INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Two classes of poliovirus vaccines were developed over four decades ago: the formalin-inactivated poliovirus vaccine (IPV) developed by Jonas Salk (20) and the live attenuated oral poliovirus vaccine (OPV) developed by Albert Sabin (26). Both vaccines elicit effective humoral immune responses that protect from poliomyelitis, but only OPV induces strong mucosal immunity and is able to prime cellular immune responses (26). Trivalent OPV has been the prevalent poliovirus vaccine used in the United States and many other countries. Its extensive use in humans has demonstrated its safety and its ability to induce long-lasting protective immunity. In addition, OPV is easy to administer orally, its low cost enables ample distribution in the developing world, and it induces both systemic humoral and cellular immunity as well as local mucosal resistance to poliovirus infection (26). In addition, quality and safety tests for OPV are well established (28).

Given these favorable characteristics of the Sabin poliovirus vaccine, recombinant poliovirus expressing foreign antigens may provide a convenient and safe vaccine vector system to induce protective immunity against diverse pathogens. Chimeric polioviruses have been constructed using several approaches (1, 2, 6, 15, 18). One of these approaches uses replicons in which the genes encoding the poliovirus capsid proteins are replaced by foreign sequence (18). This approach requires a helper virus for viral propagation, which potentially limits spread in vivo. The tight control of the spread of propagation-defective viral vectors is an attractive safety feature, but also may limit their potential to stimulate a vigorous immune response.

Our approach, in contrast, uses recombinant viruses that are able to self-propagate because they encode all viral proteins. We have inserted sequences encoding foreign antigens in frame within the poliovirus polyprotein. The inserted sequences are flanked by poliovirus protease recognition sites. Thus, initially a larger polyprotein is made, but proteolytic processing ensures production of mature and functional viral proteins plus the exogenous antigen. Vaccination of both mice and nonhuman primates with this type of recombinant poliovirus induces strong antibody and cytotoxic T-lymphocyte (CTL) responses (2, 14, 27, 31). Furthermore, inoculation of a recombinant poliovirus that expresses the C-terminal half of chicken ovalbumin (Polio-Ova) protected 100% of the immunized mice against challenge with lethal doses of a malignant melanoma expressing ovalbumin (14).

While there are many advantages in adapting common nonpathogenic viruses and well-established viral vaccines for therapeutic purposes, an important drawback of this approach is that preexisting immunity to the virus in the human population could reduce or completely abolish their therapeutic efficacy. In particular, the wide use of OPV may constrain the use of poliovirus as a vaccine vector. Indeed, this appears to be the case for other commonly used viral vectors. Numerous studies employing recombinant vaccinia virus and adenovirus vectors have demonstrated that one of the greatest challenges in the development of viral vectors is the host immune response against the vector (3, 4, 13, 19, 23, 30).

The effects of preexisting immunity to poliovirus on the efficacy of recombinant poliovirus vaccines had not previously been studied in detail. One study demonstrated that preimmunity induced by IPV does not impair the ability of poliovirus replicons (expressing the C fragment of tetanus toxin) to induce antibody responses against the foreign protein in susceptible mice (21). However, since OPV has been the prevalent vaccine in most countries, and given the different immune responses elicited by IPV and OPV, it is important to address the effects of OPV immunization on the therapeutic potential of replication-competent, recombinant poliovirus vaccine vectors.

Therefore, we investigated the effect of preexisting immunity elicited by immunization with wild-type Mahoney type 1 or live attenuated poliovirus on the outcome of vaccination with the recombinant poliovirus expressing chicken ovalbumin, Polio-Ova. We compared antibody titers, CTL activity, and ability to induce protection against lethal tumor challenge. In this experimental murine model, preexisting immunity to poliovirus has little or no effect on the ability to induce effective immunity to the foreign protein.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice and cell lines. Female poliovirus receptor (PVR)-transgenic mice (23) 6 to 8 weeks old were used for all experiments. C57BL/6-derived melanoma B16F0 (29) was obtained from the American Type Culture Collection. B16-Ova (Mo5.20.10) was constructed by transfection of B16F0 with the pAc-neo-Ova plasmid as described previously (11, 17). B16F0 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, and 1% penicillin-streptomycin (GPS). B16-Ova cells were grown in DMEM supplemented with 10% FCS, 10 mM HEPES buffer, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, GPS, 0.05 mM 2-mercaptoethanol, 60 µg of hygromycin B per ml, and 0.5 mg of G418 per ml. EL-4 cells were grown in the same medium as B16F0 cells. EG-7 cells (a subclone of EL-4 stably transfected to express ovalbumin) (18) were grown in RPMI 1640 supplemented with 10% FCS and GPS and constantly selected with G418 (0.5 mg/ml). HeLa cells were grown in suspension in Joklik's modified essential medium supplemented with 10% horse serum and GPS. HeLa cell monolayers were grown in DMEM/F12 supplemented with 10% newborn calf serum and GPS.

Viruses. Two recombinant viruses were used to vaccinate PVR-transgenic mice. Polio-Sp27 is an attenuated Mahoney type 1 virus (previously named MoV-2.11-Sp27) containing a 687-nucleotide insert coding for the p27 region of the simian immunodeficiency virus gag gene (27). Polio-Ova (MoV-2.11-Ova) contains an insert of 600 nucleotides encoding the C-terminal half of chicken ovalbumin (Ova). Ova contains the T-cell epitope SIINFEKL. Polio-Sp27 and Polio-Ova produced plaques that are somewhat smaller than those corresponding to wild-type virus. However, these recombinant viruses reach comparable titers and replicate in tissue culture with similar kinetics to wild-type virus (14, 27). Both recombinant viruses are attenuated in transgenic mice due to the foreign sequence insert. None of the mice inoculated with either virus (with up to 5 × 10⁹ PFU intraperitoneally) ever developed poliomyelitis. Wild-type Mahoney type 1 virus derived from plasmid pXpA (5) was also used to induce antipoliovirus immunity and for subsequent challenges (see Results).

Immunizations. (i) Induction of antipoliovirus immunity. Mice were inoculated twice (day 0 and day 4) intraperitoneally with 10⁶ PFU of Polio-Sp27 in 100 µl of phosphate-buffered saline (PBS). Alternatively, mice were inoculated once with 10⁶ PFU of Mahoney type 1 virus. We could not give a second dose of Mahoney type 1 because it induces paralytic poliomyelitis in a significant number of mice.

(ii) Polio-Ova immunizations. Six to eight weeks after Polio-Sp27 or Mahoney immunization, naive and preimmune mice were inoculated intraperitoneally three times, once every 4 days, with different amounts (indicated in the text and Table 1) of Polio-Ova in 100 µl of PBS.

Challenges. (i) Viral challenge. Immunized and naive mice were infected intraperitoneally with 5 × 10⁸ PFU of wild-type, pathogenic Mahoney type 1 and monitored daily for paralytic poliomyelitis.

(ii) Tumor challenge. Immunized or naive control mice were challenged by subcutaneous injection of 10⁵ ovalbumin-expressing or parental B16F0 melanoma cells in 100 µl of PBS. Melanoma cells for injection were harvested by limited trypsinization and washed once with PBS. Cells used for injection were more than 95% viable, as determined by trypan blue dye exclusion. Mice with a tumor bigger than 0.3 cm² were scored as positive.

Plaque reduction assay. Fifty microliters of preimmune or naive mouse serum (diluted 1:5 in PBS) was incubated with 10⁵ PFU of poliovirus in 50 µl of PBS at 37°C for 30 min. Controls were incubated with PBS only. The amount of remaining infectious virus after incubation with serum was then determined by plaque assays on HeLa cells (5).

Western blot analysis. Western blotting was performed essentially as described previously (31). Briefly, approximately 40 µg of partial purified poliovirus was subjected to electrophoresis through a 10% polyacrylamide gel with sodium dodecyl sulfate and analyzed by immunoblotting using serum samples diluted 1:100 in blocking buffer. Proteins were identified using a secondary antibody (anti-mouse immunoglobulin [Ig]; Amersham, Arlington Heights, Ill.) at a 1:2,000 dilution and a chemiluminescent detection system (ECL kit; Amersham) as directed by the manufacturer.

Expression of ovalbumin C terminus. The C-terminal fragment of ovalbumin with an N-terminal 6× His tag was expressed using a T7 expression plasmid (pT3CD H6-41) and purified over nickel-nitrilotriacetic acid columns (Qiagen, Valencia, Calif.). The purified protein was quantified using the Bradford assay.

ELISA. Enzyme-linked immunosorbent assay (ELISA) plates (Nunc Immuno Plate Maxisorp F96) were coated overnight with 10 µg of protein per ml in PBS at 4°C and then blocked in 4% milk-PBS (blocking buffer) for 1 h at room temperature. Plates were subsequently washed, and 50 µl of serum (diluted in blocking buffer) was added to each well and incubated for 4 h at room temperature. Bound antibody was detected using goat anti-mouse IgG conjugated to horseradish peroxidase (Southern Biotechnology, Birmingham, Ala.). Plates were developed by adding 50 µl of TMB (3,3,5,5-tetramethylbenzadine; Sigma, St. Louis, Mo.). After 30 min at room temperature in a darkened area, the reaction was stopped by adding 50 µl of sulfuric acid per well. Plates were read at 450 nm.

⁵¹Cr release assay. Spleens from immunized and naive control mice were removed and dispersed into single-cell suspensions. Splenocytes (4 × 10⁷) from each spleen were restimulated by cocultivation with 10⁶ irradiated (10,000 rad) EL-4 and EG-7 cells in upright T25 tissue culture flasks (Becton Dickinson, Franklin Lakes, N.J.) in 10 ml of complete RPMI medium. Effector cells were harvested after 5 days of restimulation, and specific cytotoxic activity of CTLs was determined by a ⁵¹Cr release assay. Briefly, restimulated effector cells were incubated for 5 h in 200 µl of cRPMI with 5 × 10^{3 51}Cr-labeled target cells at the indicated effector-to-target cell ratio. Percent specific release was calculated using the formula [(experimental release  spontaneous release)/(maximum release  spontaneous release)] × 100, where values for spontaneous and maximum release were obtained from target cells cultured in the absence of splenocytes and target cells were lysed with 2% Triton X-100, respectively. Spleens restimulated with EL-4 cells, which do not express ovalbumin, did not develop cytolytic activity (not shown). Values represent averages of triplicate wells, and variation between wells was consistently less than 5%. Statistical analysis was performed using the F test.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Induction of preimmunity. The studies described here were initiated to determine whether preexisting immunity to poliovirus induced by vaccination with live poliovirus impairs an effective immunization with recombinant poliovirus vaccines. An overview of our experimental approach is presented in Fig. 1. First, it was important to determine whether the vaccination protocol used in this study confers protective immunity to poliovirus. In order to model immunity induced in humans by live attenuated OPV, we immunized mice with sublethal doses of live wild-type Mahoney type 1 (Mahoney) or a live attenuated poliovirus (Polio-Sp27; see Materials and Methods). We used Mahoney-derived viruses to induce antipoliovirus immunity because Sabin strains do not replicate efficiently in transgenic mice. We have previously shown that two intraperitoneal immunizations with 10⁶ PFU of Polio-Ova are sufficient to induce strong immunity that protects 100% of the animals from lethal challenge with B16 melanoma cells expressing the same antigen. Thus, we used a similar protocol to ensure effective antipoliovirus immunity in mice (defined as immunity sufficient to protect against paralysis and death).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the experimental design. OVA, C-terminal half of chicken ovalbumin.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Antipoliovirus serum IgG. Sera from all mice used in this study were analyzed, but only results for 30 mice are shown. Numbers identify individual mice. Mice 1 to 10 were immunized by a single intraperitonal injection of 106 PFU of Mahoney, mice 16 to 25 received 106 PFU of Polio-Sp27 twice by the same route, and naive controls 11 to 15 and 26 to 30 were injected with PBS. (A) Sera from Mahoney-immunized mice were diluted 1:100 and analyzed by Western blot for antibodies against poliovirus capsid proteins. (B) Poliovirus neutralizing antibodies from sera of the same animals were determined by plaque reduction assay. (C) Poliovirus neutralizing antibodies from sera of mice immunized with Polio-Sp27 were determined as in panel B. Neutralizing activity is expressed as fold reduction compared to naive controls.

Effects on anti-Ova antibody titers. Next, we investigated whether preexisting immunity to poliovirus affects the ability of Polio-Ova to induce effective anti-Ova immunity. To evaluate the effect on the induction of antibodies directed against the foreign antigen, naive and preimmune mice were inoculated with Polio-Ova and bled 4 to 5 weeks after the final inoculation. The antibody titers elicited against Ova in naive and preimmune mice were indistinguishable (Fig. 3A). Titers ranged from 5 to 625. However, these relatively low titers and variability were observed in both naive and preimmune animals and may be due to the antigenic characteristics of the inserted protein (just the C-terminal half of chicken ovalbumin).

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Ova-specific antibodies and CTL. (A) Naive (triangles) and Polio-Sp27-immunized (diamonds) mice were inoculated three times with 106 PFU of Polio-Ova. Control animals were inoculated with PBS (circles). Five weeks after the first inoculation, mice were bled, and sera were analyzed for ovalbumin-specific antibodies by ELISA. Antibody titers are expressed as the reciprocal of the lowest serum dilution that gave an absorbance of 0.1 unit above the background. Each symbol represents an antibody titer for an individual mouse. (B) CTL activity measured by a 51Cr release assay. Naive mice (triangles) and mice vaccinated with Polio-Sp27 (diamonds) were inoculated with Polio-Ova as described in the text. Control mice were injected with PBS only (circles). Three weeks after the immunization, spleens from all mice were harvested and cultured with irradiated EG-7 (EL-4 cells transfected with ovalbumin). After 5 days, cells were collected and tested in 51Cr release assays on EG-7 targets (solid symbols) or on EL-4 cells as controls (open symbols). Splenocytes restimulated with EL-4 cells did not kill either target (data not shown). Results are means for five mice per group. Each spleen was tested individually. The level of CTL activity is significantly higher (dilutions 1:25 to 1:100) in the groups that were immunized with Polio-Ova (with or without prevaccination with poliovirus) than in control naive mice (P < 0.05).

Effects on anti-Ova CTL responses. To study the effects of preexisting immunity to poliovirus on the induction of ovalbumin-specific CTL responses, we immunized naive and preimmune mice with Polio-Ova and tested the specific CTL activity elicited by Polio-Ova inoculation. Control mice were immunized with PBS only. Significant antigen-specific CTL activity was detected from splenocytes derived from naive as well as preimmune mice (P < 0.05) (Fig. 3B). However, the overall CTL activity was lower in preimmune mice, indicating that preexisting immunity to the vector does, to some extent, reduce the induction of Ova-specific CTL.

Effects on protective antitumor immunity. Finally, we examined the ability of the recombinant vector to induce protective antitumor immunity in preimmune mice. Naive and preimmune mice were immunized with Polio-Ova and challenged 6 to 8 weeks later with 10⁵ B16 melanoma cells expressing chicken ovalbumin (B16-Ova). All naive mice immunized with the Polio-Ova were completely protected from lethal challenge, while control mice which had not received Polio-Ova all developed tumors. In preimmune mice (both Polio-Sp27- and Mahoney-inoculated animals), tumor protection was somewhat diminished (Table 1). Because CTL activity plays an important role in the control of tumor growth in this tumor model (11, 15a), it was likely that the reduction in CTL activity produced by preimmunity to the vector determined the reduction in tumor protection observed.

Higher vaccine doses fully restore vaccine efficacy. To determine whether we could overcome the effects of preexisting immunity by increasing the vaccination dose, we immunized mice with 10- to 20-fold-higher doses of Polio-Ova. Immunization with 10⁷ PFU protected 90% of the mice preimmunized with Polio-Sp27, while immunization with 2 × 10⁷ PFU fully restored protection to tumor in mice preimmunized with either Polio-Sp27 or Mahoney (Table 1). Thus, the effects of preexisting immunity to the poliovirus vector can be overcome by increasing the vaccination dose.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A major concern for the use of recombinant viral vaccine vectors is the potential interference of preexisting immunity to the vector with the efficacy of the recombinant vaccine. Such immunity could be induced by natural infection or by vaccination. In this study, we investigated the effects of preexisting immunity to poliovirus induced by either wild-type Mahoney or live attenuated poliovirus on the immunogenicity of recombinant poliovirus vaccine vectors. We found that serum from preimmune and naive mice showed no significant difference in anti-Ova antibody titers and that inoculation with Polio-Ova induced specific CTL in mice with preexisting immunity. However, we observed that the overall activity of CTLs in preimmune mice is lower than in naive mice. Consistent with this finding, a lower percentage of preimmune mice were protected from challenge with melanoma cells (66% of those immunized with Polio-Sp27 and 88% of those immunized with Mahoney). Interestingly, 100% tumor protection could be restored in both cases by using higher doses of the recombinant vaccine. We concluded that even in the presence of antipoliovirus immunity that effectively protects mice from poliomyelitis after lethal challenge with pathogenic virus, poliovirus vaccine vectors are able to induce both humoral and cellular immune responses against the foreign antigen.

Although we have not attempted to determine the amount of antigen produced in preimmune and naive mice, it is reasonable to assume that replication of the recombinant poliovirus and consequently the production and presentation of recombinant protein are reduced by preexisting immunity. Nonetheless, we found that the efficacy of vaccination with a recombinant poliovirus in preimmune mice could be increased in a dose-dependent manner, and 100% protection was attained. This is not the case for all viral vaccine vectors. Kundig and colleagues have shown that a 100-fold-higher dose of a vaccinia virus recombinant could only partially restore antibody responses in preimmune mice, and at these high doses, naive mice developed clinical symptoms and died (13). Thus, these data suggest that it is potentially dangerous to overcome preexisting immunity to vaccinia virus by increasing the dose of vaccination.

In contrast, we have injected naive mice with as much as 5 × 10⁹ PFU of the recombinant poliovirus vectors and never saw side effects from the vaccination.

Why does immunization with higher doses of the recombinant poliovirus vaccine result in effective vaccination? A possible explanation is that a rapid and widespread infection in the host might lead to expression and presentation of sufficient amounts of antigen, before preexisting poliovirus immunity suppresses viral replication and protein expression. It is also possible that immunity to poliovirus is not as tight as to other viruses and may not confer an absolute "sterilizing" immunity that prevents reinfection.

The effect of preexisting immunity to vaccinia virus is more severe in animals with higher antibody titers to the virus (3, 23). We determined the antipoliovirus antibody titers of all preimmune mice and integrated whether tumor occurrence correlated with high titers of antipoliovirus antibodies. We found that mice with higher neutralizing antipoliovirus antibody titers did not get tumors more frequently (data not shown). This was unexpected, since immunity to poliovirus is thought to be mostly antibody mediated, but the role of cellular immunity in protection against poliovirus infection has not been fully evaluated. In this regard, we have shown that poliovirus infection in PVR-transgenic mice elicits strong CTL responses (14, 25). However, poliovirus inhibits protein secretion and presumably also downregulates expression of major histocompatibility complex class I molecules (9, 10). Thus, the relevance of CD8⁺ T cells in poliovirus pathogenesis and immunity has not been adequately assessed and remains poorly understood.

Our findings suggest that the effects of preexisting immunity can be very different for each virus and must be studied on a case-by-case basis. Besides the type of immunity that contains the viral infections, other determinants may play a role. Factors like the kinetics of viral replication, the kinetics and type of host cell death, and the site of viral replication also vary between commonly studied vaccine vectors and are likely to influence the expression and presentation of the foreign antigen.

It is difficult to extrapolate the relevance of these studies to the human situation. Poliovirus replicates more robustly in humans than in mice, and other species-specific differences as well as different routes of infection used in mice and humans may influence the outcome of revaccination. Despite these considerations, our results can be relevant for vaccination in humans. For instance, recent research with vaccinia virus recombinants in people demonstrated that in humans, as in mice, preexisting immunity to vaccinia virus limits the effectiveness of recombinant vaccinia virus vectors (7, 24). Therefore, at least in the case of vaccinia virus, results in mice predict the outcome in humans. In addition, it is well known that following household exposure to wild-type poliovirus, 20 to 50% of naturally infected (12), 30 to 50% of OPV-vaccinated, and 90 to 100% of IPV-vaccinated persons were reinfected by wild-type poliovirus. This was indicated by excretion of virus in the gastrointestinal tract and increase in antibody titer (16). These data suggest that poliovirus can replicate in vaccinated people.

In conclusion, we have previously shown that replication-competent recombinant poliovirus vaccine vectors stimulate a broad immune response against the desired antigen in mice and a primate model system (Cynomolgus macaques) (8). Here, we show that preexisting immunity to poliovirus vectors does not seriously impair the immunogenicity of the poliovirus recombinants. If vaccination of macaques with preexisting immunity to poliovirus proves as successful as in mice, poliovirus vectors have a good chance to be effective vaccine candidates in humans previously exposed to poliovirus.