INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Cholera is an acute diarrheal disease caused by the gram-negative bacterium Vibrio cholerae. The major secreted or surface-expressed virulence factors of V. cholerae are cholera toxin and toxin-coregulated pilus (TCP) (7; reviewed in reference 9). TCP is a type 4 pilus composed of a homopolymer of 20.5-kDa TcpA pilin subunits that mediate V. cholerae colonization (8, 11, 20, 22). The TCP is allelic, i.e., there are two predominant biotype-specific derivatives of TcpA (classical and El Tor), which differ by 18.1% at the protein level in the region defined by amino acids 145 to 199 (15). TcpA and peptides derived from it are immunogenic and induce protective antibodies (Abs) when not delivered in the context of a natural infection or vaccination with intact bacteria (7, 16-19). Anti-TCP Abs when mixed with virulent V. cholerae (500 times the 50% lethal dose [LD₅₀]) and fed to infant mice provide almost complete immunity to cholera (16, 17). The regions of classical TcpA that neutralizing Abs are directed against have been partially defined. Results of several experimental approaches have indicated that domains within the C-terminal region of TcpA (amino acids 145 to 199) delineated by a single disulfide bond can induce the protective Ab response seen in animals (16, 17). Peptides Cl-4, Cl-5, and Cl-6 induce immune responses in mice that can protect 50 to 89% of infant mice against a 100LD₅₀ challenge in the infant mouse cholera model (17). Recent results (24) have further demonstrated that TcpA-derived peptides and new experimental polymer adjuvants induce Ab responses in female mice that protect their pups from V. cholerae infection. The previous studies were focused on the classical biotype of TCP. TCP of El Tor biotypes has not been analyzed for protective epitopes of peptides 4, 5, and 6. Therefore, it is of interest to compare the immunogenicities of peptides 4, 5, and 6 of the two TCP biotypes.

The effects of the immune response (Ir) gene were first described in 1965 by McDeveitt and Sella (12). They showed that a branched polymer composed of a backbone of L-lysine with side chains of DL-alanine conjugated randomly with tyrosine-glutamate when emulsified in Freund's complete adjuvant (FCA) was able to induce a robust Ab response in C57BL/6 mice (H-2^b) and only a low Ab response in CBA mice (H-2^k). This and similar phenomena were described as effects of the Ir gene. Interestingly, the (C57BL/6 × CBA)F₁ mice made intermediate responses. These results, along with results from an F₁ × C57BL/6 cross that showed a predominantly high response, suggested that a single gene was controlling the serologic response. Long-term studies using congenic mice clearly implicated the H-2 locus between K and Ss as containing the genes that controlled this phenomenon. Subsequent studies have shown that the major histocompatibility complex class II (MHC-II) molecules are critical for induction of Ab responses and that it is the associated genes in the H-2 locus that mediate the effects of the Ir gene. In the past 35 years, scientists have defined the role of class II in the induction of CD4⁺ T-cell help for Ab formation. We know that binding peptide antigens (a fragment of the B-cell-internalized antigen) in the class II binding cleft is critical for B-cell presentation of the complex to CD4⁺ T helper cells. The allelic variations in class II molecules among the H-2 types dictate the ability to bind peptides and thus the capacity to activate antigen-specific CD4⁺ T cells (2, 5, 14, 23). There is a compelling data set that indicates that the peptide and H-2 sequences must be optimized for effective presentation to induce T-cell help.

If subunit vaccines based on protein sequences are to be useful, they must be able to induce responses in a large percentage of the target population. Classical peptides 4, 5, and 6 are 24 to 26 amino acids in length. Thus, they are long enough to contain both B-cell epitopes (binds surface immunoglobulin to select a B cell) and T-cell epitopes (presented by class II to provide T-cell help). This has been demonstrated for peptides 4 and 6 in CD-1 mice, which are outbred mice of the H-2^q haplotypes (24). The issue at hand is to determine if the serologic and protective responses to peptides 4, 5, and 6 are indicative of universal T-cell epitopes for class II molecules or whether the effects of the Ir gene will be evident from limited amino acid sequences. In addition, the allelic differences between Cl and El Tor TCP biotypes could complicate a simple assignment of reactivity based on H-2 type.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Three- to 5-week-old female congenic mice were purchased from Jackson Laboratory (Bar Harbor, Maine). At 8 weeks, 4 to 6 mice per haplotype group were immunized as described below. The congenic mouse strains used were as follows: H-2^k, B10.A-H2^a H2-T18^a/SgSnJ; H-2^b, C.B10-H2^b/LilMcdJ; H-2^d, B10.D2-H2^d H2-T18^c Hc ¹/nSnJ, and H-2^k, B10.BR-H2^k H2-T18^a/SgSnJ. All mice were housed under standard conditions in the Animal Resources Center located at the Dartmouth-Hitchcock Medical Center (Lebanon, N.H.) and maintained on a basic diet of Harlan (Madison, Wis.) Teklad sterilizable rodent feed.

Materials and reagents. FCA and Freund's incomplete adjuvant (FIA) were purchased from Sigma-Aldrich (St. Louis, Mo.). Phosphate-buffered saline (PBS), blocking buffer (1× PBS, 1% bovine serum albumin, 0.05% Tween 20), wash buffer (1× PBS, 0.05% Tween 20), binding buffer (0.1 M Na₂HPO₄, pH 9.0), and stop solution (0.18 M H₂SO₄) were all prepared in-house using chemicals purchased from Fisher Scientific (Pittsburgh, Pa.) and Sigma. Methoxyflurane (Metofane) was manufactured by Schering-Plough Animal Health Corp. (Union, N.J.). Heparinized microhematocrit capillary tubes were purchased from Fisher Scientific. Seal-Rite 1.5-ml Natural microcentrifuge tubes and 0.5-ml self-standing microcentrifuge tubes with O-ring caps were purchased from USA Scientific, Inc. Perfektum 5-ml glass needle-lock tip syringes and Popper 20-gauge microemulsifying needles were purchased from Fisher Scientific. 3,3',5,5'-Tetramethylbenzidine (TMB) peroxidase substrate was purchased from Kirkegaard & Perry Laboratories (Gaithersburg, Md.). All horseradish peroxidase-conjugated secondary Abs were purchased from Southern Biotechnology Associates (Birmingham, Ala.). One-milliliter Luer-Lok latex-free syringes and 22-gauge PrecisionGlide needles used to deliver the immunogens were manufactured by Becton Dickinson and Co. and were purchased from the local stockroom. Falcon PRO-BIND 96-well U-bottom assay plates used to dilute antisera for enzyme-linked immunosorbent assay (ELISA) were purchased from Fisher Scientific.

Peptides. TcpA fragment peptides 4, 5, and 6, derived from sequences of either classical or El Tor biotypes of V. cholerae, were commercially prepared by Macromolecular Resources (Colorado State University, Fort Collins, Colo.). The peptides used in single letter amino acid code are as follows: ET-4, ADLGDFETSVADAATGAGVIKSIA; ET-5, AATGAGVIKSIAPGSANLNLTNITH; ET-6, LNLTNITHVEKLCTGTAPFTVAFGNS; Cl-4, ADLGDFENSAAAAETGVGVIKSIA; Cl-5, AETGVGVIKSIAPASKNLDLTNITH; Cl-6, LDLTNITHVEKLCKGTAPFGVAFGNS. The relative positions of these peptides are shown in Fig. 1. The lyophilized peptides were resuspended in 5.0% dimethyl sulfoxide-water-0.25× PBS to a final concentration of 0.5 mg/ml and stored at 80°C in 1.0-ml aliquots until used.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Sequences and location of the synthetic peptides within the context of amino acids 145 to 199 of classical and El Tor TcpA. Peptides 4, 5, and 6 were synthesized based on the predicted amino acid sequences from cloned classical and El Tor tcpA genes. Dashes, amino acid sequences that are shared. Peptide lengths are as follows: 24-mer for Cl-4 and ET-4, 25-mer for Cl-5 and ET-5, and 26-mer for Cl-6 and ET-6. Underlined sequences are shared between peptides 4 and 5, and 5 and 6 for the same biotype.

Immunization and serum collection. Peptide antigen emulsions were prepared by diluting peptides in 1× PBS and then mixing with adjuvant in a 1:1 ratio. Mice were administered peptides (25 µg/mouse) via intraperitoneal injection (200 µl/mouse) with FCA for the primary inoculation (day 0) and with FIA for the secondary and tertiary inoculations (days 14 and 28). Preimmune sera were collected 1 to 2 weeks before the primary immunization. The primary-response sera were collected at day +14, and the secondary- and tertiary-response sera were collected at days +28 and +42, respectively.

Serology. Costar 96-well, high-binding, flat-bottom microtiter plates (Corning Incorporated Life Sciences, Acton, Mass.) were coated with 0.5 µg of peptides in 100 µl of 0.1 M Na₂HPO₄, pH 9.0, overnight at 4°C. Plates were washed four times using a Molecular Devices (Sunnyvale, Calif.) Skan Washer 400 microplate washer with 250 µl of 1× PBS-0.05% Tween 20. Nonspecific binding was blocked using 200 µl of 1× PBS-0.05% Tween 20-1% bovine serum albumin for 1 h at room temperature. Plates were washed four more times, and 50 µl of serially twofold-diluted antiserum was added to each well and incubated at room temperature for 2 h and then overnight at 4°C. The initial dilutions were 1:500 for preimmune and primary sera, 1:1,000 for secondary sera, and 1:2,000 for tertiary sera, except for Cl-5 tertiary sera, for which a dilution of 1:2,500 was used for the analysis. Plates were washed six times, and 50 µl of horseradish peroxidase-labeled goat anti-mouse IgG1 (1 chain-specific) or IgG_2a (2a chain-specific) detector antibodies (diluted 1:4,000) was added to each well and incubated at room temperature for 2 h protected from light. Plates were washed eight times and were then developed with 100 µl of 3,3',5,5'-tetramethylbenzidine peroxidase substrate for 15 to 30 min at room temperature. The reaction was stopped with an equal volume of 0.18 M H₂SO₄. Optical densities were read using a Dynex Technologies MRX microplate reader (Thermo Labsystems, Helsinki, Finland) using Dynex Revelation, version 3.04, software at 450 nm with 630 nm as the reference wavelength.

Data analysis and statistics. All calculations and statistical analyses were performed with Microsoft Excel, version 9.0, and GraphPad Prism, version 3.0. The end point titer was defined as the reciprocal of the dilution for the last positive well for each sample after subtracting the background. Background values were defined as twice the mean optical density for all blank wells on a single microtiter plate. Wells containing the lowest dilution for prebleed samples were chosen to be blank background wells. One blank well was selected for each mouse on an individual plate.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

We selected four congenic strains of mice that represent the three most studied class II H-2 regions: H-2^b, H-2^k, and H-2^d (Fig. 2). B10.A, B10.D2, and B10.BR mice express both class II isotypes, I-A and I-E, while C.B.10 mice express I-A only. We anticipated that immunization with TcpA peptide 4, 5, or 6 would be immunogenic in mice of some of the haplotypes as the peptides have previously been shown to provide for both B- and T-cell activation in CD-1 mice (H-2^q).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Congenic mouse genotypes in the H-2 locus. H-2 congenic mice were purchased from Jackson Laboratory, and the genotypes are presented as described in the supplier material. The number of backcrosses (BC) for each strain is shown, as is the percent identity of the genome based on the number of backcrosses. B10.A mice are H-2a; C.B.10 mice are H-2b; B.10.D2 mice are H-2d; B.10Br mice are H-2k. Differences between H-2a and H-2k mice are in italics.

Cl-4 and ET-4. Following one intraperitoneal immunization of peptides in FCA and two with peptides in FIA, sera representing primary, secondary, and tertiary responses were collected and analyzed by ELISA to determine the anti-Cl-4 and anti-ET-4 titers (Fig. 3A [IgG1] and B [IgG2a]). The homologous systems were evaluated, e.g., the ET-4 peptide reacted with ET-4-specific sera. H-2^a, H-2^d, and H-2^k mice immunized with Cl-4 responded well, making IgG1-specific Abs throughout the time course of immunization (Fig. 3A). H-2^b mice were nonresponders to Cl-4, but certain mice in the group responded significantly in individual bleeds to ET-4. The IgG1 responses of H-2^d mice to ET-4 were modest, a finding which differed from that reported for H-2^d mice immunized with Cl-4.

View larger version (25K): [in this window] [in a new window]   FIG. 3.   Comparison of ELISA titers of classical and El Tor anti-peptide 4 IgG1 (A) or IgG2a (B). End point titers are shown as the reciprocal of the dilution for the last positive well for each serum analyzed. Cl-4 or ET-4 was bound to the plates in 0.1 M Na2HPO4, pH 9.0, at a concentration of 5 µg/ml. Preimmune and primary sera were diluted to 1:500, secondary sera were diluted to 1:1,000, and tertiary sera were diluted to 1:2,000. Dotted lines, baseline serum dilution. Symbols above the lines are considered positive. Means are indicated by solid lines.

Cl-5 and ET-5. H-2^a, H-2^d, and H-2^k mice responded with robust IgG1 titers to Cl-5 (Fig. 4A). H-2^b mice were again nonresponders to Cl-5. The IgG2a responses to Cl-5 of all the mice were low or absent (Fig. 4B). There was variability in H-2^d and H-2^k mice in the secondary and tertiary responses (IgG2a) to Cl-5, with not all of them responding. This was clearly different from the IgG1 responses of mice in these two haplotypes in which the overwhelming majority of the mice responded earlier and with higher titers of anti-Cl-5 IgG1. Clearly, for all mice that responded to immunization with Cl-5, the IgG1 response was induced earlier and to a higher degree than that for IgG2a.

View larger version (25K): [in this window] [in a new window]   FIG. 4.   Comparison of ELISA titers of classical and El Tor anti-peptide 5 IgG1 (A) or IgG2a (B). End point titers are shown as the reciprocal of the dilution for the last positive well for each serum analyzed. Cl-5 or ET-5 was bound to the plates in 0.1 M Na2HPO4, pH 9.0, at a concentration of 5 µg/ml. Preimmune and primary sera were diluted to 1:500, secondary sera were diluted to 1:1,000, and tertiary sera were diluted to 1:2,000, except for classical sera, which were diluted 1:2,500. Dotted lines, baseline serum dilution. Symbols above the lines are considered positive. Means are indicated by solid lines.

Cl-6 and ET-6. H-2^a, H-2^d, and H-2^k mice responded to Cl-6 with IgG1-specific Abs throughout the immunization time course, even though there was only one H-2^d mouse that had a positive response in the primary sera. As before, the H-2^b haplotype mice did not respond to the classical peptide, yet they did respond to the ET-6 peptide. H-2^b, H-2^d, and H-2^k mice made IgG1 in response to ET-6, although the response of H-2^k mice was not as high as those of H-2^d and H-2^b mice. H-2^a haplotype mice were nonresponders to ET-6 but responded to Cl-6. This is in contrast to H-2^a mice responding to ET-4 and ET-5. Immunization with either Cl-6 or ET-6 revealed a difference in the abilities of I-A^k/I-E^k mice to respond to a given peptide. H-2^a mice did not respond to ET-6 while H-2^k mice did. This was based on differences that map to the class I region of H-2.

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Comparison of ELISA titers of classical and El Tor anti-peptide 6 IgG1 (A) or IgG2a (B). End point titers are shown as the reciprocal of the dilution for the last positive well for each serum analyzed. Cl-6 or ET-6 were bound to the plates in 0.1 M Na2HPO4, pH 9.0, at a concentration of 5 µg/ml. Preimmune and primary sera were diluted to 1:500, secondary sera were diluted to 1:1,000, and tertiary sera were diluted to 1:2,000. Dotted lines, baseline serum dilution. Symbols above the lines are considered positive.

Cross-reactive B-cell epitopes in nonresponder peptide-haplotype combinations. The lack of a serologic response to a TcpA peptide by a particular haplotype of mouse could be related to a lack of B- or T-cell epitopes or both. The positive serologic results for a given biotype peptide among one or more of the haplotypes of mice indicate that there are B- and T-cell epitopes present in Cl-4, -5, and -6 and in ET-4, -5, and -6. If the Ig locus on which the expressed B-cell antigen receptor (binds B-cell epitope) repertoire is based is invariant, then the abilities to bind B-cell epitopes on TcpA peptides should be the same. Thus, the likely explanation for the lack of, or low, serologic response by a particular haplotype of mouse would be an effect of the Ir gene because of the loss of T-cell epitopes.

Immune indices of classical and El Tor peptides 4, 5, and 6. The anti-peptide 4, 5, and 6 responses of mice immunized with classical or El Tor peptides were scored as described in Materials and Methods and are shown in Table 2. The comparisons in Table 2 allow the evaluation of the immunogenicity of classical and El Tor peptides with respect to haplotypes and the relative responsiveness with respect to IgG1 or IgG2a of different haplotypes to classical and El Tor peptides 4, 5, and 6. 

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Cholera remains a disease for which a universal, highly effective vaccine is lacking (3, 4). The development of a cholera vaccine must take into account the target population and the current biotype and serotype of the endemic or pandemic organism. The success of the Bordetella pertussis subunit vaccine, which is based in part on the immunogenicity of three colonization factors and their capacity to evoke protective serologic responses, has prompted us to evaluate a similar approach for V. cholerae vaccination (11). The only well-defined colonization factor in cholera infection is TCP although it has been suggested that others have a role (18-21). A complicating factor in the use of TCP as an immunogen is the difficulty in isolating endotoxin-free TcpA from cultured bacteria.

A potential TcpA vaccine that could circumvent the LPS contamination could be based on TCP-derived peptides. TcpA peptides have been generated based on the classical TcpA amino acid sequences that span the C terminus of the TcpA subunit (18). This region of TcpA is defined by a predicted  hairpin turn with a disulfide bond that contains amino acids required for assembly of the pilus and also for interaction with either host cells or cells in the aggregated colony of bacteria (10). Corresponding peptides from regions in El Tor TcpA have not been investigated for their protective effect.

The synthesis of peptides corresponding to regions of TcpA overcomes the problem of purification, but it introduces a new problem: the potential effect of the Ir gene associated with a limited peptide sequence available for class II binding. Effects of the Ir gene are classically defined as lower immunogenicity of a protein that has been linked to the lack of peptide binding to class II molecules and thus to no or poor activation of T-cell help for Ab production. Initial studies suggested that classical peptides 4, 5, and 6 were immunogenic, but these studies did not investigate enough potential class II alleles to determine if the effects of the Ir gene would be problematic. In this study, we clearly demonstrate that the three most common H-2 haplotypes of mice manifest significant effects of the Ir gene with respect to these peptides, as evidenced by different levels of serologic responses in selected TcpA biotype and haplotype combinations. These effects are seen in both the classical and El Tor biotypes. Furthermore, there is a non-class II but H-2-linked effect on the immunogenicities of classical and El Tor peptides. This is seen in the response of H-2^a and H-2^k mice to peptides 4 or 6. In the peptide 4 system, Cl-4 is immunogenic in mice of both haplotypes, whereas only the H-2^a mice respond to ET-4. Similarly, for peptide 6, the responses by H-2^a and H-2^k mice are similar for Cl-6 but, in this comparison, H-2^k mice respond to ET-6 while H-2^a mice do not.

The canonical class II peptide binding motifs for several haplotypes are known (2, 5, 14, 23). There are preferred amino acid residues in particular positions of the class II-bound peptide, but they are not strict requirements for binding. Other amino acids in the "gestalt" of the bound peptide-class II interaction interface can account for noncanonical peptide-class II interactions that result in sufficient binding for induction of T cells. Clearly, the loss of an anchor residue resulting from the different TcpA biotype sequence could explain the responses in H-2^b mice that we report. The change in biotype sequences and the loss of effect do not explain why mice that express the same class II restriction elements would serologically respond in a similar or identical manner. A possible explanation for this effect is perhaps centered in the class I region of the H-2 locus. The H-2^a and H-2^k mice differ in this region. H-2^a mice are D^d L^d Qa-2^a, while H-2^k mice are D^k L^k Qa-2^b. MHC-I molecules are also associated with effects of the Ir gene. This has been demonstrated by several laboratories and also relates to the binding of peptides that influence the subsequent immune responses. It has been established that immunization of mice with protein antigen can result in B cells that express class I bound with peptides from the immunizing antigen (25). The B cells then become a target for cytotoxic T-lymphocyte (CTL) responses that kill the B cells and thus reduce or prevent antibody responses to the immunizing antigen. Thus, we hypothesize that, for ET-4 and ET-6, the differences in class I resulted in presented peptides that make B cells in H-2^k and H-2^a mice, respectively, targets for CTL-based elimination. Clearly, there are a number of uncharacterized genes, with the percent contributions for these genes in congenic mice being unknown because of the unmapped recombination point. If one eliminates non-immune-related genes as the explanation for the differences in response to ET-4 and CL-6, then class I and Qa-2 differences are both theoretical possibilities. Qa-2 antigens are nonclassical MHC-I molecules with the potential to bind peptides. The Qa-2 antigen is the product of the Ped (preimplantation embryo development) gene and is a glycosylphosphatidylinositol-linked cell surface protein encoded in the Q region of the mouse major histocompatibility region (13). Qa-2⁺ mice have significantly higher preimplantation embryo cleavage rates both in vivo and in vitro than Qa-2 mice. The importance of this for immune responses is unknown. However, Gould et al. recently presented evidence that Qa-2 can affect selection of intestinal intraepithelial lymphocytes (6). Qa-2 has a complex role in immunobiology. Qa-2 could serve as a restriction element (MHC-like structure with bound peptides) for subsequent interactions with T cells. The Qa-2 region differences between H-2^a and H-2^k mice are a less likely explanation for the lack of response to ET-6 because theoretically nonpermissive allele Qa-2^a is associated with a positive serologic response to that peptide by H-2^b mice. However, in the ET-4 system, the nonpermissive allele, if it is Qa-2^a, is associated with lack of a response in H-2^k mice and very low response in H-2^d mice. One would have to postulate two mechanisms for the different responses by H-2^a and H-2^k mice. Qa-2 effects could account for the lack of peptide 4 responses, and class I differences could account for the lack of a peptide 6 response.

While we have identified effects of the Ir gene for classical and El Tor TcpA peptides, it is apparent that other genes can affect the serologic response as well. The reason that H-2^a mice are such good responders is not apparent and is not related to the Ig locus as they share that with H-2^k mice. Other non-H-2 genes linked to H-2 may contribute to the differences in serologic responses. Regardless of the explanation for the differences, the effect of the Ir gene associated with the differences in the H-2 regions suggests that vaccine peptides based on classical and El Tor TcpA peptide 4, 5, or 6 may not be universally efficacious in the human population. This may be an overstatement, as the target human population is outbred at the HLA locus. The codominant expression of human class II molecules DR, DQ, and DP might allow more possible targets for classical peptide- and El Tor peptide-based binding. Thus, with more class II alleles present, it may be easier to generate a class II peptide complex to induce T-cell help for antibody production. It is, however, also true that certain proteins contain immunodominant epitopes for induction of responses, and thus, while TcpA peptide 4, 5, or 6 binding may occur at a high frequency, the peptide sequences may not be particularly immunogenic with respect to their ability to be highly expressed and thus able to activate naive T cells. There are two solutions to these issues. The most direct, and one that can take advantage of T-cell memory in the mucosal T-cell pool, would be to link classical and El Tor peptide 4, 5, or 6 to a universal protein carrier for its contribution of T-cell epitopes. A recent development by Alexander et al. could also be used by taking advantage of pan-DR epitopes that could be linked to the TcpA peptide, thus yielding a host (human)-based vaccine design based on known class II binding properties (1). Alternatively, a mixture of TcpA peptides that would more effectively cover the possible sequence solutions to MHC-II binding may be used. These solutions would mitigate the effect of the class II Ir gene but not the possible effect of class I or Qa-2 (perhaps not evident in humans, as no Qa-2 gene has been identified). The solution for the class I-based problem of the classical and El Tor peptide sequences is to clearly identify the TcpA B-cell epitopes within peptides 4, 5, and 6 that are immunogenic and protective and then use minimal B-cell epitopes as haptens associated with a universal carrier.

Clearly, the development of cholera subunit vaccines for humans will need to take into account the route of immunization that induces the most protective isotype. Whether this is secretory IgA or IgG is debatable, but the resolution of this issue may dictate the route and method of immunization with TcpA peptides. We were able to induce IgG as well as IgA (data not shown) via intraperitoneal immunization. While this form of immunization is not practical in humans, other routes such as oral or intranasal routes have been shown to be responsive to immunogens by generating IgG as well as serum IgA.