The immunogenicity and safety of respiratory syncytial virus vaccines in development: A systematic review

Abstract Background Respiratory syncytial virus (RSV) is a leading cause of acute lower respiratory infection globally. There are vaccine candidates in development, but a systematic review on immunogenicity and safety of vaccine is lacking. Methods This systematic review of RSV vaccine clinical trials was undertaken using four databases. Searches were conducted using both controlled vocabulary terms such as “Respiratory Syncytial Virus, Human,” “Respiratory Syncytial Virus Infections,” “Respiratory Syncytial Virus Vaccines,” “Immunization,” “Immunization Programs” and “Vaccines” and corresponding text word terms. The included studies were limited to clinical trials published from January 2000 to 31 December 2020. RSV infection case was defined as RSV‐associated medically attended acute respiratory illness (MAARI) or RSV infection by serologically confirmed test (Western blot) during the RSV surveillance period. We calculated the relative risk of each vaccine trial with RSV infection case. Results Of 6306 publications, 38 were included and data were extracted covering four major types of RSV vaccine candidates, these being live‐attenuated/chimeric (n = 14), recombinant‐vector (n = 6), subunit (n = 12) and nanoparticle vaccines (n = 6). For RSV infection cases, nine trials were involved and none of them showed a vaccine‐related increased MAARI during RSV surveillance season. Conclusion LID ∆M2‐2, MEDI M2‐2, RSVcps2 and LID/∆M2‐2 /1030s (live‐attenuated) were considered the most promising vaccine candidates in infant and children. In the elderly, a nanoparticle F vaccine candidate and Ad26.RSV.preF were considered as two potential effective vaccines. A promising maternal vaccine candidate is still lacking.


| INTRODUC TI ON
Respiratory syncytial virus (RSV) is one of the main causes of acute lower respiratory infection (ALRI) and commonly leads to pneumonia or bronchiolitis. 1 The pattern of RSV infection in humans shows a U-shaped age curve, with peak disease rates in those younger than 5 years and older than 65 years. 2 A recent epidemiological study on children showed an estimated 33.1 million RSV-ALRI episodes globally in 2015, which resulted in about 3.2 million hospitalisations; around 45% of the hospitalised patients were younger than 6 months old. The estimated annual number of deaths was 59 600 in children aged younger than 5 years, with 46% happening in children younger than 6 months. 3 In the elderly, several studies have shown that RSV is an important cause of illness in community-dwelling older people. 4,5 RSV may cause a similar burden of disease to nonpandemic influenza A in older age groups. 6 RSV is annually associated with around 177 000 hospitalisations and 14 000 deaths in US adults aged 65 years or older. 6 In 1955, RSV was first isolated from a chimpanzee with respiratory symptoms and designated chimpanzee coryza agent. RSV is an enveloped RNA virus and belongs to the family of Paramyxoviridae, classified within the genus Pneumovirus, and it can be separated into two major subtypes, A and B. There are four important proteins on the surface of the RSV virion, which are the attachment glycoprotein (G), the fusion (F) protein, the matrix protein (M) and the small hydrophobic (SH) protein. 7 The main human-neutralising antibody is against the F protein which enables the virus to fuse with the membrane of respiratory cells. It is highly conserved and essential for viral viability. However, the RSV virus can make a conformational change to the F protein to avoid antibody neutralisation. In contrast, the G protein focuses on the ciliated cells of the human airway; variation of it is associated with subtype classification. 8 Therefore, both of these two antigens have been targeted by novel vaccine candidates (and also by monoclonal antibodies). The function of M protein is thought to be in interaction with polymerised actin which destabilises cellular microfilaments to transport virion components in the host cells. 9 However, the function of SH protein is not yet clearly known. 10 Adverse events associated with the development of an RSV vaccine in the mid-1960s delayed the development of an RSV vaccine for decades. At that time, a formalin-inactivated (FI) RSV vaccine was being tested for protective efficacy. It failed due to worrying results. A large proportion of the study participants, who were exposed to natural RSV infection soon after vaccine recipients, developed enhanced respiratory disease (ERD) and unfortunately two of these children died. The subsequent investigation found that the FI vaccine did not produce neutralising antibodies and also failed to elicit CD8 + T cells. Instead, it induced an aggressive CD4 + T cell and cytokine response leading to ERD. 11 In 2018, Mazur et al 12 published a narrative review on RSV vaccine development. However, there has been no recent systematic review on this topic. We divided respiratory syncytial virus (RSV) vaccines under development into four major groups: particle-based, vector-based, live-attenuated or chimeric, and subunit vaccines.

| Study objective
This study has two major aims: firstly, to systematically review the medical publications on clinical trials of RSV vaccines from 1 January 2000 to 31 December 2020 and describe immunogenicity and safety data in the published journals; secondly, to evaluate the risk of RSV infection in vaccine recipients during RSV follow-up season.

| Literature searches
The initial search for this systematic review of RSV vaccine clinical trials was undertaken by a medical information specialist (CK) using the following bibliographic databases: Ovid MEDLINE, Ovid EMBASE, the Cochrane Library Database of Systematic Reviews and Cochrane Central Register of Controlled Trials. Searches were conducted using both controlled vocabulary terms such as "Respiratory Syncytial Virus, Human," "Respiratory Syncytial Virus Infections," "Respiratory Syncytial Virus Vaccines," "Immunization," "Immunization Programs" and "Vaccines" and corresponding text word terms. The searches were limited to items published from 1 January 2000. The last search was conducted on 20 January 2021.

| Screening
Items were screened using the inclusion/exclusion criteria (see Table 1) by the first author (JS). The screening was cross-checked by the senior author (RB).

| Data extraction
A data extraction form was developed by JS in consultation with coauthors (RB, PB, CK). Information extracted included "title," "name of first author," "source," "national clinical trials' number (NCT),"

TA B L E 1 Inclusion and exclusion criteria Inclusion
Clinical study of RSV vaccine used in humans with a measured outcome of immunogenicity "participants," "vaccine candidate," "study type," "outcome," and "serious adverse events." We focused on severe prognoses and decided to limit descriptions to adverse effects that were a minimum of grade 3. 13

| Evaluation of data analysis
We aimed to summarise the RSV vaccine immunogenicity based on each paper's definition of "immune-response" (described in the relevant journal papers); commonly, for instance, a ≥4-fold rise in RSVneutralising antibody (NA) in seronegative children or a ≥3-fold rise in NA in adults. Moreover, JS extracted the safety data based on the serious adverse events (SAE) presented in those papers.
The studies looked at disease prevention: a case of RSV in- This review was not prospectively registered.

| RE SULTS
A total of 6306 publications were identified through the database searches. Duplicate publications and those that were not RSV vaccine clinical trials were excluded. In total, 38 publications were included covering the four major types of RSV vaccine candidates, live-attenuated (n = 14), subunit (n = 12), vector-based (n = 6) and nanoparticle (n = 6) (see Figure 1).

| Live-attenuated/chimeric vaccines
The RSV M2-2 gene mediates the transition from transcription to RNA replication, so its deletion can be used to attenuate the virus.
Meanwhile, it still elicits a neutralising antibody response. 14

| MEDI-534
MEDI-534 is a vaccine candidate using a parainfluenza virus type 3 (PIV3) backbone genome, which was altered to express RSV F protein. 31  There was no significant difference in the side effect event rates between the vaccine and placebo groups. 33 Thirdly, a Phase 1, double-blind RCT was conducted in 49 RSV/PIV3 seronegative children aged 6-24 months. The results were better in those given multiple doses (ie 2 or 3) and at a higher dose median tissue culture infective dose (TCID50), dosage of 10 6 , but even then only about 50% responded with a ≥4-fold neutralising antibody rise so it was not a strong candidate; only one of 17 in the placebo group had a ≥4-fold rise in neutralising antibody likely due to a wild-type RSV infection. Also, a favourable immune response to PIV3 was observed.
There was no serious adverse event. 34

| BBG2Na
The BBG2Na is a subunit vaccine candidate purified from a In the control group, only 6% or fewer recipients had an NA immune response (days 30, 60 and 90). F-020 and F-024 recipients had a similar safety profile to the control group recipients, and no SAEs were considered vaccine-related (NCT02360475, NCT02753413). 39
Due to study differences, meta-analysis was not possible. However, these data did not show a significant rate of reduction (Table 2). Abbreviation: MAARI, medically attended acute respiratory illness.

| RSV infection in the subunit vaccine candidates confirmed by western blot during RSV season
Four trials were found, of which two of them were subunit vaccines while the other two were nanoparticle vaccines. The subunit vaccine candidates were PFP-3 and PFP-2. The data were collected in the children aged from 1 to 12 years in the PFP-3 study, and the infants with maternal vaccination in the PFP-2 trial. 47,64 The relative risks of RSV infection between the vaccine and placebo groups were 0.82 (95%, CI 0.56-1.22) and 0.19 (95%, CI 0.02-1.51), respectively. There was no significant reduction of RSV infection (Table 3).

| RSV infection in nanoparticle vaccine candidates confirmed by western blot during RSV season
Two RSV-F nanoparticle vaccine trials 58,59 were conducted in healthy women of childbearing age. The two relative risks were similar; 0.48 (95%, CI 0.29-0.80) was from all active vaccinees compared to placebo recipients, while 0.50 (95%, CI 0.27-0.92) was from pooled one-dose (120 µg, 60 µg) vaccinees compared to placebo recipients.
A vaccine protective effect was revealed according to the relative risk results (Table 3). In infants and children, the age of most concern is the first 6 months of their life; although they have some maternal immune protection, the risk of severe RSV infection is still high. 65 Many children ≥6 months are RSV-naive, and they are similar to infants in less than 6 months, except with a more mature immune system. 66 Almost all live-attenuated vaccine trials were in infants older than 6 monthsthey are a proxy for younger infants. Maternal vaccination is one of the best strategies for protection against RSV and avoiding ERD in infants. Ideally, boosting maternal RSV antibody level from at least 3 months prior to labour could make antibodies available for transplacental transfer. 67  The 60 µg of nanoparticle F vaccine candidate with adjuvant had a favourable immune response and its persistence was long enough to cover a whole RSV season. Therefore, nanoparticle F and Ad26.

| D ISCUSS I ON
RSV.preF could be thought as the promising vaccine candidate for older people.
Another two subunit candidates (F-020 and F-024) and a nanoparticle F vaccine candidate were conducted in women of childbearing age. All of them showed a 3-month-rise of immune antibody, which demonstrated the possibility of placental antibody transportation in the future. F-024 had less vaccination local reaction than Tdap, suggesting F-024 could be more suitable in pregnant women due to less pain from injection. Results of the relative risks of RSV infection cases in the surveillance seasons showed an acceptable protective effect for one-dose nanoparticle candidate given in healthy women of childbearing age. Moreover, the single-dose 120 µg RSV F protein vaccine with 0.4 mg adjuvant was timely and strongly immunogenic. Similar immunogenicity effects for the nanoparticle F vaccine candidate were observed between the one and two doses groups with adjuvant. In fact, one dose is more convenient and still gives a strong antibody response in women of childbearing age. It was examined in a Phase 3 trial (NCT02624947) in pregnant women in their third trimesters. 59 However, the result did not meet the expected success criterion; therefore, we still lack a promising maternal vaccination.
Limitations of this systematic review include that it is limited to publications in English only and was date limited from January 2000 to avoid duplicating the results of an earlier review. 74 Also, this review presented little information about vaccine-induced cellular immune response due to lack of relevant data. A meta-analysis on RSV vaccine candidates was not undertaken due to the heterogeneous nature of the vaccine candidates and targeting populations. There is no promising vaccine for maternal vaccination at the present time.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/irv.12850.