TEXT

Top Abstract Text References

Recombinant avirulent Salmonella vaccines that are attenuated, yet invasive, are effective vehicles for delivering heterologous antigens to the mucosal and systemic immune systems (7). The potential advantage of such a bacterial vaccine vector is its ability to replicate and express heterologous antigen in vivo. Salmonella species cause typhoid fever in their natural hosts after infection via the oral route. They invade the host by crossing the gut mucosa and colonize the gut-associated lymphoid tissue (GALT), as well as the spleen and liver. Deletion of various genes can render Salmonella avirulent while preserving different degrees of invasiveness and thus immunogenicity (7). Several attenuations have been constructed, including nutritional auxotrophs impaired in their pathways for biosynthesis of aromatic amino acids (aro mutants [2, 11, 18a, 34]), strains harboring deletions in the adenylate cyclase (cya) gene and cyclic 3',5'-AMP receptor protein (crp) genes that are required for the transcription of many genes involved in catabolite transport and breakdown (9), as well as mutants with mutation in the phoP genetic locus (15), a two-component regulatory system (phoP-phoQ) that controls the expression of genes essential for Salmonella virulence (15b, 23) and survival within macrophages (13, 14). Salmonella typhimurium causes a typhoid-like disease in mice and has been used extensively to generate attenuations which can be evaluated for their safety and immunogenicity in the murine model.

We have shown that vaccination with attenuated salmonellae can lead to specific antibody responses in the genital tract of mice (19, 31, 32) and at least one human female volunteer (25). This suggests that Salmonella-based vaccines could be used to control infection by human genital pathogens. The high-risk human papillomavirus (HPV) types, most commonly type 16 (HPV16), are etiologically linked to over 90% of cervical cancers (3). Cervical cancer is the second leading cause of cancer deaths in women worldwide, encouraging the development of a prophylactic vaccine to prevent genital infection by these viruses. Recently, we have demonstrated that intranasal immunization of mice with live recombinant S. typhimurium PhoP^c bacteria expressing HPV16 virus-like particles (VLPs) resulted in HPV16-specific conformational and neutralizing immunoglobulin A (IgA) and IgG in genital secretions (26). However the PhoP^c strain contains a point mutation in the phoQ gene (15a, 16, 23a) and can revert at high frequency (24). Therefore, we decided to evaluate the HPV16-specific antibody responses elicited by S. typhimurium strains harboring attenuations (PhoP and cya crp/cdt), which are more suitable for a human vaccine (Ty800 [18] and 4632 [25, 35]). To avoid bias due to the different genetic background (SR-11) of the existing cya crp strains (9), new isogenic strains were constructed.

Expression of HPV16 L1 and HPV16 VLPs in the four isogenic strains. 4989 (cya crp) and 4990 [cya (crp-cdt)] were constructed in the ATCC 14028 wild-type strain as the recipient by methods described by Zhang et al. (41). Plasmid pFS14nsdHPV16-L1 (26) was electroporated (29) into the four S. typhimurium isogenic strains: i.e., the PhoP^c (CS022 [24]) and PhoP (CS015 [22]) strains and 4989 and 4990 (this paper). Western blot analysis of bacterial lysates revealed only minor differences in the levels of HPV16 L1 expression (57-kDa band in Fig. 1A) among the different strains examined. Densitometric analysis of the Western blot bands obtained from five different preparations of each recombinant strain are shown in Fig. 1B. The maximal difference observed in HPV16 L1 expression was 2.5-fold between 4990 and 4989. We were not able to detect HPV16 L1 in the supernatant of Salmonella cultures, suggesting that none of these strains secreted the HPV16 VLP antigen. We have previously shown that the anti-HPV16 L1 antibodies measured with our enzyme-linked immunosorbent assay (ELISA) are conformational and only recognize assembled HPV16 VLPs (26). It was therefore important to establish whether, despite similar levels of HPV16 L1 expression, the assembly of VLPs could differ between the recombinant strains. For all Salmonella strains, HPV16 VLPs could be extracted by sonication of the bacteria, suggesting that they were not located in inclusion bodies. To determine the amount of VLPs assembled in Salmonella, we developed a sandwich ELISA by using two monoclonal antibodies that recognize conformational epitopes on HPV16 VLPs (H16.E70 and H16.V5; kindly provided by N. D. Christensen, Hershey, Pa. [6]). The amounts of VLPs produced in preparations of the different recombinant strains were determined by using VLPs produced in insect cells as standards (21, 26), but only minor differences were measured (Fig. 1C).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Expression of HPV16 L1 and VLP production in the four recombinant isogenic Salmonella strains. Salmonellae were grown overnight at 37°C and lysed by boiling in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer containing 5% SDS. Bacterial lysate equivalent to 3 × 107 CFU was separated by SDS-PAGE, and an immunoblot with an anti-HPV16 L1 monoclonal antibody (MAb), CAMVIR-1, is shown (A). The 57-kDa protein band identified as L1 is indicated by an arrow. Scanning of the L1 protein bands obtained in five independent experiments was performed by using NIH Image software. The results are shown as the means of the pixel densities of the L1 protein bands expressed in units per 1011 CFU (B). VLP production was assessed in overnight cultures. After sonication, the bacteria were centrifuged to discard the debris, and the supernatants were analyzed for the presence of HPV16 VLPs. The amounts of VLPs were determined by sandwich ELISA. Microtiter plates were coated with an anti-HPV16 VLP conformational MAb, H16.E70, another biotinylated anti-HPV16 VLP conformational MAb, H16.V5, was used as the secondary antibody, and HPV16 VLPs produced in insect cells (26) were used as a standard. The results are shown as the means of VLPs in micrograms per 1011 CFU measured in five independent experiments (C). Error bars in panels B and C indicate standard errors.

HPV16 VLP-specific antibody responses after nasal immunization with four S. typhimurium attenuated isogenic strains. Our previous results involving nasal immunization with the PhoP^c strain expressing HPV16 VLPs demonstrated the efficacity of this route of immunization in contrast to the oral route (26). Moreover, pilot experiments with administration of the other Salmonella strains expressing HPV16 VLPs by the oral route did not result in the induction of anti-HPV16 VLP antibodies (data not shown). Therefore, nasal immunization was used in the following experiments. Nasal immunization under anesthesia and sampling of female BALB/c mice were performed as described previously (19, 26). Four mice per group were immunized with 20 µl of inoculum (ca. 10⁷ CFU, which corresponds to approximately 5 ng of HPV16 VLPs) of the different recombinant strains at weeks 0 and 12. Sampling of blood, saliva, and genital secretions was performed at weeks 0, 2, 4, 6, 12, 15, and 22. The anti-HPV16 VLP and anti-lipopolysaccharide (LPS) IgG and IgA titers were determined by ELISA (26) and are shown in Fig. 2. Surprisingly, only the PhoP^c HPV16 strain elicited sustained levels of HPV-specific antibodies in serum and secretions. PhoP HPV16 did not elicit any detectable HPV-specific antibodies, while 4989 HPV16 induced transiently low anti-HPV16 VLP IgG levels in serum and 4990 HPV16 induced variable low anti-HPV16 IgG levels in serum. Control ELISAs using purified HPV16 VLPs in carbonate buffer as the coating antigen did not reveal the presence of antibodies against unfolded HPV16 VLPs induced by any of the recombinant strains examined (data not shown). All recombinant strains induced anti-LPS antibodies at variable levels, but 4990 required a booster immunization in order to do so.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Anti-HPV16 VLP and anti-LPS systemic IgG and mucosal IgA after nasal immunization with the four Salmonella HPV16 strains. Groups of four 6-week-old BALB/c female mice were immunized with ca. 107 CFU at weeks 0 and 12 (indicated by arrows). Data are expressed as the geometric means (log10) of reciprocal dilutions of specific IgG in serum and specific IgA per microgram of total IgA in secretions. Error bars indicate the standard errors of the means.

The four S. typhimurium attenuated isogenic strains expressing an unrelated viral antigen can induce specific antibodies after nasal immunization. The immunogenicity of recombinant 4989 and 4990, as well as that of the PhoP strain, has not been tested previously. Therefore, we tested the antibody responses elicited by these attenuated isogenic strains expressing the hepatitis B nucleocapsid HBc, an unrelated viral antigen that we have extensively used in the past (19, 30). The plasmid expressing HBc, pFS14PS2 (28), was electroporated in the four isogenic strains, and female BALB/c mice were intranasally immunized with 10⁷ CFU at week 0. Sampling was performed at weeks 4, 7, and 11, and the presence of anti-HBc antibodies was determined by ELISA as described previously (19). In contrast to the results obtained with HPV16, all HBc recombinant strains induced HBc-specific antibodies in serum and secretions (Fig. 3), although the PhoP strain appeared to induce lower anti-HBc IgG levels in serum.

View larger version (23K): [in this window] [in a new window]   FIG. 3.   Anti-HBc IgG in serum and IgA in secretions after nasal immunization with the four isogenic Salmonella HBc strains. Groups of four 6-week-old BALB/c mice were immunized with 107 CFU at week 0. Data are expressed as the geometric means (log10) of reciprocal dilutions of specific IgG in serum and specific IgA per microgram of total IgA in secretions. Error bars indicate the standard errors of the means.

Recovery of the recombinant S. typhimurium 2 weeks after nasal immunization. In order to obtain insights into the mechanisms that underlie the different antibody responses observed after nasal immunization with the different recombinant strains, we analyzed the survival of the bacteria and the maintenance of the plasmid encoding the foreign antigens in various organs. Survival experiments after oral immunization with various attenuated S. typhimurium strains have been performed, and the results have been used as indicators for the probable immunogenicity of a given strain (7, 9, 12, 30). Survival of bacteria after nasal immunization was only partially examined previously (1, 26, 31). We have shown that in anesthetized mice, one-third of the inoculum reaches the lungs, while the rest is swallowed (1). Here, a pilot experiment using 5 × 10⁷ CFU of the PhoP^c HBc strain as a nasal inoculum revealed that bacteria reached the lungs, cervical lymph nodes, Peyer's patches, and spleen, from which they disappeared after 40 days (Fig. 4). About 2 weeks after nasal immunization, the numbers of bacteria peaked or plateaued in all of the organs examined. Therefore, we intranasally immunized eight groups of three female BALB/c mice with ca. 10⁷ CFU of each of the recombinant S. typhimurium strains and examined survival in lung, cervical lymph nodes, Peyer's patches, and spleen after 2 weeks. The organs were processed as previously described (26), and recovered bacteria were analyzed on medium containing or not containing ampicillin in order to discriminate between the Salmonella strains still harboring the plasmid encoding the heterologous antigen and the Salmonella strains that had lost the plasmid. As we have already described for PhoP^c HPV16 (26), the plasmid encoding the HPV16 antigen was unstable in vivo in all of the strains examined, although at various levels (Table 1). This probably reflects the toxicity of the L1 molecule and a lack of antibiotic selection in vivo. In contrast, the plasmid encoding HBc was stable in all strains except the PhoP one. The total numbers of salmonellae recovered after immunization with identical strains expressing either the HPV16 or the HBc antigens were very similar. However, the survival of the bacteria in the various organs differed, depending on the strain analyzed and irrespective of the antigen expressed. As expected, few or no bacteria were recovered from the spleen of mice immunized with the PhoP strain (15) or with 4990, a strain that bears a deletion in cdt, a gene responsible for colonizing deep tissue (8, 10, 20). However, we were unable to find any correlation between the number of salmonellae localized in a given organ and the antibody responses induced. For instance, the highest numbers of salmonellae still harboring the HPV16 plasmid were found with 4989 and not with the PhoP^c strain. An exception might be the lowest response, by the PhoP HBc strain, which might correlate with the instability of the HBc plasmid in that particular strain.

View larger version (11K): [in this window] [in a new window]   FIG. 4.   Recovery of S. typhimurium PhoPc HBc after nasal immunization. Groups of three 6-week-old BALB/c female mice were immunized with 5 × 107 CFU at day 0 and were sacrificed at different days postimmunization. The different organs were processed and plated on agar as previously described (26). Data are expressed as the geometric mean (log10) CFU per organ. The level of detection (20 CFU/organ, or log10 1.3) is indicated by a horizontal dashed line, and values below the level of detection are shown below that line. Error bars indicate the standard errors of the means.