRYSFKPMPLaR) behaved as full agonists relative to natural C5a with potencies and C5aR binding affinities comparable to or greater than YSFKPMPLaR. In contrast, the construct in which the functionally important C-terminal carboxyl group of the C5a agonist moiety was blocked with the HBsAg CTL epitope (YSFKPMPLaRRRIPQSLDSWWTSL) was significantly less potent in umbilical artery contraction (Table I⇓), inactive in MPO release from PMNs (Table II⇓), and bound poorly to the C5aR (Table III⇓) relative to both natural C5a and YSFKPMPLaR. Similarly, the construct in which a Gly residue blocked the C-terminal carboxyl group of the agonist moiety (IPQSLDSWWTSLRRYSFKPMPLaRG) was significantly less potent than its C5a-active counterparts in umbilical artery contraction (Table I⇓) and bound with significantly less affinity to the C5aR (Table III⇓). This construct was unable to induce a full response relative to natural C5a in MPO release from PMNs (Table II⇓).
Immunogen activity in smooth muscle contraction of human umbilical artery
Immunogen activity in MPO release from human PMNs
Immunogen binding affinity for C5aRs on human PMNs
CTL responses are induced in mice when the Ld MHC class I-restricted peptide, S28–39, of HBsAg is covalently attached to the N terminus of the C5a agonist via an Arg-Arg linkage
Initial experiments were designed to evaluate CTL induction by the free Ld MHC class I-restricted peptide S28–39 (IPQSLDSWWTSL), the same peptide with two Arg residues added to the C terminus (IPQSLDSWWTSLRR), the free C5a agonist (YSFKPMPLaR), and admixtures of the above peptides. Of particular interest was the evaluation of C5a-active constructs in which the HBsAg CTL epitope was covalently attached either directly to the N terminus of the C5a agonist (IPQSLDSWWTSLYSFKPMPLaR) or through the double-Arg, protease-sensitive linker (IPQSLDSWWTSLRRYSFKPMPLaR). The administered amounts of the latter two constructs were adjusted to reflect amounts equal (by weight) to that of the free HBsAg CTL epitope based on relative molecular mass. Mice in each group received two injections, spaced 21 days apart, of the indicated construct. Mice in each group were tested for splenic CTL activity at day 42 as described in Materials and Methods. As shown in Table IV⇓, only the group injected with the double-Arg-linked construct (IPQSLDSWWTSLRRYSFKPMPLaR) exhibited a significant CTL response against the P815S-transfected target cells.
Percent specific lysis of 51Cr-labeled P815S target cells from mice immunized with various HBsAg CTL epitope/C5a agonist peptides
C5a agonist activity is necessary for the induction of Ag-specific CTL responses
To evaluate the necessity of C5a agonist activity in the induction of Ag-specific CTL responses, mice were immunized with C5a-active and C5a-inactive HBsAg CTL epitope-containing constructs. C5a-active constructs were generated by the covalent attachment, via the protease-sensitive, double-Arg linker sequence, of the HBsAg CTL epitope to the N terminus of the C5a agonist (IPQSLDSWWTSLRRYSFKPMPLaR). C5a-inactive constructs were generated by blocking the functionally important carboxyl group on the C-terminal Arg of the C5a agonist with either the HBsAg (YSFKPMPLaRRRIPQSLDSWWTSL) or a Gly residue (IPQSLDSWWTSLRRYSFKPMPLaRG). The double-Arg-containing HBsAg CTL epitope (IPQSLDSWWTSLRR) was used as a control. As shown in Fig. 1⇓, only mice that were immunized with the C5a-active construct (IPQSLDSWWTSLRRYSFKPMPLaR) generated an Ag-specific CTL response. The phenotype of the effector cells responsible for the in vitro cytolytic activity in these experiments was determined to be CD8+ as judged by the ability of a rat anti-mouse lyt 2.2 mAb (2.43, TIB210; American Type Culture Collection) (18) to almost completely inhibit (>90% inhibition) the cytolytic process. In contrast, a rat anti-mouse L3T4 mAb (GK1.5, TIB207; American Type Culture Collection) known to block CD4+ T cell activity (19) had no effect on the in vitro cytolysis induced by the active constructs (data not shown).
HBsAg-specific CTL response induced only by C5a-active constructs. BALB/c mice received three s.c. injections at 21-day intervals using 50 μg doses of the HBsAg-Ld MHC class I-restricted peptide (S28–39) synthesized with two Arg residues appended to the C-terminal end (IPQSLDSWWTSLRR), the double-Arg-linked C5a-active construct (IPQSLDSWWTSLRRYSFKPMPLaR), and the double-Arg-linked C5a-inactive constructs (IPQSLDSWWTSLRRYSFKPMPLaRG and YSFKPMPLaRRRIPQSLDSWWTSL). Pooled spleen cell suspensions were prepared from two mice in each group 14 days following the third injection. The cell suspensions were cultured in the presence of the S28–39 peptide (75 nM) for 4 days and used as effector cells against 51Cr-labeled P815S (HBsAg transfectant) or P815 targets. Percent specific lysis of 51Cr-labeled P815S cells at various E:T ratios is shown and represent the means of triplicate determinations. Lysis of 51Cr-labeled P815 cells at an E:T ratio of 50:1 was <5% (not shown). These data are representative of three separate experiments.
CTL responses are induced only by C5a-active constructs containing a protease-sensitive linker sequence between epitope and C5a agonist
To evaluate the requirement for a protease-sensitive linkage between the HBsAg CTL epitope and the C5a agonist for CTL induction, mice were immunized with C5a-active constructs in which the HBsAg CTL epitope was covalently attached directly to the N terminus of the C5a agonist (IPQSLDSWWTSLYSFKPMPLaR) or separated by protease-sensitive linker sequences (IPQSLDSWWTSLRRYSFKPMPLaR and IPQSLDSWWTSLRVRRYSFKPMPLaR). As noted previously, the double-Arg (RR) sequence is sensitive to cleavage by proteases of the subtilisin family and other trypsin-like proteases (15). The RVRR sequence is a motif recognized by the intracellular protease furin (16). The results shown in Fig. 2⇓ indicate that of the three C5a-active constructs, only those containing the protease-sensitive linker sequence between HBsAg CTL epitope and the C5a agonist were capable of inducing an Ag-specific CTL response in mice. Although the RVRR-containing construct was able to elicit a specific CTL response that was significantly above background levels, the response did not exhibit an enhanced magnitude of lysis or more rapid kinetics of induction when compared with that of the RR-containing construct (data not shown). This finding suggests that intracellular furin likely plays a less significant role, or perhaps no role, in the processing of the epitope-C5a agonist construct than other intracellular proteases with specificity for basic or dibasic residues.
HBsAg-specific CTL responses induced only by the C5a-active, protease-sensitive-linked constructs. BALB/c mice received three s.c. injections at 21-day intervals using 50 μg doses of the C5a-active constructs in which the HBsAg-Ld MHC class I-restricted peptide (S28–39) was covalently attached directly to the N terminus of the C5a agonist (IPQSLDSWWTSLYSFKPMPLaR), spaced by two Arg residues (IPQSLDSWWTSLRRYSFKPMPLaR), or spaced by the furin protease-sensitive sequence RVRR (IPQSLDSWWTSLRVRRYSFKPMPLaR). Spleen cell suspensions were prepared from two mice in each group 14 days following the third injection. The cell suspensions were cultured in the presence of the S28–39 peptide (75 nM) for 4 days and used as effector cells in 51Cr-release assays against P815S or P815 targets. Percent specific lysis of 51Cr-labeled P815S is shown and represent the means of triplicate determinations. Lysis of 51Cr-labeled P815 cells by these effector cells was <5% at an E:T ratio of 50:1 (not shown).
Abs to the HBsAg CTL epitope or the C5a agonist are not produced by immunization with the HBsAg CTL epitope-C5a agonist constructs
Previous studies showed that immunization of mice in the presence of additional adjuvant with MUC1 epitope-C5a agonist (4) and opioid receptor epitope-C5a agonist constructs (5) induced Ab responses to the MUC1 and opioid receptor epitopes and the full-length, intact proteins. Thus, it was of interest to determine whether sera from mice immunized with the control peptides and epitope-C5a agonist constructs contained Ab directed against any of the peptides. Mice were bled at various times following injection and the sera from all groups were tested by ELISA for reactivity with YSFKPMPLaR, IPQSLDSWWTSL, IPQSLDSWWTSLRR, and IPQSLDSWWTSLRRYSFKPMPLaR. These analyses failed to show binding to either the C5a agonist or the CTL epitopes (ODs equal to normal mouse serum at a 1:50 dilution of the serum). However, sera taken from mice following three injections of IPQSLDSWWTSLRRYSFKPMPLaR yielded an ELISA titer of 1:1600 against the immunizing peptide, but did not bind the free peptides IPQSLDSWWTSLRR or YSFKPMPLaR. These results suggest that the C5a-active, RR-containing construct contributes to the formation of a neo-B cell epitope that is presented via the class II pathway.
Discussion
The results of this study further support the ability of the conformationally biased C5a agonist YSFKPMPLaR to serve as an effective molecular adjuvant, in this case by inducing Ag-specific CTL responses against a well-defined T cell epitope derived from the HBsAg. The CTL responses were CD8+ and were observed only in mice that were immunized with C5a-active constructs in which the HBsAg CTL epitope was covalently attached to the N terminus of YSFKPMPLaR (i.e., IPQSLDSWWTSLRRYSFKPMPLaR and IPQSLDSWWTSLRVRRYSFKPMPLaR). This arrangement leaves the biologically important conformational features in the C-terminal region of YSFKPMPLaR free to interact with C5aRs expressed on the cells involved in the immune response and underscores the necessity of C5a agonist activity in the generation of the observed CTL responses. However, the presence of C5a agonist activity in the epitope-C5a agonist constructs alone was not sufficient in generating a HBsAg-specific CTL response. This was indicated by the lack of a CTL response in mice immunized with IPQSLDSWWTSLYSFKPMPLaR, despite the fact that this construct behaved as a full agonist of C5a. It is noteworthy that this C5a-active construct lacked a protease-sensitive linker sequence separating the epitope moiety from the C5a agonist moiety that was present in the two C5a-active constructs that generated a CTL response—either the double-Arg (RR) or the furin protease-specific sequence (RVRR). That CTL responses were observed only in mice immunized with C5a-active constructs that contained these protease-sensitive sequences between the epitope and C5a agonist moieties supports the concept that during the internalization of the C5aR/ligand complex an intracellular cleavage event may separate the epitope from the agonist to facilitate the entry of the epitope into intracellular Ag presentation pathways. It should be noted that the failure of the free HBsAg peptide epitope to elicit a CTL response could be attributable to degradation after injection and internalization. Thus, it might be considered that the attachment of the epitope peptide to the N terminus of the RR-containing C5a agonist peptide might, in part, reduce the sensitivity of the epitope to degradative effects. Such stabilization of the CTL epitope could, in part, contribute to the effectiveness of the colinear RR-containing constructs. However, it is unlikely that such a phenomenon represents the sole mechanism involved because blocking the C terminus of the C5a agonist moiety in these constructs abrogates their ability to elicit a CTL response, thereby indicating an essential role for C5a agonist activity in the observed responses.
It is also noteworthy that the C5a-active constructs IPQSLDSWWTSLRRYSFKPMPLaR and IPQSLDSWWTSLRVRRYSFKPMPLaR induced robust CTL activity after a 2° boost in the absence of any added adjuvant. This observation suggests that the C5a agonist moiety is capable of eliciting the T cell help necessary to induce the observed CD8+ CTL response. It is likely that this T cell involvement emanates from the ability of the C5a agonist moiety to induce the release of immunopotentiating cytokines from C5aR-bearing APCs with which the epitope-C5a agonist constructs interact. This supposition is supported by the fact that C5a has been shown to induce the synthesis and release of IL-1β, IL-6 (20), IL-8 (21), and IL-12 (E. L. Morgan, unpublished observations) from human monocytes and IL-1β, IL-6, IL-8, IL-12, TNF-α, and IFN-γ from human dendritic cells (E.L. Morgan, unpublished observations). Therefore, the C5a agonist moiety appears capable of both targeting the attached epitope to C5aR-bearing APCs and eliciting the appropriate immunopotentiating activity. Thus, the YSFKPMPLaR moiety of the constructs used in these immunizations can be viewed as a molecular entity that embodies adjuvant properties characteristic of both a “targeting vehicle” and an “immunomodulator.” Finally, mice immunized with YSFKPMPLaR-containing constructs displayed no outward physical signs that would be characteristic of a C5a-mediated anaphylactic response. This in vivo use of YSFKPMPLaR and lack of associated toxicity is consistent with the response-selective activities that have been observed in vitro (6, 7, 8).
The results described herein are consistent with a mechanism that was proposed for the induction of Ag-specific Ab responses induced by B cell epitope-YSFKPMPLaR constructs (4). The YSFKPMPLaR moiety of the HBsAg constructs interacts with C5aRs expressed on the surface of APCs to induce the synthesis and release of cytokines that activate T cells. Following C5aR activation and cytokine release, the C5aR/ligand complex internalizes, allowing intracellular proteases to separate the HBsAg epitope from the C5a agonist by cleaving at the double-Arg (RR) or furin-specific sequence (RVRR) that separate these two moieties. The HBsAg epitope then associates with MHC class I determinants that are subsequently expressed on the APC surface. While this mechanism of CTL induction by the C5a agonist-containing constructs remains speculative, it may involve a novel pathway of exogenous MHC class I Ag presentation such as that described by Rock and colleagues using both in vitro and, more recently, in vivo systems (22, 23). Because it had been generally assumed that class I-mediated Ag presentation involved the generation of peptides from endogenously synthesized proteins, the finding that extracellular soluble proteins could be taken up by professional phagocytes (macrophages and dendritic cells) processed in the cytoplasm or perhaps endosomes to yield antigenic peptides, which are presented in association with MHC class I molecules, is of considerable significance. As emphasized recently by Raychaudhuri and Rock (24), presentation of extracellular Ags would be expected to be most efficient when they are particulate in nature and, consequently, are more susceptible to phagocytosis by macrophages and dendritic cells. In the case of the C5a agonist constructs, we can postulate that targeting to the C5aR on such cells might accomplish a similar enhancement of presentation in the MHC class I pathway. Such a proposal seems especially tenable in light of the recent findings that subunits of several bacterial toxins, especially anthrax toxin, when coupled to protein and peptide Ags, are capable of effecting internalization of the Ags and delivering them into the class I presentation pathway with resultant Ag-specific CTL production (25, 26, 27, 28). Finally, it is noteworthy that a proteosome-independent, furin-dependent viral Ag-processing pathway where cleavage occurs in the Golgi or post-Golgi secretory pathway has been recently described (29). Again, this finding suggests that intersection of internalized Ags/peptides with elements of the anterograde secretory pathway (endosomal or trans-Golgi region), as may occur with C5aR/ligand complexes, could result in processing events and association with unoccupied class I molecules that are in transit through this pathway.
Although an exogenous pathway of intracellular processing of the epitope-RR-C5a agonist constructs appears to be a plausible mechanism, it is also possible that the HBsAg epitope peptide is introduced onto MHC class I determinants expressed on the surface of the APC. Thus, the colinear peptides containing the RR-C5a agonist moiety could bind to the surface of APCs and, after proteolytic cleavage of the scissile linkage, the HBsAg CTL epitope could displace lower-affinity endogenous peptide in cell-surface class I molecules. At present, we cannot assess the relative contribution of extracellular and intracellular processing events in CTL induction mediated by the C5a agonist moiety. Further in vitro experiments are being designed to address this issue.
In contrast to formulating peptide and/or protein Ags as particulates or as toxin subunit-conjugates or other derivatives, the C5a agonist peptide, rather than serving as an inert carrier, might provide the added benefit of delivering immunopotentiating signals. Accordingly, such constructs, containing either covalently linked peptides or proteins, might be of particular benefit in those situations where the target proteins or peptides are nominally immunogenic irrespective of the delivery vehicle or construct employed. In the case of peptides, a further advantage of this technology is that the C5a agonist constructs are relatively simple to produce and do not require recombinant technologies and associated protein purification methodologies or specialized formulation procedures.
The Ag-specific responses to well-defined T cell and B cell epitopes observed in our studies support the potential use of YSFKPMPLaR and other response-selective C5a agonists as molecular adjuvants for inducing a defined spectrum of humoral and/or cellular responses against peptide, protein, and, possibly, nonprotein Ags. Such a possibility would provide a broad-based adjuvant/delivery technology that would be applicable to a number of infectious and oncologic diseases in either prophylactic or therapeutic settings.
Acknowledgments
We thank Dr. Fulvio Perini of the Eppley Institute Molecular Biology Core Laboratory for amino acid compositional analyses, Mr. Jon Ward of Corixa for expert technical assistance with immunizations and CTL assays, and the Protein Structure Core Facility, Department of Biochemistry, University of Nebraska Medical Center for mass spectrometry analyses.
Footnotes
-
↵1 This work was supported by grants from the National Institutes of Health Care (CA36727), Corixa, and the National Health and Medical Research Council (to S.M.T.).
-
↵2 Address correspondence and reprint requests to Dr. Sam D. Sanderson, Eppley Institute for Research in Cancer and Allied Diseases, 600 South 42nd Street, 986805 Medical Center, Omaha, NE 68198-6805. E-mail address: sdsander{at}unmc.edu
-
↵3 Abbreviations used in this paper: MUC1, human mucin type 1; PMN, polymorphonuclear leukocytes; MPO, myeloperoxidase; HBsAg, hepatitis B surface Ag; MR, maximum release; SR, spontaneous release.
- Received September 20, 1999.
- Accepted March 1, 2000.
- Copyright © 2000 by The American Association of Immunologists
References
In this issue
