T- and B-Cell Immune Responses of Patients Who Had Undergone Colectomies to Oral Administration of Salmonella enterica Serovar Typhi Ty21a Vaccine

Study design.Nine adult patients (two of whom were women), aged 36 to 56 years (mean age, 45 years), who had undergone colectomies due to ulcerative colitis, were recruited from the regular follow-up program for patients with inflammatory bowel disease at the Department of Surgery of the Sahlgrenska University Hospital in Göteborg. Continence surgery had been performed 1 to 20 years earlier (mean, 11.3 years) by the construction of a pelvic pouch with an ileoanal anastomosis. All patients had had very similar amounts of small bowel resected. The maximal extent of the small bowel resection was limited to 10 cm of the distal ileum. All patients were generally in good health and had had no episodes of acute pouchitis or signs of extraintestinal manifestations of ulcerative colitis for the last 3 years preceding the study. None of the patients had been taking anti-inflammatory drugs or antibiotics 1 month prior to the study period, nor did they take any during the study period. The only drugs regularly administered to the patients who had undergone colectomies were loperamide (four patients) and vitamin B₁₂ (three patients). Seventeen healthy volunteers (10 of whom were women), aged 20 to 66 years (mean age, 35 years), were recruited and served as controls to the patients who had undergone colectomies. None of the individuals had a history of typhoid fever or had previously been vaccinated against typhoid fever. All individuals agreed to participate in the study, which was undertaken with due approval from the Human Research Ethical Committee at the Medical Faculty, Göteborg University.

Each individual received, 2 days apart, a total of three doses of a live attenuated S. enterica serovar Typhi Ty21a vaccine (Vivotif; Swiss Serum and Vaccine Institute Berne, Bern, Switzerland). The vaccine was ingested orally in the form of enteric-coated capsules containing at least 2 × 10⁹ bacteria per capsule.

Collection of specimens.To obtain peripheral blood mononuclear cells (MNCs), 50 ml of heparinized venous blood was collected from six of the patients who had undergone colectomies and from 10 of the healthy volunteers prior to the first immunization (day 0), and then 7 or 8 days after the onset of vaccination. Serum specimens were obtained from nine patients who had undergone colectomies and 17 healthy volunteers before the first immunization and then 7 or 8 days and/or 14 days after the first vaccine dose. From the patients, ileostomy fluids were also obtained on days 0, 7 or 8, 21, and 35 after the onset of immunization. In addition, intestinal biopsies were collected from the ileal pouches of three of the patients on days 0 and 8. The ileostomy fluids were collected within 3 h after the last emptying of the reservoir. A 50-ml portion of fluid was immediately chilled on ice, centrifuged, and treated with enzyme inhibitors as previously described (18). The ileostomy fluids were frozen at −70°C until used. Biopsies were collected from the ileal pouch by using a gastrointestinal fibroscope and a pair of biopsy forceps (FB 24 K; Olympus, Stockholm, Sweden). On each occasion, 15 pinch biopsies, 1 to 2 mm in diameter, were collected.

Cell preparation.Peripheral blood MNCs were isolated by density gradient centrifugation on Ficoll-Paque (Pharmacia Biotech AB, Uppsala, Sweden), and intestinal MNCs were isolated from ileal pouch biopsies by using an enzymatic dispersion technique as previously described (23). A pool of 15 ileal pouch biopsies yielded 0.6 × 10⁶ to 4.8 × 10⁶ MNCs. After isolation, the blood and intestinal MNCs were washed three times with phosphate-buffered saline (PBS) and resuspended in culture medium (Iscove's medium supplemented with 5% heat-inactivated AB⁺ serum, 1% l-glutamine, and 1% gentamicin). CD4⁺ and CD8⁺ cells were isolated from the blood MNCs by positive selection by using immunomagnetic techniques as described previously (22). The purity of the blood T-cell populations was monitored by incubation with anti-CD4 fluorescein isothiocyanate and anti-CD8 phycoerythrin (both from Becton-Dickinson, San Jose, Calif.) followed by analysis with a flow cytometer (FACSCalibur, Becton-Dickinson). The purified populations were found to contain at least 97% CD4⁺ cells and 94% CD8⁺ cells.

Cell culture.To obtain antigen-presenting cells (APCs), the CD4⁺- and CD8⁺-depleted blood MNCs remaining after the magnetic separation were plated in round-bottomed 96-well plates (NUNC, Roskilde, Denmark) at 10⁵ cells per well and incubated for 2 h at 37°C. The nonadherent cells were then removed, the wells were washed twice with PBS, and 10⁵ cells per well of the two different T-cell subpopulations were added followed by the addition of 2 × 10⁶ heat-killed serovar Typhi Ty21a bacteria. The heat-killed Ty21a bacteria were obtained by incubating viable vaccine strain bacteria which had been subcultured on horse blood agar plates in an 80°C water bath for 1 h (22). Phytohemagglutinin (PHA) (5 μg ml⁻¹; Murex diagnostics Ltd, Temple Hill, United Kingdom) was used as a positive control, and culture medium alone was used as a negative control. The plates were incubated at 37°C and 5% CO₂ for 5 days. Two days after initiation of the cultures, 100 μl of the culture medium supernatant was removed and replaced by 100 μl of fresh culture medium. The supernatants (three or six wells) were pooled and stored at −70°C until analyzed for cytokine content. After another 3 days, the wells were pulsed with 0.5 μCi of [³H]thymidine (Amersham, Arlington Heights, Ill.) per well for 8 h, and then the plates were frozen at −20°C. Later on the plates were thawed, and the cells were harvested onto nylon filters and analyzed for proliferative responses by using a scintillation counter. For each specimen, the median counts per minute for a set of unstimulated cultures were subtracted from the median counts per minute for a corresponding set of cultures stimulated with heat-killed Ty21a vaccine strain bacteria. A greater-than-twofold increase in net counts per minute value between pre- and postimmunization specimens was considered a significant proliferative response.

Cytokine detection.The concentration of IFN-γ in ileostomy fluids and cell supernatants from peripheral blood and intestinal MNCs was determined by a recently described enzyme-linked immunosorbent assay (ELISA) method (22). Ileostomy fluids were also analyzed for cytokine concentrations of interleukin-4 (IL-4) and transforming growth factor β with reagents (DuoSet) from R&D Systems (Abingdon, United Kingdom) and for IL-10 with reagents (Opt-EIA) from BD PharMingen (San Diego, Calif.); these assays were performed according to the recommendations of the manufacturers. In each instance the ratio of the amount of released IFN-γ in cell culture supernatants stimulated with heat-killed Ty21a bacteria to the amount stimulated with PHA was calculated. A greater-than-twofold increase in IFN-γ production between pre- and postimmunization specimens was defined as a significant response.

Antibody determinations.Levels of total IgA in ileostomy fluids were determined by a modified microplate ELISA as previously described (3, 25). The antibody responses to Salmonella in serum (IgA and IgG) and in ileostomy fluid (IgA) were analyzed by a modified whole-cell bacterium ELISA (14, 22). Briefly, flat-bottomed polystyrene microtiter plates (NUNC) were coated with 0.01 mg of poly-l-lysine (Sigma) ml⁻¹, and then live S. enterica serovar Typhimurium (strain SL2404; kindly provided by Bruce Stocker, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, Calif.) (10⁷ bacteria per well) were added. The SL2404 strain expresses the same O antigens (i.e., O9 and O12) as the vaccine strain Ty21a and the wild-type S. enterica serovar Typhi (14). The bacteria were fixed in the plates by centrifugation followed by the addition of 0.5% glutaraldehyde (Sigma). Nonspecific binding sites of the plates were blocked with 10% fetal calf serum in PBS. After incubating serial dilutions of pre- and postvaccination specimens for 2 h, horseradish peroxidase-labeled rabbit anti-human IgA or IgG antibodies (Jackson Immunoresearch Laboratories, Inc., West Grove, Pa.) were added, and the plates were further incubated for 90 min followed by the addition of o-phenylenediamine substrate for 30 min. The antibody titers were determined as the interpolated dilutions of the samples giving an absorbance at 492 nm of 0.4 above background. A greater-than-twofold increase in titer between pre- and postimmunization samples was chosen to signify seroconversion (14). The specific anti-Salmonella O9 and O12 IgA activity in ileostomy fluid was determined by dividing the IgA anti-Salmonella titer by the total IgA concentration (in micrograms per milliliter) of the sample. A greater-than-twofold increase in the ratio of mean IgA anti-Salmonella titer to total IgA concentration between pre- and postimmunization specimens was regarded as a significant response (11, 14).

T-cell immune responses.The T-cell immune responses in peripheral blood after oral immunization with a live attenuated S. enterica serovar Typhi Ty21a vaccine were studied for six patients and 10 healthy volunteers by measuring the proliferative responses and IFN-γ production after stimulation of purified CD4⁺ and CD8⁺ cells with heat-killed Ty21a bacteria. Three doses of vaccine failed to induce any proliferative responses among the various T-cell subsets from the patients, and increases in IFN-γ production were found only among the stimulated CD8⁺ cells from three (50%) of the patients (Table 1 in CD8⁺ cells (P = 0.034 versus patients; Fisher's exact test, two-sided), and for 3 of them a proliferative response was also observed for CD4⁺ cells; the magnitude of the responses among responders was 20-fold for CD8⁺ cells and 3-fold for CD4⁺ cells. Increases in IFN-γ production were observed for CD4⁺ and CD8⁺ T cells from three and six of the healthy volunteers, respectively (Table 1 and Fig. 1).

The mucosal T-cell response in the distal ileum of the patients who had undergone colectomies was also assessed by obtaining ileal pouch biopsy specimens from three of them. The low cell yield did not allow the use of immunomagnetic cell sorting techniques. No proliferative responses could be detected after vaccination, and stimulation of the intestinal MNCs with heat-killed Ty21a bacteria did not result in any significant increases in IFN-γ production (data not shown). Low levels of IFN-γ, IL-4, IL-10, and transforming growth factor β were detected in some of the ileostomy fluids obtained from the patients before immunization, but there were no appreciable changes in cytokine content seen after administration of Ty21a vaccine.

B-cell immune responses.In accordance with the T-cell immune responses seen in peripheral blood after vaccination, the antibody responses in serum were also weaker in the patients than in the healthy volunteers (Table 2). Thus, increases in Salmonella-specific IgA and/or IgG antibody titers were observed for five (56%) of nine patients and for 15 (88%) of 17 healthy volunteers. The magnitudes of the Salmonella IgA as well as the IgG antibody responses were significantly lower for the patients than for the healthy volunteers (Table 2).

To evaluate the ability of the Ty21a vaccine to induce a mucosal B-cell response in the intestines of patients who had undergone colectomies, vaccine-specific IgA antibodies in ileostomy fluids were also measured. Three doses of vaccine induced significant increases in the ratio of Salmonella-specific IgA antibody titers to total IgA concentrations in six (67%) of nine patients, and the maximal geometric mean increase for responders was 12-fold. Peak intestinal IgA antibacterial responses were seen on days 21 to 35 after the onset of vaccination for five of the patients and on day 8 for one patient (Table 3).