Research ArticleInfectious Disease

Chimpanzee adenovirus– and MVA-vectored respiratory syncytial virus vaccine is safe and immunogenic in adults

See allHide authors and affiliations

Science Translational Medicine  12 Aug 2015:
Vol. 7, Issue 300, pp. 300ra126
DOI: 10.1126/scitranslmed.aac5745

RSV vaccine cows infection

Respiratory syncytial virus (RSV) causes a severe lower respiratory tract disease that affects both children and the elderly. Vaccines have shown promise in rodents and nonhuman primates, but it remains unclear if these models reflect human RSV infection. Now, two papers by Taylor et al. and Green et al. translate one vaccine strategy first into calves, which are natural hosts of bovine RSV, and then into humans in a phase 1 clinical trial. A prime-boost strategy protected against upper and lower respiratory tract infection and pulmonary disease induced by heterologous bovine RSV challenge in calves and demonstrated safety and immunogenicity in humans. These data support further trials to test vaccine efficacy in human patients.


Respiratory syncytial virus (RSV) causes respiratory infection in annual epidemics, with infants and the elderly at particular risk of developing severe disease and death. However, despite its importance, no vaccine exists. The chimpanzee adenovirus, PanAd3-RSV, and modified vaccinia virus Ankara, MVA-RSV, are replication-defective viral vectors encoding the RSV fusion (F), nucleocapsid (N), and matrix (M2-1) proteins for the induction of humoral and cellular responses. We performed an open-label, dose escalation, phase 1 clinical trial in 42 healthy adults in which four different combinations of prime/boost vaccinations were investigated for safety and immunogenicity, including both intramuscular (IM) and intranasal (IN) administration of the adenovirus-vectored vaccine. The vaccines were safe and well tolerated, with the most common reported adverse events being mild injection site reactions. No vaccine-related serious adverse events occurred. RSV neutralizing antibody titers rose in response to IM prime with PanAd3-RSV and after IM boost for individuals primed by the IN route. Circulating anti-F immunoglobulin G (IgG) and IgA antibody-secreting cells (ASCs) were observed after the IM prime and IM boost. RSV-specific T cell responses were increased after the IM PanAd3-RSV prime and were most efficiently boosted by IM MVA-RSV. Interferon-γ (IFN-γ) secretion after boost was from both CD4+ and CD8+ T cells, without detectable T helper cell 2 (TH2) cytokines that have been previously associated with immune pathogenesis following exposure to RSV after the formalin-inactivated RSV vaccine. In conclusion, PanAd3-RSV and MVA-RSV are safe and immunogenic in healthy adults. These vaccine candidates warrant further clinical evaluation of efficacy to assess their potential to reduce the burden of RSV disease.


Respiratory syncytial virus (RSV) causes annual epidemics of respiratory infection throughout life, with young infants and the elderly especially susceptible to developing severe disease. There is only supportive care for cases of infection. Despite decades of research effort, there remains no licensed vaccine for the prevention of severe disease, and the use of palivizumab monoclonal antibody prophylaxis is limited to high-risk infants only. The peak incidence of disease requiring hospitalization is in the first few months of life (1). Infants with bronchiolitis account for up to 18% of winter pediatric admissions (2), and infection by RSV is responsible for almost 80% of cases (3). Two-thirds of infants are infected by RSV in the first year of life (1), and 2 to 3% of primary infections require admission to hospital (4, 5). About 6% of these admissions will require management on dedicated pediatric intensive care units (6). Worldwide, RSV disease in children under the age of 5 years account for an estimated 33.8 million lower respiratory tract infections, 3.4 million hospitalizations, and up to 200,000 deaths annually (7). RSV-associated deaths are almost exclusive to resource-poor areas of the world, where RSV is second only to malaria in all-cause infant mortality between 1 and 12 months of age (7, 8). Cumulative RSV exposure produces an immune response capable of protection against severe lower respiratory tract disease but not protection from infection. Fifty percent of infants suffer at least one RSV reinfection by their second birthday (1), and there is increasing evidence for causality between RSV infection in infancy and subsequent wheezing and possibly asthma in later life (9, 10). Healthy adults can expect a 7 to 9% annual risk of infection with mild symptomatic disease (11, 12), and severe immune suppression can reestablish a risk of developing severe disease (13). Later in life, senescence of the immune system and comorbid conditions place the elderly at risk of developing severe RSV disease, and estimates of hospital burden and mortality from RSV in the elderly vary but may be comparable to seasonal influenza (1416).

The high rate of emergency admissions, the lack of universal and cost-effective preventative measures, and the magnitude of seasonal disease incidence maintain RSV as a major priority for vaccine development (17). Concerns over safety and an incomplete understanding of the immune correlates of protection have hampered efforts to develop such a vaccine. The formalin-inactivated RSV vaccine candidate (FI-RSV), tested nearly 50 years ago, led to enhanced respiratory disease (ERD) upon RSV exposure. ERD, which had a propensity to manifest in the youngest infants, caused 80% to become hospitalized over the subsequent RSV transmission season (compared to 5% in controls) and two fatalities in FI-RSV recipients (1821). Animal challenge experiments and postmortem lung histology from the infant fatalities implicated ERD as a vaccine-primed T helper cell 2 (TH2)–associated immunopathology after exposure to natural RSV infection and have cast a long shadow over subsequent vaccine design and development. Subunit vaccine formulations have therefore remained in development for seropositive target populations and as maternal vaccines. Intranasal (IN) live attenuated RSV vaccine candidates, whose development also started in the 1960s, continue to be evaluated in humans including RSV-naïve infants but have been troubled by nasal congestion (young infants being obligate nasal breathers), genetic instability, vaccine virus transmission, and limited immunogenicity (2226).

Genetically modified chimpanzee-derived adenovirus (ChAd) and modified vaccinia virus Ankara (MVA) viral vectors are safe and highly successful biological platforms that induce robust immune responses when used as genetic vaccine carriers for several infectious diseases and cancer (27, 28). We have generated an artificial, consensus-based, RSV antigen including a soluble fusion (F) protein for the induction of neutralizing antibodies and the conserved nucleocapsid (N) and matrix (M2-1) internal proteins to elicit T cell immunity. The replication-defective viral vectors PanAd3 and MVA encode this antigen as a vaccine. PanAd3 is a hexon group C ChAd and one of the most immunologically potent adenoviral vectors tested in rodents and primates (29). PanAd3 has not been tested in humans, but its sequence is very similar to other group C adenoviral vectors, including human Ad5, human Ad6, and the chimpanzee adenovirus 3 (ChAd3), which have been used extensively in clinical trials and returned good safety and strong immunological potency in humans (30, 31). In developing this approach toward an RSV vaccine in humans, homologous and heterologous combinations of PanAd3-RSV, including IN vaccination route, and MVA-RSV were tested in preclinical models. The genetic vaccines elicited RSV-specific neutralizing antibodies and T cell immunity in nonhuman primates and protective efficacy in challenge experiments in rodents with human RSV and in young seronegative calves with bovine RSV (32, 33). Of critical importance in both rodent and bovine challenge models was the absence of immunopathology associated with ERD after vaccination, with the calf model acting as a translational model for the development of a vaccine for the pediatric population. All regimens fully protected the lower respiratory tract from bovine RSV infection in the calf, and heterologous combinations resulted in sterilizing immunity in both upper and lower respiratory tracts (33). Here, we report the translation of this preclinical research into a first-in-man clinical trial in healthy adult volunteers to test the safety and immunogenicity of these vaccine candidates administered in four different prime/boost combinations, including IN delivery.


Forty-two healthy adult volunteers were selected for testing different prime/boost combinations of vaccine in an open-label, dose escalation study design

The vaccination schedules that defined each study group and the baseline physical and demographic characteristics of volunteers within each group are shown in Table 1 and fig. S1. In each experimental group, the first two enrolled volunteers received a lower dose of PanAd3-RSV [5 × 109 viral particles (vp)] and MVA-RSV [1 × 107 plaque-forming units (pfu)]. The remaining volunteers received a target dose of each vaccine, which was a 10-fold higher dose of PanAd3-RSV (5 × 1010 vp) and MVA-RSV (1 × 108 pfu).

Table 1. Definition of each study group by prime/boost vaccine combination and the baseline physical characteristics of volunteers enrolled into each group.

Prime vaccines were delivered by IM or IN spray, and all boost vaccines were delivered by IM injection. Recorded details include the age at enrollment in years and the body mass index (BMI). A CONSORT (Consolidated Standards of Reporting Trials) flow diagram from recruitment to completion of the trial and further information on the study population are available in the Supplementary Materials.

View this table:

Study volunteers were self-selected members of the public responding to recruitment material that invited an expression of interest to take part in the trial. We received 374 expressions, and 101 were potentially eligible and invited for face-to-face physician screening. From these, 40 eligible volunteers were recruited to the study according to protocol-defined inclusion and exclusion criteria (see table S1 and fig. S2). Two volunteers withdrew after receiving the prime dose vaccination for reasons unrelated to the vaccine and were replaced as per protocol, resulting in a total number of 42 volunteers enrolled into the trial. A total of 82 doses of vaccine were administered according to the protocol-defined groups, and 418 of 433 (96.5%) scheduled visits were attended within the protocol-defined windows after vaccination.

PanAd3-RSV and MVA-RSV appear safe in healthy adult volunteers

There were two severe adverse events, each considered unrelated to the vaccine, and are described in table S2. Overall, 18,406 of 19,027 (96.7%) of all expected safety data points were collected for analysis. Common adverse events were local site reactions typical to vaccines given by intramuscular (IM) injection. These events were self-limiting and generally mild to moderate in severity (Fig. 1). Only a few volunteers reported one or more solicited adverse event that lasted more than 1 week after vaccination, and all adverse events reached full resolution. IM MVA-RSV caused a greater frequency, severity, and duration of pain and other local reactions relative to IM PanAd3-RSV. There were two recorded fevers. One was from a volunteer 3 days after IN PanAd3-RSV prime and was concurrent with an influenza-like illness that developed after vaccination and the detection of rhinovirus on nasal sample polymerase chain reaction (PCR). The second fever occurred after IM PanAd3-RSV boost. There were no cumulative adverse events with repeated doses of IM PanAd3-RSV for group 2 volunteers. Volunteers who received IN PanAd3-RSV reported very few adverse events within 1 week of vaccination.

Fig. 1. Frequency of the maximum severity solicited adverse event, oral temperature, and size of local injection site reactions within 1 week of vaccination.

The number of volunteers is represented across the x axis without distinction between low-dose and target-dose recipients; n = 10 or 11 for events after prime and n = 10 for events after boost because of withdrawals. Volunteers reported subjective symptoms as none, mild (does not interfere with routine activities), moderate (interferes with routine activities), and severe (unable to perform routine activities). Redness, swelling, and induration at the site of injection used the maximal recorded diameter of any reaction for objective severity grading. Redness and induration were graded as none (0 to 2 mm), mild (3 to 50 mm), moderate (51 to 100 mm), and severe (≥101 mm). Swelling was graded as none (no visible reaction), mild (1 to 20 mm), moderate (21 to 50 mm), and severe (≥51 mm). Fever was graded as none (≤37.6°C), mild (37.6° to 38.0°C), moderate (38.1° to 39.0°C), and severe (≥39.1°C). Overall, 5587 of 5593 (99.9%) expected data points for solicited adverse events within 1 week after vaccination were collected for analysis. The only missing data were for temperature recordings. Sore throat reactions were not a solicited symptom, although they occurred as an unsolicited event in 5 of 21 IN primed volunteers. Tabular data is available in the Supplementary Materials, table S4.

Unsolicited adverse event reporting identified that 5 of 21 recipients of IN PanAd3-RSV suffered short, mild, and self-limiting sore throat reactions within 1 week of vaccination. No sore throat reactions were reported, or were required to be reported, after IM prime or boost. Nasal samples collected 3 days after IN PanAd3-RSV did not detect any shedding of vaccine virus. Adverse events detected from safety blood sampling and visit observations were generally mild, and a drop in hemoglobin was detected in volunteers from all groups after vaccination, likely related to regular phlebotomy. Two transient drops within 1 week of boost vaccination (one after IM PanAd3-RSV and the other after IM MVA-RSV) were clinically significant and possibly related to the vaccine because there was a concurrent drop in other hematological indices. All measurements returned to normal at the next sample collection 1 week later (table S3). A vaccine-related, clinically nonsignificant, and momentary rise in C-reactive protein above 10 mg/liter was noted 3 days after IM vaccination in three volunteers after IM PanAd3-RSV prime (maximum, 48.4), nine volunteers after MVA-RSV boost (maximum, 56.8), and one volunteer after IM PanAd3-RSV boost (maximum, 11.2).

Influenza-like illnesses reported shortly after vaccination were not due to RSV infection

Unscheduled visits were provided for volunteers reporting influenza-like illnesses throughout the study to identify RSV infections that could potentially affect the measured immune responses. Nasal swabs were collected and analyzed by PCR for respiratory viral infections. A total of 30 unscheduled visits were performed. Eleven of these unscheduled visits were within 5 weeks after either the prime or boost vaccine, and 9 of 11 tested PCR-positive for rhinovirus, with the other two cases failing to identify any pathogen. Overall, 9 of 42 (21%) healthy adult volunteers suffered a symptomatic rhinovirus infection within 1 month of receiving one of the vaccines, with no increased risk attributed to a particular vaccine or route of immunization (three after IN PanAd3-RSV prime, two after IM PanAd3-RSV prime, three after IM MVA-RSV boost, and one after IM PanAd3-RSV boost). Three nasal swab samples were positive for RSV by PCR at different times after vaccination and during the transmission season (two volunteers from group 1, one volunteer from group 3). The first RSV detection was 10 weeks after boost, and in all cases, RSV detection was concomitant with the detection of other respiratory viruses. The only immunological assay performed after the detection of these RSV infections was the final assessment of antibody-mediated RSV neutralization titers at week 34.

Serum RSV neutralizing antibody titers increased in response to vaccination with PanAd3-RSV and MVA-RSV

A plaque reduction neutralization assay (PRNA) was used to investigate the effect of vaccination on the functional antibody response to RSV (Fig. 2 and table S5). Baseline titers are representative of the background immune response to RSV from cumulative natural exposure in young healthy adults, with a geometric mean titer (GMT) for all 42 volunteers of 3191 [95% confidence interval (CI), 2415 to 4217]. Anti-RSV GMTs rose in response to IM PanAd3-RSV prime from 2771 (95% CI, 2199 to 3495) at baseline and peaked at 4817 (95% CI, 3731 to 6218) for groups 1 and 2 at both doses combined 4 weeks later. In contrast, serum anti-RSV GMTs remained indistinguishable from baseline levels after IN PanAd3-RSV prime, 2801 (95% CI, 2094 to 3747) at baseline and 2547 (95% CI, 1801 to 3602) 8 weeks later.

Fig. 2. The RSV neutralizing antibody in response to vaccination.

Data for all volunteers at both doses of vaccine, summarized as the geometric mean titer with 95% CI for each study group [group 1 (•), group 2 (•), group 3 (•), and group 4 (•)]. Responses after prime (P) are grouped by the route of PanAd3-RSV administration [IM (•) or IN (•)]. At week 4, the IM PanAd3-RSV boost (B) was administered to group 2 (•). At week 8, IM MVA-RSV and IM PanAd3-RSV boost (B) vaccines were given to the remaining volunteers. The results of individual volunteers are presented in the Supplementary Materials. The antibody titer 4 weeks after IM PanAd3-RSV prime was significantly elevated from baseline (P < 0.001, paired t test), but not after the IN route (P = 0.816, paired t test). For volunteers who received IM prime, the titers 4 weeks after boost were not statistically significant from pre-boost titers (group 1 between week 8 and week 12, P = 0.152; group 2 between week 4 and 8, P = 0.872; paired t tests). Final measures of serum neutralizing antibody titers were statistically indistinguishable from baseline in all groups (group 1, P = 0.316; group 2, P = 0.416; group 3, P = 0.587; group 4, P = 0.152; paired t tests).

Serum GMTs from IM primed volunteers were observed to be lower at later recorded time points, indicating a waning of antibody titers toward baseline by weeks 30 and 34 despite the administration of a booster vaccine at week 4 (for group 2) or week 8 (for group 1). For volunteers primed with IN PanAd3-RSV, we observed a 4-week lag after IM boost before anti-RSV GMTs rose. Titers rose from 2540 (95% CI, 1506 to 4281) to 3562 (95% CI, 1718 to 7388) after IM MVA-RSV boost (group 3) and from 3015 (95% CI, 1824 to 4984) to 4196 (95% CI, 2936 to 5998) after IM PanAd3-RSV boost (group 4). Subsequent time points recorded a waning of neutralizing antibody titers toward baseline levels at or before week 34. In all study groups, there did not appear to be any relation between the dose of vaccine and the magnitude of the immune response (fig. S3). Volunteers with RSV-confirmed infections were not excluded from the final measure of neutralizing antibody titers at week 34, which was performed 8 and 16 weeks after RSV detection for two of three cases and on the same day as the detection for the final case. For these individuals, there was no increase in titers from the last recorded titer.

Antibody-secreting cells appear in blood in response to PanAd3-RSV and MVA-RSV

To investigate cellular immunogenicity of the vaccines in each group and the antibody response in more detail, we proceeded to measure the anti-F specific immunoglobulin G (IgG) and IgA plasma B cell [antibody-secreting cell (ASC)] response 7 days after vaccination using a dual IgG and IgA ex vivo enzyme-linked immunospot (ELISpot) assay (Fig. 3 and table S5). Detectable responses were found in 1 of 26 and 5 of 32 volunteers for anti-F IgG and IgA, respectively, at baseline. One week after IM PanAd3-RSV prime, we could detect circulating anti-F IgG ASCs in 19 of 19 volunteers and anti-F IgA ASCs in 18 of 20 volunteers with a median of 92 and 31 spots per million peripheral blood mononuclear cells (PBMCs), respectively. Fewer volunteers made detectable responses to IN PanAd3-RSV prime (8 of 17 and 8 of 18 for anti-F IgG and IgA ASCs, respectively) and these were of a lower magnitude (median of 4 and 5 spots per million PBMCs for IgG and IgA, respectively).

Fig. 3. Ex vivo B cell (ASC) response to vaccination.

Fresh PBMCs were collected for analysis at baseline and 1 week after prime and 1 week after boost vaccinations and subjected to a dual-color ELISpot. The responses are represented by a scatterplot of ASC spots per million PBMCs after HSA background subtraction. (A and B) The anti-F specific IgG ASC response (A) and the anti-F specific IgA ASC response (B) to vaccination. The greatest ASC responses were detected after administration of the first IM vaccine. Overall, 13 of 50 (26%) plates for anti-F IgG and 10 of 50 (20%) plates for anti-F IgA were rejected because of contamination or laboratory error. A total of 90 of 214 (73%) and 95 of 124 (77%) data points were available for the analysis of anti-F IgG and anti-F IgA ASC responses, respectively. Study groups: group 1 (), group 2 (), group 3 (•), and group 4 (•). Combined groups: by route of PanAd3-RSV prime administration, IM (•) and IN (•).

The hierarchy of responses to prime was reversed for the boost response. Volunteers primed with IN PanAd3-RSV after boost made comparable responses to IM PanAd3-RSV–primed volunteers irrespective of whether the boost was IM MVA-RSV or IM PanAd3-RSV. At 7 days after boost, responses could be detected in 9 of 9 volunteers, with a median of 217 and 27 spots per million PBMCs for anti-F IgG and IgA ASCs, respectively, in group 3 (IM MVA-RSV), and in 9 of 9 and 8 of 9 volunteers, with a median of 109 and 27 spots per million PBMCs for anti-F IgG and IgA ASCs, respectively, in group 4 (IM PanAd3-RSV). In contrast, responses could be detected in 4 of 5 and 3 of 5 volunteers, with a median of 38 and 5 spots per million PBMCs for anti-F IgG and IgA ASCs, respectively, in group 1 after IM MVA-RSV boost, and in 2 of 9 and 0 of 9 volunteers, with a median of 0 spots per million PBMCs, in group 2 after IM PanAd3-RSV boost (fig. S4).

PanAd3-RSV and MVA-RSV expand interferon-γ T cell responses in healthy volunteers

To further characterize the cellular immune response to vaccination, we examined the interferon-γ (IFN-γ) T cell response before and after vaccination using an ex vivo ELISpot assay (Fig. 4 and table S5). This assay used four peptide pools encompassing the whole amino acid sequence of the vaccine antigen. At baseline, PBMCs from 19 of 33 (58%) volunteers had detectable responses to at least one peptide pool. Analysis of fresh PBMCs collected 2 weeks after prime recorded that the frequency of subjects showing RSV-specific IFN-γ T cell responses had increased to 16 of 18 (89%) and 16 of 19 (85%) after IM and IN PanAd3-RSV, respectively (fig. S5). Consistent with the after prime ASC response, the magnitude of the IFN-γ T cell response was greater 2 weeks after IM prime (geometric mean of 306 spots per million PBMCs; 95% CI, 199 to 471) than after IN prime (geometric mean of 123 spots per million PBMCs; 95% CI, 68 to 224).

Fig. 4. The ex vivo T cell IFN-γ response to vaccination.

Fresh PBMCs were collected for ex vivo IFN-γ ELISpot analysis at baseline, 2 weeks after prime and before boost, and 1 week after boost. Cells were stimulated overnight with peptide pools Fa, Fb, M, and N being representative of the vaccine antigens. (A) The results for each group, presented as a scatterplot of the summed response for each volunteer [(Fa + Fb + M + N) − (4 × DMSO)]. The red line denotes the geometric mean. (B) Individual responses to the separate peptide pools linked between before vaccination and after prime (P) and boost (B). Empty circles denote volunteers who received the lower dose (n = 2 per group). Overall, 13 of 68 (19%) plates failed because of contamination or laboratory error, resulting in the loss of 30 of 163 (18%) samples. There was a disproportionate loss of group 2 pre-boost samples. A further five peptide responses from three volunteers were rejected with a triplicate variance greater than 10. Study groups: group 1 (•), group 2 (•), group 3 (•), and group 4 (•). Combined groups: by route of PanAd3-RSV prime administration, IM (•) and IN (•).

IFN-γ T cell responses were comparable to prevaccination responses before boost at week 4 (group 2) and week 8 (groups 1, 3, and 4). The heterologous PanAd3-RSV prime/MVA-RSV boost generated the highest magnitude and breadth of RSV-specific T cell responses with a 10-fold increase over the baseline recorded 1 week after boost, irrespective of the route of prime. The geometric mean responses reached 1643 (95% CI, 1152 to 2344) and 1588 (95% CI, 1077 to 2342) spots per million PBMCs in groups 1 and 3, respectively. The response to IM PanAd3-RSV boost was 598 (95% CI, 437 to 820) and 400 (95% CI, 211 to 758) spots per million PBMCs in groups 2 and 4, respectively. The kinetics of individual responses within each study group showed no effect of vaccine dose on the magnitude of responses after prime or boost vaccination, and IFN-γ T cell responses after boost were distributed over the peptide pools covering F, N, and M2-1 vaccine antigens. Most of the T cell responses were directed to the F protein, possibly because of the larger size of this antigen (524 amino acids) as compared to N (391 amino acids) and M2-1 (256 amino acids) (fig. S6).

PanAd3-RSV and MVA-RSV expand both CD4+ and CD8+ T cell responses

Next, we tested the functional phenotype of vaccine-induced T cells by intracellular cytokine staining (ICS) and fluorescence-activated cell sorting (FACS) analysis on frozen PBMCs stimulated with the same peptide pools used in the IFN-γ ELISpot (see Fig. 5 and fig. S7). In keeping with the ELISpot data, we observed low levels of CD8+ and CD4+ T cell responses at baseline. There was a small but consistent increase in IFN-γ–producing CD4+ T cells, of a magnitude consistent with the ELISpot data, and responses to all peptide pools. This was seen most clearly in groups 1 and 3 that were boosted with IM MVA-RSV. Similar data (in terms of magnitude, breadth, and group responses) were observed on analysis of CD8+ T cells, indicating that the IFN-γ responses were balanced between CD4+ and CD8+ T cells. Similar data were obtained for the CD4+ and CD8+ IFN-γ–secreting cell populations by analysis of total response pooled across the four proteins. Using a threshold for positivity of 0.02% to detect a T cell response (34, 35), we observed an increase in the number of CD4+ and CD8+ IFN-γ responses, again most obviously in groups 1 and 3 (fig. S8). In terms of TH2 responses, we did not observe any responses against these peptide pools at either time point by parallel analyses of interleukin-5 (IL-5) secretion, although reactivity from positive control stimuli was observed (see the Supplementary Materials, figs. S9 to S11).

Fig. 5. CD4+ and CD8+ IFN-γ responses at baseline and 1 week after boost by ICS.

Empty circles are low-dose vaccine recipients (n = 2 per group). Within each group, the baseline response (left) is matched with the response 1 week after boost (right, at week 5 for group 2 and week 9 for the other groups). Overall, the responses to Fa and Fb peptide pools were greater, with similar responses to N and fewer responses to M. The overall CD4+ and CD8+ responses were greatest after MVA-RSV boost compared to baseline. Study groups: group 1 (•), group 2 (•), group 3 (•), and group 4 (•). The frequency of responses is presented in fig. S8.

To further explore the phenotype of vaccine-induced immune responses, we measured the production of cytokines by a cytometric bead array (CBA) using supernatants from the IFN-γ ELISpot of three volunteers in each target-dose study group at baseline and 1 week after boost. No detectable IL-2, IL-4, IL-10, and tumor necrosis factor–α (TNFα) responses were observed above background, and IL-6 and IL-17 production was detected mostly in response to N peptide pool stimulation with no consistency in changes across baseline or after vaccination (fig. S12).

Anti-PanAd3 neutralizing antibody titers were detectable at baseline and were efficiently boosted by IM PanAd3-RSV

At baseline, 29 of 40 (73%) volunteers had detectable neutralizing antibody titers to the PanAd3 adenoviral vector (Fig. 6), and a higher proportion of volunteers than expected, 15 of 40 (38%), recorded a titer >200 (29). There was no correlation between preexisting vector antibody titers and volunteer age (fig. S13). Baseline titers were higher in volunteers allocated to receive prime with the IN spray. After administration of the prime vaccine, it was evident that IM PanAd3-RSV induced a significant rise in circulating PanAd3 neutralizing antibody where IN PanAd3-RSV did not. There appeared to be no correlation between the anti-PanAd3 titers before vaccination and any of the measured immunological responses after vaccination (fig. S14).

Fig. 6. Vector neutralizing antibody (anti-PanAd3) titers before prime and before boost vaccination.

Anti-PanAd3 titers were measured for the 40 volunteers who completed the trial. No prescreening of anti-PanAd3 titers was performed before enrollment and study group allocation. (A) Scatterplot of the anti-PanAd3 titer from before prime (baseline) and before boost vaccine. The lower limit of detection for the assay was a titer of 18, and titers ≤18 were arbitrarily assigned a titer of 9. (B) Fold change in anti-PanAd3 neutralizing antibody after IM and IN PanAd3-RSV prime. The red bar denotes the geometric mean. Study groups: group 1 (•), group 2 (•), group 3 (•), and group 4 (•). Combined groups: by route of PanAd3-RSV prime administration, IM (•) and IN (•).


Development of an effective RSV vaccine remains a high public health priority for infants, the elderly, and immunocompromised adults. However, an incomplete understanding of the immune correlates of protection and concerns over vaccine immunopathology have slowed progress toward meeting this need. Here, we report the first human data on the safety and immunogenicity of viral vectored vaccines expressing RSV proteins.

PanAd3-RSV and MVA-RSV appeared to be safe in this small population of healthy adults. Local adverse events at sites of vaccine injection were common, especially after MVA-RSV, consistent with the experience from other adenovirus- and MVA-vectored vaccine preparations (31). IN PanAd3-RSV caused a proportion of volunteers to develop a mild, self-limiting sore throat shortly after vaccination. The reason for this is unknown (it was not reported or required to be reported after vaccination) and may be a nonspecific effect of the mucosal immune response. None of the nasal samples from volunteers contained detectable vaccine virus at day 3 after IN PanAd3-RSV vaccination. The immunobiology of RSV is complex and can be both protective and harmful as observed after FI-RSV vaccination. In humans, FI-RSV immunopathology featured nonneutralizing antibody and a mononuclear cellular lung infiltrate in postmortem lung histology (20, 21, 36). The human cytokine response after FI-RSV has not been characterized but is inferred from animal challenge data. Similar pulmonary immunopathology was observed to RSV G glycoprotein candidate vaccines in the mouse challenge model (37). Although the mechanisms of disease are different for FI-RSV and G glycoprotein vaccine ERD, the mouse model was characterized by the involvement of CD4+ cells (with a loss of CD8+ cell inhibition) and the production of IL-4, IL-5, and IL-13 cytokines (3740). The absence of ERD-associated patterns of response to vaccination in the respective ICS and CBA assays in the current study, together with the preclinical results obtained with PanAd3-RSV and MVA-RSV vaccination followed by RSV challenge, provides important support for the further development of this approach. It is important to note that the assays presented here were performed after vaccination and not after natural RSV exposure, and our clinical study population had previous exposure to RSV, which was a significant factor that conferred protection against ERD in older infants after FI-RSV (41). The absence of detectable TH2 cytokine responses in a small number of volunteers after boost, although encouraging, does not fully discount the induction of undesirable vaccine immunogenicity in other populations after vaccination followed by natural infection. This will require continued evaluation in the elderly and throughout age de-escalating development of these vaccines in seropositive children and infants toward RSV-naïve infants.

RSV disease is propagated within the host by the release of virus (targets of neutralizing antibody) from infected cells and by cell-to-cell transfer (targets of cellular immunity). Safe and optimal vaccine-induced protection from severe disease might therefore require the induction of desirable humoral and cellular RSV-specific immune responses. Serum neutralizing antibodies alone constitute a significant obstacle for RSV and mitigate the risk of developing severe disease. Passive immune prophylaxis (palivizumab) for high-risk infants can reduce hospitalization by 45 to 55% (42, 43), and these circumstances demonstrate how RSV F protein–specific antibodies alone can sometimes be sufficient to confer protection from severe disease. Although adult sera contains high titers of naturally acquired F- and G-specific neutralizing antibody, as measured at baseline, IM administered PanAd3-RSV and MVA-RSV were able to induce up to a twofold rise in neutralizing titer after vaccination. A similar fold change in neutralizing antibody titers was observed after a promising RSV nanoparticle vaccine, now entering late-stage clinical evaluation as a maternal vaccine candidate (44). High RSV-specific neutralizing antibody titers from natural exposure can persist into later life (45), though, according to some reports, lower titers are associated with the development of severe disease in the elderly (4648). Vaccine-induced F-specific antibodies might therefore fulfill functional and biologically relevant roles in protection against RSV in infants and the elderly. The neutralizing antibody response wanes from a few weeks after the first IM vaccination, but the relevance of this for a pediatric seronegative population cannot be anticipated. Further preclinical studies in calves could model the longevity of vaccine-induced immune responses in naïve populations. A clear signal of vaccine takes after IM PanAd3-RSV, and later, IM MVA-RSV was observed in all subjects by the detection of circulating anti-F IgG and IgA ASCs, which were undetectable at baseline. The mean anti-F IgG ASC response 7 days after IM PanAd3-RSV prime was 149 spots per million PBMCs (±SD 136.1), which compares with postinfection clinical data in elderly adults who, about 7 days into symptomatic RSV infection, recorded a mean anti-F IgG ASC response of 200 spots per million PBMCs (±SD 256) (49). The absence of detectable ASC responses 7 days after boost in IM primed volunteers, and not IN primed volunteers, may be an effect of vector neutralizing antibody or may imply differences in the kinetics and magnitude of ASC responses after primary and secondary immunizations. In other vaccine trials, the peak ASC response after a booster vaccine appeared around day 7 for rabies, capsular group C meningococcal, pneumococcal, and oral cholera vaccines (5053). Furthermore, where a second influenza dose is not indicated in healthy adults and the elderly, it remains a requirement for naïve infants for an effective immune response, indicating that favorable and demonstrable immunogenicity after two IM injections remains possible in seronegative infants when it was not observed in an adult population.

In vivo viral neutralization by antibody is supported by cellular immune components to protect from severe RSV disease. T cell deficiencies in infants and T cell immune senescence in the elderly confer a substantial risk of developing severe disease (45, 46, 54, 55). It seems likely that an optimal vaccine for the elderly should also reestablish T cell effectiveness, and a heterologous prime/boost regimen may be preferred for optimal restoration of RSV-specific immunity. For other vaccine antigens, the heterologous prime-boost vector combination induces strong antigen-specific T cell responses (30), and consistent with these observations, we showed that PanAd3-RSV prime/MVA-RSV boost was able to induce robust RSV-specific T cell responses independent of the route of priming. This indicates additional immunogenicity from a boost vaccine, within the context of a previous vaccine “prime” in naturally primed healthy adults, and is of potential value in considering the development of prime/boost combinations for seropositive children and the elderly where impaired T cell responses have been associated with severity of disease. We did not seek to identify de novo priming of naïve T cells after prime or after boost, but in several volunteers, who later developed clear responses after vaccination, we were not able to detect T cell reactivity to RSV peptide pools at baseline. Phenotypic analysis of the RSV-specific T cells showed that baseline CD4+ responses were generally below 0.1%, which was within the range for healthy adults reported elsewhere (0.05 and 0.3%) (56), and that baseline CD8+ T cell responses were similar to those against human leukocyte antigen B7 (HLA-B7)–restricted epitopes to RSV nucleoprotein (57). Analysis of the cell population frequencies 1 week after PanAd3-RSV or MVA-RSV boost showed an expansion of CD4+ and CD8+ T cells producing IFN-γ. A similar RSV-specific TH1-biased CD4+ T cell response was observed after immunization of PanAd3-RSV prime/MVA-RSV boost in relevant preclinical models such as nonhuman primates and seronegative calves (32, 33). The identification of RSV epitopes for CD4+ and CD8+ cells remains an active area of investigation (58), but in the context of influenza, a positive correlation was reported between the frequency of memory IFN-γ CD4+ cells and protection from clinical symptoms (59).

All our immunology end points indicate that the IN prime did not induce detectable immune responses in peripheral blood for many volunteers. However, responses measured in blood are not representative of the overall immune response to vaccination or infection. The vaccination regimens selected for clinical evaluation were determined from challenge experiments in rodents and seronegative calves, where they appeared safe and capable of inducing sterilizing immunity without FI-RSV–associated pulmonary pathology (32, 33). The regimens based on IN vaccination showed improved protective efficacy in the upper respiratory tract and were therefore included in the human trial. Preclinical experiments in mice demonstrated that the PanAd3 vector, used as an influenza vaccine candidate delivered IN, induced greater IgG antibody responses in bronchoalveolar lavage samples and greater CD8+ IFN-γ T cell responses in the lungs compared to the same vaccine given by IM injection, which generated greater responses in the spleen (60). When applied as a vector for RSV antigen, IN PanAd3-RSV induced lower levels of systemic RSV-specific T cells than did IM PanAd3-RSV, but IN prime generated comparable levels of immunity in the lung (32). Selective accumulation of memory T cells in the lung against respiratory pathogens like RSV has been described in humans (61), and pulmonary T cells may have contributed substantially to protection afforded by IN PanAd3-RSV in animal challenge studies. The only licensed IN vaccine, the live attenuated influenza vaccine (LAIV, FluMist), protects despite its limited ability to induce humoral and cellular effector responses in blood compared to the injectable trivalent inactivated vaccine (TIV) (62, 63), signifying a crucial role for local mucosal immunity in protection from respiratory pathogens. Recent data from healthy adults infected with RSV under experimental challenge conditions implicated an important role in RSV-specific IgA mucosal immune responses (as well as IgA B cell memory immunity) and a putative correlate for protection against infection (64). Although one obstacle facing IN live attenuated RSV vaccine candidates has been their limited capacity to induce serum antibody in seronegative infants, these infants were able to significantly restrict, by 100-fold, the replication of a “second challenge” subsequent vaccines dose (25). Therefore, IN PanAd3-RSV prime may have generated desirable immune responses at mucosal sites. The mucosal immune response to IN PanAd3-RSV in humans remains potentially important for protection against infection of the upper respiratory tract and supports the use of an IM boost combination (PanAd3-RSV or MVA-RSV), even in naturally primed vaccinees, to induce RSV-specific immune responses in blood for the purpose of supporting protection from severe disease in the lower respiratory tract.

An alternative consideration is that IFN-γ T cell responses to IM MVA-RSV, in the case of group 4, were independent of the vaccine priming route in naturally exposed adult volunteers. Previous use of single-dose MVA-vectored vaccines, or use of these vectors for priming, has not proved sufficiently immunogenic in formulations developed for malaria, tuberculosis, or hepatitis C virus (HCV) because of the presumed need for effective priming by an alternative vaccine or natural exposure (31, 65, 66). In the context of RSV, this opens the possibility of IM MVA-RSV acting as a single-dose vaccine candidate for populations with past exposure to RSV, and this is currently being explored in an extension to the clinical trial in healthy adults aged 60 to 75 years (, NCT01805921). These data from healthy younger adults support the evaluation of prime/boost combinations in older adults, and the trial extension also includes the combinations used in group 1 (IM PanAd3-RSV/MVA-RSV, 8 weeks apart), group 2 (IM PanAd3-RSV/IM PanAd3-RSV, 4 weeks apart), and group 3 (IN PanAd3-RSV/IM MVA-RSV, 8 weeks apart.

The potential limiting effect of preexisting and de novo induced anti-vector neutralizing antibody on the magnitude of immune responses to viral vectored vaccines remains an area of active investigation. Anti-Ad5 neutralizing antibody titers >200 were associated with lower immune responses to an Ad5-vectored HIV vaccine candidate (67). The PanAd3 vector was selected on the basis of low prevalence of neutralizing antibodies in humans and potent immunogenicity in animal models. The same criteria were applied to select successful ChAd vectors for HCV, malaria, and Ebola antigens in humans (ChAd63 and ChAd3) (29, 30, 68). Our study volunteers were excluded at screening if they had, at any time, received another adenoviral or MVA-vectored vaccine, and yet the proportion of volunteers with preexisting neutralizing anti-PanAd3 titers >200 exceeded the 3% expected from earlier estimates from the U.S. and European populations (29). The source of these antibodies remains unclear, but the PanAd3 hexon protein results in its classification within subgroup C of human adenovirus (29) and therefore the potential for cross-neutralizing antibodies from other adenovirus exposure. The predictive value of in vitro neutralization assays on adenovirus-vectored vaccination has been a matter of debate (69). Whereas our data do not identify a clear effect from anti-PanAd3 neutralizing antibody on immune responses to vaccination, the group sizes are small and the immune responses we report are from robustly pre-primed individuals. Cumulative natural exposure to RSV produces highly functional antibodies that may mask loss of vaccine immunogenicity caused by vector neutralization. Data from RSV-infected adults showed no correlation between the magnitude of the RSV-specific ASC response, RSV neutralizing antibody titers, and anti-F IgG antibody titers in serum (49). Furthermore, anti-AdHu5 neutralizing antibody had no effect on T cell responses to AdHu5-vectored vaccines (70), implying that the mechanisms of vector antibody interference with vaccine responses may not be readily measured using standard assays. The study was not designed to infer the protective efficacy of these vaccines, and we detected three mild RSV infections (concomitant with other respiratory viruses) that occurred within the RSV transmission season 18, 10, and 26 weeks after the last vaccination. Infection rate estimates in healthy adults by volunteer-reported (and subjective) influenza-like illnesses would not capture very mild or subclinical RSV exposure and the potential that this has to alter RSV-specific immunity after vaccination. There was no nonvaccinated control arm to this study and the sample size is small, which makes it difficult to evaluate the finding of three confirmed RSV infections, though this frequency of infection is consistent with estimates of the annual risk of symptomatic RSV infection in young healthy adults (11, 12). In healthy adults, vaccine safety and protective efficacy could be further explored using controlled human challenge experiments. However, whereas sterilizing immunity may be desirable, this is not achieved with wild-type RSV infection and is therefore a high bar for a vaccine. The key driver for immunization is prevention of death, severe disease, and hospitalization.

This phase 1 study had several limitations inherent in all early-phase studies resulting from the statistical and interpretative constraints of small group sizes, though the data presented here do nevertheless provide useful data supporting the safety and immunogenicity of these RSV vaccines and their further development. PanAd3-RSV and MVA-RSV were tested in an open-label study design, which might have led to adverse event reporting bias by the volunteers. It is not possible in this study to draw conclusions about the protective efficacy of these RSV vaccines, which would require investigation in a much larger sample size and a control group, followed either by natural exposure or deliberate infection using a human challenge model. Furthermore, the effects of previous exposure to RSV and preexisting RSV immunity make extrapolation of these data to vaccine performance in young infants difficult to predict, and future studies in this and older age groups will be required. Despite these limitations, these data do indicate that RSV vaccine candidates PanAd3-RSV and MVA-RSV are safe, well tolerated, and induce desirable immune responses in healthy adult volunteers, which supports the further investigation of this approach. RSV viral vectored vaccines are now under evaluation in healthy older adults aged 60 to 75 years and will enter a pediatric development program later this year (EudraCT number 2014-005333-31).

In conclusion, we report the successful transition of genetic vaccine technology for RSV from preclinical investigation to phase 1 safety and immunogenicity in humans. Vaccine immunogenicity was generated despite the presence of vector neutralizing antibodies and achieved immune responses above the background of immune responses to RSV derived from repeated seasonal exposure. The immune response to vaccination included a rise in serum RSV neutralizing antibody titers and was supported by anti-F IgG- and IgA-secreting B cells and TH1 (IFN-γ)–responsive T cells. The safe and potent immunogenicity of PanAd3-RSV and MVA-RSV observed in this study warrant their further clinical development toward target populations in need of an RSV vaccine, and represent a new and exciting development in over 50 years of RSV vaccine research.


Study design

RSV001 was an open-label, dose escalation, single-site, phase 1 clinical trial in 42 healthy adult volunteers aged 18 to 50 years. The primary objective was to characterize the safety and tolerability of four prime/boost vaccination regiments, as depicted in Table 1. The terms “prime” and “boost” are conventional and used here to indicate the first and second dose of vaccine. The terms are inherited from previous adenoviral and MVA-vectored vaccine research in immunologically naïve subjects, and our population was already primed from repeated natural exposure. Volunteers were assigned to study groups by sequential allocation with the first two volunteers in each study group assigned to receive the low dose of each vaccine. One month after the last low-dose prime, an analysis of the safety data was submitted for approval from the data safety monitoring committee (DSMC) to proceed to low-dose boost vaccinations and to commence target-dose vaccination schedules of the remaining volunteers. Halting rules for dose escalation included whether any volunteer experienced a severe adverse event related to the vaccine, or if two or more volunteers experienced severe adverse events that were clinically significant and had reasonable possibility of being related to the vaccine. Each volunteer was invited to attend 12 visits over 34 weeks for study groups 1, 3, and 4, and 11 visits over 30 weeks for group 2. Vaccinations were performed between May and the first week of November 2013 to minimize the potential for natural subclinical boosting of RSV-specific immune responses in the immediate postvaccination period. Each volunteer was followed up for 6 months after boost, and the last volunteer completed the study in May 2014. The trial was conducted in accordance with the clinical trial protocol, the principles of the Declaration of Helsinki, and the International Conference on Harmonization (ICH) Good Clinical Practices standards.

Clinical trial authorization was granted by the UK Medicines and Healthcare Products Regulatory Agency (ref 35082/0003/001-0001). Ethical approval and amendments were granted by National Ethics Research Service (NRES) Berkshire (ref 13/SC/0023). The trial was performed by the University of Oxford at the Centre for Clinical Vaccinology and Tropical Medicine, Oxford, and was monitored by the Clinical Trial Research Governance department of the University of Oxford.


Each vaccine was a replication-defective, genetically modified organism engineered to deliver the F, N, and M2-1 RSV proteins. The generation of PanAd3-RSV and MVA-RSV and the results of preclinical evaluation are described elsewhere (33). In brief, the genetic insert for both PanAd3 and MVA vectors constituted a single synthetic DNA fragment encoding all three proteins. Upon transfection into a mammalian cell, cleavage of a foot and mouth disease virus 2A region between the F and N and M2-1 released a soluble truncated F protein (devoid of the transmembrane region) into the supernatant while the N and M2-1 proteins remain intracellular. Deletion of the E1 and E4 loci of PanAd3 rendered the vaccine vector replication defective and MVA naturally cannot replicate in mammalian cells. Clinical grade PanAd3-RSV and MVA-RSV vaccine products were manufactured under good manufacturing practice conditions by Advent and Impfstoffwerk Dessau-Tornau. Safety and characterization tests were performed on well-defined stages of the production processes of the two vaccine products and comply with the European Pharmacopoeia. Testing included sterility, endotoxins, residual host cell DNA and proteins, genome sequencing, and extensive screening for extraneous virus contamination. In PanAd3-RSV, the absence of replication-competent adenovirus was verified. The biological activity (potency) of PanAd3-RSV vaccine product was quantified by anti-hexon immune staining that quantitatively measures infectious virus titer, quantitation of total vector particle concentration by quantitative PCR, a calculated determination of total versus infectious particle ratio, and determination of expression of RSV transgene by Western blot. Testing of the MVA-RSV vaccine product included virus titer, identity by PCR, purity by PCR (free of nonrecombinant MVA), and expression of RSV transgene by Western blot. The stability of PanAd3-RSV and MVA-RSV vaccine products, at the recommended storage condition of ≤−60°C, has been followed since product manufacturing and for the entire duration of the clinical trial. Both PanAd3-RSV and MVA-RSV vaccine products met product stability specifications throughout this period. The methods for stability testing included sterility with some characterization tests as part of the stability monitoring plan. Potency-related tests (Western blot, vector particle concentration, and infectious virus titer) were performed at each time point because these parameters are considered to be main indicators of the stability of the vaccine products.

Vaccines were stored cryopreserved at the trial site in monitored −80°C freezers until use. Each vaccine was granted use under Genetically Modified Organisms (Contained Use) Regulations 2000 by the Oxford University Hospitals NHS Trust Genetic Modification Safety Committee (ref GM462.11.64). Doses of vaccine were prepared by diluting the concentrated product with 0.9% sterile saline solution to the required concentration and volume. The low dose and target dose for PanAd3-RSV were 5 × 109 and 5 × 1010 vp, respectively. The low dose and target dose for MVA-RSV were 1 × 107 and 1 × 108 pfu, respectively. PanAd3-RSV was given either by IM injection of 0.5-ml vaccine product to the nondominant deltoid muscle or by IN spray of 0.15-ml volume to each nostril in the sitting position using a syringe attached to an LMA MAD Nasal needle-free drug delivery system (LMA). MVA-RSV was administered by IM injection of 0.5-ml vaccine product to the nondominant deltoid muscle only. All IM injections used the Z technique to avoid tracking of the vaccine through the needle track.

Study participants and eligibility criteria

Male and female participants were self-selected healthy volunteers aged 18 to 50 years responding to open invitation to the trial. Volunteers provided informed consent in writing before any study procedures. Potential volunteers were excluded if they had any history of significant organ or system disease, any known or suspected alteration in immune function (including IgA deficiency and autoimmune disease), previous receipt of a simian adenoviral or MVA-vectored vaccine of any kind, or any other significant disease or disorder that presented potential for risk, could influence the results, or impair the participant’s ability to participate in the study. Further details of eligibility criteria are set out in the Supplementary Materials. Eligible volunteers were considered enrolled at the point of receiving the first dose of vaccine.

Objectives and end-point measures

The primary objective was to investigate the safety and tolerability, and secondary objective was the characterization of immunogenicity, of the replication-defective ChAd and MVA vectors expressing RSV F, N, and M2-1 in healthy adult volunteers. Primary end-point measures were the frequency and severity of solicited and unsolicited local and systemic adverse events within 1 week after vaccination, safety bloods (full blood count and differential, serum renal and liver biochemistry, and C-reactive protein and amylase) and visit observations (pulse, respiratory rate, and blood pressure) obtained at all visits. Adverse events were graded using modified US Food and Drug Administration and Division of AIDS criteria. For volunteers primed with PanAd3-RSV by IN spray, an additional nasal sample was obtained 3 days later to detect vaccine virus shedding. An independent DSMC acting in accordance to a prespecified charter provided safety oversight for the duration of the trial and formal approval before dose escalation and boosting. Immunogenicity assays are detailed below, and sample processing and analyses of all immunology data were performed observer-blinded by use of a randomly generated laboratory identifier. The end points of the clinical trial were prospectively selected.

Sample processing

Blood samples were collected in heparinized tubes (400 μl of heparin per 50 ml of whole blood) for assays that required PBMCs. PBMCs were isolated within 6 hours of sample collection from a 1:1 mix of heparinized blood and R0 (RPMI with penicillin/streptomycin and l-glutamine, stored at 4°C) by density centrifugation through Lymphoprep (Alere). An aliquot of PBMCs was immediately used for fresh ELISpot assays, and the remainder was cryopreserved in Recovery Cell Freezing Medium [10% dimethyl sulfoxide (DMSO) and calf serum, Invitrogen]. Serum samples were obtained by centrifugation of whole blood collected in clotted tubes and then cryopreserved.

PanAd3 vaccine virus shedding assay

Nasal swab samples were collected in viral transport medium (see Materials and Methods) 3 days after IN vaccination and were analyzed by WuXi AppTec Inc. Specimens were inoculated onto monolayers of human embryonic kidney (HEK) 293 and A549 cells to detect both replication-competent and replication-incompetent adenovirus by the presence of virus-induced cytopathic effects (CPEs) [14-day in vitro assay, and immunofluorescence detection (IFA) as confirmatory assay]. Briefly, each test article was thawed at 37° ± 2°C, vortexed, and filtered through 0.8- and 0.2-μm filters. Filtered test articles were inoculated in 24-well plates and 35-mm dishes of HEK-293 and A549 cells and were incubated for 50 to 70 min in a humidified atmosphere of 37°C with 5% CO2. The inoculum was removed, the monolayers were washed with phosphate-buffered saline (PBS), and fresh culture medium was added. Dish cultures were visually monitored for CPEs for 14 days. On day 3 or 4, 24-well plates were processed for IFA using a monoclonal antibody against hexon adenovirus. The nasal swab samples were considered positive if positive immunofluorescence and/or CPEs consistent with viral infection were observed in the test article–inoculated HEK-293 or A549 cultures.

Anti-PanAd3 neutralization assay

PanAd3 neutralizing antibody titers at baseline and before boost were assayed as previously described using a secreted alkaline phosphatase (SEAP) assay (71). Briefly, HEK-293 cells were seeded (3.5 × 104 cells per well) in a 96-well plate for 2 days. SEAP-expressing PanAd3 was preincubated for 1 hour at 37°C alone or with serial dilutions (1:18, 1:72, 1:288, 1:1152, and 1:4608) of heat-inactivated serum from trial volunteers and added to the 95 to 100% confluent HEK-293 cells for 1 hour at 37°C, and the supernatant was then removed and replaced with 10% fetal calf serum (FCS) in Dulbecco’s modified Eagle’s medium. SEAP activity in the supernatant was measured after 22 to 26 hours, using the chemiluminescent substrate (CSPD) from Phospha-Light kit (Tropix) following the manufacturer’s instructions. Light signal output expressed as relative light units was measured 45 min after the addition of the CSPD substrate using a luminometer (EnVision 2102 Multilabel Reader, PerkinElmer). The neutralization titer was defined as the reciprocal of sera dilution required to inhibit SEAP expression by 50%, compared to the SEAP expression of virus infection alone.

Plaque reduction neutralization assay for the detection of neutralizing antibody

RSV strain A2 (50 pfu) was mixed with doubling dilutions of heat-inactivated sera over the range of 1:20 to 1:10,240. This mixture was incubated for 1 hour to facilitate the neutralization reaction before adding to a confluent layer of HEp-2 cells that had been seeded onto 96-well plates at 3 × 104 cells per well. The plates were then incubated for 60 hours at 37°C with 5% CO2 and 95% humidity. Cells were then fixed using cold acetone/methanol (80:20%, v/v), and RSV plaques were detected by immunostaining with 3-amino-9-ethylcarbazole (AEC). The neutralizing titer was defined as the sera dilution at which 50% of plaques survive and was calculated using the Spearman-Karber method.

Dual-color ex vivo ELISpot assay for the detection of anti-F IgG and anti-F IgA ASCs

MultiScreenHTS HA plates (MSHAN4510, Millipore) were coated with F protein antigen (5 μg/ml) (Sino Biological Inc.), human serum albumin (10 μg/ml) (HSA, Sigma), tetanus toxoid protein (5 μg/ml) (Statens Serum Institut), and polyvalent goat anti-human immunoglobulins (10 μg/ml) (Caltag). After being washed, the plates were blocked with 1% skimmed milk for 45 min at 37°C before PBMCs (100 μl per well) in R10 [RPMI, 10% fetal bovine serum, 2 mM l-glutamine, streptomycin (50 μg/ml), and 50 U penicillin] were added at a starting dilution of 2 × 106/ml and incubated overnight at 37°C with 5% CO2 and 95% humidity. After being washed, the plates were developed with anti-human IgG–fluorescein isothiocyanate (FITC) (Sigma) and anti-human IgA-biotin (AbD Serotec). After washing, anti–FITC–alkaline phosphatase (AP) (Sigma) and streptavidin–horseradish peroxidase (AbD Serotec) were added for 30 min at room temperature. After the final washes, the AEC substrate (Sigma) was added for 30 min at room temperature, the plates were then washed with distilled water, and Vector Blue substrate (100 μl per well) (Vector Laboratories) was added for 10 min at room temperature before a final wash in distilled water. The plates were dried overnight and then read using Autoimmun Diagnostika (AID version 5.0), and responses were measured as antigen-specific spots per million PBMCs with the HSA background subtracted. A positive response was defined as any detection of spots above the HSA background.

Ex vivo T cell ELISpot assay for the detection of IFN-γ

The plates were coated overnight with mouse anti-human IFN-γ and clone 1-D1K (Mabtech) in Dulbecco’s PBS. Plates were blocked before the addition of RSV peptide pools (50 μl per well) consisting mainly of 15-mer sequences with 11–amino acid overlaps and covering the sequence of proteins F, N, and M2-1 (JPT Peptide Technologies). The 269 peptides were dissolved in 100% DMSO and arranged in four pools designated as Fa (N-terminal half of the F protein, 64 peptides), Fb (C-terminal half of the F protein, 64 peptides), N (95 peptides), and M (46 peptides). The concentration of the four pools was adjusted to 0.3 mg/ml single peptide in the mixture and was used in the ELISpot assay at a final concentration of 3 μg/ml of each peptide. DMSO (Sigma) was used as a negative control, and cytomegalovirus (CMV) cell lysate, FEC (mixed HLA class I–restricted peptides from flu, Epstein-Barr virus, and CMV), and concanavalin A (Sigma) acted as positive controls. PBMCs (50 μl per well) were added to peptide wells in triplicate at a concentration of 4 × 106/ml and were incubated overnight at 37°C with 5% CO2 and 95% humidity. Detection was done with anti-human IFN-γ, clone 7-B6-1, biotin conjugate (Mabtech), and an anti-biotin AP conjugate (Vector Laboratories) with 5-bromo-4-chloro-3′-indolyphosphate p-toluidine salt/nitro-blue tetrazolium chloride substrate (BCIP/NBT 1-Step solution, Pierce). IFN-γ–producing cells were counted using AID software version 5.0. The mean + 4 SD of the DMSO response from all samples identified a cutoff whereby individual samples with background DMSO values ≥55 spot-forming cells per million PBMCs were excluded from the analysis. Samples were also excluded from the analysis if no spots were detected in any positive control well. Calculation of triplicate well variance was applied as described elsewhere (72), and a threshold of 10 was applied for exclusion from analysis. A response was considered positive when peptide pool responses were (i) >55 spots per million PMBCs and (ii) greater than 3× DMSO background for the individual. A subject was considered a positive responder if reactivity against at least one of the four RSV peptide pools was observed using these criteria.

PBMC intracellular cytokine staining

Frozen PBMCs were thawed in 9 ml of thawing medium (RPMI with 10% FCS, 1% penicillin streptomycin, and glutamine), resuspended in 1 ml of serum-free CTL-Wash (Cellular Technology Ltd.) wash buffer with 100 μl of deoxyribonuclease, and rested in an incubator overnight (5% CO2, 37°C) in 4 ml of R10 before plating at a concentration of 1 × 106 cells per well in a 96-well tissue culture plate. DMSO, Fa, Fb, M, N, FEC, phorbol 12-myrstate 13-acetate (Sigma)/ionomycin (Sigma) and phytohemagglutinin peptides (Sigma) were added at 1 μl to each well with brefeldin A (1:100 dilution of 0.5 μg/ml stock), and the plate was incubated overnight (5% CO2, 37°C). After 1× PBS wash, fluorochrome-conjugated monoclonal surface staining antibodies L/D-APC (allophycocyanin)–Cy7 (Life Technologies Ltd.) in PBS solution were added for 20 min before washing and adding 1% FACS Fix (1 ml of formaldehyde in 36 ml of PBS) for 20 min. Plates were washed with permeabilization buffer (eBioscience Inc.) and resuspended in permeabilization buffer for 20 min. Fluorochrome-conjugated monoclonal antibodies were added in permeabilization buffer solution, which included CD3-efluor450 (Affymetrix, eBioscience Inc.), CD4-APC (BioLegend), CD8-Viogreen (Miltenyi Biotec), IFN-γ–FITC, IL-2–PerCP (peridinin chlorophyll protein)–Cy5.5, TNFα–PE (phycoerythrin)–Cy7, and IL-5–PE. These antibodies were allowed 25 min of incubation with cells before wash, and samples were spun at 1500 rpm for 5 min and resuspended in 200 μl of PBS. FACS was performed using a MACSQuant flow cytometer (Miltenyi Biotec) and analyzed using FlowJo software (version X0.7 for Mac). Responses were background DMSO–subtracted, and a threshold of 0.02% was applied to define a positive T cell response (34, 35).

Cytokine quantification by cytometric bead array

Cytokine quantification was performed with a BD Cytometric Bead Array (CBA) Human TH1/TH2/TH17 Cytokine kit, using supernatants from the ex vivo IFN-γ ELISpot. Thirty-five microliters of the pooled peptide triplicate supernatant from the DMSO, Fa, Fb, M, and N wells was mixed with a 5-μl aliquot of each cytokine capture bead (human IL-2, IL-4, IL-6, IL-10, TNF, IFN-γ, and IL-17A) and 35 μl of detection reagent (PE-conjugated antibody) for 3 hours at room temperature and protected from light. Eight hundred microliters of wash buffer was then added, and each sample was centrifuged at 200g for 5 min. The supernatant was discarded, and the bead pellet was resuspended in 200 μl of wash buffer. Cytokine detection was performed using an LSRII FACS machine (BD), BD FACSDiva software (version 6.0 for Windows), and FlowJo software (version X0.7 for Mac).

Detection of respiratory viral infection by PCR from nasal swabs

Nasal samples were collected using a midturbinate swab and the Copan Universal Transport Medium kit (UTM-RT mini, Copan Diagnostics Inc.) according to the manufacturer’s instructions. Viral diagnostics were performed by PCR for RSV, influenza A, parainfluenza 1/2/3, rhinovirus, coronaviruses, adenovirus, metapenumovirus, enterovirus, parechovirus, bocavirus, and Mycoplasma pneumoniae.


The purpose of the study was to characterize the safety and immunogenicity of different prime/boost combinations of vaccine, and therefore the analyses were descriptive in nature. There was no prespecified hypothesis on which to power the study, and preplanned analyses did not include hypothesis testing. Statistical analyses of the data have thus been kept to a minimum, and results were instead presented as descriptive statistics using graphical presentations. Analyses were based on the intention-to-treat population that included all participants with any data. Appropriate comparative statistics, as annotated in the text, and the generation of P values are post hoc analyses. A two-sided P value of <0.05 was considered statistically significant.

Graphs and analyses were generated using GraphPad Prism version 6.0 for Mac (GraphPad Software), Stata version 13.1 (StataCorp LP), SPSS (Statistical Package for the Social Science) version 21 for Mac (IBM Corporation), and SAS (Statistical Analysis Software) version 9.3 (SAS Institute).


Fig. S1. Ethnicity and baseline physical and demographic characteristics of the 42 enrolled volunteers.

Fig. S2. CONSORT flow diagram for recruitment and completion of the study.

Fig. S3. Kinetics of the serum neutralizing antibody response to vaccination for each individual.

Fig. S4. Supplementary ASC data derived from ELISpot.

Fig. S5. Supplementary T cell IFN-γ data derived from ELISpot.

Fig. S6. The distribution of peptide pool responses before and after vaccination.

Fig. S7. The gating strategy used to quantify immune response from ICS FACS data.

Fig. S8. The number of CD4+ and CD8+ IFN-γ responses after boost.

Fig. S9. CD4+ and CD8+ IL-2 responses before and after vaccination.

Fig. S10. CD4+ and CD8+ TNFα responses before and after vaccination.

Fig. S11. CD4+ and CD8+ IL-5 responses before and after vaccination.

Fig. S12. IL-6 and IL-17 responses before and after vaccination.

Fig. S13. Serum PanAd3 neutralizing antibody titers and volunteer age.

Fig. S14. Correlation plots between anti-PanAd3 neutralizing antibody titers before prime and boost vaccination and the magnitude of subsequent immune responses.

Table S1. Summary of the eligibility criteria for study volunteers.

Table S2. Serious adverse event (SAE) descriptions.

Table S3. Hematological changes within 1 week after boost vaccination; two volunteers with clinically significant concurrent drops in hematological indices.

Table S4. Supplementary safety data.

Table S5. Supplementary data on the RSV-specific immune response to vaccination.


  1. Acknowledgments: We thank the support and active input of the data safety monitoring committee, which was chaired by S. Gordon with S. Faust, S. Paulus, and C. Yap (statistician). We thank V. Ammendola for setting up the shedding assay. M. Voysey provided statistical input to the protocol and data analysis and wrote the statistical analysis plan. We wish to acknowledge the research staff at the Oxford Vaccine Group and thank the many volunteers who were willing to contribute to this research. Funding: This study was supported and sponsored by ReiThera SRL (formerly Okairos SRL), which was acquired by GlaxoSmithKline Biologicals SA during the trial, the NIHR Oxford Biomedical Research, and salary support for C.S. and P.K. (WT 091663MA) from the Wellcome Trust. Author contributions: C.A.G. was the lead physician; K.H. was the lead research nurse; K.S.T. was the trial statistician; C.A.G., E.S., R.C., P.K., A.N., A.J.P., C.T., A.F., S. Colloca, S. Capone, and A.V. designed the study/protocols; C.J.S., A.J.T., C.M.d.L., M.D.S., L.S., and S.D.M. optimized and performed the assays; C.A.G., K.S.T., E.S., C.J.S., S. Capone, A.V., and P.K. performed data analysis; C.A.G., B.A., and A.J.P. provided clinical safety oversight throughout the trial; C.A.G., E.S., A.V., S. Capone, A.N., P.K., and A.J.P. wrote the manuscript; A.J.P. was the chief investigator. All authors had input into the manuscript and have approved the manuscript for publication. Competing interests: A.J.P. has previously conducted clinical trials of vaccines on behalf of Oxford University funded by GlaxoSmithKline Biologicals SA and ReiThera SRL but does not receive any personal payments from them. A.J.P. is the chair of the UK Department of Health’s (DH) Joint Committee on Vaccination and Immunisation (JCVI), but the views expressed in this manuscript do not necessarily represent the views of JCVI or DH. A.V., R.C., and A.N. are named inventors on patent applications covering RSV antigen expression system (WO 2012/089833). The remaining authors declare that they have no competing interests. Data and materials availability: RSV001 was registered with and EudraCT (ref NCT01805921 and 2011-003589-34, respectively).
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article