This phase 2a, observer-blind, randomized, single-center, controlled study was conducted at the KEMRI-Wellcome Trust in Kilifi, Kenya, between August 2016 and March 2017. The study enrolled healthy adults aged 18–45 years, fulfilling protocol inclusion and exclusion criteria, willing to comply with study procedures and signing, or thumb printing, the informed consent form for study participation. Females of child bearing potential were enrolled only if they agreed to use an effective birth-control method prior to and during the study. However, any potential pregnancies during the trial were to be reported and, if possible, their outcome was to be monitored. Any condition potentially interfering with the ability to participate in the study or with the study results, or causing additional risk by participation in the trial was an exclusion criterion. Individuals were also excluded if they had any of the following conditions: progressive or severe neurological disorders, seizures, previous Guillain–Barré syndrome, history of reactive arthritis, hepatitis B infection, HIV or HIV-related disease, autoimmune disorders, known or suspected impairment/alteration of the immune system, known bleeding diathesis (or any condition that may be associated with prolonged bleeding time), serious chronic or progressive disease, malignancy or lympho-proliferative disorder, history of allergy to vaccine components or of substance or alcohol abuse within the past 2 years, participation in clinical trials with other investigational product within 28 days prior to screening, receipt of vaccines containing meningococcus A, C, W, Y, tetanus, diphtheria or pertussis antigens within 12 months before screening, receipt of any other vaccine within 4 weeks prior to screening or plan to do so during the study, receipt of blood/plasma products within 12 weeks prior to first study vaccination, body mass index >30 kg/m², laboratory confirmed case of disease by S. sonnei, and breast-feeding.

Individuals with an absolute neutrophil count (ANC) <1.8 × 10⁹/L for the initial 18 participants (first study cohort) or <1.0 × 10⁹/L, if recommended by an independent Data Safety Monitoring Board (DSMB) for the remaining study population (second cohort) at screening or with previous history of benign ethnic neutropenia/drug-related neutropenia, prior use or likelihood to use neutropenic drugs were not enrolled. Inclusion criteria were re-assessed for all participants, prior to each study vaccination.

Enrolled individuals were randomized in a 1:1:1 ratio, to receive two vaccinations with either the investigational or the control vaccines at day (D) 1 and D29. Investigational groups received the S. sonnei 1790GAHB vaccine, used at two different OAg/protein doses, while participants in the control group were randomized to receive a quadrivalent meningococcal conjugate vaccine (MenACWY; Menveo, GSK) at D1 and a vaccine against tetanus, diphtheria, and acellular pertussis (Tdap; Boostrix, GSK) at D29 (Figure 2).

Randomization was performed by a validated internet-based system. The study was observer-blind, due to the different presentations of the investigational and control vaccines. Designated unblinded trained and qualified staff prepared or administered the study vaccines, but was not involved in the evaluation of the participants for safety or in the collection of study data.

The study was monitored by the DSMB. Eighteen participants were initially enrolled in the first study cohort, received first vaccination and were followed up, according to study procedures, for 7 days. A summary of all safety data and listings including hematology, blood chemistry and urine dipstick/urinalysis test values were provided to the DSMB. After reviewing all safety data, the DSMB recommended that enrollment could be completed using the published African consensus ANC threshold of 1.0 × 10⁹/L (18) as inclusion criterion and the severity grading system proposed by the Division of Acquired Immunodeficiency Syndrome (19), which takes into consideration also the ethnic differences in ANC, be used for classification of postvaccination neutropenia. The DSMB was also consulted for any potential safety issue reported during the trial.

The informed consent form and the study protocol were reviewed and approved by the KEMRI Scientific and Ethics Committee, the Kenyan Pharmacy and Poisons Board, and the Oxford Tropical Research Ethics Committee prior to study start. The trial was designed and conducted in agreement with the ICH Harmonized Tripartite Guidelines for Good Clinical Practice, applicable local regulations and the Declaration of Helsinki and was registered at ClinicalTrials.gov (NCT02676895).

The primary objective was to evaluate the safety profile of two vaccinations in healthy adults with two different dose levels of 1790GAHB in a Shigella-endemic country. The secondary objective was to assess the immunogenicity of the investigational vaccine, as measured by anti-S. sonnei LPS serum immunoglobulin G (IgG) levels, at 28 days after each vaccination.

The 1790GAHB vaccine consisted of S. sonnei 1790-GMMA (approximately 11.8 µg OAg/200 μg total protein per milliliter) adsorbed to Alhydrogel (0.7 mg Al³⁺/mL) in tris-buffered saline, was available as a liquid formulation in single-dose vials with 0.7 mL of injectable suspension and did not contain any preservative. A 0.5-mL dose containing 1.5/25 μg of OAg/protein was obtained by bed-side mixing, by dilution with Alhydrogel in tris-buffered saline (0.7 mg Al³⁺/mL).

Each 0.5 mL dose of MenACWY contained 10 µg of serogroup A oligosaccharide and 5 µg of each of serogroups C, W, and Y oligosaccharides conjugated to 32.7–64.1 µg of CRM₁₉₇.

Each 0.5 mL dose of Tdap contained 2.5 Lf diphtheria toxoid, 5 Lf tetanus toxoid, 8 µg pertussis toxoid, 8 µg filamentous hemagglutinin, 2.5 µg pertactin, and 0.5 mg aluminum hydroxide.

At each vaccination, a 0.5-mL vaccine dose was administered intramuscularly into the deltoid area of the non-dominant arm.

Participants were observed at 30 and 60 min after each vaccination for any adverse event (AE). Local (injection site pain, erythema, and induration) and systemic (arthralgia, chills, fatigue, headache, malaise, myalgia, and fever) solicited AEs were recorded on diary cards by study personnel performing daily home visits for 7 days following each vaccination. Unsolicited AEs occurring within 28 days following vaccination were reported by the participants and documented by the investigator, during follow-up clinic visits carried out at 7 and 28 days after each vaccination or in the course of unscheduled visits. Solicited AEs continuing beyond the 7-day period following each vaccination were reported as unsolicited AEs.

All AEs were graded for severity by the investigator. Erythema, induration, and swelling of 25–50 mm, 51–100 mm, and >100 mm and fever as axillary temperatures of ≥38.0–38.9, ≥39.0–39.9, and ≥40.0°C were graded as mild, moderate, and severe, respectively. All other local and systemic AEs, if present, were classified as: mild (present but not interfering with activity), moderate (interfering with activity), and severe (preventing daily activity).

Serious AEs (SAEs), AEs of special interest (AESIs), and AEs leading to withdrawal from the study were collected for the entire duration of the study.

The relationship between study vaccination and unsolicited AEs, medically attended AEs, any new onset of chronic disease, AEs leading to withdrawal, and SAEs were also assessed by the investigator. As defined in the clinical protocol, reactive arthritis and neutropenia were AESIs (the former being a general concern for enteric pathogen vaccinations and the latter due to occurrence of similar events during the phase 1 trials evaluating the 1790GAHB vaccine) and, if present, were to be reported as SAEs.

Blood samples for hematology (white blood cells, red blood cells, hemoglobin, hematocrit, platelets, eosinophils, basophils, neutrophils, monocytes, and lymphocytes) and clinical chemistry (total bilirubin, aspartic aminotransferase, alanine aminotransferase, γ-glutamyl transferase, lactic dehydrogenase, alkaline phosphatase, glucose, blood urea nitrogen, creatinine, sodium, and potassium) testing, and urine dipstick samples (for the assessment of glucose, proteins, pH, ketones, nitrites, and blood levels) were collected at 7 and 28 days post-each vaccination. Urinalysis (white blood cells, red blood cells, casts, and bacteria) was performed if urine dipstick showed deviations from normal values. Laboratory measurements were assessed by the investigators and any abnormality considered as clinically significant was reported as an AE.

All study participants with a neutropenia (ANC < 1.8 × 10⁹/L for adults in the first cohort and ANC < 1.0 × 10⁹/L for individuals from the second cohort), occurring at 7 days after each vaccination, had additional blood draws for repeated complete blood count on a weekly basis until resolution of neutropenia. Occurrence of ANC < 0.5 × 10⁹/L, after the first vaccination, was an exclusion criterion for second vaccination.

Blood samples for immunogenicity assessments were collected prior to first vaccination and 28 days post-each vaccination (Figure 2). The sera were kept frozen at −80°C at the KEMRI-Wellcome Trust laboratory until shipment to the GSK Clinical Laboratory Sciences (Marburg, Germany), for serologic testing; for each participant, one aliquot of serum was stored in the clinical site laboratory and one aliquot was used for immunogenicity analyses.

Anti-S. sonnei LPS serum IgG was measured by an enzyme-linked immunosorbent assay (ELISA), as previously described (15).

Seroresponse to vaccination was defined as an increase in the anti-S. sonnei LPS serum IgG level of at least 50% for participants with pre-vaccination levels >50 ELISA units (EU) or an increase of at least 25 EU for participants with pre-vaccination levels ≤50 EU.

In the absence of a correlate of protection, the median anti-S. sonnei LPS serum IgG following vaccination was compared to the median level in convalescent patient sera from individuals infected with S. sonnei, as previously reported (20). Postvaccination levels of 121 EU were estimated to correspond to the median end point titer of 1:800 measured for the convalescent sera with an ELISA method by Cohen et al. (16).

A total of 72 participants were planned to be enrolled in the study. No formal statistical sample size was calculated, due to the descriptive nature of the study objectives.

Safety analyses were carried out in all participants from the exposed full analysis set who received at least one study vaccination and had safety data. The number and percentage of participants with AEs, SAEs, AESIs, and deviations from normal ranges of safety laboratory data after vaccination was calculated.

Serologic analyses were performed on participants from the full analysis set who had available ELISA data. The ELISA antibody concentrations were logarithmically transformed (base 10). For each group, geometric mean concentrations (GMC) and their 95% confidence intervals (CIs) were computed by exponentiating (base 10) the mean and 95% CIs of the log10 ELISA concentration. ELISA concentrations below the limit of detection were set to half that limit for the purposes of analysis.

The number and percentage of participants with seroresponse and with postvaccination levels ≥121 EU for anti-S. sonnei LPS serum IgG at 28 days post-each vaccination was calculated together with 95% Clopper–Pearson CIs.

Additionally, geometric mean ratios (GMR) were computed for GMC at 1 month after first and second vaccination versus baseline levels (D1). The GMR and 95% CIs were constructed by exponentiating the mean within-subject differences in log-transformed titers and the corresponding 95% CIs.

A total of 152 adults were screened, 74 were enrolled and randomized, 72 received at least one study vaccination, and 66 completed the study. Main reasons for the 78 screening failures were: not fulfillment of inclusion/exclusion criteria (n = 52) and lack of interest to further trial participation, despite being eligible (n = 15). Of the two individuals who were not vaccinated after randomization, one declined vaccination and the other was erroneously randomized, after expiry of the allowed 28-day window for screening. The primary reasons for study discontinuation were AEs (n = 4), administrative reason (n = 1), lost to follow-up (n = 1), protocol deviation (n = 1), and consent withdrawal (n = 1) (Figure 2). The overwhelming majority of the participants were male (88–91% in each group) and of Black origin (≥95% across all groups). Baseline characteristics were well-matched across all vaccine groups (Table 1).

Following the first vaccination, pain was the only reported solicited local AE, in 20 (91%), 25 (96%), and 10 (42%) participants in the 1.5/25 μg, 5.9/100 μg, and control groups, respectively (Table 2). Post-second vaccination, pain was reported by 15 participants in each of the 1.5/25 μg (68%) and 5.9/100 μg (65%) groups, compared with 17 (81%) in the control group, while induration was only reported by 1 (5%) participant in the control group. All reported pain was mild to moderate (Figure 3). No severe local reactions were recorded (Table 2).

The most frequently reported solicited systemic AE was headache, reported by 5 (23%), 10 (38%), and 8 (33%) participants following first vaccination and 4 (18%), 5 (22%), and 4 (19%) participants following the second vaccination, in the 1.5/25 μg, 5.9/100 μg, and control groups, respectively (Table 2). The incidence of all systemic AEs seemed to decrease following the second vaccination and severe reactions were only reported in 1 (4%) participant in the 5.9/100 μg (severe headache occurring 6 h after the first vaccination and only lasting for that day) (Table 2).

Unsolicited AEs following any vaccination were reported by 19 (86%), 24 (92%), and 19 (79%) participants in the 1.5/25 μg, 5.9/100 μg, and control groups, respectively. Possibly or probably related unsolicited AEs were reported by 13 (59%) participants in the 1.5/25 μg and 16 (62%) in the 5.9/100 μg groups, compared to 12 (50%) in the control group. Most of these AEs were post-immunization reactions continuing beyond the 7-day collection window following vaccination (Table 3). Fever was very seldom observed (Table 3) and was never ≥39°C; no participant reported the use of analgesics/antipyretics within 24 h prior to each vaccination.

During the trial, 10 episodes of neutropenia fulfilling the protocol definition of AESI were reported in two (9%), three (12%), and one (4%) participants in the 1.5/25 μg, 5.9/100 μg, and control groups, respectively. These episodes were all considered probably or possibly related to vaccination, two of them occurred in the 1.5/25 μg group (one mild and one moderate), five in the 5.9/100 μg group (four mild and one moderate), and three in the control group (one mild, one moderate, and one severe). All cases were transient (i.e., recovery within 7 days and by study end) and, except for one participant in the 5.9/100 μg group, who experienced upper respiratory tract infection (D1) and 38.1°C fever (D8), were also asymptomatic, as confirmed by daily home visits during the 7 days postvaccination. Eight of these episodes occurred in 5 of the 18 individuals from the first cohort (screened and monitored using the ANC Western threshold of 1.8 × 10⁹/L), while the remaining two episodes occurred in one of the 54 participants in the second cohort (screened and monitored using the local threshold of 1.0 × 10⁹/L). No SAEs occurred in the study. However, as defined in the study protocol the AESIs were reported as SAEs. No case of reactive arthritis was recorded.

Four participants were prematurely withdrawn due to unsolicited AEs: three in the 5.9/100 μg group (two cases of neutropenia and one case of bone tuberculosis) and one in the control group (γ-glutamyl transferase increase). One participant, who should have been excluded from the second vaccination due to neutropenia, was inadvertently not withdrawn and experienced moderate neutropenia at D57.

Few laboratory abnormalities were considered as clinically significant by the investigators. These were: an increase in alkaline phosphatase levels in one participant in the control group; γ-glutamyl transferase increase in one individual in the control group, and one in the 5.9/100 μg group; decreased hematocrit and hemoglobin levels in one participant in the 5.9/100 μg group; low ANC for one participant in group 5.9/100 μg group; a low platelet count for one participant in the control group; and one increase in platelet count in the 5.9/100 μg group. Following urinalysis, increased leukocyte levels for two, three, and one participants in the 1.5/25 μg, 5.9/100 μg, and control groups, respectively, and high erythrocyte level for one individual in each of the 1.5/25 and 5.9/100 μg groups were considered clinically significant. There were no pregnancies, hospitalization or deaths reported in the study.

Pre-vaccination anti-S. sonnei LPS IgG GMC varied between 971 and 1,196 EU across all groups (Figure 4A). At 28 days post-first vaccination, antibody levels increased 2.10- and 4.43-fold from baseline values in the 1.5/25 and 5.9/100 μg groups, respectively, but no increase was observed in the control group (Figures 4A,B). At 28 days after the second vaccination, anti-S. sonnei LPS IgG GMC further increased in the 1.5/25 μg group, but not in the 5.9/100 μg group; however, changes observed from D29 to D57 were not statistically significant (Figure 4A).

In the 1.5/25 μg group, the seroresponse was 68% after the first vaccination and 90% after the second vaccination, whereas in the 5.9/100 μg group, seroresponse was 96% after both the first and second vaccination (Figure 5A).

At baseline, the percentages of participants with anti-S. sonnei LPS IgG ≥ 121 EU were 95, 92, and 88% for the 1.5/25 μg, 5.9/100 μg, and control group, respectively. At 28 days following each vaccination, all participants in the investigation groups achieved anti-S. sonnei LPS IgG ≥ 121 EU, while in the control group 86 and 90% of participants had this level at 28 days post-first and second vaccination, respectively (Figure 5B).

Reverse cumulative distribution curves for anti-S. sonnei LPS antibody levels pre-vaccination and following each vaccination with the 1790GAHB vaccine were compared to antibody levels in convalescent patients as shown in Figure 6. The review of individual immunogenicity results showed that data from five serum samples (out of the 210 collected during the study and shipped for testing), all obtained at D29, were clinically implausible. Most likely, the root cause was human error during sample labeling, executed in the same days for 2 and 3 of the samples, respectively; however, as no definite proof of an error could be established, the original data have been used for the analyses. Additional analyses were performed excluding the potentially invalid results and the interpretation of immunogenicity results did not change.