INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Microbiological surveys in areas where diarrheal disease is endemic implicate Shigella species as etiologic agents in at least 20% of diarrheal cases. Shigella flexneri 2a is usually the most prevalent species and serotype in these areas (8, 14, 30). Shigellae are extraordinarily adept intestinal pathogens, as evidenced by their small infectious doses (7). Shigella infection is usually transmitted by the fecal-oral route and can be manifested as uncomplicated watery diarrhea. A more definitive manifestation of shigellosis is dysentery, i.e., frequent passage of small-volume stools with gross blood, mucus, and fecal leukocytes. Constitutional symptoms (e.g., fever, rectal tenesmus, and headache) also characterize severe disease. Colonoscopy of patients infected with either Shigella dysenteriae or S. flexneri reveals diffuse erythema, focal hemorrhages, and inflammatory changes resembling ulcerative colitis. Rectal biopsies taken during the early stages of infection reveal aphthoid lesions overlying small lymphoid follicles (23). These clinical findings are consistent with experimental observations made with the rabbit ileal loop model, suggesting that Shigella initiates intestinal infections by invading the follicle-associated membranous cells (28).

Experiments employing polarized epithelial cells as a model of the intestinal epithelium suggest that shigellae invade enterocytes through the basolateral membrane. Internalized bacteria subsequently spread within infected cells by organizing host cell actin into a cytoskeleton-based motor (10, 22). Genetic analysis has shown that this spreading phenotype is dependent upon a plasmid-borne virulence gene designated icsA (2) or virG (22). The 120-kDa protein expressed by this plasmid-carried gene acts as a recruiter for cytosolic nucleators of filamentous actin (10). This actin is concentrated at the distal poles of septating shigellae, and the resulting comet-like tails provide a motive force for the bacteria within the cytoplasm of infected epithelial cells (36). The mobilized bacteria impinge on the inner face of the host cell plasma membrane, and they spread into contiguous epithelial cells via membrane protrusions (10). Intragastric challenge of rhesus monkeys, a primate model of intestinal shigellosis, demonstrates that icsA-mediated intercellular spread of shigellae is a key step in pathogenesis. For example, endoscopy of asymptomatic animals challenged with an icsA mutant of S. flexneri serotype 5 reveals only scattered nodular abscesses rather than the hemorrhagic ulcerations and diffuse mucosal inflammation seen in animals challenged with the virulent parent strain (32). icsA mutants are also avirulent in the Sereny guinea pig keratoconjunctivitis model of suppurative Shigella infection (13, 26). Deletion of the iuc chromosomal locus (encoding aerobactin) partially attenuates S. flexneri in the Sereny guinea pig test and in the rabbit ileal loop model (24). Intragastric inoculation with either an icsA single mutant or an icsA iuc double mutant protects rhesus monkeys against subsequent challenge with the virulent S. flexneri 5 parent strain (32, 33).

Epidemiological and clinical studies indicate that an episode of shigellosis elicits substantial immunity against subsequent disease caused by the same Shigella serotype (6, 8, 15, 19). In a rational approach to Shigella vaccine development, we and others have constructed genetically attenuated vaccines designed to establish asymptomatic infections that induce protective immune responses. However, the ideal balance of safety and efficacy in attenuated Shigella vaccines has been elusive (11, 15-18, 20, 25). Current research suggests that icsA iuc mutants could serve as attenuated Shigella vaccines, and the SC602 icsA iuc S. flexneri 2a candidate was constructed to test this combination of attenuating mutations in volunteers. Here we describe a preliminary dose selection study, two expanded dose selection studies, and an efficacy (challenge) study of SC602 that were performed in the clinical inpatient ward of the U.S. Army Medical Research Institute for Infectious Diseases (USAMRIID), Ft. Detrick, Md.

(Parts of this work were previously presented at the 96th General Meeting of the American Society for Microbiology, 19 to 23 May 1996, and at the 98th General Meeting of the American Society for Microbiology, 17 to 21 May 1998.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Vaccine construction and manufacture. S. flexneri 2a strain 454, from the Centre National de Reference des Shigelles, Unité des Entérobactéries, Institut Pasteur, was the SC602 progenitor. The icsA iuc double mutant was constructed in the Unité de Pathogénie Microbienne Moléculaire, Institut Pasteur, as described previously (1). The iuc mutation was generated by recombination of iuc::Tn10 into the chromosome by using phage P1 transduction. Spontaneous excision of the tetracycline resistance gene, and its flanking regions including the iuc locus, was selected by growth on fusaric acid medium. The icsA gene was inactivated by double recombination with a kanamycin resistance (Km^r)-sucrose sensitivity (sacB) cartridge carrying flanking regions of icsA. Deletion of the Km^r-sacB cartridge was selected by growth on sucrose, and the resistant clones were screened for retention of the invasive phenotype in HeLa cells. An isolate designated SC602 had suffered a deletion of the entire icsA gene along with substantial flanking sequences (total deletion is approximately 10 kb). This SC602 isolate was expanded into a master cell bank and was manufactured as a lyophilized product under current good manufacturing procedures at the Walter Reed Army Institute of Research pilot vaccine production facility in Forest Glen, Md. The product was dispensed as a 5-ml fill in 50-ml serum bottles and was stored at 80°C. The reconstituted product yielded 5 × 10¹⁰ to 1 × 10¹¹ CFU per vial (with approximately 30% viability). This organism was invasive for tissue culture cells, and preclinical studies demonstrated its safety and efficacy in guinea pig and rhesus monkey models. Clinical trials of SC602 were conducted under a Food and Drug Administration investigational new drug application.

Subject selection. Volunteers were recruited from the local community, and written, informed consent was obtained under protocols approved by internal review boards within USAMRIID. Potential volunteers were excluded if they reported previous exposure to shigellae; were allergic to quinolones; had any significant gastrointestinal abnormality; were pregnant; were HLA B27, human immunodeficiency virus, or hepatitis B surface antigen positive; were currently being treated with antibiotics, theophylline, iron, zinc, histamine H₂-receptor antagonist blockers, or proton pump inhibitors; or had a febrile illness within 48 h of admission. Because of the theoretical possibility of vaccine excretion after release from the ward, food handlers, day care workers, and individuals who live with a child less than 2 years of age were also excluded.

Vaccination, challenge, and safety assessment. Volunteers fasted for 90 minutes before and after vaccination (and before and after challenge in the subsequent efficacy study). Lyophilized SC602 vaccine was reconstituted in sterile, deionized water and was diluted in phosphate-buffered saline to achieve the target number of CFU in 1-ml volumes. This inoculum was mixed with 30 ml of sodium bicarbonate buffer (2 g of NaHCO₃ per 150 ml of sterile, deionized water) and was ingested by each volunteer 2 min after ingestion of 120 ml of the sodium bicarbonate solution (19). Placebo controls received sodium bicarbonate buffer with no added bacteria. The challenge inoculum, containing approximately 10³ CFU of virulent S. flexneri 2a strain 2457T, was prepared at the Center for Vaccine Development, University of Maryland School of Medicine, and was administered with sodium bicarbonate as described previously (19). All subjects who were vaccinated or challenged with S. flexneri were treated with ciprofloxacin (500 mg, twice daily for 5 days), and passage of two consecutive stools with no cultivable S. flexneri was a prerequisite for discharge. Statistical significance between study groups was determined by a two-tailed Fisher's exact test.

Laboratory methods. A measured sample from each collected stool (or a rectal swab if no stool was passed within 24 h) was suspended in buffered glycerol saline, diluted in phosphate-buffered saline, and plated for quantitative colony count on Hektoen enteric agar (Difco Laboratories, Detroit, Mich.). Non-lactose-fermenting colonies were identified as S. flexneri 2a by slide agglutination in homologous antiserum (Difco). Ten colonies had to test negative in 2a antiserum before a Hektoen enteric agar plate was recorded as negative for S. flexneri. Randomly selected S. flexneri 2a isolates were confirmed to be the SC602 icsA deletion mutant by Southern blotting of bacterial DNA extracted and digested with EcoRI and SalI. The blotted DNA fragments were hybridized with a radiolabeled probe consisting of a nick-translated PCR product amplified from an internal portion of the icsA structural gene. The enzyme-linked immunospot assay (35) was used to enumerate immunoglobulin A (IgA), IgG, and IgM antibody-secreting cells (ASC) per 10⁶ peripheral blood lymphocytes (PBL) in samples obtained on days 0, 5, 7, and 9. The means plus 3 standard deviations (SD) of numbers of ASC recognizing S. flexneri 2a lipopolysaccharide (LPS) on day 0 were 5.6 (IgA), 7.1 (IgG), and 7.6 (IgM). Antibody responses against S. flexneri 2a LPS and Shigella invasion plasmid antigen (Ipa) proteins (27) were assessed as IgM, IgA, and IgG enzyme-linked immunosorbent assay (ELISA) titers in serum collected on days 0, 7, 14, and 28. Titers of antibody against Ipa proteins were determined by endpoint dilution with seroconversion defined as a fourfold rise in titer (12). Titers of antibody against S. flexneri 2a LPS were determined in serum and urine samples by using endpoint titers derived from a linear regression analysis of eight doubling dilutions by using adjusted optical densities of 0.3 for serum and 0.1 for urine (3). Secretory IgA (sIgA) recognizing S. flexneri 2a LPS in urine was quantified by ELISA, and the titer was adjusted for urine concentration by using the creatinine concentration as a divisor (4). Antibody titers in urine collected on days 7, 14, and 28 were considered significant if there was a fourfold increase in titer and if the peak titer exceeded mean day 0 values by 3 SD.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

SC602 dose selection studies. Thirty-three subjects (aged 19 to 46 years) were enrolled in the initial, placebo-controlled dose selection trial. Eighteen subjects received the SC602 vaccine (Table 1) and fifteen received sodium bicarbonate placebo. The objective of this trial was to determine the maximal tolerated vaccine dose. The first group of volunteers was inoculated with a single dose of 10² CFU and was treated with ciprofloxacin on day 3 postvaccination. Five additional cohorts (three vaccinees and three placebo controls per group) were inoculated according to a double-blinded, placebo-controlled protocol. Cohorts 2 and 3 received doses of vaccine on days 0 and 3. Cohorts 2 and 3 showed that the first inoculation of SC602 achieved adequate intestinal colonization; therefore, groups 4, 5, and 6 received only one dose of vaccine. On day 8 postvaccination, or earlier if clinically indicated, ciprofloxacin treatment was initiated. Doses of 10² to 10⁷ CFU were well tolerated in that no vaccine recipients developed diarrhea. However, transient fever was observed in 20% of these vaccinees, including one subject in group 2 who had ingested 10⁴ CFU. A control volunteer in the same group also had fever, and one control volunteer in group 3 had diarrhea. Severe headache (7% of vaccinees), moderate headache (20%), and moderate abdominal cramping and loss of appetite (7%) were reported by vaccinees in groups 2 through 5. Moderate headache was reported by 16% of controls. Within the first day after vaccination with 2.9 × 10⁸ CFU, two of three vaccinees in group 6 experienced diarrhea, fever, and severe intestinal and constitutional symptoms. These subjects were treated with ciprofloxacin on day 1. This initial study established 10⁸ CFU as a reactogenic endpoint. The subsequent, expanded dose selection trials assessed the safety and immunogenicity of a maximal tolerated dose defined as being 100-fold below the experimentally determined reactogenic dose.

Laboratory evaluation of dose selection studies. Robust and prolonged intestinal colonization by S. flexneri 2a was observed in all volunteers who had ingested the SC602 vaccine. For example, 100% of volunteers who had ingested 2 × 10⁶ CFU excreted shigellae for 7 days, and 57% excreted the organisms until treatment began 12 days later. Likewise, 92% of the volunteers who had ingested 10⁴ CFU shed shigellae until treated on day 8. The peak excretion of vaccine was 10⁴ to 10⁶ CFU/g of stool regardless of the dose ingested. The first stools yielding S. flexneri 2a were passed by 93% of volunteers within 24 h of ingesting 10⁶ CFU of SC602, and 100% of volunteers excreted these organisms within 12 h of ingesting 10⁸ CFU. In both cases, symptoms of shigellosis coincided with vaccine excretion. In contrast, only 42% of volunteers who ingested 10⁴ CFU excreted S. flexneri 2a within 24 h. The proportion of excretors gradually increased to 58% on day 2, 83% on day 3, and 91% on day 4.

Clinical evaluation of phase 2b efficacy trial. The dose selection trials established 10⁴ CFU of SC602 as a relatively safe vaccine dose that induces measurable immune responses in a majority of volunteers. Subjects who received this dose were eligible to volunteer for an efficacy trial that was scheduled 8 weeks after inoculation. The seven subjects who volunteered were readmitted to the inpatient ward along with seven unvaccinated controls. All volunteers were treated with ciprofloxacin on day 5, or sooner if they met the clinical criteria for shigellosis. Following experimental challenge, the symptoms of dysentery, fever, and severe shigellosis were confined to the unvaccinated control group (Table 4). The absence of these symptoms in vaccinees allowed statistical differentiation of the two groups on the basis of fever and severe shigellosis (P = 0.005). Volunteer H in the control group (Table 3) failed to excrete the challenge organism, and this individual had no symptoms of shigellosis. Interestingly, this volunteer was the only subject who had a significant number of circulating ASC against S. flexneri 2a LPS at the time of challenge (9 IgA and 14 IgM/10⁶ PBL). With the exception of volunteer H, all controls had severe shigellosis, as illustrated by a mean of eight diarrheal stools per day during the acute phase of disease and a mean total of 11 diarrheal stools (Table 4).

Immune correlates of protection in the phase 2b efficacy trial. Immune correlates of vaccine efficacy against diarrhea and severe shigellosis included a significant IgA ASC response and a threefold or greater rise in serum IgA antibody against S. flexneri 2a LPS (Table 3). Other correlates of protection against all symptoms included urinary sIgA responses against 2a LPS in addition to IgG ASC and IgG serum responses. Subjects B/D and J/G evidenced only IgM ASC responses against the S. flexneri 2a LPS after vaccination, and these volunteers experienced mild diarrhea after challenge, even though both were protected from severe shigellosis. Secretory IgA is locally produced and is actively transported into the colon rather than into the peripheral circulation (29); therefore, it is possible that protective levels of secretory IgA could be present on the colonic epithelium of vaccinees who did not demonstrate measurable IgA responses in the peripheral circulation. For example, subject S/N had substantial IgM and IgA ASC responses after vaccination, but no serum antibody response was detected. Although S/N passed some diarrheal stools, S. flexneri 2a was not excreted by this vaccinee after challenge, suggesting a substantial degree of immunity.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Site-directed genetic mutation of Shigella has allowed construction of stable, attenuated, candidate vaccines that retain the invasive phenotype. These vaccines are designed to initiate abortive intestinal infections, efficiently delivering protective Shigella antigens through follicle-associated membranous cells into the underlying gut-associated lymphoid tissue without eliciting clinical shigellosis. For the present study, the SC602 S. flexneri 2a candidate vaccine was attenuated by deletion of the icsA gene, resulting in the loss of intracellular motility and the intercellular spreading phenotype. The secondary aerobactin mutation (iuc) is not sufficient to attenuate shigellae for vaccine use, but it may moderate the reactogenicity of SC602.

Others have attenuated candidate vaccines with an aro mutation (16) or with combinations of aro and icsA mutations (virG) (20). We have evaluated an aro hybrid of Escherichia coli K-12 and S. flexneri 2a (18) that also has some characteristics of an icsA (virG) mutant (25). Auxotrophic aro mutants require PABA (para-aminobenzoic acid) for intracellular growth (16), and they do not survive intracellularly since PABA is not a constituent of mammalian cytosol. The aro mutants are clearly attenuated and demonstrably safe at doses of 10⁶ or 10⁷ CFU (16). When the dose of aro vaccines was increased to 10⁸ or 10⁹ CFU, however, a significant proportion of volunteers suffered intestinal or systemic reactions within the first 24 h. These reactions, which included diarrhea and fever, were similar to those observed after volunteers ingested 10⁶ CFU of SC602.

In contrast to the 10⁶-CFU dose of SC602, a 10⁴-CFU dose was not associated with serious adverse events, although one subject developed transient fever and one volunteer passed occasional diarrheal stools (13% reactogenicity). Mild to moderate constitutional and intestinal symptoms were also experienced by some subjects, but these subclinical symptoms did not affect normal activities. In the context of previous studies with aro mutants, the reactogenicity of a 10⁴-CFU dose of SC602 is roughly comparable to a 7 × 10⁸-CFU dose of the aroD hybrid EcSf2a-2 and to a 10⁷-CFU dose of the SFL1070 aroD mutant (16). The reactogenicity of this dose also falls within the spectrum of reactions reported after ingestion of 10⁸ and 10⁶ CFU of CVD1203 aroA virG double mutant (20). A significant immune response against S. flexneri 2a LPS was detected in approximately two-thirds of volunteers ingesting any of the above vaccines. Only the EcSf2a-2 and the SC602 vaccines have been evaluated for efficacy by challenge with virulent S. flexneri 2a 2457T. The former vaccine elicited no significant protection against fever, dysentery, or diarrhea (17, 18). In contrast, SC602 gave significant protection against fever and severe shigellosis. The mild diarrhea that was experienced by some of the challenged vaccinees did not inhibit normal activities and did not indicate antibiotic treatment. Solid protection against severe shigellosis, with occasional mild diarrhea or fever, was also reported in experimental rechallenge studies of volunteers who had experienced previous clinical shigellosis (15, 19, 29).

The mean time to excretion of S. flexneri 2a by vaccinees who ingested 10⁴ CFU of SC602 was 72 h, while the mean time to excretion by volunteers who ingested 10⁶ CFU was only 22 h. We speculate that the lower vaccine inoculum allows gradual invasion of the colonic epithelium by SC602 while limiting the cumulative inflammatory response to a subclinical threshold in most subjects. Of the five volunteers who excreted substantial numbers of shigellae within 24 h of ingesting a 10⁴-CFU dose, two experienced presumptive vaccine reactions (fever of 100.9°F or mild diarrhea). In contrast, none of the vaccinees with delayed colonization had clinical symptoms. An expanded outpatient safety trial of SC602 has recently been performed to further assess the safety and immunogenicity of the 10⁴ CFU dosage (5). In this trial, six volunteers (19%) experienced mild diarrhea or fever lasting less than 24 h. These data suggest a narrow window of safety for icsA vaccines, and further attenuation of SC602 would be desirable if the efficacy of this strain can be maintained.

One of the characteristics of SC602 that sets it apart from the aro vaccines is the persistent colonization that is achieved even after ingestion of a 10⁴-CFU dose of vaccine. For example, all volunteers were excreting shigellae 8 days after vaccination in the clinic-based studies described above, and the mean time of colonization in a subsequent community-based phase 1 trial was 12 days (34). In contrast, the auxotrophic aro vaccines are usually excreted for 5 days or less (16, 18, 20). We judge the hazard of secondary fecal-oral transmission of SC602 to be minimal. Secondary spread of S. flexneri 2a from an adult index case of shigellosis occurs in only 5% of households (37), and no transmission to placebo controls was detected in the clinical trials of SC602 presented here. In addition, excretion of SC602 by Bangladeshi adults in ongoing phase 1 trials has been scarcely detectable, suggesting that secondary spread of the vaccine in areas of endemicity will also be unlikely. It should also be noted that the reactogenicity profile of live Shigella vaccines is ameliorated in partially immune adolescents and adults living in developing countries (21), and this is also the case with SC602.

Naturally acquired immunity against shigellosis is species specific (8), and it is anticipated that a multivalent product, or a series of vaccinations with different icsA mutant vaccines, would be required to make a significant impact on levels of shigellosis in developing countries where the disorder is endemic. Candidate icsA vaccines have been described for S. dysenteriae 1 (9) and Shigella sonnei (13), and phase 1 trials of these serotypes are planned. Although the initial consumers of commercially produced vaccines against shigellosis would be the 20 million civilian and military travelers who visit developing countries annually, the simple and economical manufacturing process required for these live vaccines would make local production in developing countries feasible. The single-dose regimen suggested by the present clinical trials is a distinct advantage for vaccine delivery to children at risk of shigellosis in areas of endemicity. By eliminating the need for medical intervention and antibiotic usage (31) in vaccinated populations, icsA vaccines could prove to be practical public health tools for the prophylactic control of shigellosis in areas where shigellosis is endemic.