![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
and Stephen R. Attridge*
Department of Molecular Biosciences, The University of Adelaide, Adelaide, South Australia 5005, Australia
Received 29 July 2002/ Returned for modification 12 September 2002/ Accepted 22 October 2002
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A comparison of experimental protocols used in the initial conflicting reports identified at least two variables which might influence the outcome of vector priming. Bao and Clements (2) used the aroA Salmonella enterica serovar Dublin vector strain SL1438, the auxotrophy of which is presumed to restrict colonization and persistence within the murine gut-associated lymphoid tissue (GALT). Priming with this vector did not compromise responses to a passenger antigen, LT-B, subsequently delivered by the same strain. In contrast, Attridge et al. (1) used S. enterica serovar Stanley as a vector strain; this is naturally attenuated for mice but colonizes Peyer's patches for several weeks. Priming with S. enterica serovar Stanley completely abolished the serum response to foreign antigen (K88) normally induced by oral immunization with recombinant S. enterica serovar Stanley. Perhaps the more prolonged GALT colonization by serovar Stanley induces stronger antivector immune responses, with a consequently greater adverse effect on the immunogenicity of the recombinant construct administered later. Consistent with this interpretation, the impact of vector priming was significantly reduced in animals given a lower priming dose of bacteria (1). To examine the relevance of the vector strain's propensity to colonize the GALT, an aroA mutation was introduced into the S. enterica serovar Stanley vector previously used in our laboratory. After in vitro and in vivo characterization, the resulting mutant was used to prime mice which were subsequently immunized with S. enterica serovar Stanley-K88.
The second obvious difference between the two initial conflicting reports lies in the foreign antigen being delivered to the GALT. Bao and Clements (2) used LT-B, a strong mucosal immunogen, which is localized in the periplasm but is likely to be released upon contact with intestinal factors (9). In contrast, earlier experiments in our laboratory were conducted with the Escherichia coli pilus protein K88, which is located on the surface of the recombinant vector, as the foreign antigen (1). These antigens therefore differ both in their intrinsic immunogenicity and in their location within the vaccine construct. To evaluate the importance of the antigen as a determinant of the consequences of vector priming, the immunogenicity of recombinants expressing each antigen was compared in control and vector-primed mice. This experiment was performed with both aroA S. enterica serovar Dublin and S. enterica serovar Stanley as vectors.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Growth media. Wt and recombinant S. enterica serovar Stanley were grown in CBT (Casamino Acids-vitamin B1-tryptophan) broth as described before (1), with 100 µg of ampicillin (AP) per ml as required. aroA S. enterica serovar Stanley was grown in CBT medium with the addition of supplements 2,3-dihydroxybenzoic acid (DHB; Sigma) and para-aminobenzoic acid (pABA; Sigma) to final concentrations of 2 µg/ml. Since aroA S. enterica serovar Dublin did not grow in CBT, Luria broth with supplements was used to culture this strain for assessment of Peyer's patch colonization. Bacteria recovered from GALT were plated on agar with AP or kanamycin (KM, 50 µg/ml) as appropriate. To determine the effect of priming with aroA S. enterica serovar Dublin on the subsequent delivery of K88 or LT-B by the same strain, bacteria were grown in tryptic soy broth (Difco) as described by Bao and Clements (2).
Construction and characterization of an aroA mutant of S. enterica serovar Stanley. The aroA gene of S. enterica serovar Stanley was amplified by using PCR primers 2934 (GAGAGTTGAGTTTCATG; binds to aroA from nucleotide -14 to +3; start codon underlined) and 2936 (AGAAGACTTAGGCAGGCG; binds to aroA from nucleotide 1279 to 1298; stop codon underlined) designed from the sequence of the S. enterica serovar Typhimurium aroA gene (accession number Y10355). The resulting PCR product (ca. 1.2 kb) was ligated into pGEM-T Easy (Promega) and mutated by introducing the apha-3 nonpolar KM resistance cartridge (13) into a unique EcoRV restriction site found ca. 600 bp from the start codon of aroA. The aroA::apha-3 fragment (ca. 2 kb) was liberated by digestion with NsiI and SphI, separated by agarose gel electrophoresis, and purified by using a QIAquick gel extraction kit (Qiagen). This was then ligated into pCACTUSmob (C. Clark, our department), which had been digested with PstI and SphI (NsiI and PstI yield compatible cohesive ends). pCACTUSmob is a suicide vector which carries a chloramphenicol resistance gene and a temperature-sensitive replicon which is inoperative at 42°C but supports plasmid replication at 30°C. It also carries a mob region, which allows plasmid transfer via conjugative pili, and the sacB gene from Bacillus subtilis, which produces toxic sugar polymers when bacteria are grown in the presence of sucrose (10). pCACTUSmob carrying aroA::apha-3 was electroporated into S17-1, facilitating its subsequent mobilization into S. enterica serovar Stanley. The protocol for allelic exchange mutagenesis was then followed as previously described (15).
Chromosomal DNA from putative aroA mutants was isolated by phenol-ether extraction (15) and subjected to PstI digestion and agarose gel electrophoresis. Southern hybridization was then performed at high stringency with aroA and apha-3 digoxigenin-UTP-labeled (Roche) gene probes as described by Sambrook et al. (17) to confirm the presence of a chromosomal aroA mutation. Bound probe was detected using BM chemiluminescence Blotting Substrate (POD) from Boehringer Mannheim.
A low-copy-number plasmid carrying a minimal aroA fragment was constructed for attempted complementation of the aroA mutation in S. enterica serovar Stanley. pGEM-T Easy carrying the aroA gene was digested with EcoRI, liberating a DNA fragment of ca. 1.2 kb, which was then ligated into EcoRI-digested pWSK30 (20). The resulting construct (pCVSA7) and the control plasmid were electroporated into various aroA mutants via EX2000.
Animal experiments. Female BALB/c mice were obtained from the Central Animal House of our university and were 9 to 10 weeks old at the commencement of experiments. Preparation of mice, feeding of mice with bacteria, and assessment of Peyer's patch colonization potential were performed as previously described (1). All animal experiments were approved by the Animal Ethics Committee of our university.
Assessment of immune responses. Serum and fecal pellet supernatant (FPS) samples were prepared at various time points and respectively titrated for immunoglobulin G (IgG) and IgA antibodies. Serum was prepared from blood collected under ether anesthesia and was stored at -20°C. Nine fresh mouse fecal pellets were collected in 600 µl of buffer on ice in plastic reaction tubes preblocked (overnight at 4°C) with 1% bovine serum albumin-phosphate-buffered saline (BSA-PBS); the collection buffer was PBS containing 0.1 mg of soybean trypsin inhibitor (Sigma) per ml, 1% (wt/vol) BSA, 25 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 50% (vol/vol) glycerol, adapted from the work of Butterton et al. (3). Pellets were homogenized with blunted Pasteur pipettes and left at 4°C for 4 h. Samples were then spun at 4°C (Eppendorf 5417R centrifuge, 13,000 rpm), and clarified supernatants were stored in aliquots in fresh (preblocked) tubes at -20°C.
Antibody responses were assayed by enzyme-linked immunosorbent assay (ELISA) as described previously (8), but with modifications. For the titration of antibodies to K88, ELISA trays (Polysorp F96; Nunc) were sensitized overnight at 4°C with purified pili (isolated as previously described [8]); wells received 100 µl of a 1-µg/ml solution in TSA buffer (Tris-saline-azide; 132 mM NaCl, 25 mM Tris-HCl [pH 7.5], 0.05% sodium azide, and MilliQ water). For the titration of antibodies to lipopolysaccharide (LPS), purified lyophilized LPSs isolated from S. enterica serovar Typhimurium (O-4,5,12; the same O-antigen serotype as S. enterica serovar Stanley) and S. enterica serovar Enteritidis (O-9,12; the same O-antigen serotype as S. enterica serovar Dublin) were obtained from Sigma. These were held at 4°C as 20-mg/ml stocks in sterile MilliQ water and diluted to a final concentration of 20 µg/ml in carbonate buffer (3.2 mM Na2CO3, 6.7 mM NaHCO3, 0.5 µM MgCl2; pH 9.6) for sensitization of ELISA trays (overnight at 4°C).
The following day, the sensitizing solution was discarded and the trays were washed three times with wash buffer (0.1% [vol/vol] Triton X-100, 5 mM Tris, 150 mM NaCl in MilliQ water, pH 7.6). After the trays had been banged dry, wells were blocked with 200 µl of blocking solution (1% BSA-PBS) for ca. 90 min at 37°C. Test samples were initially diluted 1:3 (FPSs) or 1:20 (sera) and then subjected to fivefold serial dilutions (by transferring 25 µl into 100 µl) in blocking solution. Each tray also contained one row set aside for the titration of a standard antiserum (or FPS) to check assay performance. Trays were incubated for 4 h at 37°C prior to washing as described above. Secondary antibody for the detection of IgG in serum samples was alkaline phosphatase-conjugated goat anti-mouse IgG (Kirkegaard and Perry Laboratories Inc.); 100 µl of a 1:50,000 dilution in blocking solution was dispensed into each well, and the trays were left overnight at 4°C. After three washes in wash buffer, 100 µl of substrate solution was added to each well; this was prepared by dissolving Sigma 104 alkaline phosphatase substrate tablets in assay buffer (1 M diethanolamine, 0.05% [wt/vol] sodium azide, 5 mM MgCl2, and MilliQ water, pH 9.8; one tablet per 5 ml of buffer). After incubation at 37°C for 3 h, optical density (OD) was measured with either a Molecular Dynamics Biolumin 960 or a Labsystems Multiskan Ascent ELISA plate reader, with a primary filter setting of 410 nm and a secondary filter setting of 620 nm. OD readings were adjusted by subtracting the secondary filter and averaged blank readings from the primary filter reading, and sample titers were determined as the reciprocal dilution yielding an endpoint of an OD at 410 to 620 nm (OD410-620) of 0.15.
The detection of (specific) IgA in FPS samples required a two-layered detection system, the first being biotin-conjugated goat anti-mouse IgA (Sigma). This was diluted 1:5,000 in blocking solution, dispensed at 100 µl per well and incubated overnight at 4°C. After washing, trays were incubated at 30°C for 90 min with an avidin-peroxidase conjugate (Extravidin [Sigma]; 1:3,500 dilution in blocking solution, 100 µl per well). Trays were washed again and incubated for 3 h at 30°C with Sigma Fast o-phenylenediamine dihydrochloride substrate tablets (Sigma) dissolved in distilled water (as recommended by the manufacturer; 100 µl per well). Color development was stopped by the addition of 50 µl of 3 M HCl per well, and OD was measured by using the aforementioned plate readers with a primary filter setting of 485 nm. After the averaged blank readings had been subtracted, titers were determined as the reciprocal FPS dilution yielding an endpoint of an OD485 of 0.15.
Due to the variability in immunoglobulin content between FPS samples, antibody activity was expressed as an ELISA titer per microgram of total IgA. The total IgA concentration of each FPS was estimated in a separate assay as follows. ELISA trays were sensitized overnight at 4°C with 5 µg of goat anti-mouse IgA (Kirkegaard and Perry Laboratories Inc.) per ml diluted in TSA buffer (100 µl/well). Trays were washed and blocked as described above. FPS samples were titrated in duplicate in fourfold falling dilutions (beginning at 1:100) in blocking solution. A standard curve was included in each tray and was constructed by using purified mouse myeloma IgA (MOPC 315, kindly provided by P. Ey of our department); this was titrated in duplicate in threefold falling dilutions from a starting concentration of 1.6 µg/ml. Trays were then left at 37°C for 4 h, washed, incubated with the biotin-labeled and avidin-peroxidase conjugates, and developed as described above. The IgA concentration of each sample was determined by selecting several OD values which fell within the linear range of the standard curve, estimating IgA concentration, and multiplying by the appropriate dilution factor.
Measurement of LT-B release from recombinant Salmonella.
Estimation of LT-B release from recombinant Salmonella was performed by a method modified from that published by Hunt and Hardy (9). Overnight Luria broth cultures of the strains of interest were subcultured 1 in 20 into 20 ml of fresh medium containing conalbumin (iron chelator; Sigma) at 450 µg/ml and grown to an OD600 of 
0.6; additional subcultures into medium without conalbumin were set up as negative controls. At this stage trypsin (100 µg/ml; Sigma) and bile salts (1.4 mM sodium glycholate, 0.7 mM sodium deoxycholate, 1.2 mM sodium glycochenodeoxycholate, 2.8 mM sodium taurocholate, 2.4 mM sodium taurochenodeoxycholate, 1.4 mM sodium taurodeoxycholate) (Sigma) were added to mimic the environment of the human small intestine. Negative control cultures did not receive trypsin or bile salts. Two samples (1 ml each) were taken from cultures at hourly intervals and centrifuged (4°C, 15,000 rpm, Eppendorf 5417R centrifuge). Supernatants were transferred to clean tubes, and soybean trypsin inhibitor (Sigma; to 120 µg/ml) and PMSF (Sigma; to 0.25 mM) were added to inhibit protein degradation. Cell pellets were resuspended in 1 ml of a solution containing the same protease inhibitors (50 mM Tris-HCl [pH 7.8], 5 mM MgCl2, 0.25 mM PMSF, 120 µg of soybean trypsin inhibitor per ml) and subjected to sonication on ice (Branson Sonifier cell disruptor fitted with a microtip; 3 pulses at 1 min per pulse; output control setting, 3). Samples were stored at -20°C until assayed for LT-B content by an ELISA inhibition assay, using the protocol described by Hone et al. (8).
Statistical analysis. ELISA endpoint titers and total IgA concentrations were calculated and plotted by using GraphPad Prism (GraphPad Software, Inc.), assigning titers of 20 and 1 (the limits of detection) to negative sera and FPSs, respectively. Probability values were determined by using the (two-tailed) Student's t test function in Microsoft Excel (Microsoft Corporation) for comparisons of two sets of data. When more than two groups were compared, one-way analysis of variance (95% confidence interval; GraphPad Software, Inc.) was performed; applying the Dunnett's or Bonferroni posttest to compare selected pairs of treatment groups yielded identical probability values. Where appropriate, data were log10 transformed prior to these calculations.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In vivo characterization of Salmonella aroA mutants. An experiment was performed to examine the impact of the aroA mutation on the in vivo behavior of S. enterica serovar Stanley and to determine whether the anticipated reduction in GALT-colonizing potential could be overcome by expression of a minimal aroA fragment. BALB/c mice were separated into five groups, one of which received the wt vector. Consistent with the dosing regimens used by Attridge et al. (1) and Bao and Clements (2), mice in groups 2 and 3 were fed aroA S. enterica serovar Stanley with either one dose of ca. 109 CFU or two doses of ca. 1010 CFU (on days 0 and 4). Smaller groups were included to ascertain whether the aroA mutation could be complemented by pCVSA7 or pWSK30 in vivo; mice in these groups were fed a single dose of ca. 109 CFU. At various times after immunization, Peyer's patches were excised from the small intestine, homogenized, and plated onto XLD medium (Oxoid). Homogenates recovered from mice fed the aroA mutant were also plated on XLD medium with KM (to check the anticipated in vivo stability of the aroA mutation), while those from mice receiving complemented mutant were also plated on XLD medium with KM and AP (to assess plasmid stability).
Wt S. enterica serovar Stanley colonized the Peyer's patches at levels comparable to those previously seen in our laboratory (1), reaching a maximum burden of ca. 104 CFU on days 5 to 8 and still present in each of the eight mice sacrificed at days 34 and 40 (Fig. 1A). The colonization profile of the aroA S. enterica serovar Stanley mutant was surprisingly similar. After a single dose of ca. 109 CFU, the recoveries of mutant bacteria were significantly lower (P < 0.05) than those of wt bacteria only on days 5, 8, and 27. Using the alternative dosing schedule of Bao and Clements, even greater numbers of aroA S. enterica serovar Stanley organisms were recovered (Fig. 1A). For both groups 2 and 3, replicate aliquots of homogenate plated on XLD with or without KM confirmed that the apha-3 cartridge was stably maintained (data not shown).
|
The high recoveries of aroA S. enterica serovar Stanley prompted an examination of the GALT-colonizing potential of SL1438, which has not been reported previously. Two groups of mice were fed SL1438 by using the alternative dosing regimens used with aroA S. enterica serovar Stanley to define and compare the colonization profile of this mutant. GALT persistence of the aroA S. enterica serovar Dublin was similar to that of the S. enterica serovar Stanley auxotroph until day 8 (Fig. 1B). From this point on, the former was cleared more quickly, such that recoveries of the S. enterica serovar Dublin mutant at day 14 were similar to those of the S. enterica serovar Stanley mutant at day 34 (Fig. 1).
Immune responses to recombinant Salmonella in vector-primed mice. Despite the unexpected GALT-colonizing potential of the aroA S. enterica serovar Stanley mutant, an experiment was set up to compare the consequences of priming mice with ca. 109 CFU of either wt or aroA bacteria, with respect to the capacity of such animals to subsequently respond to a foreign antigen orally delivered by the same (wt) vector. Groups of mice were primed with wt or aroA S. enterica serovar Stanley on day -70, while controls received sodium bicarbonate only; a fourth group was given the aroA mutant according to the dosing schedule used by Bao and Clements. On day 0, all mice were orally immunized with S. enterica serovar Stanley-K88, and then serum and FPS samples were prepared at various time points for the assessment of antibody responses to K88 and vector LPS. It was of particular interest to determine whether the hyporesponsiveness to K88 previously observed in vector-primed mice (1) extended to the intestinal response.
Serum anti-K88 and anti-LPS responses in (wt) vector-primed and control mice were comparable to those seen previously (1). Control mice developed a sustained primary serum IgG response to K88 following immunization with S. enterica serovar Stanley-K88; anti-K88 titers were uniformly high, particularly from day 52 onwards, when each mouse (n = 8) had a serum ELISA titer of >104 (Fig. 2A). In contrast, priming with wt S. enterica serovar Stanley reliably blocked the induction of antibody to K88; in general these mice had serum IgG titers of <100. These responses were significantly different (P < 0.01 from days 39 to 83). A similar pattern of antibody responses was observed in the GALT (Fig. 2C). Although the intestinal IgA response to K88 seen in control mice was less dramatic and more variable than the serum IgG response, this too was completely inhibited by prior exposure to the vector. The IgA responses in control and vector-primed mice were significantly different on days 39, 52 (P < 0.01), and 66 (P < 0.05).
|
5,000, representing a 250-fold increase above background and typical of responses seen in unprimed controls; six other mice showed titers in serum of 
200, typical of the responses seen in animals primed with wt S. enterica serovar Stanley; and five mice showed intermediate titers. Evidence of priming was indicated by elevated serum and intestinal anti-LPS antibody levels on day -1 in mice preimmunized with either wt or aroA S. enterica serovar Stanley and by the occurrence of secondary responses, particularly in serum (Fig. 2B and D). Although the latter were similar in strength and kinetics among animals primed with one low or two high doses of aroA S. enterica serovar Stanley, priming with two doses of the auxotrophic vector resulted in higher IgA anti-LPS titers upon exposure to S. enterica serovar Stanley-K88 (but significantly so only on day 39; P < 0.01). Together with the more consistent suppression of the gut anti-K88 response in this group, this raised the possibility that mice given two doses of aroA S. enterica serovar Stanley might have been more effectively primed, prompting an examination of the primary anti-LPS responses generated by the two schedules used for vector priming.
Anti-LPS responses induced by oral immunization with wt or aroA S. enterica serovar Stanley. Three groups of mice were immunized with wt or aroA S. enterica serovar Stanley by using the dosing regimens employed in the previous experiment. As a comparison, a fourth group was immunized with two doses of ca. 1010 CFU of SL1438 spaced 4 days apart; vector priming with this strain has been reported not to compromise subsequent responses to recombinant Salmonella expressing LT-B (2). Serum and FPS samples were collected at various time points, and anti-LPS responses were measured by ELISA.
Mice fed wt S. enterica serovar Stanley produced serum anti-LPS responses that were significantly greater than those fed aroA S. enterica serovar Stanley at either one dose of 109 CFU (P < 0.01 on days 46 and 70) or two doses of 1010 CFU spaced 4 days apart (P < 0.05 on days 46 and 70) (Fig. 3A). Responses in mice primed with one dose of the aroA mutant were barely detectable; those given two high doses developed weak responses which were only fivefold greater than background levels at day 70. Intestinal IgA responses to LPS were weak and did not differ (Fig. 3B). Thus, the data from this experiment did not reveal any major difference in antivector responses between the two aroA immunization protocols.
|
Immunogenicity of S. enterica serovar Stanley-LT-B in vector-primed mice. To evaluate the importance of the foreign antigen delivered by recombinant Salmonella as a determinant of the immune responses elicited in vector-primed mice, S. enterica serovar Stanley was transformed with pJC217, the plasmid used to specify LT-B production in EL23 (2). To examine whether S. enterica serovar Stanley-LT-B and EL23 would be likely to release LT-B to similar extents in vivo, an in vitro experiment was performed based on a study by Hunt and Hardy (9). Cultures were incubated (in duplicate) in the absence (controls) or presence of physiological concentrations of bile salts and conalbumin, in an attempt to gauge the effects of intestinal factors on the release of LT-B. Supernatants and sonicated cell pellets were prepared from samples taken at hourly intervals, and LT-B production was quantified by ELISA inhibition assay. The two recombinant Salmonella constructs behaved similarly, producing and secreting similar levels of LT-B in the presence or absence of intestinal factors (data not shown).
An experiment was designed to assess the immunogenicity of S. enterica serovar Stanley-LT-B in mice previously primed with wt S. enterica serovar Stanley. Vector-primed and control mice were fed 109 CFU of S. enterica serovar Stanley-LT-B on day 70, and serum and FPS samples were prepared at various intervals for determination of anti-LT-B and anti-LPS responses. As shown in Fig. 4A, control mice developed a strong primary serum anti-LT-B response following exposure to S. enterica serovar Stanley-LT-B. In comparison, vector-primed mice developed weaker serum anti-LT-B responses, which were significantly different from those in the control group on days 29 and 42. The variability of the anti-LT-B serum responses in vector-primed mice is apparent from a scatter plot analysis (Fig. 5A). On days 29, 42, 56, and 70, four of the five vector-primed mice had the lowest serum anti-LT-B titers of the 10 mice sampled. On days 56 and 70, however, the fifth mouse in this group showed the highest antibody titer, explaining the large SDs obtained. When these same groups were included as part of a subsequent, larger experiment, a more consistent inhibition of the serum anti-LT-B response was seen in vector-primed mice. The mean titers were 14-, 33-, and 75-fold lower than those seen in controls on days 26, 46, and 70 after administration of S. enterica serovar Stanley-LT-B; these differences were highly significant (P < 0.01) at all three time points (data not shown).
|
|
20-fold increase above background (Fig. 4C and 5B). The response seen in vector-primed mice was significantly impaired, again illustrating the impact of preexisting immunity on responses of the murine GALT. The efficacy of vector priming was shown by the elevated anti-LPS titers in both serum and gut (Fig. 4B and D). Delivery of LT-B or K88 by aroA S. enterica serovar Dublin in vector-primed mice. To continue our investigation of the significance of the vector strain and of the foreign antigen as determinants of the impact of vector priming, it was decided to repeat the experiment of Bao and Clements (2) but to extend the study to include delivery of K88 as an alternative foreign antigen. It was first necessary to construct a derivative of the aroA S. enterica serovar Dublin vector which expresses K88; this was done by electroporating plasmid pFM205 into SL1438. Plasmid pFM205 carries a truncated K88 pilus biosynthetic operon which is constitutively expressed; from this was derived the plasmid used to specify K88 biosynthesis from S. enterica serovar Stanley (1). Since pJC217 and pFM205 are both pBR322 based, plasmid copy numbers were equivalent in the two constructs and the Apr marker allowed selection for plasmid retention in vitro but not in vivo. Slide agglutination, immunoblot analysis, and ELISA inhibition assay confirmed the expression of K88 by SL1438-K88 (data not shown).
Two groups of BALB/c mice were orally primed with SL1438, delivering two doses of 1010 CFU at days -70 and -66 (2), with additional mice set aside as controls. At days 0 and 4, paired groups of unimmunized and primed mice were boosted with 1010 CFU of either SL1438-K88 or EL23. Serum and FPS samples were collected at various time points and anti-LT-B, anti-K88, and anti-LPS responses were monitored by ELISA.
Serum and mucosal anti-LT-B responses (Fig. 6, graphs A1 and A3) in primed mice were equivalent to (if not greater than) those in control mice, confirming the results of Bao and Clements (2). In contrast, serum anti-K88 responses (Fig. 6, graph B1) were significantly reduced in vector-primed mice, the mean antibody titers being consistently 10-fold lower. The intestinal anti-K88 response was only slightly reduced. Serum and mucosal anti-LPS profiles (Fig. 6, graphs A2, A4, B2, and B4) confirmed effective vector priming in SL1438-dosed mice.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
An aroA S. enterica serovar Stanley mutant was constructed in order to examine the significance of the vector strain's GALT colonization potential as a determinant of the consequences of vector priming. This mutant, whose auxotrophy could be reversed by provision of a complementing aroA+ gene, displayed surprisingly strong GALT colonization. Although attenuated aroA Salmonella was originally thought to be unlikely to replicate in vivo, recent studies have shown that aroA S. enterica serovar Typhimurium persists in the Peyer's patches for at least 21 days (7) and is even lethal in certain knockout mouse strains (19). Stocker (19) has suggested that the in vivo persistence of attenuated aroA Salmonella may be due to a leeching of (diet- or microflora-derived) DHB and pABA precursors from the gut lumen.
Our aroA S. enterica serovar Stanley mutant was clearly attenuated, however, and stimulated weaker anti-LPS responses than its parent (Fig. 3). Perhaps as a consequence of generally weaker antivector responses, about 60% of animals primed with (either one or two doses of) the mutant were able to produce serum antibodies to K88 when given boosters of S. enterica serovar Stanley-K88; no responses were observed in animals primed with the wt vector (Fig. 2). At the level of the gut, some anti-K88 responses were observed in mice primed with a single dose of the auxotrophic vector but not in those given two higher doses. Together the results with the aroA vector suggest that a slight reduction in GALT colonization potential has led to weaker antivector immunity and allowed some responses to the subsequently presented K88.
When results obtained with vectors of different serogroup are compared, however, no consistent pattern is apparent. Although the aroA S. enterica serovar Dublin vector colonizes the Peyer's patches for a significantly shorter period than aroA S. enterica serovar Stanley, it induces much stronger serum IgG and intestinal IgA anti-LPS responses (Fig. 1 and 3). Despite this, Bao and Clements (2) found that priming with the former did not compromise the immunogenicity of LT-B subsequently delivered by the same vector, a finding confirmed and extended in the present study. Our experience with K88 and LT-B indicates that hyporesponsiveness following vector priming is more likely to be observed with wt S. enterica serovar Stanley than aroA S. enterica serovar Dublin (Fig. 2, 4, and 6), although the basis for this remains unclear. Whereas the former is unable to cause systemic infection in mice, SL1438 was derived from a parent strain virulent for both calves and mice. The two vectors also have different LPSs, and since this structure is a critical determinant of the in vivo behavior of Salmonella strains in mice (12), it would not be surprising if different response profiles are induced. Although aroA S. enterica serovar Dublin elicited stronger anti-LPS responses, these were not accompanied by stronger responses to the passenger antigens. Primary gut responses to K88 and LT-B were similar with both vectors, but serum responses were actually higher with the S. enterica serovar Stanley vector. Further studies will be needed to elucidate the basis for the inconsistent data obtained with different vectors in the context of vector priming.
Clearly, the consequences of priming with a particular vector strain are also determined by the nature of the foreign antigen it subsequently delivers. With both of the vectors used in the present study, the impact of vector priming was more severe if the antigen presented later was K88 rather than LT-B. Mice primed with wt S. enterica serovar Stanley failed to produce significant responses to K88, with serum antibody titers being 1,000-fold lower than those of controls. If the foreign antigen was LT-B, however, primed mice did show weak antibody responses, but levels in serum were still 10- to 30-fold lower than control levels. When the aroA S. enterica serovar Dublin vector was used, primed mice showed 10-fold-lower serum antibody responses to K88, while responses to LT-B were not reduced at all. The reason for this antigen-related difference is unclear. While LT-B is regarded as a potent immunogen, ELISA titers of antibody to K88 were at least as high, in both serum and gut. The different localization of the foreign antigens within the recombinant vectors may be significant.
Since our initial study there have been two further reports of preexisting antivector immunity compromising the immunogenicity of recombinant Salmonella vaccines. Roberts et al. (16) found that prior exposure to strains of homologous or heterologous serotype reduced the immunogenicity and protective efficacy of a Salmonella construct expressing FrgC (fragment C of tetanus toxin). The vectors were aroA aroD mutants, and both serum IgG and intestinal IgA responses were impaired; the impact of prior immunity was greater when the mice were primed with the homologous vector strain. Of interest was the finding that vector-primed mice showed stronger serum (though not intestinal) responses to FrgC if the foreign antigen was cloned downstream of the htrA promoter, rather than the nirB promoter. This suggests a strategy for reducing the impact of preexisting antivector immunity and identifies the nature of the foreign antigen expression construct as another determinant of the outcome of vector priming. Kohler et al. (11) also reported reduced serum IgG and salivary IgA responses to foreign antigen presented by recombinant Salmonella in vector-primed mice.
In contrast, but in support of the findings of Bao and Clements (2), Whittle and Verma (21) found that responses to a viral B-cell epitope located in the flagellar subunits of an aroA S. enterica serovar Dublin vector were actually enhanced in mice previously exposed to the vector strain. This result is not necessarily inconsistent with demonstrations of the negative impact of vector priming, however, if this phenomenon is caused by epitopic suppression (1). According to this hypothesis, immunization with the vector would have expanded clones of B cells responsive to flagellar antigens, and these would be available for presentation of the viral epitope on administration of the recombinant construct. Consistent with this interpretation is the finding that it is possible to generate secondary responses to foreign antigens delivered by Salmonella, if the recombinant construct is used for priming as well as boosting (8a, 11). However, this is different from the situation that would confront authorities wishing to instigate mass immunization programs in developing regions where Salmonella is endemic.
In conclusion, the present findings indicate that both the vector strain and the foreign antigen contribute to the observed consequences of vector priming. With three of the four vector-antigen combinations tested, serum IgG and intestinal IgA responses were significantly impaired in animals previously exposed to the vector strain. These experiments confirm that preexisting antivector immunity poses a serious challenge for the recombinant Salmonella vaccine strategy.
|
ACKNOWLEDGMENTS |
|---|
This study was supported by a grant to S.R.A. from the World Health Organization, through its Global Programme for Vaccines and Immunization. C.J.V. acknowledges support through a postgraduate scholarship from The University of Adelaide.
|
FOOTNOTES |
|---|
Present address: Center for Vaccine Development, University of Maryland School of Medicine, Baltimore, Md.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| 1. | Attridge, S. R., R. Davies, and J. T. LaBrooy. 1997. Oral delivery of foreign antigens by attenuated Salmonella: consequences of prior exposure to the vector strain. Vaccine 15:155-162.[CrossRef][Medline] |
| 2. | Bao, J. X., and J. D. Clements. 1991. Prior immunologic experience potentiates the subsequent antibody response when Salmonella strains are used as vaccine carriers. Infect. Immun. 59:3841-3845. |
| 3. | Butterton, J. R., E. Ryan, R. A. Shahin, and S. B. Calderwood. 1996. Development of a germ-free mouse model of Vibrio cholerae infection. Infect. Immun. 64:4373-4377.[Abstract] |
| 4. | Cardenas, L., and J. D. Clements. 1992. Oral immunization using live attenuated Salmonella spp. as carriers of foreign antigens. Clin. Microbiol. Rev. 5:328-342. |
| 5. | Dougan, G. 1994. The molecular basis for the virulence of bacterial pathogens: implications for oral vaccine development. Microbiology 140:215-224.[Medline] |
| 6. | Dougan, G., C. E. Hormaeche, and D. J. Maskell. 1987. Live oral Salmonella vaccines: potential use of attenuated strains as carriers of heterologous antigens to the immune system. Parasite Immunol. 9:151-160.[Medline] |
| 7. | Dunstan, S. J., C. P. Simmons, and R. A. Strugnell. 1998. Comparison of the abilities of different attenuated Salmonella typhimurium strains to elicit humoral immune responses against a heterologous antigen. Infect. Immun. 66:732-740. |
| 8. | Hone, D., S. Attridge, L. van den Bosch, and J. Hackett. 1988. A chromosomal integration system for stabilization of heterologous genes in Salmonella based vaccine strains. Microb. Pathog. 5:407-418.[CrossRef][Medline] |
| 8. | Huang, Y., G. Hajishengallis, and S. M. Michalek. 2001. Induction of protective immunity against Streptococcus mutans colonization after mucosal immunization with attenuated Salmonella enterica serovar Typhimurium expressing an S. mutans adhesin under the control of in vivo-inducible nirB promoter. Infect. Immun. 69:2154-2161. |
| 9. | Hunt, P. D., and S. J. Hardy. 1991. Heat-labile enterotoxin can be released from Escherichia coli cells by host intestinal factors. Infect. Immun. 59:168-171. |
| 10. | Kaniga, K., I. Delor, and G. R. Cornelis. 1991. A wide-host-range suicide vector for improving reverse genetics in gram-negative bacteria: inactivation of the blaA gene of Yersinia enterocolitica. Gene 109:137-141.[CrossRef][Medline] |
| 11. | Kohler, J. J., L. B. Pathangey, S. R. Gillespie, and T. A. Brown. 2000. Effect of preexisting immunity to Salmonella on the immune response to recombinant Salmonella enterica serovar Typhimurium expressing a Porphyromonas gingivalis hemagglutinin. Infect. Immun. 68:3116-3120. |
| 12. | Liang-Takasaki, C.-J., N. Grossman, and L. Leive. 1983. Salmonellae activate complement differentially via the alternative pathway depending on the structure of their lipopolysaccharide O-antigen. J. Immunol. 130:1867-1870.[Abstract] |
| 13. | Ménard, R., P. J. Sansonetti, and C. Parsot. 1993. Nonpolar mutagenesis of the ipa genes defines IpaB, IpaC, and IpaD as effectors for Shigella flexneri entry into epithelial cells. J. Bacteriol. 175:5899-5906. |
| 14. | Mooi, F. R., F. K. de Graaf, and J. D. van Embden. 1979. Cloning, mapping and expression of the genetic determinant that encodes for the K88ab antigen. Nucleic Acids Res. 6:849-865. |
| 15. | Morona, R., L. van den Bosch, and P. A. Manning. 1995. Molecular, genetic, and topological characterization of O-antigen chain length regulation in Shigella flexneri. J. Bacteriol. 177:1059-1068. |
| 16. | Roberts, M., A. Bacon, J. Li, and S. Chatfield. 1999. Prior immunity to homologous and heterologous Salmonella serotypes suppresses local and systemic anti-fragment C antibody responses and protection from tetanus toxin in mice immunized with Salmonella strains expressing fragment C. Infect. Immun. 67:3810-3815. |
| 17. | Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. |
| 18. | Schodel, F., and R. Curtiss III. 1995. Salmonellae as oral vaccine carriers. Dev. Biol. Stand. 84:245-253.[Medline] |
| 19. | Stocker, B. A. 2000. Aromatic-dependent Salmonella as anti-bacterial vaccines and as presenters of heterologous antigens or of DNA encoding them. J. Biotechnol. 83:45-50.[CrossRef][Medline] |
| 20. | Wang, R. F., and S. R. Kushner. 1991. Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100:195-199.[CrossRef][Medline] |
| 21. | Whittle, B. L., and N. K. Verma. 1997. The immune response to a B-cell epitope delivered by Salmonella is enhanced by prior immunological experience. Vaccine 15:1737-1740.[CrossRef][Medline] |
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| J. Bacteriol. | J. Virol. | Eukaryot. Cell |
|---|
| Microbiol. Mol. Biol. Rev. | Clin. Vaccine Immunol. | All ASM Journals |
|---|
); 1.3 x 109 CFU of aroA S. enterica serovar Stanley on day 0 (
); 1.5 x 1010 CFU of aroA S. enterica serovar Stanley on day 0 and 1.2 x 1010 CFU of aroA S. enterica serovar Stanley on day 4 (
). (B) Mice were fed aroA S. enterica serovar Dublin, either 1.2 x 109 CFU on day 0 (
) or 9.4 x 109 CFU on day 0 and 9.5 x 109 CFU on day 4 (
). Histograms show log10 bacterial burden (GM ± SD; n = 4) in Peyer's patches; lines show limit of detection (20 bacteria).
, 1.5 x 109 CFU on day -70) or aroA S. enterica serovar Stanley (
, 1.1 x 1010 CFU day 0 and 1.0 x 1010 CFU on day 4). Serum IgG (A) and FPS IgA (B) responses to LPS are shown as log10 ELISA titers (GM ± SD; n = 5), the latter being standardized per milligram of IgA. For clarity, SDs are not shown for one group. The horizontal line shows the limit of detection of serum ELISAs at a titer of 20.
