CZ20003851A3 - Attenuated mutated Salmonella strain and typhoid vaccine - Google Patents

Abstract

Attenuated mutant strains of Salmonella typhi stably expressing the Vi antigen are described, as well as a typhoid vaccine containing these mutant strains; this vaccine can also be used as a live vector or DNA vaccine.

Description

WEAKENED MUTATED STRAIN OF ALMONELLA AND TYPHOID VACCINE

TECHNICAL AREA

The proposed invention describes attenuated Salmonella mutants that stably express the Vi antigen on their surface, as well as typhoid fever vaccines containing this substance, live vector vaccines containing this substance, and DNA-based vaccines containing this substance.

STATE OF THE ART

Antigen Vi

The Vi antigen is a capsular polysaccharide that was first described in Felix et al., Lancet, 227:186-191 (1934). This capsular polysaccharide is expressed by Salmonellae such as Salmonella typhi, Salmonella paratyphi C and Salmonella dublin, as well as in Citrobacter freundii. Structurally, the Vi antigen is a linear polymer of P-4,2-deoxy-2N-acetylgalactouronic acid with variable O-acetylation (Daniels et al., Infect. Imun., 57:3159-3164 (1989)). The presence of this antigen on the surface of Salmonella typhi is directly proportional to the reduction in serum lysis, complement activation and phagocytosis in vitro (Looney et al., J. Lab., Clin. Med., 108:506-516 (1986)). The Vi antigen may therefore serve as a shield protecting Salmonella typhi against the immune system.

Chromosomal region viaB and regulation of Vi antigen expression

Three distant chromosomal loci, viaA, viaB, and ompB, are most likely essential for the expression of the Vi antigen (Johnson et al., J. Bacteriol., 90: 302-308 (1965) and Snellings et al., J. Bacteriol., 145:1010-1017 (1981)). Of these, the viaB locus is always present in Vi antigen-positive strains and probably encodes genes encoding enzymes essential for the synthesis of the Vi antigen (Hashimoto et al., J. Bacteriol., 175:4456-4465 (1993), Virlogeux et al., Microbiol., 141:3039-3047 (1995)). The viaB locus consists of 11 open reading frames (ORFs) (Fig. 1A), of which the vipA and vipB genes encode enzymes that synthesize the nucleotide sugar of the Vi polysaccharides and 5 vex genes (vexA-E) that are most likely responsible for the translocation of the Vi antigen (Hashimoto et al., (1993) see above). The first open reading frame of the viaB region, vipR (Fig. 1A), is a positive transcriptional regulator of its own expression as well as the expression of vipA, vipB, orf4, vipC and probably other vipR-mock genes (Hashimoto et al., J. Bacteriol., 178:1430-1436 (1996)). In addition, the promoter upstream of vipR also controls the transcription of vipA and

vipB (encoding the structural units of the Vi antigen), forming an operon within the viaB region. Another chromosomal region, the ompB operon, contains the ompR - envZ genes and plays an important role in the expression of the Vi antigen as a transcriptional regulator of viaB (Pickard et al., Infect Immun., 62:39843993 (1994)). The ompR - envZ region forms part of the adaptive mechanisms of Escherichia coli to high osmolarity conditions. In Salmonella typhi, the Vi antigen is osmotically regulated and ompR is essential for its expression (Pickard et al., supra).

Relationship between Vi antigen expression and immunoprotection

The capsular polysaccharide Vi of Salmonella typhi is a virulence factor and protective antigen in humans* (Felix et al., (1934) supra). A purified form of the Vi polysaccharide is a licensed parenteral typhoid vaccine which provides an acceptable level of protective immunization after a single dose of inoculation, with protection being provided by serum IgG antibodies (Acharya et al., New England Journal of Medicine, 317:1101-1104 (1987) and Klungman et al., Lancet 2:1165-1169 (1987)). In contrast, the attenuated strain of Salmonella typhi Ty21a, as a licensed oral vaccine, does not express the Vi antigen and therefore does not induce serum antibodies against the Vi antigen, yet a vaccine containing Ty21a is able to provide a sufficient level of protection (Wahdan et al., J. Infect. Dis., 145:292-296 (1982) and Levine et al., Lancet, 1:1049-1052 (1987a)). It is generally believed that cell-mediated immune mechanisms provide sufficient protection in this case. Several new attenuated strains of Salmonella typhi expressing the Vi antigen in vitro have failed to elicit serum antibody production against the Vi antigen when administered as oral vaccines in vivo, yet they are capable of eliciting high titers of O and H antibodies (Tacket et al., J. Infect Dis., 163:901-904 (1991), Tacket et al., Vaccine., 10:443-446 (1992a), Tacket et al., Infect. Immun., 65:452-456 (1997)). The most likely explanation for this phenomenon is that the expression of the Vi antigen is strongly regulated by osmotic stimuli (Pickard et al., supra). It is likely that antigen expression is disrupted when the bacterium is not extracellular in the blood or other body fluid (e.g., bile).

In the proposed invention, it is postulated that if the expression of the Vi antigen is sufficient, it leads to sufficient stimulation of serum IgG antibodies against the Vi antigen and thus to an overall improvement in protection against typhoid fever. In the proposed invention, it is also postulated that stable expression will improve the immune response to the foreign antigen expressed by the live attenuated strain.

Salmonella contained in the vaccine and DNA vaccine.

• · • · »··· «4

Target population for typhoid fever

Typhoid fever is a very rare disease in modern industrialized countries where the population is supplied with treated, bacteriologically tested water and where sewage systems dispose of human fecal waste. In contrast, in less developed countries where the population lacks these amenities, typhoid fever is often endemic and from a public health perspective usually represents the most important enteric disease of school-age children (Levine et al., Pediatr. Infect. Dis. J. 8:374-381 (1989a)). Systematic clinical, epidemiological and bacteriological surveillance of typhoid fever in relation to trials of proposed vaccines has been established with high accuracy to estimate the incidence of typhoid fever in many populations ( Levine et al., Lancet, 336:891-894 (1990), Black et al., Vaccine, 8: 81-84 (1990), Levine et al., (1987a) supra, Ferreccio et al, J. Infect. Dis., 159:766-769 (1989) and Simjunek et al., Lancet, 338:1055-1059 (1991)). The incidence rate detected was far higher than would have been expected from the results of unsystematic surveillance.

Systematic surveillance also shows a surprising frequency of clinically mild but bacteremic typhoid infections among children and toddlers in endemic areas (Ferreccio et al, J. Pediatr., 104:899-901 (1984)).

In addition to school-age children in less developed countries, other groups at increased risk for typhoid fever are tourists and clinical microbiologists. Among tourists from the United States, the risk is highest in countries along the Pacific coast of South America and on the Indian subcontinent (Ryan et al., Rev. Infect. Dis., 11:1-8 (1989) and Matieu et al., Arch. Intern. Med., 154:1713-1718 (1994). In addition, clinical microbiologists, including those from developed countries, are exposed to increased levels of Salmonella typhi in the workplace and thus represent a high-risk group (Blaser et al, J. Clin. Microbiol., 13:855-858 (1981)).

Multidrug-resistant strains of Salmonella typhi

Since about 1990, sporadic cases and localized epidemics of typhoid fever have occasionally occurred in the Middle East, northeast Africa, and south and southeast Asia. These epidemics have been caused by strains of Salmonella typhi carrying plasmid-encoded genes conferring resistance to trimethoprim-sulfamethoxazole, as well as resistance to chloramphenicol (Gupta et al., Pediatr Infect. Dis., 13:124-140 (1994) and Rowe et al., Lancet, 336:1065-1066 (1990)). Therapy against these strains requires the use of quinolone antibiotics, such as orally administered ciprofloxacin, or third-generation cephalosporins, such as parenterally administered ceftriaxone. However, these antibiotics are too expensive for developing countries. In contrast to previously discovered antibiotic-resistant strains that caused previously locally limited sporadic epidemics, such as in Mexico from 1972 to 1973 (Olarte et al., • ·

Antimicrob. Agents Chemother., 4:597-601 (1973)) and in Peru from 1979 to 1981 (Goldstein et al., J. Infect. Dis., 2:261-266 (1986)) and which have essentially disappeared and the resistant strains have been replaced by susceptible strains, multidrug-resistant strains of Salmonella typhi that emerged in the Middle East and the Indian subcontinent around 1990 are still the dominant strains in this region. In addition, these strains have spread and become dominant in northeast Africa (Mikhail et al., Trans. R. Soc. Trop. Med. Hyg., 83:120 (1989)) and southeast Asia (Vinh et al., Antimicrob. Agents Chemother., 40:958-961 (1996)).

Salmonella typhi resistant to previously available antibiotics seems to have become a part of life. It is therefore important to note that typhoid fever is no longer a disease that can be easily and cheaply treated with oral antibiotics. Moreover, health authorities may continue to base their typhoid fever control programs on earlier treatment regimens. Complications of typhoid fever leading to death are increasingly common, mainly due to inadequate, delayed and inappropriate antibiotic therapy (Singh, Indian Pediatr., 28:329-332 (1991) Bhutta, Ann. Trop. Paediatr., 16:299-306 (1996) and Gupta, supra).

This spread of resistant strains of Salmonella typhi poses a significant health risk in many less developed countries, and therefore the development of an improved orally administered vaccine is once again in the foreground. This immunoprevention can be used in children in areas with a high incidence of this disease, specifically in areas where Salmonella typhi strains are highly resistant to antibiotics, as well as for tourists and clinical microbiologists.

Previously used typhoid vaccines

A. Inactivated whole-cell parenteral vaccine

A heat-inactivated, phenol-fixed whole-cell vaccine was developed at the end of

19th century (Wright et al., Br. Med. J., 1:256-258 (1897), Pfeifer et al., Dtsch. Med. Wochescr. 22:735-737 (1896)) and was shown in a randomized trial sponsored by the World Health Organization (WHO) in the 1960s to provide adequate protection (51 to 67% vaccine efficacy) lasting up to 7 years (Ashcroft et al., Am. J. Hyg., 79:196-206 (1964), Ashcroft et al., Lancet, 2:1056-1060 (1967), Yugoslav Typhoid Commission, Bull. WHO, 30:623-630 (1964) and Hejfec, Bull. WHO, 34:321-339 (1966). Unfortunately, this vaccine is completely unsuitable because it causes a completely unsuitable reaction at an unacceptably high rate and therefore can never become a widely used vaccine (Levine, Typhoid Feaver Vaccines, In: Vaccines, 2nd Ed., Ed. Plotkin et al., Philadelphia, PA, W. B. Saunders (1994)). In control tests, approximately 25% of the probands who received the thermally • · φ · · « · · · 9 Φ · · Φ • · · Φ ΦΦΦΦ • ΦΦΦ · ·····

5··· ·· · · · · · ··· ·Φ ··· ΦΦ·· ·· ·Φ inactivated, phenol-fixed whole-cell vaccine developed disease with systemic symptoms and about 15% were unable to work or attend school after vaccination (Ashcroft et al., (1964) see above, Yugoslav Typhoid Commission (1964) see above and Hejfac, see above).

For these reasons, this vaccine has been replaced by two newer vaccines, an orally administered vaccine containing the live Ty21a strain (Vivotif®) and a parenterally administered vaccine containing purified capsular polysaccharide Vi (TyphiViM®) (Levine et al., see above). These two new vaccines provide protection comparable to the whole-cell vaccine described above, but are much better tolerated.

B. Parenterally administered vaccine containing Vi polysaccharide

As mentioned above, in the mid-1930s, the Vi antigen, a polysaccharide capsule of Salmonella typhi, was discovered (Felix et al., (1934, see above). This antigen was also found to be present on most bacterial strains isolated from the blood of patients suffering from typhoid fever and is probably the virulence factor of Salmonella typhi in mice. Its presence protects the O antigen from agglutination by O antisera (Felix et al., J. Hyg. Cambridge, 49:92-109 (1951) and Felix et al., J. Hyg. 35:421-427 (1935)). Antibodies to the Vi antigen were found to play a significant role in protection against typhoid fever and it was recommended that phenol-inactivated whole-cell parenteral typhoid vaccines be replaced by alcohol-inactivated whole-cell parenteral vaccines. This is because the Vi structure is better preserved in vaccines treated in this way. antigen and thereby give rise to a higher titer of antibodies against the Vi antigen in animals and humans (Felix, BMJ, 1:391-395 (1941)).

However, in randomized clinical trials in Yugoslavia, heat-inactivated and phenol-fixed whole-cell vaccines were significantly more effective than alcohol-inactivated vaccines (Yugoslav Typhoid Commission, Bull. WHO, 26:357-369 (1962)). For the same reason, acetone-inactivated whole-cell parenteral vaccines were further tested (Landy et al., Am. J. Public Health, 44:1572-1579 (1954)). In two independent clinical trials conducted in the 1960s in Guyana and Yugoslavia, acetone-inactivated whole-cell parenteral vaccines showed the best protection rate of all vaccines previously tested (Yugoslav Typhoid Commission, (1962) see above). It has also been confirmed that these exceptionally good characteristics of the acetone-inactivated whole-cell parenteral vaccine are due to the better preservation of the Vi antigen (Wong et al., J. Infect. Dis., 125:360-366 (1972)), or the higher titer of antibodies against the H antigen which this vaccine provides.

I • · · · • · · • · · · • · · ···· ··

A key problem has been the purification of Vi antigens using non-denaturing techniques (Levine et al., Infect. Immun., 12:1290-1294 (1975) and Robbins et al., J. Infect. Dis., 150:436-449 (1984)).

The immunogenicity of two preparations containing undenatured Vi antigen, prepared at the National Institutes of Health (Bethesda, Maryland) and at the Institut Merieux, (Lyon, France) were tested (Tacket, Vaccine, 6:307-308 (1988)). Both preparations induced high titers of antibodies to the Vi antigens in about 90% of recipients. Furthermore, it has been shown that antibodies to the Vi antigens induced by the vaccine persist for at least 3 years (Tacket et al (1988) supra). Antibody levels similar to those induced by parenteral vaccines containing purified Vi antigen are only observed in nature in chronic carriers of typhus. In contrast, most patients recovering from typhoid fever do not have such high titers of peo- bodies in their blood (Lanata et al., Lancet, 2:441-443 (1983), Losonsky et al., J. Clin. Microbiol., 25:2266-2269 (1987)).

Two randomized, controlled clinical trials were conducted to determine the efficacy and safety of vaccines containing purified Vi antigen at the Institut Merieux. Both vaccines were well tolerated (Acharya et al., supra; Klugman et al., supra). In Nepal, a single intramuscular dose containing 25 pg of Vi antigen provided 72% protection for at least 17 months against culture-verified strains of typhoid fever in adults and children (Klugman et al., supra).

Similar results were achieved in clinical trials in South Africa in school children, where the same dose of Vi antigens provided 64% protection for at least 21 months against culture-verified strains of typhoid fever (Klugman et al., see above).

In addition to better efficacy and safety in children and adults, the advantage of the vaccine containing purified Vi antigen is that it guarantees adequate protection even after a single dose. Its disadvantage compared to the attenuated vaccine containing Ty21a is the necessity of parenteral or intramuscular administration.

C. Previously used attenuated strains serving as a live oral vaccine

1. Streptomycin-dependent strains of Salmonella typhi

The first promising strains for use as live orally administered vaccines against typhoid fever were the Streptomycin-dependent (SmD) strains of Salmonella typhi developed in the late 1960s and early 1970s (Du Pont et al., Antimicrob. Agents Chemother., 10:236-239 (1970) and Levine et al., J. Infect. Dis., 133:424-429 (1976)). When administered in several separate doses containing about 10 ¹⁰ colony-forming units (cfu) per dose, these SnD strains were well tolerated and showed about 80% protection against experimental infection in volunteers (DuPont et al., supra). However, when volunteers were immunized with a lyophilized preparation that had been previously dissolved in water for administration in liquid form,

no protection was achieved (Levine et al., (1976) see above). As a result, further tests with SmD strains were stopped.

2, Strain Ty21a, a patented orally administered live cell vaccine

Ty21a, an attenuated strain of Salmonella typhi that is safe and can serve as an orally administered live cell vaccine, was prepared in the early 1970s by chemical mutagenesis of the pathogenic strain of Salmonella typhi Ty2 (Germanier et al., J. Infect. Dis., 141:533-558 (1975)). A characteristic mutation in this bacterial strain that is thought to be responsible for the attenuation is the inactivation of galE (which encodes UDP-galactose 4-epimerase, an enzyme involved in lipopolysaccharide synthesis), and the inability to form the Vi polysaccharide. While Ty21a is very well tolerated, as has been demonstrated in placebo-controlled clinical trials (Gilman et al., J. Infect. Dis., 136:717-723 (1997), Wahdan et al., (1982) supra, Wadhan et al., Bull. WHO, 58:469-474 (1980)), it is not yet clear which of the mutations is responsible for the stable, highly attenuated phenotype of this vaccine. Results from three placebo-controlled studies that used active survival techniques to determine the reactogenicity of Ty21a in adults and children demonstrated that adverse reactions were not observed more frequently in vaccine recipients compared to placebo recipients, for any symptom or sign. Similarly, in large-scale trials involving approximately 530,000 school-age volunteers in Chile, 32,000 school-age volunteers in Egypt, and approximately 20,000 children and adults in Indonesia, no signs of adverse effects from this vaccine could be identified (Ferreccio et al., (1984), supra; Levine et al., (1987a) supra; Simunjuntak et al., supra; Wahdan et al., (1982), supra; and Wadhan et al., (1980) supra).

Control trials with Ty21a vaccine show that the formulation of the vaccine, the number of doses administered, and the timing of administration significantly influence the degree of protection that can be achieved (Black et al., supra; Ferreccio et al., (1984) supra; Levine et al., (1987a) supra; Wadhan et al., (1980) supra). In the first randomized, placebo-controlled clinical trials with Ty21a in Alexandria, Egypt, 6- to 7-year-old children received three doses of lyophilized vaccine that was resuspended in a diluent (each time just before administration) (Wadhan et al., (1980) supra). To neutralize gastric acids, the children swallowed a 1.0 g NaHCCL tablet a few minutes before oral administration of the vaccine or placebo. During three years of survival, Ty21a provided 96% protection against confirmed typhoid infection (Wadhan et al., (1982) supra).

Newer formulations, which have been commercially available since the mid-1980s, contain the lyophilized vaccine in enteric-coated, acid-resistant capsules. In randomized, placebo-controlled clinical trials in Santiago de Chile, these enterically administered vaccines, in three doses one week apart, provided 67% efficacy for the first three years after administration (Levine et al., (1987a) supra), and 63% protection for the following seven years. Four doses of Ty21a in enteric-coated capsules given eight days apart showed a much better protection rate than three doses (Ferreccio et al., (1984), supra). When the Ty21a strain was patented by the US Food and Drug Administration in late 1989, a 4-dose regimen given on alternate days was recommended, while other countries still used a 3-dose immunization protocol. The disadvantages of this vaccine include the lack of knowledge of the molecular basis of its attenuation, its poor immunogenicity, and the most significant disadvantage is the need for at least three separate doses to ensure adequate protection. Another disadvantage of the Ty21a vaccine is that it does not induce serum IgG antibodies against the Vi antigen.

3. Correlation of protection after immunization with an attenuated vaccine

Although the patented oral vaccines containing live cells of the Ty21a strain are only weakly immunogenic and require three to four separate doses, their efficacy is surprisingly long-lasting, lasting for 5 to 7 years. The results of two immunological techniques confirmed the correlation of the protection achieved with different formulations and different immunization protocols with Ty21a in clinical trials. These include seroconversion of serum IgG antibodies to antigen 0 (Levine et al., Rev. Infect, Dis., 1 l(suppl 3):S552-S567 (1989b)) and determination of the number of intestinal IgA anti-antigen 0 secreting cells (ASCs) detected among peripheral blood mononuclear cells (PBMCs) (Kantele, Vaccine, 8:321-326 (1990), Tacket et al., Infect Immun., 60:536-541 (1992b, Van de Verg et al., Infect Immun., 58:2002-2004 (1990)). Identification of these results as immunological correlates of protection provides an invaluable tool for use in preclinical trials when evaluating the properties of novel attenuated strains of Salmonella typhi for use as oral vaccines.

V. New generations of attenuated strains of Salmonella typhi usable as oral vaccines

Scientists from laboratories around the world are looking for new suitable strains for use in vaccines that will be well tolerated, like Ty21a, but far more immunogenic so that a single oral dose of this new vaccine is able to provide sufficient protective immunity. Finally, some suitable strains were prepared by inactivating genes encoding various biochemical pathways, regulatory systems, haet shock proteins, regulatory genes and probable virulence factors (Curtis et al, Infect. Immun., 55:3035-3043 (1987), supra, Edwards et al., J. Bacteriol., 170:39991-3995 (1984), Hohman et al., J. Infect. Dis., 173:1408-1414 (1996a), Hone et al., Infect Immun., 56:1326-1333 (1988), Hone et al., Vaccine, 9:810-816 (1991), Levine et al, J.

Biotechnol., 44:193-196 (1996)). One attempt has been to increase the immunogenicity of the Ty21a strain by restoring its ability to express the Vi antigen on its surface (Cryz et al., Infect Immun., 57:3863-3868 (1989) and Tacket et al (1991), supra). The relative attenuation potential has usually been determined by administering Salmonella typhimurium strains carrying these mutations to mice and comparing the results with the isogenic wild-type strain. The mutations introduced into various recombinant variants of Salmonella typhi strains that were eventually administered to humans in clinical trials as potential new vaccines are summarized in Table 1.

Mutated gene Vaccine strain Parental wild strain Clinical phenotype Immunological phenotype galE, via EX462 You2 unweakened immunogenic aroA, purA 541 You CDC1080 too weak low immunogenicity aroA, purA, Vi 543You CDC1080 too weak low immunogenicity aroC, aroD CVD 908 You2 weakened immunogenic aroC, aroD, htrA CVD 908-htrA You2 weakened immunogenic cya,crp X3927 You2 insufficiently weakened immunogenic cya, crp, cdt X4073 You2 weakened immunogenic phoP / phoQ Ty800 You2 weakened immunogenic

Important conclusions and results of clinical trials with the most important attenuated strains of Salmonella typhi that play an important role in the development of other live vaccines are summarized below.

A. Vi ⁺ Ty21a strain

A variant of the Ty21a strain was prepared by introducing viaB (encoding the structural genes necessary for the synthesis of Vi polysaccharides) from the wild-type Ty2 strain into the Ty21a genome, thereby achieving expression of the capsular polysaccharide Vi (Cryz et al., supra). The resulting Vi ⁺ Ty21a strain was administered to adults in North America in a single dose in a liquid suspension containing 5.0 ^x105 , 5.0 ^x107 , 5.0 ^x109 cfu in buffer, while another test group received 3 doses of 5.0 ^x109 cfu in buffer (each every other day) (Tacket et al., (1991) supra). The Vi ⁺ Ty21a strain was very well tolerated, and most subjects who received three doses showed a marked increase in serum IgG and mucosal IgA antibodies against the O antigen of Salmonella typhi. However, in none of the tested individuals was there a detectable increase in serum anti-Vi IgG antibodies, nor an increase in the number of cells secreting mucosal IgA anti-Vi antibodies (Tacket et al., (1991) see above).

·· ·· • · · · • · ·

• · · • · · • · ·

B. EX462

An attempt to construct a new variant of the Ty2 strain using recombinant DNA techniques, with the aim of creating a new attenuated variant to achieve a similar effect to that achieved by chemical mutagenesis in the creation of the Ty21a strain (Hone et al., (1988) see above).

Specifically, a deletion mutation in galE was used to prevent the formation of Vi polysaccharides, but other mutations present in the Ty21a strain resulting from non-specific chemical mutagenesis were not used. The resulting strain, EX462, was completely inadequately attenuated and caused typhoid-like symptoms in several recipients. These observations suggest that the combination of the galE mutation and the inability to produce Vi antigen alone is not the main reason for the attenuation of the Ty21a strain.

C. Tribes 541 Ty and 543 Ty

The concept of preparing auxotrophic mutations of Salmonella with inactivated genes encoding enzymes involved in the biosynthetic pathways of aromatic amino acids has been widely popularized (Edwards et al., supra, and Hoiseth et al., Nature, 292:238-239 (1981)). These mutations make Salmonella nutritionally dependent on substrates (para-aminobenzoic acid and 2,3-dihydrooxybenzoate) that are not available in sufficient quantities in mammalian tissue, resulting in a live vaccine but completely suppressed in its ability to replicate.

The prototypes of aroA', strains 541Ty and 543Ty (Vi' variant 541Ty) were prepared from CDC1080, a wild strain obtained from the Centers for Disease Control (Edwards et al., supra). The pathogenicity of the CDC1080 strain has never been directly tested in volunteers. Most other volunteers have been based in attempts to construct a new vaccine on the wild strain Ty2, the original strain from which the Ty21a vaccine is derived (Germanier et al., supra). The pathogenicity of the Ty2 strain has been determined in volunteers (Hornick et al., N. Engl. J. Med., 283:686-691 and 739-746 (1970)).

Strains 541 Ty and 543Ty also carry a deletion mutation in purA that results in a specific dependence on adenine (or adenine-containing substances such as adenosine) supply (Edwards et al., supra). A third mutation is a vhisG mutation that results in a dependence on histidine supply, which does not affect virulence but provides an additional biochemical feature by which the vaccine strain can be distinguished from wild strains of Salmonella typhi.

Strains 541Ty and 543Ty are generally well tolerated at doses up to 5.0 x 10 ¹⁰ cfu in phase 1, but are less immunogenic than Ty21a in stimulating an antibody response (Levine et al., J. Clin. Invest., 79:888-902 (1987b)). For example, only 11% of individuals had detectable serum IgG class antibodies to the 0 antigen (Levine et al., (1987b) supra).

• ·· ♦ ·

D. Strain CVD908

One vaccine strain was tested as highly immunogenic after a single oral dose in phase 1 human clinical trials using freshly harvested cells of strain CVD908 (Tacket et al., (1992b), supra and Tacket et al., (1992a) supra) which carry a precise deletion mutation in aroC and aroD (Hone et al., (1991) supra). This CVD908 vaccine was the first genetically engineered Salmonella typhi vaccine that was sufficiently immunogenic, well tolerated and suggested that a live, orally administered vaccine could be prepared in a single dose. At a very well tolerated dose of 5.0 x 10 ⁷ cfu, 92% of individuals vaccinated with this vaccine demonstrated IgG 0 antibody seroconversion and also showed activation of the immune system in the intestinal mucosa (IgA ASCs) (Tacket et al., (1992a) see above).

1. Cellular immune response after immunization with the attenuated vaccine CVD908

Clinical trials with CVD908 were characterized by intensive testing of cell-mediated immune responses (CMI) (Stein et al., J. Immunol., 155:3987-3993 (1995) and Sztein et al., J. Infect Dis., 170:1508-1517 (1994)). When administered to healthy adults, the CVD908 strain induces a cellular immune response against Salmonella typhi antigens, including cytokine production and a proliferative response to heat- and phenol-inactivated Salmonella typhi cells and purified flagella (Sztein et al., (1994) supra).

Since Salmonella typhi is an intracellular pathogen, it is believed that cytotoxic lymphocyte responses may play an important role in suppressing the progression of typhoid infection by killing Salmonella typhi-infected cells. Techniques for determining cytotoxic activity were used to determine whether immunization of adult volunteers with attenuated Salmonella typhi CVD908 elicited a cytotoxic response in blood cells capable of killing Epstein-Barr virus (EBV)-transformed autologous B-lymphocytes infected with wild-type Salmonella typhi (Sztein et al., (1995) supra). Cytotoxic activity was determined using peripheral blood mononuclear cells isolated before and 14 and 29 days after the first immunization. Peripheral blood mononuclear cells were either used immediately in the CTL technique or were expanded in vitro for 6 to 8 days in the presence of Salmonella typhi-infected Epstein-Barr virus (EBV)-transformed autologous cells before measuring the CTL response. Using this technique, CTL effector cells lyse Epstein-Barr virus (EBV)-transformed autologous cells infected with Salmonella typhi. Specific CTL activity was observed in peripheral blood mononuclear cells 14 days after immunization and the following 7 to 8 days of expansion in vitro in the presence of Salmonella typhi-infected Epstein-Barr virus (EBV)-transformed autologous cells. ···· ·« ·· 99 ·· • · · · ·

9 9 9 9

4 4 · * >

• · · 9 · ··· 9· ··· transformed autologous cells. Peripheral blood mononuclear cells isolated 29 days after immunization show comparable or higher CTL activity than that observed in cells isolated 14 days after immunization. The effector cells are a standard population of CD8 ⁺ , MHC-I, cytotoxic lymphocytes (Sztein et al., (1995) supra). The observation that immunization with attenuated strains of Salmonella typhi induces an increase in circulating CD8 ⁺ cytotoxic effector cells capable of killing autologous cells infected with Salmonella typhi supports the idea that the cytotoxic response plays a key role in suppressing the progression of typhoid fever infection by killing bacteria-infected cells.

2. Disadvantages of using the attenuated vaccine strain CVD908

A possible disadvantage observed in the first phase of clinical trials is that about 50% of individuals who were orally administered the vaccine at a dose of 5.0 10 ⁷ cfu and 100% of individuals who were orally administered at a dose of 0.010 7 cfu were infected.

administered 5.0 10 cfu of vaccine showed strong signs of vaccinemia, with vaccine organisms being isolated from blood samples taken one or more times between days 4 and 8 after administration. Blood samples were taken systematically from these individuals one hour after vaccine administration and then on days 2, 4, 5, 7, 8, 10, 14, 20, 27 and 60. None of the blood samples were positive before day 4 and after day 8. Vaccinemia appeared to be without clinical sequelae (i.e., not associated with fever) and was only short-lived and resolved spontaneously without the use of antibiotics. However, the possible consequences of vaccination with this attenuated strain of Salmonella typhi in a large population where immunocompromised individuals are likely to occur have not been evaluated.

Another disadvantage of vaccination with the CVD908 strain is the development of fever in a small percentage of individuals after administration of a high dose of vaccine (Tacket et al., (1992a), supra).

As an alternative, additional mutations were introduced into the genome of strain CVD908, which were intended to maintain its good tolerability and high immunogenicity, but to suppress the development of vaccinemia or fever.

E. CVD908 - htrA

It was found that inactivation of the htrA gene encoding a stress protein, which is also a serine protease, attenuated a wild-type strain of Salmonella typhimurium in a mouse model (Chatfield et al., Microbial Pathogenesis, 12:145-151 (1992a)). Furthermore, mice immunized orally with Salmonella typhimurium carrying a mutation in htrA were protected against subsequent infection with a lethal dose of Salmonella typhimurium (Chatfield et al., (1992a) see above). A deletion mutation was therefore introduced into the htrA strain of CVD908, creating a new strain, CVD908-htrA (Levine et al., (1996) see above). A single dose of the CVD908-htrA strain was administered orally to three • 4 • ·4 • 44 · • 4

4 ·

4

4 4 4 4 4 4

4 4 4

4 4 4

4 4 4

4 4 4

44 groups of individuals received 5.0 x 10 ⁷ (N=7), 5.0 x 10 ⁸ (N=7), 5.0 x 10 ⁹ (N=7) cfu (Tacket et al., (1997) see above). The CVD908 - htrA strain was very well tolerated, as was the parental CVD908 strain. Only one of the 22 individuals developed a mildly elevated temperature, which was detected by routine tests and was not accompanied by any other disorders or complications. Two of the 22 tested individuals developed diarrhea (Tacket et al., (1997) see above).

The immune response was robust, with 20 of 22 subjects (91%) having significantly elevated serum IgG antibodies to the antigen, and intestinal IgA antibody-producing cells against antigen 0 being detectable in 100% of the subjects tested. These immunological findings are essentially consistent with the results of phase 1 clinical trials in subjects immunized with comparable doses of the CVD908 strain (Tacket et al., (1992a) supra). The main difference was the absence of vaccinemia. While in the case of strain CVD908, vaccinemia was detectable in 10 of 18 individuals who received 5.0 x 10 ⁷ or 5.0 x 10 ⁸ cfu, none of the 22 individuals who received 5.0 x 10 ⁷ ' ⁹ cfu of the well-tolerated, highly immunogenic vaccine containing CVD908-htrA (p < 0.001) (levine et al., (1987b) supra, Tacket et al., (1992a), supra, and Tacket et al., (1997) supra). Strain CVD908 - htrA also elicits a strong cell-mediated immune response that is comparable to that to strain CVD908 (Tacket et al., (1992a), supra, and Tacket et al., (1997) supra). Thanks to these emerging data, the CVD908 - htrA strain was admitted to phase 2 clinical trials, so that clinical applicability and immunogenicity could be verified in a larger group of individuals, including children.

F. Strains with mutations in the cya, crp or cya, crp, cdt genes

In Salmonella, the genes cya (encoding adenylate cyclase) and crp (cyclic AMP receptor protein) represent important regulatory systems that affect many genes and operons (Curtiss et al., Dev. Biol. Stand., 82:23-33 (1994)). Strains of Salmonella typhi carrying deletion mutations in the cya and crp genes are attenuated relative to the parental wild-type strains, and oral immunization of mice with these strains protected the immunized individuals against infection with virulent strains of Salmonella typhimurium (Curtis et al., (1987) supra).

A new vaccine strain, X, was prepared, which is a cya', crp' double mutant of the Salmonella typhi strain Ty2 (Curtis et al., (1994) supra). In phase 1 clinical trials, it was shown that strain X ³⁹²⁷ , although attenuated relative to the wild-type strain, was insufficiently attenuated to serve as a live orally administered vaccine in humans, as some individuals developed typhoid symptoms associated with high fevers after administration (Tacket et al., (1992b), supra). Some individuals also developed vaccinemia (Tacket et al., (1992b), supra).

• 4

4· ·· • 4 4 • 4

44*4 ·· ·· *·

4 4 4 · « 4 4 4

4 4 · · ·

4444

To achieve a higher degree of attenuation, a third deletion mutation was introduced into the genome of the double mutant strain X ³⁹²⁷ , namely in the cdt gene (Curtis et al., (1994) see above). The cdt gene affects the penetration of Salmonella from the lymphoid tissues of the intestine into the deeper organs of the reticuloendothelial system, such as the liver, spleen and bone marrow (Kelly et al., Infect Immun., 60:4881-4890 (1992)). The resulting strain X ⁴⁰⁷³ with a triple mutation in cya', crp, cdt was administered to adults in North America, with a single dose of buffer containing 5.0 χ 10 ⁵ , 5.0 χ 10 ⁶ , 5.0 χ 10 ⁷ , 5.0 χ 10 ⁸ cfu (Tacket et al., (1997) see above). This strain was well tolerated, except for one patient who received 5.0 x 10 ⁷ cfu and developed diarrhea (Tacket et al., (1997) see above). None of the tested individuals developed vaccinemia. All 5 individuals who received 5.0 x 10 ⁸ cfu had significantly increased serum IgG antibodies to the 0 antigen as well as intestinal mucosal cells producing mucosal IgA to the 0 antigen (Tacket et al., (1997) see above).

G. Strains with mutations in phoP/phoQ

Two new strains of Salmonella typhi carrying deletion mutants in phoP/phoQ have been prepared (Hohman et al., (1996a) supra, Hohman et al., Vaccine, 14:19-24 (1996b)). Strain Ty445, which carries a deletion mutation in aroA, was evaluated as too attenuated and only minimally immunogenic (Hohman et al., (1996b) supra). In contrast, strain Ty800, derived from Ty2, carrying deletion mutations only in phoP/phoQ, was generally well tolerated and immunogenic when administered at doses of 10 ⁷ cfu to 10 ⁱ⁰ cfu in a phase I clinical trial to a small number of subjects 11 (Hohman et al., (1996b) supra). At higher doses, diarrhea occurred in 1 in 3 subjects. It is difficult to compare the immune response of individuals immunized with strains CVD908-htrA and X ⁴⁰⁷³ , as some immunological tests were different and even if the same technique was used for testing (e.g., the technique for detecting ASCs producing IgA against the 0 antigen), the differences may be due to different techniques between the two laboratories.

H. Summary

The CVD908 strain is well tolerated but causes vaccinemia in approximately 50% of individuals immunized with a dose of 10 ⁷ cfu. Mild but demonstrable diarrhea was observed in approximately 10% of individuals immunized with the CVD908-htrA, Ty800 or X ⁴⁰⁷³ strains. The immune response to the X ⁴⁰⁷³ strain appears to be weaker than that observed after oral immunization with CVD908, CVD908-htrA and Ty800 (see Table 1). There are therefore four attenuated strains of Salmonella typhi that have completed phase 1 clinical trials, each of which has its own disadvantages and the interest is to prepare new attenuated strains suitable for vaccination. Furthermore, none of these strains induces serum antibodies against the Vi polysaccharides.

·· ·<

« Φ · * • · « φ φ · · • · ·

Φ··Φ ·· ·· Φ·

φ. 9 9 Φ • Φ Φ · • · · · • Φ · Φ φ· ♦ ·

VI. Use of attenuated Salmonella strains as live vectors for the expression of foreign genes encoding protective antigens of other pathogens and presentation of these antigens to the immune system

Live vectors serving as vaccines, also called "carriers" or "live antigen expression systems", represent a wide field of interest in vaccinology (Levine et al, Microecol. Ther., 19:23-32 (1990); Morris et al, Gastroenterol., 103:699-702 (1992); Barletta et al, Res. Microbiol., 141:931-940 (1990), (Dougan et al, Para. Immunol., 9:151-160 (1987); Curtiss et al, Curr. Top. Microbiol. Immunol., 146:35-49 (1989)). In these systems, live viral or bacterial vaccines are modified so that they express protective antigens of a foreign pathogenic microorganism on their surface and present them to the immune system to induce a protective immune response. Live bacterial vectors, which Attenuated strains of Salmonella (Levine et al (1990), supra; Morris et al, supra; Dougan et al, supra; and Curtiss et al (1989), supra), Bacille Calmette Guerin (Barletta et al, supra), Yersinia enterocolitica (Van Damme et al, Gastroenterol., 103:520-531 (1992)), V. cholerae Ol (Viret et al, Mol. Microbiol., 7:239-252 (1993)), and E. coli (Hale, Res. Microbiol., 141:913-919 (1990)) have been tested, among others. Each has some advantages or disadvantages.

Attenuated strains of Salmonella typhi are of interest as live vectors for humans because they can be administered orally, colonize lymphoid tissues associated with the digestive tract as well as the reticuloendothelial system, and elicit a broad spectrum of immune responses (including serum antibodies, mucosal sIgA, various cell-mediated immune responses including classical CTL, and various forms of antibody-mediated cellular cytotoxicity) (Conry et al, Gene Therapy, 2:5965 (1995), Cox et al., J. Virol., 67:5664-5667 (1993); Davis et al, Proc. Natl. Acad. Sci., USA, 93:7213-7218 (1996a); Davis et al, Vaccines, 96:111-116, Brown et al, (Eds.), Cold Spring Harbor Laboratory Press (1996b); Tacket et al (1992a), supra; Gonzalez et al, J. Infect. Dis., 169:927-931 (1994); Sztein et al (1994), supra). Furthermore, Salmonella is easily genetically manipulated and a number of foreign antigens have already been introduced into its genome. Theoretically, orally applicable vaccines against various infectious diseases could be prepared by stably introducing and expressing foreign genes encoding protective antigens in well-tolerated and highly immunogenic strains of Salmonella typhi (Levine et al., (1990) supra).

VII. Salmonella with mutations in the purine metabolic pathway

A. Older reports of Salmonella strains with mutations in the purine metabolic pathway

Mutations in the purA and purB genes of Salmonella (thereby disrupting the synthesis of adenine-containing nucleotides) (see Fig. 2) cause such severe attenuation (McFarland et al., Microb. Pathogen., 3:129-141 (1987)) that these strains become ineffective in eliciting a protective immune response in animals (O'Callagham et al., Infect Immun., 56:419-423 (1988)) and are only weakly immunogenic in humans (Levine et al., (1987b) supra). Conversely, mutations in any of the genes encoding proteins of the standard purine metabolic pathway (i.e. purF, purG, purC, purHD) that disrupt the biosynthesis of both purine nucleotides (see Fig. 2) or in the guaB or guaA genes that disrupt the biosynthesis of guanine nucleotides (see Fig. 2) have been described as being significantly less debilitating (McFarland et al., supra), these strains inducing death in mice at doses as low as 10 ² cfu (McFarland et al., supra).

The above observations suggest that:

i) mutations affecting enzymes of the standard purine metabolic pathway are only slightly attenuating (McFarland et al., see above).

ii) Mutations affecting the synthesis of adenine nucleotides are highly attenuating (i.e. purA, purB) (McFarland et al., supra, O'Callagham et al., supra, Levine et al., (1987b) supra), however mutations affecting the synthesis of guanine-containing nucleotides are only minimally attenuating (McFarland et al., supra).

B. Unanticipated and unique observation according to the proposed invention

In line with the above findings, it was unexpectedly discovered that specific gua auxotrophy in Salmonella typhi provides a high degree of attenuation. This attenuation is due, in part, to reduced invasiveness (a property never previously described in other Salmonella strains attenuated by mutations in metabolic pathways) as well as to reduced intracellular replication of Salmonella typhi mutants designated ÁguaB-A. This finding was unexpected, as Salmonella dublin and Salmonella typhimurium carrying the insertional Tn5 mutation affect guanine nucleotide synthesis and have been described as only minimally attenuated (McFarland et al., supra. Another unexpected feature was that in addition to the high degree of attenuation, the ÁguaB-A mutant of Salmonella typhi strain CVD915 described here was highly immunogenic in a mouse model. This observation is unexpected, given previous results where strongly attenuated strains with mutations inactivating purA and purB were very poorly immunogenic (Edwards et al., supra). Another unexpected finding was that strain CVD915 induces a very strong immune response to recombinant antigens expressed on its surface. This observation is unexpected, given previous results where strongly attenuated strains were only weak live vectors for heterologous recombinant antigens (Tackete et al. (1997, see above).

• ·

In addition, the CVD915 strain is able to deliver a eukaryotic expression vector ^carrying a recombinant foreign antigen (DNA vaccine) and elicit an unexpectedly strong immune response. The possibility of delivering a eukaryotic expression vector in the form of a plasmid has been described for attenuated strains of Shigella vectors (Sizemore et al., Science, 270:299-302 (1995)) and using attenuated Salmonella typhimurium vectors (Darji et al., Cell, 91:765-775 (1997)). Shigella is an invasive organism and its success in delivering DNA vaccines is probably due to its ability to secondarily leave the phagosome (Sansonetti et al., Infect Immun., 51:461-469 (1986)). However, Salmonella does not leave the phagosome. Therefore, the observations presented here are unexpected and unique in the current state of the art.

The present invention describes a unique way in which genetically engineered attenuated Salmonella typhi maintains stable expression of Vi antigens. The Salmonella mutants of the present invention are attenuated primarily by mutations in metabolic pathways related to de novo guanine synthesis.

ESSENCE OF THE INVENTION

The object of the proposed invention is to describe attenuated strains of Salmonella that stably express the Vi antigen on their surface.

Another object of the present invention is to describe a method for achieving stable expression of Vi antigens.

Another object of the proposed invention is to describe Salmonella strains with attenuated mutations in de novo guanine synthetic pathways.

Another object of the proposed invention is to describe a method for introducing deletion mutations into the gueB and guaA genes in Salmonella.

Another object of the proposed invention is to describe orally or intranasally (i.e. mucosally) applicable vaccines against typhoid fever.

Another object of the present invention is to describe novel Salmonella strains that can serve as live vectors for foreign genes cloned from other pathogens such that this expression and the corresponding immunization raise a protective immune response against these pathogens from which the foreign antigens are derived.

Another object of the present invention is to describe Salmonella strains that are useful as live vectors for the delivery of DNA vaccines.

These and other objects of the proposed invention, which are set forth in the detailed description, have one thing in common, they all utilize attenuated Salmonella mutants that stably express the Vi antigen on their surface.

OVERVIEW OF IMAGES IN THE DRAWINGS

Figures 1A and 1B show the viaB operon and schematically illustrate the replacement of the wild-type osmotically regulated vipR promoter in the viaB operon with the sacB-Neo cassette (Fig. 1A) and the replacement of the sacB-Neo cassette by the insertion of the strong stable promoter P _tac to ensure strong stable expression of Vi antigens (Fig. 1B).

Figure 2 shows the de novo purine biosynthetic pathway and the relationship to the aromatic metabolic pathway. In Figure 2, the dashed arrows indicate pathways where individual steps are not shown. The enzymes are represented by their genes. The superscript letters represent selected strains with a mutation in the gene involved in this reaction. The strains represented are: a: purF1741::TnlO S. dublin SL5437 (McFarland et al, supra); b: purG876::TnlO S. dublin SL5436 (McFarland et al, supra); c: purC882::TnlO S. dublin SL5435 (McFarland et al, supra); d: purH887::TnlO S. dublin SL2975 (McFarland et al, supra); e: AguaB-A S. flexneri 2a CVD 1204 and AguaB-A, wire S. flexneri 2a CVD 1205 (Noriega et al, Infect. Immun., 64:3055-3061 (1996)) and ÁguaB-A Salmonella typhi CVD 915 (present invention); f: aroD25::TnlO S. flexneri Y SFL114 (Lindberg et al, (1990) supra), SFL124 (Li et al, supra), S. flexneri 2a 1070 (Kamell et al (1995) supra); g: ΔaroC, ΔaroD S. typhi CVD 908 (Hone et al (1991), supra); h: hisG46 DEL407, aroA554::Tnl0 S. typhimurium SL3261 (Hoiseth et al, supra); i: ÁaroA S. flexneri 2a CVD 1201 and AaroA, wire CVD 1203 (Noriega et al, Infect. Immun., 62:5168-5172 (1995); and j: ÁaroA, AhisG, ApurA S. typhi 541 Ty (Levine et al (1987b), supra).

For Figure 2, the following abbreviations are given:

PRPP: 5-phosphoribosyl-1-3-pyrophosphate; PRA: 5 - phosphoribosylamine; GAR 5'-phosphoribosyl-1-glycinamide; FGAR: 5'-phosphoribosyl-N-formylglycinamide; FGAM: 5'-phosphoribosyl-N-formylglycinamidine; AIR: 5'-phosphoribosyl-5-aminoimidazole; CAIR: 5'-phosphoribosyl-5-aminoimidazole-4-carboxylate; SAICAR: 5'-phosphoribosyl-4-(N-succinocarboxamide)-5-aminoimidazole; AICAR: 5'phosphoribosyl-4-carboxamide-5-aminoimidazole; FAICAR: 5'-phosphoribosyl-4-carboxamide-5 formamidoimidazole; IMP: inosinic acid; Hx: hypoxanthine; G: guanine; and A: adenine.

Figures 3A to 3G show the DNA sequence encoding the guaA and guaB genes of Escherichia coli (guaA and guaB genes of Escherichia coli (SEQ ID NO:1) (GenBank No. M10101-M10102), (Thomas et al, Gene, 36:45-53 (1985); Tiedeman et al, J. Biol. Chem., 260:8676-8679 (1985); Tiedeman et al, Nucleic Acids Res., 13:1303-1316 (1985)), which is highly homologous to the guaA and guaB genes of Salmonella species as well as some restriction endonuclease sites useful in preparing deletion mutants according to the present invention.

• ·

4

Figure 4 schematically illustrates the orientation of the guaB-A operon, as well as the location of restriction sites for endonucleases useful in the preparation of deletion mutants according to the present invention. The guaB-A operon is amplified by PCR and the fusion of segments of the guaB-A operon by PCR results in the replacement of the AguaB-A allele. It is also shown that cloning the arsenite operon (ars) in the middle of the AguaB-A allele results in the formation of the AguaB-A::P _ta c-ars cassette. This deletion cassette is then cloned into a suicide plasmid and replaced by the wild-type guaB-A allele in Salmonella typhi strain Ty2, resulting in the formation of the AguaB-A Salmonella typhi strain designated CVD915.

Figure 5 is a schematic representation of the pcDNA3/tetC construct (a DNA vaccine encoding the C fragment of tetanus toxin under the control of the human cytomegalovirus (CMV) promoter). Figure 6 shows the antibody response in mice against the C fragment of tetanus toxin after intranasal immunization with the AguaB-A CVD915 strain expressing the C fragment of tetanus toxin, or carrying a eukaryotic expression cassette encoding the C fragment of tetanus toxin.

Figure 7 shows the proliferative response of mouse lymphocytes against tetanus toxin fragment C after intranasal immunization with the AguaB-A CVD915 strain as such, or as a tetanus toxin fragment C expression vector or carrying a eukaryotic expression cassette encoding tetanus toxin fragment C.

Figure 8 shows the proliferative response of mouse lymphocytes against Salmonella typhi flagella after intranasal immunization with the AguaB-A CVD915 strain as such, or as an expression vector of the tetanus toxin C fragment or carrying a eukaryotic expression cassette encoding the tetanus toxin C fragment.

Figure 8 shows the antibody response of mouse lymphocytes against Salmonella typhi flagella after intranasal immunization with the AguaB-A strain CVD915 or the AaroC, AaroD, AhtrA strains of Salmonella typhi CVD908-htrA.

Figure 8 shows the antibody response against Salmonella typhi LPS in mice after intranasal immunization with the AguaB-A strain CVD915 or the AaroC, AaroD, AhtrA strains of Salmonella typhi CVD908-htrA.

The present invention describes attenuated strains of Salmonella for use, among other things, as orally administrable typhoid vaccines, as live vectors and DNA vaccines expressing foreign antigens. As used herein, the term "foreign antigen" describes any foreign Salmonella antigen. The present invention also describes an attenuated mutant of Salmonella, wherein said mutant is characterized by stable expression of the Vi antigen.

It is appropriate that the chromosomal genome of Salmonella be modified by substitution in the highly regulated vipR promoter with a strong replacement promoter to ensure stable

expression of Vi antigen in attenuated strains. The specific strong promoter used in the present invention is not limited. Examples of such useful strong promoters include P _tac (de Boer et al, Proc. Natl. Acad. Sci., 80:21-25 (1983)), P _ac (De Boer et al, supra), P _trc (De Sutter et al, Gene, 141: 163-170 (1994)), P _O i _ac and P _pp (De Sutter et al, supra; Guisez et al, Protein Expr. Purif., 2:249-258 (1998))

The specific strain of Salmonella typhi used as the starting strain in the present invention is not limited in any way.

In the examples below, the wild strain Ty2 is used as the vaccine against Salmonella typhi.

Ty2 is a highly virulent strain of Salmonella available from a number of sources, including the Centers for Disease Control and Prevention (CDC), Walter Reed Army Institute of Research, University of the Healthy Sciences collections, the Pasteur Institute in Paris, and Imperial College (England).

Other strains of Salmonella typhi have been used to prepare suitable attenuated vaccines, such as: strain CDC 1080 (Levine et al., (1987b) see above), which was obtained from Stanford University or from the CDC, and strain ISP1820 (Hone et al., (1991) see above), which can be obtained from the Center for Vaccine Development, University of Maryland.

In the proposed invention, the chromosomal genome of Salmonella typhi is usually modified by deletion or other modification of the guaB-A operon, thereby blocking the de novo biosynthesis of guanine nucleotides. Better yet, broadly defined, a non-polar mutation in the guaB-A operon is suitable to inactivate the metabolic pathway of the enzyme IMP dehydrogenase (encoded by the guaB gene) and GMP synthase (encoded by the guaA gene). As a result of these two mutations, Salmonella is unable to synthesize GMP de novo and consequently GDP and GTP nucleotides (see Fig. 2), which significantly limits its multiplication in mammalian tissues.

In vitro, the Salmonella AguaB-A mutants of the present invention are unable to grow in minimal medium unless supplemented with guanine. In tissue cultures, a sharp decrease in invasiveness of Salmonella ÁguaB-A mutants was found. These Salmonella ÁguaB-A mutants are able to obtain guanine to some extent from mammalian host tissues. However, its assimilation into Salmonella requires its incorporation as a nucleotide precursor into the guanine metabolic pathway before dephosphorylation to nucleoside by periplasmic nucleotidases. Since this nucleotide is available in the mammalian tissue environment, the weakening of the strain due to blocking de novo guanine synthesis is due to the inefficiency of alternative guanine pathways or for reasons still unknown.

The guaA gene, encoding GMP synthase, is 1575 bp in size (see Fig. 3A to 3G). The size of the intracistronic inactivating deletion in guaA mutants can range from 1 to 1575 bp. More preferably, from 100 to 1575, if the deletion is in the appropriate reading frame. This deletion can also be made by • 4

44· · 4 4 extends beyond the guaA gene, i.e. an extracistronic deletion downstream of guaA can also affect the transcription of this gene and thus inactivate it. However, this is not the most suitable option.

The guaB gene, encoding IMP dehydrogenase, is 1533 bp in size (see Fig. 3A to 3G). The size of the intracistronic inactivating deletion in guaB mutants can range from 1 to 1533 bp. More preferably, from 100 to 1533 bp, if the deletion is in the appropriate reading frame. This deletion can also be made so that it extends beyond the guaB gene, i.e. an extracistronic deletion downstream of guaB can also affect the transcription of both genes (guaA and guaB) and thus inactivate them. However, this is not the most suitable option.

Deletions can be made in the guaA gene using 2 suitable restriction sites within the guaA gene, for example sites for Seal or AccI and EcoRV or SalI, Smali, SphI or XmaI (Figs. 3A to 3G and 4) or by site-directed mutagenesis using oligonucleotides (Sandbrook et al., Molecular Cloning, A Laboratory Manual, Eds., Cold Spring Harbor Publications (1989)).

Deletions can be made in the guaB gene using 2 suitable restriction sites within the guaB gene, for example the Bell, BsmI, PvuII, Scall, SspI or XhoII sites (Figs. 3A to 3G and 4) or by site-directed mutagenesis using oligonucleotides (Sandbrook et al., Molecular Cloning, A Laboratory Manual, Eds., Cold Spring Harbor Publications (1989)).

Additionally, any combination of deletion sites located simultaneously in the guaB and guaA genes can be used to obtain the desired deletion according to the present invention, examples of which include AccII, AflII, Ahall, Alwl, Asul, Bani, BcnI, Bspl286, BspMII and Ddel (Fig. 3A to 3G).

Inactivation of the guaA and/or guaB genes can also be achieved by inserting foreign DNA using the above-mentioned restriction sites, or by site-directed mutagenesis using oligonucleotides (Sandbrook et al., Molecular Cloning, A Laboratory Manual, Eds., Cold Spring Harbor Publications (1989)), so as to disrupt transcription of the guaA and/or guaB genes. The typical size of this insertion, which inactivates the guaA and guaB genes, can be from 1 bp to 100 kbp, although insertions smaller than 100 kbp are most suitable. This insertion can be made anywhere in the region encoding the guaA or guaB genes and or in the promoter.

Other methods of inactivating the guaA and/or guaB genes are the transfer of deletionally or insertionally inactivated guaA and/or guaB genes from another microorganism into the Salmonella genome, transposon-induced deletions, and imprecise removal of DNA insertions. The latter two methods are rather techniques for introducing deletions spanning the guaA and guaB genes and are therefore not the most suitable. The mutations in gua according to the proposed invention are useful for the development of non-pathogenic variants of Salmonella typhi useful as orally administered vaccines.

Suitable Salmonella strains for use as vaccines can be prepared, e.g. by adding additional attenuating mutations, impairing the expression of the aro locus products. This can be achieved

9 by deleting part of at least one aro gene (aroA, aroC, aroD) in the Salmonella chromosome, thereby rendering the strain dependent on PABA, a substrate that is not available in mammalian cells, such as Salmonella typhi strain CVD908 (Hone et al., (1991) supra).

In another variant, the above-described factors of the proposed invention are combined in a typhoid vaccine, which comprises:

(A) a pharmaceutically effective amount of Salmonella typhi mutants, wherein said mutants stably express the Vi antigen on their surface (B) a pharmaceutically acceptable carrier or diluent

In another variant according to the proposed invention, the above-described factors are combined in a live vaccine which comprises:

(A) a pharmaceutically effective amount of Salmonella typhi mutants, wherein said mutants stably express a Vi antigen on their surface and wherein said mutant encodes and expresses a foreign antigen under the control of a prokaryotic promoter (i.e. bacterial expression of a foreign antigen) (B) a pharmaceutically acceptable carrier or diluent

The foreign antigen that is inserted into Salmonella as a live vector is not essential for the subject matter of the present invention. The attenuated strains of Salmonella according to the present invention can be used as live vectors for immunization against enteric pathogens (i.e. bacteria, viruses, etc.), sexually transmitted pathogens, pathogens of acute respiratory tract diseases or pathogens that pass through the mucosa and cause systemic diseases (e.g. meningococcus). It can also be used against various types of parasitic diseases, such as Plasmodium falciparum, leishmania, Entamoeba histolytica and Cryptosporidium. Examples of antigens of enteric pathogens that can be expressed on live strains of Salmonella according to the present invention are:

(a) fimbrial colonization factor of enterotoxigenic Escherichia coli and in particular the B subunit of the heat-labile enterotoxin (which does not cause intestinal secretion but elicits a neutralizing antitoxin) (b) fimbrial antigen and/or intimin of enterohaemorrhagic Escherichia coli together with the B subunit of Shiga toxin 1 or 2 (or mutant Shiga toxin 1 or 2) (c) Shigella 0 antigen and the Shigella VirG invasive plasmid (d) neutralizing antigens of rotaviruses (e) urease, colonization antigens and other protective antigens from Helicobacter pylori (f) inactivated Clostridia difficile toxin or antigens derived from it on ·

Examples of sexually transmitted disease antigens that can be expressed on live Salmonella strains according to the present invention are:

(a) pili and outer membrane proteins of Nesseria gonorrhoae (b) proteins from Chlamydia trachomatis.

Examples of antigens of pathogens causing acute respiratory tract diseases that can be expressed on live Salmonella strains according to the present invention are:

(a) a mutant diphtheria toxin with a replaced NAD binding domain that lacks toxic activity and induces the production of a neutralizing antitoxin (b) Bordetella pertussis antigens including a fusion protein containing a truncated SI subunit of pertussis toxin fused to the C fragment of tetanus toxin, mutant pertussis toxin, filamentous hemagglutinin and pertactin (c) F and G glycoproteins of respiratory syncytial virus (d) capsular polysaccharide of Haemophilus influenzae type b (e) capsular polysaccharide of groups A and C of Neisseria meningitidis and outer membrane proteins of group B of Neisseria meningitidis.

Examples of parasite antigens that can be expressed on live Salmonella strains according to the present invention are:

(a) circumsporozoite protein (CSP), liver stage antigen-1 (LSA-1), SSP-2 (also known as TRAP) and Exp-1 from Plasmodium falciparum, (b) erythrocyte asexual stage antigen from Plasmodium falciparum, including MSP-1, MSP-2, SERA, AMA-1 and SREHP, (c) sexual stage antigen from Plasmodium falciparum, such as Pfs25 and gp63 from Leishmania (d) serine-rich protein from Entamaeba histolytica (SREHP).

In another variant, the above-described factors of the present invention are combined in a DNA vaccine, which comprises:

(A) a pharmaceutically effective amount of Salmonella typhi mutants, wherein said mutants stably express on their surface the Vi antigen and comprise a plasmid that encodes and expresses, using a eukaryotic promoter in eukaryotic cells, a foreign antigen (B) a pharmaceutically acceptable carrier or diluent

A detailed description of the construction and use of DNA vaccines is described in U.S. Patent Application Serial No. 08/433,790, filed May 3, 1995, which is incorporated herein by reference.

The specific foreign antigen to be used in the DNA vaccine of the present invention is not critical to the present invention. Examples include a wide range of antigens from pathogens such as influenza (JUSTewicz et al, J. Virol., 69:7712-7717 (1995); Fynan et al, Int. J.

Immunopharmacol., 17:79-83 (1995)), lymphocytic choriomeningitis virus (Zarozinski et al, J. Immunol., 154:4010-4017 (1995)), human acquired immunodeficiency virus (Shiver et al, Ann. NY. Acad. Sci., 772:198-208 (1995)), hepatitis B virus (Davis et al, Vaccine, 12:15031509 (1994)), hepatitis C virus (Lagging et al, J. Virol., 69:5859-5863 (1995)), rabies virus (Xiang et al, Virology, 209:569-579 (1995)), Schistosoma (Yang et al, Biochem. Biophys. Res. Commun.,212:1029-1039 (1995), Plasmodium (Sedegah et al., Proc. Natl. Acad. Sci., USA, 91:9866-9870 (1994)) Mycoplasma (Barry et al., Nature, 377:632-635 (1995)). The decision whether to express a foreign antigen in Salmonella (using a prokaryotic promoter in a live vaccine) or in Salmonella-infected cells (using a eukaryotic promoter in a DNA vaccine) depends on which system elicits a better immune response of the desired type in animal tests and/or in clinical trials and/or may depend on the glycosylation of the antigens and whether this glycosylation is essential for eliciting a protective immune response and/or may depend on the correct tertiary conformation of the antigens, which can only be achieved in one expression system.

In vaccines according to the present invention, the pharmaceutically effective amount of the mutants according to the present invention to be administered may depend on the age, weight, sex of the subject and the route of administration. In general, the immunogenic dose should be about 10 ² cfu to about 10 ^1θ cfu. More preferably, from 10 ⁶ cfu to 10 ⁹ cfu for oral administration, when the vaccine is administered in the form of capsules, or resuspended in a buffer to protect the attenuated bacteria from the acidic pH of the stomach. For intranasal administration in the form of drops or aerosol, the optimal dose is 10 ² cfu to 10 ⁷ cfu. The specific pharmaceutically acceptable carriers are not essential to the present invention, and are conventional substances and compositions used in the art.

Examples include buffers that protect against the acidic pH of the stomach, such as citrate buffer (pH = 7.0) containing sucrose, bicarbonate buffer (pH = 7.0) alone (Levine et al., (1987b) supra and Black et al., supra) or bicarbonate buffer (pH - 7.0) containing ascorbic acid, lactose and possibly aspartame (Levine et al., Lancet, 11:467-470 (1988)). Examples of carriers include proteins, e.g., collected from milk, sugars, e.g., sucrose, or polyvinylpyrrolidone.

Mutant strains of the present invention can be stored at -70°C when in Luria medium (DIFCO) containing 30% to 50% by weight of glycerol.

The following examples are provided for illustration only and should not be construed as limitations on the invention.

* · « · « « 44 44 · · 4 4

444 4 ♦ 4444

44·· · 4·····

AC ··· · 4 4 4 4 4

4444 4· 444 4444 44 ··

EXAMPLES OF EMBODIMENTS OF THE INVENTION

EXAMPLE 1. Construction of the ÁguaB-A deletion cassette pJWlOl

DNA segments containing the 5' end of the guaB gene and the 3' end of the guaA gene were amplified by PCR using genomic DNA of Salmonella typhi strain Ty2 as a template and then fused in a second PCR reaction to generate the ÁguaB-A allele (Fig. 4). The method used to construct the ÁguaB-A allele is the same as that used to construct the ÁguaB-A allele for ÁguaB-A of Salmonella flexneri strain CVD 1204 2a (Noriega et al., (1996) supra and U.S. Patent Application Serial No. 08/629,600 filed April 9, 1996 (unaccepted), which is incorporated herein by reference).

In the first PCR reaction (Fig. 4) the primers used to amplify the 5' end of the guaB-A allele were:

'-CGAGCTCGCGAGCTCGGTAAAGTACCAGTGACCGGAAGCTGGTTGCGT-3' (sequence no. 2) and

-GGGCCCGGGGGATCCTCAACCGACGCCAGTCACGATACGAGTTGTACAGAT-3' (sequence no. 3) and the primers used to amplify the 3' end of the ÁguaB-A allele were:

-TGAGGATCCCCCGGCCCGGCTACGCGCAGGTTGAAGTCGTAAACGACAGC-3' (sequence no. 4) and '-GCTCTAGAGCTCTAGAGCTCATTCCCACTCAATGGTAGCTGGCGGCTT-3' (sequence no. 5).

The internal primers (sequences 3 and 4) were introduced with in-frame stop codons (TGA and TCA) upstream of two unique restriction sites (ApaI and BamHI) (underlined) that were added to introduce foreign genes into the chromosomal ÁguaB-A allele. This stop signal will prevent translational fusion of ÁguaB with ÁguaA products or ÁguaB products with products of foreign genes inserted into the ÁguaB-A allele. In addition, a sequence is introduced into these internal primers that serves as a homologous region (in bold) to which the 5' and 3' PCR products will fuse in the second PCR reaction (thus creating the ÁguaB-A allele) (Fig. 4).

In the second PCR reaction, the 5' and 3' segments were fused using extreme primers (sequences 2 and 5) and the resulting fusion products were amplified in the same reaction. In this reaction, the appropriate homologous region (ApaI - BamHI) after cooling effectively fuses to the 5' and 3' segments, which at this point can serve as their own primers and/or templates for Taq polymerase, depending on which of the DNA strands they anneal with. To achieve this fusion, the PCR program is adjusted so that the first 15 cycles reach the renaturation temperature (1 °C / 8 seconds from 40 °C to 50 °C + 50 °C for 2 minutes) followed by 15 cycles of the renaturation temperature of 55 °C at which the new ÁguaB-A allele is amplified. This involved 900 bp of the guaB-A operon that were not amplified, resulting in the mutant ÁguaB-A operon (Fig. 4).

• ·

The extreme primers (SEQ ID NOs. 2 and 5) were designed to introduce unique restriction sites (SstI and XbaI) (underlined), which were used to clone the AguaB-A allele into a unique restriction site in the temperature-sensitive, pSCIO1-based (Blomfield et al., Mol. Microbiol., 5:1447-1457 (1991)) suicide plasmid pFM307A, creating the plasmid pFM307:: AguaB-A. Plasmid FM307A (6.3 Kbp) was prepared by replacing the 4.3 Kbp Nael - Nael fragment containing the cam gene (chloramphenicol resistance) in pIB307 (Blomfield et al., Mol. Microbiol., 5:1447-1457 (1991)) with a 3.8 Kbp BamHI - BamHI fragment (blunted with Klenow polymerase) containing the SacB - Kan cassette (contributing to kanamycin resistance and sucrose selection) from pIB279 (Blomfield et al., Mol. Microbiol., 5:1447-1457 (1991)). In addition, it was convenient to insert a non-antibiotic selection marker in the middle of the AguaB-A atk allele in order to:

(i) facilitate the replacement of the AguaB-A allele with the very strong guaB-A allele in the Ty2 strain (ii) introduce a selectable trait that will allow identification of this vaccine in the clinic

Therefore, the 5.0 Kbp R factor R773 operon conferring arsenic resistance (ars) was isolated as a HindIII fragment from pUM1 (Chen et al., J. Bacteriol., 161:758-763 (1985)) (kindly donated by Dr. Rosen of Wayne State University, Detroit, MI) and cloned under the control of the P _tac promoter in pKK223-3 (Pharmacia, Piscatway, NJ). The blunt-ended P _tac -ars Nael-Dral segment was then cloned into the ApaI site (blunted by T4 polymerase) of the AguaB-A allele in pFM307::AguaBA, generating pJW101 (Fig. 4).

B. Suicide deletion of a cassette-borne mutation

Plasmid pJW101 was used to introduce pFM307::AguaB-A into the wild-type Salmonella typhi strain Ty2 by homologous recombination as described by Blomfield et al., supra, Noriega et al., (1995) supra and Noriega et al., (1996) supra, except that sodium arsenite (SIGMA) was present in the culture medium throughout the procedure.

C. Mutant testing using DNA probes and PCR

New bacterial clones were grown at 37°C overnight on Luria agar plates supplemented with 6.0 μΜ sodium arsenite or on minimal medium. Colonies growing on arsenite-supplemented L agar but not on minimal medium were transferred to type 541 filter paper (Whatman, Maidstone, England) and blotted as described by Gicquelais et al., J. Clin. Microbiol., 28:2485-2490 (1990) and probed with γΡ ³² labeled 40 bp oligonucleotides: 5'-GGGCGGCCTGCGCTCCTGTATGGGTCTGACCGGCTGTGGT-3' (SEQ ID NO: 6) corresponding to the deleted portion of the wild-type guaB-A allele (negative probe). Clones that did not hybridize with the γΡ ³² labeled 40 bp oligonucleotides were collected. Salmonella typhi AguaB-A clones are unable to grow on minimal medium unless 10 mg guanine/L is supplemented. The deletion mutation of the guaB-A operon and the P _tac -ars insertion into the AguaB-A allele were confirmed by Southern blot hybridization with P ³² labeled probes for the AguaB-A operon, the ars operon, and the guaB-A negative probes described above. One Salmonella typhi AguaB-A clone was randomly selected and designated CVD915. Strain CVD915 was submitted to the American Type Cell Culture Collection on May 4, 1998, under ATCC accession number 202115.

Strain CVD915 grows on medium supplemented with 6.0 μΜ sodium arsenite, where none of the strains such as Ty2, nor other strains of Salmonella, Shigella or Escherichia coli that have been tested grow.

EXAMPLE 2. Invasiveness and intracellular growth in tissue cultures

Invasiveness and intracellular growth were assayed using the gentamycin technique, which was a slight modification of the technique previously described in Tartera et al., Infect Immun., 61:3084-3089 (1993). Briefly, a monolayer of Henle407 cells in a 24-well plate was infected in triplicate with wild-type strain Ty2, AguaB-A CVD915, AaroC, AaroD CVD908, or AaroC, AaroD, AhtrA CVD908 - htrA at a ratio of 50:1 for 90 minutes, and then any bacteria remaining outside the cells were killed with gentamycin for 30 minutes, the cells were washed (this time point was designated time = 0), and then the cells were further incubated with 20 pg/ml gentamycin. At 0 h, 4 h and 22 h, the test samples were lysed with sterile water and the dilution series were incubated overnight at 37 °C on LB agar supplemented with 10 μg guanine per liter. The results are shown in Table 2 below.

Invasiveness and intracellular growth on Henle407 cells

Intracellular cfu ^and tribe genotype 0 hours 4 hours 22 hours You2 wild type 6.7 x 10 ³ 3.1 x 10 ⁴ 8.8 x 10 ⁴ CVD915 AguaB-A 3.3 x 10 ² 2.9 x 10 ² 2.5 x 10 ² CVD908 AaroC, AaroD 5.8 x 10 ² 1.3 x 10 ³ 3.5 x 10 ³ CVD908-htrA AaroC, AaroD, AhtrA 1.3 x 10 ³ 3.4 x 10 ³ 1.3 x 10 ³

^and units forming colonies » · · · · · · ♦ · oo · · · * · * ·

Ζδ ·«·« ·· *·· ···· ·*

As shown in Table 2 above, strain CVD915 had an invasive ability and intracellular growth significantly lower than that of wild-type strain Ty2 and comparable to that of strain CVD908-htrA. That is, as shown in Table 2, in two different experiments, wild-type strain Salmonella typhi Ty2 effectively invaded Henle407 cells and multiplied in them thirteen-fold in 22 hours. Strain AaroC, AaroD CVD908 had a significantly weaker intracellular proliferation, i.e. six-fold in 22 hours.

As can also be seen from Table 2, the AguaB-A strain CVD915 has a significantly lower invasiveness rate on Henle cells than the wild-type parental strains CVD908 and CVD908-htrA and its intracellular growth, which is zero, is the same as that of the CVD-htrA mutant.

EXAMPLE 3: Comparison of the safety of the Ty21A, CVD908 and CVD915 vaccine strains

The attenuation of Salmonella typhi strain AguaB-A CVD915 was verified relative to wild-type strain Ty2 and other attenuated strains using the hog-mucin technique in mice (Powe et al., J. Biolog. Standar., 8:79-85 (1980)). Mice were infected intraperitoneally with 0.5 ml of a solution containing the various strains in a decimal dilution series and 10% hog mucin. The mice were then monitored for mortality for a further 72 hours and the LD50 of each group was calculated by linear regression analysis. The results are shown in Table 3.

Table 3.

tribe LD50 in cfu You2 1.4 x 10 ² CVD908 4.4 x 10 ⁶ CVD915 7.7 x 10 ⁷ Ty21a 1.9 x 10 ^x

As shown in Table 3 above, strain CVD915 has an LD50 value that is more than 5 orders of magnitude higher than that determined for wild-type strain Ty2 and between those determined for strains AaroC, AaroD CVD908 and the chemically mutated, non-invasive strain Ty21a.

EXAMPLE 4: Use of strain CVD915 as a live vector for delivery of foreign protein and plasmid DNA vaccine for expression of foreign antigen in eukaryotic cells

A mouse model of intranasal immunization with genetically engineered AaroC, AaroD strains of Salmonella typhi (CVD908) carrying the C fragment of tetanus toxin (C fragment of TT) was described in Galen • 9 ·· ·· • in · • 9 9 ·

9 9

9999 99

999 9999

99

9 9 9

9 9 9

9 9 9

9 9 9

9» 99 et al., Vaccine 15:700-708 (1997). The constructs used for this assay were plasmids carried by strain CVD908 encoding the unfused fragment C of tetanus toxin under the control of an anaerobically activatable prokaryotic promoter derived from nirB or the strong lpp promoter. It was found that:

(i) intranasal immunization of mice with this construct resulted in high titers of neutralizing serum antibodies against tetanus toxin (ii) immunized mice were then completely protected against a 150% lethal dose of tetanus toxin that killed all control mice (iii) cells obtained from the spleen and cervical lymph nodes of intranasally immunized mice 6 weeks after immunization exhibited a strong proliferative response to Salmonella typhi (Hd flagella and phenol-inactivated whole cells of Salmonella typhi), as well as to tetanus toxin fragment c, or whole tetanus toxin.

These findings were further extended by testing immunogenicity in mice with novel attenuated strains of Salmonella typhi carrying plasmids containing genes encoding foreign antigens (i.e., tetanus toxin C fragment) under the control of a prokaryotic or eukaryotic promoter.

A. Immune Response Induced by Immunization with Attenuated Salmonella Typhi Strains Carrying Prokaryotic and Eukaryotic Expression Vectors Encoding Tetanus Toxin Fragment C The observation that an immune response can be induced by direct immunization with plasmid DNA (Ulmer et al., Science, 259:1475-1749 (1993)) led to the testing of several nucleic acid vectors encoding foreign antigens as potential vaccines. The use of DNA vaccines has been shown to elicit both strong antibody and cellular immune responses when plasmids encoding a number of viral antigens, such as NP or influenza A virus (Ulmer et al, supra). Because of these earlier discoveries, several nucleic acids encoding foreign proteins have been tested as potential vaccines. DNA vaccination has been tested extensively, particularly with influenza virus antigens. Protective immune responses have been demonstrated in mice (Fynan et al, Proc. Natl. Acad. Sci., U.S.A., 90:11478-11482 (1993a); Justewicz et al (1995), supra; Justewicz et al, Virol., 224:10-17 (1996)), ferrets (Webster et al, Vaccine, 12:1495-1498 (1994)), chickens (Fynan et al, DNA Cel 1 Biol., 12:785-789 (1993b); Fynan et al (1993a), supra) and primates (Donnelly et al, Nat. Med., 1:583-587 (1995)) following intramuscular or epidermal (gene gun) immunization with a eukaryotic expression vector encoding influenza virus antigens. Immune responses against some other antigens of other viruses (Cox et al, supra; Davis et al (1996a), supra, Donnelly et al, supra; Lagging et al, supra; Wang et al, Virol., 211:102-112 (1995); Zarozinski et al, supra); parasites (Doolan et al, J. Exp. Med., 183:1739-1746 (1996); Gardner et al, J. Pharm: Sci., • ·

85:1294-1300 (1996); Hoffman et al, Vaccine, 12:1529-1533 (1994); Sedegah et al, supra; Yang et al., supra); bacteria (Anderson et al, Infect. Immun., 64:3168-3173 (1996); Barry et al, supra; Huygen et al, Nat. Med., 2:893-898 (1996); Luke et al, J. Infect. Dis., 175:91-97 (1997); Tascon et al, Nat. Med., 2:888-892 (1996)); tumor cells (Conry et al, Cancer Res., 54:1164-1168 (1994); Conry et al (1995), supra; Wang et al, Human Gene Therapy, 6:407-418 (1995)) has also been induced by DNA vaccines in various animal models. These studies help to understand the nature of the immune response induced by DNA vaccines. For example, a strong antibody response was induced in mice against the surface antigens of Hepatitis B virus (HbsAg) after administration of a single dose of a plasmid encoding this antigen (Davis et al., Hum. Mol. Genet., 2:1847-1851 (1993), Michel et al., Proc Nati. Acad. Sci., USA, 92:5307-5311 (1995)). In these studies, the maximum IgG antibody titer was observed 4 to 8 weeks after immunization and remained at this level for 6 months. A DNA vaccine encoding this antigen has also been shown to be effective in inducing antibody and MHC I cytotoxic responses in mice that do not normally respond to purified HBsAg (Davis et al., (1996b), supra, and Schrimback et al, J. Virol., 69:5929-5934 (1995)). Thus, at least in the case of HBsAg, a DNA vaccine has been shown to overcome genetic limitations in mice and to elicit a better preventive immune response than parenteral administration of purified protein.

Tetanus toxin (TT) is a potent neurotoxin synthesized by Clostridium tetani. Protective immunity against tetanus in humans and animals is measured by serum neutralizing antibody titers (Bizzini, Clostridium tetani, Gyles et al., (Eds), Iowa State University Press, Ames, IA, pp. 97-105 (1993) and Habig et al., Vaccines and Immunotherapy, Cryz (Ed), Pergamon, New York, pages 13-19 (1991)). Thus, a good protective immune response can be achieved by parenteral immunization with inactivated tetanus toxin or its nontoxic derivatives. The determination of the efficacy of this vaccine is a well-established procedure and is described in the European and British Pharmacopoeias (Coster et al., Lancet, 345:949-952 (1995) and Sanchez et al, Trans. R. Soc. Trop. Med. Hyg., 89:542-545 (1995)). The level of antibodies to tetanus toxin has traditionally been determined by the toxin neutralization technique in mice. In recent years, ELISA tests have been developed that are faster and cheaper and correspond well to the results of the neutralization test (Manghi et al., J. Immunol. Methods, 168:1724 (1994), Melville-Smith, Diptheria and Tetanus Antitoxins, Wreightt et al., (Eds.), ELISA in the Clinical Laboratory Service, London, pp. 136-147 (1990)). The titer of antibody to the toxin in serum is normally determined in IU/ml (where 0.1 IU/ml of human serum is sufficient to protect against tetanus toxin challenge) for comparison with the international standard anti-TT serum (WHO). The efficacy of the vaccine (expressed in IU/ml) can therefore be compared between different types of vaccines. Currently, tetanus vaccine is prepared by formaldehyde inactivation • ·

tetanus toxin from Clostridia tetani, which is a very lengthy and technically demanding task. Therefore, it is desirable to search for other possible detoxified variants of tetanus toxin for vaccination. TT consists of a 15 kDa protein containing a 50 kDa light chain (L) disulfide-linked to a heavy (H) chain (Helting et al., J. Biol. Chem., 252:187-193 (1977a) and Nieman et al., Molecular Biology of Clostridial Neurotoxins, Alouf Ed., Sourcebook of Bacterial Protein Toxins, Academie Press, London (1991)). The toxicity of this protein is mediated by the light chain, a zinc-dependent protease (Schiavo et al., EMBO J. 11:3577-3583 (1992)), which prevents release of the inhibitor from the neuron by proteolysis of synaptobrevin (Schiavo et al., Nature (London), 359:832-835 (1992)). The heavy (H) chains initiate penetration of the toxin across the presynaptic membrane (Helting et al, J. Biol. Chem. 252:194-197 (1977b) Morris et al., J. Biol. Chem., 255:6071-6076 (1980)). Digestion of the toxin molecule with papain yields a 50 kDa polypeptide corresponding to the C-terminal end of the H-chain and a 100 kDa molecule corresponding to the L chain attached to the N-terminal end of the H chain (Helting et al, (1977a) supra). This 50 kDa polypeptide, designated TT fragment C (FC), is nontoxic but binds gangliosides (Halpern et al., (1980) supra, Moris et al., (1980 supra) and has the ability to bind some proteins (Schiavo et al, FEBS. Lett. 290:227-230 (1991)). In earlier studies, vaccination of animals with FC obtained by proteolysis of native toxin was shown to be protective against subsequent administration of a lethal dose of tetanus toxin (Helting et al, (1977a) supra). In addition, studies with monoclonal antibodies demonstrated that neutralizing epitopes lie on this molecule (Kenimer et al., Infect. Immun. 42:942948 (1983)). The FC molecule was therefore a suitable candidate for the preparation of alternative vaccines against tetanus toxin.

Recent results show that intramuscular immunization with a plasmid encoding the C fragment of tetanus toxin under the control of the (Helting et al, (1977a) supra) promoter/enhancer (pcDNA3/tetC) induces the production of high titers of serum antibodies against fragment C, stimulates a proliferative response and the production of interferon γ. In addition, these mice have been shown to be resistant to a lethal dose of tetanus toxin (Anderse et al., supra).

In an effort to further investigate whether the immunogenicity and protective capacity of pcDNA3/tetC could be further optimized, a series of experiments were initiated to deliver this construct into host cells using attenuated Salonella typhi strains. This would be a completely new delivery system for DNA vaccines against bacterial and other infectious diseases.

A description of these experiments, including a description of the construction of pcDNA3/tetC, is described below.

• · • ·

1. Gene encoding fragment C of tetanus toxin

The sequence of the gene encoding tetanus toxin is well described in the art (Eisel et al, EMBO J., 5:2495-2502 (1986); Fairweather et al, J. Bacteriol., 165:21-27 (1986); Fairweather et al, Nucleic. Acids Res., 14:7809-7812 (1986)). Plasmid vectors encoding the native form of fragment C under the control of P _tac have previously been shown to induce low levels of FC production in Escherichia coli (3 to 4% of the time) (Makoff et al., Bio/Technology, 7:1043-1046 (1989)). In order to increase the expression level, the codon transcription of the sequence encoding fragment C (which is A and T rich) was optimized for Escherichia coli. The resulting 99% artificial gene encoding fragment C (tetC) increased expression levels (11 to 14% tep) in Escherichia coli (Makoff et al., Nucleic Acids Res., 17:10191-10202 (1989)). This led to testing of other expression systems with the intention of increasing expression levels and producing larger amounts of fragment C for vaccination. Preparations containing recombinant fragment C isolated from yeast (Clare et al., Biotechnology (NY), 9:455-460 (1991) Romanos et al, Nucleic Acids Res., 19:1461-1467 (1991)) and insect cells in culture (Charles et al., Infect immun., 59:1627-1632 (1991)) have also been found to be effective and capable of inducing protective immunity against tetanus toxin after parenteral immunization. Interestingly, as in the case of Escherichia coli, the native Clostridia tetani gene encoding fragment C could not be expressed in Saccharomyces cerevisiae and a gene with an optimized codon sequence from pTETtacll5 had to be expressed. The artificial fragment C, which was codon-optimized for expression in Escherichia coli, appears to be coincidentally optimized for expression in eukaryotic cells.

2. Construction of pTETnirl5

The stable high level of expression of a foreign antigen in vivo can be disrupted by events such as loss of the encoding plasmid or the inability to present this antigen to the immune system. Therefore, several procedures are used to stabilize the plasmid in vivo. One of these options is to use in vivo inducible promoters, such as the nirB promoter. This promoter originally comes from Escherichia coli, where it controls the expression of an operon encoding, for example, the NADH-dependent nitrate reductase gene, the nirB gene. Its expression is induced by nitrates and changes with the oxygen pressure in the environment, reaching maximum activity in an anaerobic environment. Modified versions of the nirB promoter have been prepared that induce expression when the bacteria grow in an atmosphere with limited oxygen supply, or when they enter a eukaryotic cell in vitro. Homologous Col-El plasmids expressing fragment C with the nirB and/or tac promoters have been prepared. These strains showed a high ability to induce a protective response against tetanus toxin after oral immunization. Two doses of the strain with the sca promoter provided complete protection, while the strains with the snirB promoter induced a sufficient protective immune response after the first dose. This suggests that the nirB construct is better and more stable in vivo than the tac constructs (Chatfield et al., Biotechnology, 10:888 (1992b)).

3. Construction of pcDNA3/tetC

Construction and characterization of the pcDNA3/tetC plasmid was performed according to the instructions provided in Anderson et al., supra, which is incorporated herein by reference.

4. Immunization with Salmonella typhi CVD915 carrying pcDNA3/tetC or pTETnirl5

As mentioned above, studies were conducted to test the ability of attenuated Salmonella typhi strains to deliver prokaryotic and eukaryotic expression vectors encoding foreign antigens, such as the C fragment of tetanus toxin, to the mammalian immune system and thus induce a protective immune response. To compare this system with traditional “naked” DNA immunization, the immune response elicited was also compared for the pcDNA3/tetC plasmid alone.

Plasmids pcDNA3/tetC and pTETnirl5 were introduced into CVD915 by electroporation. The structure of the purified plasmids was confirmed by agarose gel separation of restriction endonuclease digestion products: pcDNA3 and pcDNA3/tetC with HindIII and XbaI and pTETnirl5 with EcoRI and BamHI. The expression of fragment C was confirmed by SDS-page electrophoresis and immunoblotting. Bacterial samples were grown to stationary phase and 1.0 ml aliquots were collected by centrifugation and dissolved in 200 μΐ SDS buffer. These samples were then used for SDS-page electrophoresis and Western blotting. Mouse monoclonal antibody against fragment C and horseradish peroxidase-labeled goat antibody against mouse immunoglobulin were used for detection. Each sample was used at a concentration of approximately 1.0 x 10 ⁸ cfu/well.

The intranasal route was used for immunization with both the live vector CVD915 carrying both plasmids CVD915- pcDNA3/tetC and CVD915- pTETnirl5. For immunization with the pure plasmid, intramuscular administration (pcDNA3/tetC) was used. Balb/C mice were immunized according to the schedule shown in Table 4:

group number of mice immunogen method of immunization first dose second dose 1 10 CVD915- pcDNA3/tetC intranasally 5.9 x 10 ⁹ cfu 11 x 10 ⁹ cfu 2 10 CVD915-pTETnirl5 intranasally 5.0 x 10 ^y cfu 6.3 x 10 ⁹ cfu 3 10 CVD915 intranasally 6.5 x 10 ⁹ cfu 5.25 x 10 ⁹ cfu 4 10 pcDNA3/tetC intramuscularly 98 pages 95 pg 5 5 PBS intranasally 25 pl 5.0 μΐ

5. Antibody response

Blood was collected from the animals from the retroorbital vein according to the following protocol:

(a) before immunization (b) 14 days after immunization 36 days after immunization (c) 14 days after second dose 28 days after second dose 42 days after second dose days after second dose (animals killed)

The total amount of antibodies against fragment C in serum was determined by ELISA. In a simplified manner, recombinant fragment C was used to coat ELISA plates at a concentration of 4.0 pg/ml. Sera were diluted in a two-fold series to eight concentrations. The amount of antibodies expressed as IU was compared with a previous calibration of a standard serum. The standard serum was prepared by immunizing mice with adsorbed tetanus toxoid four times consecutively at two-week intervals. The sera were pooled and titrated in an in vivo neutralization test in mice according to the procedure described in the European Pharmacopoeia. The value of the international neutralization units was determined according to the International Standards for Tetanus Antitoxin. The results are shown in Fig. 6.

As shown in Fig. 6, although the CVD915-pcDNA3/tetC induced somewhat higher antibody production, very high titers of serum antibodies against fragment C (1000 to 2000 times higher than the protective level of antibodies against tetanus toxin in humans) were observed after immunization and a single secondary dose of both CVD915-pTETnirl5 and CVD915-pcDNA3/tetC constructs. The observed serum antibody level against fragment C reached up to 20 IU/ml, which corresponds to the end of the dilution series of about 1:50,000. In contrast, relatively low levels of antibodies against fragment C (about 50 times the protective level of antibodies against tetanus toxin in humans) were observed after immunization with the pure plasmid pcDNA3/tetC. Antibody levels against fragment C reached a maximum on the 50th day after the first immunization dose and remained at this level for at least 92 days after, which was the last time period monitored. No response was detected in mice immunized with the CVD915 strain alone.

6. Cell-mediated immune response: proliferation technique Days after immunization, mixtures of cell suspensions from cervical lymph nodes, mesenteric lymph nodes and spleens from 5 to 7 mice from all the above groups were prepared.

A cell suspension was also prepared from inguinal lymph nodes isolated from mice immunized with pure pcDNA3/tetC plasmid alone. The following average yields were obtained:

cervical lymph nodes: mesenteric lymph nodes: inguinal lymph nodes spleen:

6.6 x 10 ⁶ cells/mouse 10.1 x 10 ⁶ cells/mouse

1.6 x 10 ⁶ cells/mouse

52.5 x 10 cells/mouse

For the measurement of proliferation activity, cells were transferred to complete RPMI medium (RPMI containing 10% (v/v) fetal calf serum, glutamine and antibiotics) and incubated at a density of 2.0 x 10 ⁵ cells/well in triplicates in 96-well plates in the absence or presence of antigen. The following soluble antigens were used for this study: bovine serum albumin (BSA), tetanus toxin C fragment and highly purified Salmonella typhi flagella at concentrations of 0.02, 0.2 and 2.0 pg/ml. After five days of incubation at 37°C in 5% CO2, H thymidine was added to the culture and the culture was harvested after 18 hours.

The rate of thymidine incorporation was determined by liquid scintillation counting. The results shown in Figures 7 and 8 are reported as the stimulation index (S.I.).

As shown in Figure 7, immunization with the CVD915(pTETnirl5) and CVD915 (pcDNA3/tetC) constructs, as well as with the pure pcDNA3/tetC plasmid, induces the formation of activated lymphoid cells that proliferate in response to fragment C of tetanus toxin. This proliferative response was observed in cell populations isolated from cervical lymph nodes and spleen, but not in cells isolated from mesenteric lymph nodes. A specific proliferative response was also observed against tetanus toxin. No increase in proliferative activity against fragment C or tetanus toxin (> 3 S.I.) was observed in cells isolated from mice immunized with CVD915 or PBS.

As shown in Figure 8, proliferative responses to Salmonella typhi flagella were observed in cell populations isolated from mice immunized with CVD915, CVD915(pTETnirl5) and CVD915(pcDNA3/tetC) constructs. As shown in Figure 8, this response was observable in all populations isolated from mice immunized with CVD915 and in lymphoid cells isolated from mice immunized with CVD915(pTETnirl5) and CVD915(pcDNA3/tetC) constructs. A significantly lower proliferative response was observed in splenocytes and mesenteric lymph node cells isolated from mice immunized with CVD915(pTETnirl5) and CVD915(pcDNA3/tetC) constructs. Due to technical difficulties, no data are presented regarding the proliferative response to Salmonella typhi flagella in mice immunized with the pcDNA3/tetC plasmid alone. Interestingly, cells isolated from the inguinal lymph nodes of mice immunized i.m. with the pcDNA3/tetC plasmid alone showed no specific proliferative response to any of the antigens tested. No proliferative response to BSA was observed in any of the tested groups.

7. Comparison of serum IgG antibodies against Salmonella LPS and against Salmonella typhi flagella induced by intranasal immunization with CVD915 and CVD908-htrA

One of the most important parameters in the development of a new vaccine is to compare the immune response elicited (e.g. CVD915) with the best proposed vaccine (e.g. CVD908-htrA) for which the greatest amount of experimental data is available. A group of 10 Balb/C mice were immunized intranasally with 10 ^1θ cfu of the attenuated CVD915 or CVD908htrA strains, in two doses 36 days apart. Mice were bled before immunization (day 0) and on days 35, 55 and 95 (CVD915 only). Anti-LPS and anti-flagellar antibody titers were determined by ELISA as described by Tacket et al (1977), supra. Briefly, ELISA plates were coated with 5.0 μg of each antigen, and serum samples were tested in 8 dilutions in a two-fold series. The antibody titer was determined as ELISA units/ml, which is defined as the inverse dilution ratio that has an absorbance of 0.5 at 492 nm. The results are shown in Fig. 9 and 10.

As shown in Fig. 9 and 10 in the case of LPS and Salmonella flagella, very close titer values of specific antibodies were measured in both cases.

EXAMPLE 5. Preparation of attenuated strains of Salmonella typhi stably expressing the Vi antigen

A. Preparation of Salmonella typhi strains (CVD916 and CVD909) stably expressing the Vi antigen In order to shift the expression of Vi antigens to a stable osmotically regulated site, it was necessary to target the vipR promoter (Fig. 1A). It is likely that the vipR and ompR-envZ products achieve their regulatory activity by binding to the upstream region of vipR (Hashimoto et al., (1996) see above). Thanks to this, it is possible, as described in the proposed invention, to replace the vipR promoter with another strong promoter, e.g. P _tac , to suppress the suppressive effect of the vipR promoter and thereby achieve control over the expression of Vi antigens. The P _tac promoter is stable in Salmonella species because these organisms lack laql . Stable expression of Vi antigens in strains related to CVD915 and CVD908-htrA was therefore ensured in this way.

In the first phase, a deletion was introduced into vipR and a positive or negative selection marker, e.g., a sacB-neo cassette, was inserted to replace the vipR promoter region (Fig. 1A). Specifically, a 601 bp segment just upstream of the -35 region of the vipR promoter was amplified from genomic DNA of wild-type Salmonella typhi Ty2 strain using the following primers:

'-GGGGGAGCTCAATTCTGCAAACCAGCCCTGTACCATCAAGTTCATA-3' (Sequence No. 7) and '-CCTCATCCCGGGCCCGAGTCCACCTGCACAATTCATTGTTTGTACCTATC-3' (Sequence No. 8).

The first 601 bp amplified (segment A) were cloned using the pGEM-T Vector System kit (Promega, Madison, WI) to generate the pGEM-T::fragment A construct. This system includes an open plasmid (pGEM-T) with a 3'-T overhang to the insertion site, which improves the efficiency of ligation of PCR products. The pGEM-T::fragment A construct is then digested with Sstl and BamHI and cloned into the Sstl and BamHI sites upstream of P _tac in pBS::P _tac (this plasmid is obtained by cloning the P _tac promoter into the BamHI-EcoRI sites of pBluescript (Stratagen)), resulting in the pBS:: segment AP _tac .

At the same time, a 770 bp segment of vipR (including the vipR initiation codon) (segment B) was amplified from the same genomic DNA of strain Ty2, using the primers: 5'-GCAGGTGGATCCGGCCCGGGATGAGGTTTCATCATTTCTGGCCTCCGAATGAT ATC-3' (sequence no. 9) and

5'-ATCCTTGAATTCGGGGATCCTACTAAAATTTTATATTTACAAAGTTAATTCTAG GT-3' (SEQ ID NO: 10). With these primers, unique restriction sites Sstl, EcoRI, BamHI and Smali (underlined) were introduced for cloning purposes. The amplified segment B was first cloned into pGEM-T using the pGEM-T kit from Vector System, resulting in pGEM-T::segment B. This plasmid was digested with EcoRV and SalI and the resulting fragment was cloned into the EcoRV and SalI sites of pBS::segment AP _tac just downstream of P _tac . The resulting plasmid, called pBS::P _tac -vipR, contains the AP _tac -segment B segment and a 167 bp deletion was introduced in the -35, -10 and ribosomal binding regions of the vipR promoter. In the construction of pBS::P _tac -vipR, the deletion in the vipR promoter was effectively replaced by a 257 bp insertion containing -35, -10 and the ribosomal binding region of the tac promoter.

At the same time, another cassette was prepared by inserting sacB-neo between segment A and segment B, as described above. The original fragments A and B were fused by PCR using both fragments as templates and oligonucleotides according to sequence no. 7 and 10 as primers, as described by Noriega et al., (1996) see above. The resulting amplified fused fragment A-B was cloned into pGEMT::fragment A - fragment B. The sacB-neo sequence was obtained by digestion of pIB729 with SmaI (Blomfield et al., see above) and inserted into the SmaI site of pGEM-T::fragment A - fragment B. The resulting plasmid was designated pGEM-T::vipR::sacB-neo. This plasmid was digested with SstI, effectively removing the vipR::sacB-neo allele, which was then cloned into the SstI site of the suicide vector pJG14, generating pJG14::vipR::sacB-neo. Plasmid pJG14 is a temperature-sensitive, pSCIOl-derived, chloramphenicol-selectable, suicide plasmid (Galen et al., 96th General Meeting, American Society for Microbiology, Abstract, page 529-H260 (1996)). In addition, the SstI-digested vipR::sacB-neo allele was cloned into the SstI site of the suicide vector pKIN1701 (Hone et al., (1991) supra), generating pKT::vipR::sacB-neo. Salmonella typhi strain CVD915 was transfected by electroporation with pJG14::vipR::sacB-neo. At the same time, strain CVD908-htrA was electroporated with pKT::vipR::sacB-neo. Homologous recombination occurred between pJG14::vipR::sacB-neo and the vipR gene of Salmonella typhi strain CVD915, using the technique described in Noriega et al., (1996, supra), and between pKT::vipR::sacB-neo and the vipR gene of Salmonella typhi strain CVD908-htrA, using the technique described in (Hone et al., (1991) supra), except that double crossover mutants were removed by isolating kanamycin-resistant, chloramphenicol-sensitive clones. The resulting strains, related to Salmonella typhi strains CVD915 and CVD908-htrA, did not express the Vi antigen due to an insertion/deletion in the vipR allele.

In the second step, the stable P _tac promoter was replaced by sacB-neo insertion in the vipR locus of the new Salmonella typhi mutant strains CVD915 and CVD908-htrA (Fig. 1B). Specifically, the P _tac -vipR segment in pBS::P _tac -vipR was cloned into the BamHI-EcoRI site of pJG14, generating pJG14::P _tac vipR. The plasmid pJG14::P _tac -vipR was then used to replace the P _tac with the sacB-neo cassette in the new Salmonella typhi mutant strains CVD915 and CVD908-htrA by homologous recombination as described above. Separation of the two crossover mutants was performed by negative selection using sucrose toxicity of sacB, and reverse selection for kanamycin resistance. The resulting strain derived from CVD915 was designated CVD916 and was deposited in the American Type Culture Collection in May 1998, under ATCC number 202116. The new strain derived from CVD908-htrA was designated CVD909 and was deposited in the American Type Culture Collection in May 1998, under ATCC number 202117.

Genotyping confirmation of the P _tac insertion in both strains was performed using PCR, which confirmed its insertion at the correct site.

B. Stable expression of Vi antigen by strains CVD916 and CVD909

Phenotyping of stable expression of the Vi antigen on the new strains was performed by agglutination of bacteria grown at different osmolarities in the presence of a commercial anti-VI antibody (Difco). The results are shown in Table 5, below.

Table 5. Agglutination using specific anti-Vi serum

NaCl tribe CVD915 CVD916 CVD908-htrA CVD909 You2 0.17 M +++ +++ +++ +++ +++ 0.3M +++ +++ +++ +++ +++ 0.5M - +++ +/- +++ - 0.6M - +++ - +++ - 0.7M - +++ - +++ -

As shown in Table 5, the wild-type strain Salmonella typhi Ty2 and the attenuated strains CVD915 and CVD908-htrA, the expression of Vi antigens is highly dependent on the osmolarity of the medium (determined by the concentration of NaCl). In contrast, the expression of Vi antigens in strains CVD916 and CVD909 is high, stable and does not show any changes in osmolarity.

EXAMPLE 6. Immune response to stably expressed Vi antigen

Experimental groups of 6-week-old Balb/C mice were immunized intranasally with 1.0 x 10 ¹⁰ cfu of CVD915 or CVD916. Mice were bled before and 30 days after immunization and their sera were frozen at -20°C until testing. Antibody titers against LPS and Salmonella flagella in the serum of immunized mice were determined by ELISA, as described by Tacket et al., (1997) supra. The results are shown in Table 6 below.

Table 6. Geometric mean titer of serum IgG antibodies to Salmonella typhi antigens after immunization with a single dose of strains CVD915 and CVD916

Specific IgG antibodies against See LPS H tribe day 0 day 30 day 0 day 30 day 0 day 30 CVD915 21 46 19 640 20 197 CVD916 26 * 109 17 747” 21 _ Ht 320

* p = 0.033 ** p = not significant

As shown in Table 6, immunization with strain CVD916 induces antibodies against the Vi antigen to a significantly higher extent than with strain CVD915. The immune response against other Salmonella typhi antigens (LPS and H) was similar in both strains compared. The results demonstrate that stable expression of the Vi antigen enhances the immune response against this antigen without affecting the immune response against other antigens present on Salmonella typhi cells.

Although the present invention has been described in great detail with reference to specific embodiments of the present invention, it will be apparent to any person skilled in the art that numerous modifications may be made without departing from the scope of the present invention.

Claims (18)

PATENT CLAIMS An attenuated mutant Salmonella strain stably expressing a Vi antigen and whose vipR promoter is replaced by a strong stable promoter to ensure stable and stable expression of viaB. The attenuated mutated Salmonella strain of claim 1, wherein the mutant is a mutated Salmonella typhi strain. The attenuated mutant Salmonella strain according to claims 1 and 2, wherein the vipR promoter of said mutant strain is replaced by a promoter from the group P 1, P 1, P ₀ and _ac and Pi _pp , in such a way as to ensure stable viaB expression. An attenuated mutated Salmonella strain according to claims 1 and 2, wherein said mutated strain is unable to produce de novo guanine-containing nucleotides due to a mutation in the guaB-A operon. The attenuated mutant Salmonella strain according to claim 4, wherein the mutation described above in the guaA-B operon is a deletion mutation and is located either in the guaA gene, guaB gene, or in both the guaA and guaB genes. The attenuated mutated Salmonella strain according to claim 5, wherein the deletion mutation a described above is located simultaneously in both the guaA and guaB genes. The attenuated mutant Salmonella strain of claim 6, wherein the phenotype of said mutant strain is aro ‘. The attenuated mutated Salmonella strain of claim 2, wherein the mutated Salmonella typhi strain is derived from a parent Salmonella typhi strain CVD915 (ATCC No. 202115). The attenuated mutant Salmonella strain of claim 2, wherein the mutated Salmonella typhi strain is Salmonella typhi strain CVD916 (ATCC No. 202116) or Salmonella typhi strain CVD909 (ATCC No. 202117). The attenuated mutated Salmonella strain of claim 1, wherein the mutated strain encodes and expresses the foreign antigen. The attenuated mutated Salmonella strain of claim 1, wherein the mutated strain comprises a plasmid that encodes and expresses the foreign antigen in a eukaryotic cell. <· <<<<<<<<<<<<<<<<<<<<<<<<< 12. Typhoid vaccine, comprising: (A) a pharmaceutically effective amount of an attenuated mutated Salmonella typhi that stably expresses a Vi antigen on its surface and whose vipR promoter is replaced with a strong stable promoter to ensure stable and stable expression of viaB (B) a pharmaceutically acceptable solvent or carrier The vaccine of claim 12, wherein the vipR promoter of this mutant strain is replaced with a promoter from the group of P _tac , P _trc , Poiac and Pi _pp , in such a way as to ensure stable expression of viaB. The vaccine of claims 12 or 13, wherein said mutated strain is unable to produce de novo guanine-containing nucleotides due to a mutation in the guaA-Β operon. The vaccine of claim 14, wherein the mutation described above in the guaA-Β operon is a deletion mutation and is located either in the guaA gene, guaB gene, or in both the guaA and guaB genes. The vaccine of claim 15, wherein the deletion mutation α described above is located simultaneously in both the guaA and guaB genes. The vaccine of claim 16, wherein the phenotype of said mutated strain is aro '. The vaccine of claim 12, wherein said mutated Salmonella typhi strain is derived from a parent Salmonella typhi strain CVD915 (ATCC No. 202115).