Respiratory syncytial virus (RSV) causes significant disease burden in older adults. MVA-BN-RSV is a novel poxvirus-vectored vaccine encoding internal and external RSV proteins.
In a phase 2a randomized double-blind, placebo-controlled trial, healthy participants aged 18 to 50 years received MVA-BN-RSV or placebo, then were challenged 4 weeks later with RSV-A Memphis 37b. Viral load was assessed from nasal washes. RSV symptoms were collected. Antibody titers and cellular markers were assessed before and after vaccination and challenge.
This was a phase 2a, randomized, double-blind, placebo-controlled trial to assess safety, immunogenicity, and efficacy of MVA-BN-RSV vaccine against infection with the Memphis 37b strain of respiratory syncytial virus subtype A (RSV-A). The vaccine is based on the MVA vector and genetically engineered to encode the RSV F, G(A), G(B), N, and M2-1 proteins [29–31]. The vaccine was produced at Bavarian Nordic A/S (Kvistgård, Denmark) according to Good Manufacturing Practice standards and used at a nominal titer of 5 × 10E8 infectious units per 0.5 mL dose; placebo was an equal volume of Tris buffered saline solution. RSV-A Memphis 37b at a dose of 4.5 log10 plaque-forming units was utilized as the challenge virus. This virus challenge trial was conducted by hVIVO Services Limited at the Queen Mary BioEnterprises Innovation Centre in London, UK, from January to November 2021. The challenge model has been previously described [34, 35]; for the schedule of trial activities, see the Supplementary Material.
All trial-related procedures were conducted in accordance with Good Clinical Practice and the provisions of the Declaration of Helsinki and were approved by the Office for Research Ethics Committees Northern Ireland before trial initiation.
Subjects provided written informed consent before participating in the trial. They were healthy adults between 18 and 50 years of age expected to be susceptible to RSV based on screening nAb titers in the lowest population quartile for the previous year. For full eligibility criteria, see the Supplementary Material.
Subjects were randomized 1:1 to be vaccinated intramuscularly with MVA-BN-RSV or placebo. Subjects used a memory aid card to collect information on local (pain, erythema, swelling, induration, and pruritus) and systemic (pyrexia, headache, myalgia, chills, nausea, and fatigue) reactions up to 7 days after vaccination. Blood was collected for antibody titers and cellular markers before vaccination, 7 days after vaccination for cellular markers, and 14 days after vaccination for antibody titers.
Subjects were admitted to a quarantine unit 2 days before being inoculated intranasally with challenge virus (approximately 4 weeks after vaccination) and observed there for 12 days following challenge. Subjects rated their experience of 13 symptoms on a diary card twice on the first day of quarantine, once on the last day, and 3 times each day in between. Most symptoms were scored in severity from 0 to 3; shortness of breath and wheezing had a fourth severity category for symptoms at rest. Details of symptom scoring is provided in the Supplementary Material. Nasal washes were obtained twice a day from the second to eleventh day after challenge and once on the final quarantine day. Blood was collected for antibody titers and cellular markers before challenge and 5 and 10 days after challenge.
Subjects were followed after discharge from quarantine. Blood was collected for antibody titers and cellular markers at 4 weeks after challenge and approximately 6 months after vaccination.
Reports of unsolicited adverse events (AEs) were collected throughout the trial, from informed consent to trial end, and followed until resolution.
A validated reverse transcriptase quantitative polymerase chain reaction (qRT-PCR) assay was used both to determine whether RSV-A Memphis 37b could be detected and to measure the amount of challenge virus present (lower limit of quantitation [LLOQ] was defined as a cycle threshold [Ct] value of 3.9, which equated to 2.8 log10 copies/mL) in nasal washes. Results below the LLOQ were set to 0. Nasal washes also were cultured for replicating virus, and the results were measured by plaque assay; the LLOQ was 2.0 log10 plaque-forming units/mL, and results below the LLOQ were set to 0. Refer to Supplementary Material for details.
Serum samples were analyzed by enzyme-linked immunosorbent assay (ELISA) for titers of RSV-specific immunoglobulins A (IgA) and G (IgG) and by plaque reduction neutralization test (PRNT) for neutralizing titers of RSV subtype A- and subtype B-specific antibodies. A double-color enzyme-linked immunospot (ELISpot) was used to enumerate peripheral blood mononuclear cells (PBMCs) producing interferon-γ (IFN-γ) and interleukin 4 (IL-4) in response to stimulating pools representing the surface proteins F, G(A), and G(B); the internal proteins M2-1 and N; and whole RSV. These methods have been previously described [30, 31]. PRNT titers were calibrated to the first World Health Organization International Standard [36].
Incidence by qRT-PCR in each treatment group was calculated as the proportion of subjects infected from day 2 to end of quarantine, according to 3 a priori definitions: (1) infection confirmed by qRT-PCR with RSV detectable (at or above the lower limit of detection [≥LLOD]) in samples from at least 2 consecutive days and symptomatic as evidenced by 1 or more clinical symptoms of grade ≥2; (2) detectable infection with symptoms as above or with 1 or more symptoms of any grade from 2 different categories in the symptoms scoring system; or (3) detectable infection regardless of symptoms. Vaccine efficacy was defined as (1 − incidence ratio) × 100%.
Infection existent in viral culture was similarly defined a priori but based on the presence of quantifiable RSV (≥LLOQ). Vaccine efficacy using culture-confirmed infection was calculated post hoc.
Also post hoc, it was recognized that other recently published human challenge trials [14, 15] included definitions of infection using quantifiable (≥LLOQ) rather than detectable (≥LLOD) qRT-PCR measures, as detectable measures were too sensitive, and respiratory symptoms did not add sufficient specificity to the definitions of infection. Infection definitions and vaccine efficacy calculations therefore were expanded to include corollary definitions based on quantifiable qRT-PCR measures (see section Efficacy Results).
Statistical analyses were performed using SAS 9.4 (SAS-Institute) by Venn Life Sciences. The primary and secondary efficacy analyses were performed in the per protocol population, which included all participants who were vaccinated, challenged, and had nasal washes at least until day 10 of quarantine. The intent-to-treat (challenge) population, which included all vaccinated and challenged participants, was used for supportive analyses on efficacy end points. Safety end points were analyzed on all vaccinated participants (safety population).
Summary statistics were calculated for demographic and baseline characteristics and efficacy and safety end points. For viral load area under the curve (AUC) by qRT-PCR, confirmatory testing of treatment differences was done with a 1-sided Wilcoxon rank sum test. Descriptive treatment comparisons of peak viral load, peak viral culture, sum of total symptom scores, and peak total symptom score were performed with 2-sided Wilcoxon rank sum tests. For the incidence of RSV-A Memphis 37b infection detected by qRT-PCR and quantified by qRT-PCR, and incidence of infection confirmed by virus culture, descriptive treatment comparisons were performed using the Fisher exact test. The Wilson score method was used to calculate confidence intervals for proportions. For safety data, unsolicited adverse events were coded using the Medical Dictionary for Regulatory Activities, version 24.0.
A total of 74 participants were randomized, 36 to receive MVA-BN-RSV and 38 to receive placebo in a blinded fashion. All but one were vaccinated and are included in safety analyses. Ten participants, 5 from each treatment group, did not proceed with RSV-A Memphis 37b inoculation. Of the 63 participants inoculated (intent-to-treat [challenge] analysis set), 1 participant in each group did not have viral load samples collected through day 10 (study withdrawals); therefore 61 participants are included in the per protocol analysis set (see participant disposition in Figure 1). The baseline characteristics (Table 1) and medical history of the participants were similar between the MVA-BN-RSV and placebo groups (median age 26.5 and 25.0 years, female sex 36.1% and 45.9%, white race 91.7% and 89.2%, respectively).
As shown in Table 2, the primary end point of RSV-A Memphis 37b viral load AUC (log10 copies × hour/mL) from nasal washes as determined by qRT-PCR was lower in the MVA-BN-RSV group (median = 0.00, interquartile range 0.00 to 53.44) than in the placebo group (median = 49.05, interquartile range 0.00 to 999.94) (P = .017). Mean viral load over time is shown in Figure 2A; it diverged 2 days after challenge and remained divergent for all of quarantine. Similarly, viral load AUC by quantitative virus culture was lower in the MVA-BN-RSV group (P < .001); Figure 2B shows that this difference occurred primarily from day 3 to day 7. In addition to viral load AUCs, which were measures of disease summed over time, peak viral load measures by qRT-PCR demonstrated less disease acuity in the MVA-BN-RSV group (median = 0.00) than in the placebo group (median = 3.45) (P = .032). Likewise for clinical symptoms, both symptoms summed over time (P = .004; Table 2 and Figure 2C) and peak symptom scores were lower in those vaccinated with MVA-BN-RSV, with symptom scores diverging day 4 through day 9.
Incidence of RSV-A Memphis 37b infection was investigated using several definitions, as described in the methods; the results are presented in Table 2. Under the a priori definitions based upon viral load at or above the qRT-PCR LLOD, vaccine efficacy ranged from a low of 9.6% for infection confirmed by laboratory measure only to 58.7% for infection confirmed both by laboratory measure and by the presence of at least 1 RSV symptom of grade ≥2, and the differences in infection incidence between the treatment groups were not statistically significant. When infection was defined instead by viral load ≥ LLOQ, efficacy ranged from 51.8% to 79.3%, and differences in symptomatic infection were statistically significant. Finally, when defined by virus culture results alone, vaccine efficacy was 88.5% (P = .0125); addition of clinical symptoms to the definition did not improve efficacy.
RSV-specific humoral responses are shown in Figure 3A–D and Supplementary Table 2. In the MVA-BN-RSV group, the largest increases in geometric mean titers (GMTs) from baseline to 2 weeks after vaccination were seen in IgA (264.3 to 994.0) and IgG (815.4 to 3268.2); neither immunoglobulin increased in response to challenge. The placebo group experienced smaller increases in response to challenge, which peaked at 28 days in IgA (202.1 to 561.2) and IgG (737.9 to 1762.7). The MVA-BN-RSV group had higher IgG and IgA GMTs than the placebo group at the end of the study, but both groups remained above baseline. For nAbs against RSV-A and RSV-B, the MVA-BN-RSV group increased from baseline to 2 weeks after vaccination (287.7 to 600.2 IU/mL and 174.8 to 277.8 IU/mL, respectively); they had no response to challenge. The placebo group's response from quarantine admission to postchallenge peaked later (day 28) but at higher levels (276.6 to 751.2 IU/mL and 181.0 to 393.9 IU/mL) compared to the MVA-BN-RSV group. Both groups’ nAb titers were greater than baseline at the end of the study.
Solicited local AEs (Table 3) were reported in 88.9% of participants receiving MVA-BN-RSV and 37.8% of those receiving placebo; the most common local reaction was injection site pain, and 3 participants reported grade 3 pain. The median duration of pain in the MVA-BN-RSV group was 4.0 days; the maximum was 7 days. Solicited systemic AEs were also common but were reported somewhat more frequently than local reactions by the placebo group as well. Grade 3 fatigue was reported by 4 MVA-BN-RSV recipients, 1 of whom also reported grade 3 myalgia, and by 1 placebo recipient. The median duration of all systemic AEs was 1.0 days. Unsolicited AEs in the vaccination phase (within 29 days postvaccination) were reported by a comparable proportion of participants in the MVA-BN-RSV and placebo groups. Most were considered unrelated to study vaccination, although 1 case of fatigue and increased body temperature in the MVA-BN-RSV group and 1 case of pruritic rash in the placebo group were attributed to vaccination. Unsolicited AEs in the postchallenge phase also were reported by similar proportions of the MVA-BN-RSV and placebo groups. No events were considered related to study vaccination. Two SAEs of myocarditis (1 event of moderate severity in the MVA-BN-RSV group that started 10 days after challenge and 1 event of mild severity in the placebo group that started 7 days after challenge; Table 3) were considered probably related to RSV-A Memphis 37b inoculation. Both were diagnosed by a cardiologist from electrocardiogram and cardiac enzyme findings with no or minimal symptoms and were sent to hospital for observation and follow-up. Tests returned to normal, and the subjects completed follow-up.
MVA-BN-RSV vaccination resulted in lower viral load and symptom scores, fewer confirmed infections, and induced humoral and cellular responses.
Respiratory syncytial virus (RSV) remains a common, recurring infectious disease. Infants are understood to be at high risk from RSV; it is estimated that nearly 120 000 children younger than 5 years die of RSV each year globally [1, 2], mostly in the developing world. Older adults are the second highest risk group, with substantial RSV disease burden [3–5]. In high-income countries, an estimated 1.62% of adults ≥60 years of age develop acute RSV infections annually, 0.15% are hospitalized, and approximately 33 000 die of RSV-related causes [4]. Underlying comorbidities likely drive the risk of severe outcomes in older adults [5–7]. Therefore, this population is an important target for preventive measures, as older adults in the developed world may spend decades at risk of both hospitalization and death from RSV.
In the 1960s, research in children with a formalin-inactivated RSV vaccine was halted when it was recognized that those immunized frequently experienced enhanced disease upon later natural infection [8–10]. This slowed the pace of RSV vaccine research for decades, as scientists investigated potential mechanisms for this phenomenon [11]. Modern RSV vaccine research has mostly focused on delivery of RSV proteins, rather than whole virus, either as a subunit vaccine or using a viral vector. In particular, vaccine development efforts have aimed to elicit neutralizing antibodies (nAbs) to the surface fusion (F) protein that promotes syncytium formation in respiratory epithelium. Two candidate vaccines for older adults containing the F protein in its postfusion conformation were unsuccessful in phase 2 and 3 trials [12, 13]. Four other vaccines delivering the F protein in its prefusion conformation (preF) have reported positive results from phase 2 challenge trials [14, 15] and/or from pivotal phase 3 trials [16–18]. None of these vaccines in older adults appears to elicit the imbalanced T helper 2 (Th2)-mediated immune response thought to be involved in RSV vaccine-enhanced disease [19, 20].
MVA-BN-RSV is a novel vaccine aimed at broad immunogenicity, inducing both humoral and cellular responses to multiple RSV proteins. It utilizes the nonreplicating, modified vaccinia Ankara (MVA) virus, which has been widely and safely tested and used as a smallpox/mpox vaccine [21–27] and in recombinant form against Ebola [28]. The MVA-BN-RSV recombinant vaccine encodes not only the F protein (expressed as both pre- and post-F) but also surface glycoproteins from the 2 RSV subtypes, G(A) and G(B), that facilitate viral attachment to airway ciliated epithelial cells and 2 internal proteins (nucleoprotein [N] and transcription elongation factor [M2-1]) [29–31]. The F and G proteins are the main targets of RSV nAbs, but this immune response to natural infection is not durable [32, 33]. The N and M2-1 proteins, highly conserved among different RSV subtypes, were included in the recombinant vaccine to promote cytotoxic T-cell responses.
MVA-BN-RSV may have an advantage over other candidate vaccines that rely on nAbs to a single protein. The vaccine has induced humoral and cellular immune responses in animal models [29] and in early clinical trials [30, 31] without safety concerns. The safety and immunogenicity of MVA-BN-RSV are further tested and efficacy is examined for the first time in this human challenge trial report.
RSV remains a ubiquitous infectious disease, and a vaccine against it has been an elusive goal through much of 60 years of research on the immunology of RSV. Human challenge trials are a unique method in clinical RSV research to study the ability of a vaccine to prevent infection under artificial, controlled exposure conditions with a viral strain that causes mild to moderate upper respiratory disease [34, 35]. Human challenge trial results are not definitive, however, as research must confirm vaccine efficacy under conditions of natural infection and in the intended older adult population. Furthermore, phase 3 trials must investigate the prevention of severe disease, not merely the reduction of viral load and symptom scores, although their correlation with disease severity in natural infection has been demonstrated [37, 38].
In this human challenge trial, vaccination with MVA-BN-RSV clearly resulted in lower viral load AUCs and sums of total symptom scores with similar shapes of the curves (Figure 2). The picture with infection prevention, used to measure vaccine efficacy, was somewhat more complex. The vaccine did not prevent detectable qRT-PCR–confirmed infection alone, as there was little difference between the treatment groups in this regard. However, the importance of detectable RSV in the absence of symptoms or positive culture and whether it represents clinically relevant RSV infection is unclear, as qRT-PCR can detect RNA from nonreplicating virus [39]. When infection was defined as the presence of both quantifiable viral load and clinical symptoms, vaccine efficacy was nearly 80%. When the presence of live, replicating virus by culture was examined, only 1 subject who received MVA-BN-RSV had a positive culture; vaccine efficacy estimates were 85% and 89%. All culture-positive subjects in both groups were symptomatic by one or both definitions, confirming the ability of positive culture to denote infection clinically relevant to the participant, not just to disease transmission. Whether the MVA-BN-RSV vaccine can prevent clinical infection or rather protect against infection developing into severe disease will be determined by the phase 3 trial.
As useful as the human challenge model is, it carries known but small risks. Previous experience shows an expected but low frequency of non–life-threatening myocarditis after challenge. Two cases occurred in this study, at 7 and 10 days after challenge. The case timing, as well as experience with the MVA-BN vector, suggest these cases were more likely related to challenge virus than trial vaccination, and with 1 case per group, no meaningful conclusions can be drawn on potential differences in the case characteristics. In the development program of the MVA-BN vector, which included more than 10 000 volunteers [22–27], only 1 doubtful case of pericarditis [22] was reported. During the mpox vaccination campaigns, inflammatory cardiac disorders were reported at a frequency of <1:100 000 doses, roughly in line with expected background incidence [40]. Also of note, severity and duration of reactogenicity was in line with previous research using MVA-BN and MVA-BN-RSV [22, 41].
Regarding a further challenge inherent in this study design, it is noteworthy that even in the placebo group, less than half of subjects had quantifiable virus by qRT-PCR, and less than a quarter had quantifiable culture results, despite eligibility criteria intended to select for susceptibility to RSV infection. This, along with a small sample size, made it more difficult to detect differences between the treatment groups, and vaccine efficacy confidence intervals are wide.
Scientific understanding of how immune response associates with prevention of either infection or severe disease is limited. Neutralizing antibodies to the F and G proteins are often used as a primary measure of immune response, as higher titers have been associated with reduction in disease [42, 43]. Results from this study provide evidence that protective immune responses to RSV go beyond nAbs, as nAb GMTs following vaccination were lower than observed elsewhere [14, 15], but high vaccine efficacy was still observed. This may be attributable to robust serum IgA and to cellular responses that were observed for all peptide pools, particularly GMSFUs that remained elevated over baseline for M2-1, N, G(A), and G(B).
The MVA-BN-RSV vaccine appears to represent a mode of action broader than other vaccine candidates focused on the production of neutralizing antibodies to the preF protein. Dependence of RSV vaccines on the activity of neutralizing antibodies against a specific epitope of one protein conformation may be risky, as such reliance may provide selective pressure for the development of mutant viruses capable of neutralizing antibody escape [44, 45]. In fact, the monoclonal antibody suptavumab failed in a phase 3 clinical trial to reduce RSV hospitalizations or lower respiratory tract infections in infants, a result attributed to epitope mutations found on circulating RSV-B strains [45]. Having a vaccine that provides multiple targets for both humoral and cellular responses may protect against the consequences of such genetic pressure.
MVA-BN-RSV vaccination resulted in significantly lower viral load AUC by qRT-PCR after challenge with RSV-A Memphis 37b compared to placebo. Vaccination also resulted in fewer infectious virus particles by culture, lower symptom scores, and vaccine efficacy in the range of 79.3% to 88.5% against infection after challenge confirmed by symptoms and quantifiable viral load measures or by positive cultures. Humoral and cellular responses support broad immunogenicity of the vaccine. Injection site pain was the most common adverse event. Phase 3 evaluation of MVA-BN-RSV to determine clinical efficacy in an older adult population has commenced (NCT05238025).
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Register for a free account at Science Publication and you will be able to read the full publication