key: cord-0021625-op614djm
authors: Zhao, Wei; Zhang, Peng; Bai, Shuang; Lv, Min; Wang, Jian; Chen, Weixin; Yu, Qingzhong; Wu, Jiang
title: Heterologous prime-boost regimens with HAdV-5 and NDV vectors elicit stronger immune responses to Ebola virus than homologous regimens in mice
date: 2021-09-30
journal: Arch Virol
DOI: 10.1007/s00705-021-05234-4
sha: 7c2edeeb6b33b40e364845f45ca71f31dcb860d3
doc_id: 21625
cord_uid: op614djm

The 2014 Ebola outbreak in West Africa resulted in more than 11,000 deaths, highlighting the need for a vaccine. A phase I clinical trial of a human adenovirus type 5 vector-based Ebola virus (EBOV) vaccine (HAdV-5-MakGP) showed that a homologous prime-boost regimen with HAdV-5 vaccine elicited a robust humoral response but a weak cellular immune response. Due to pre-existing anti-vector immunity, boosting with the same vaccine did not increase the magnitude of the cellular immune response, which contributes significantly to protection against EBOV infection. Here, we generated a recombinant Newcastle disease virus (NDV), based on the LaSota vaccine strain, expressing the GP protein of the EBOV variant Makona (rLS/EB-GP) using reverse genetics technology. The humoral and cellular immune responses to this vaccine candidate in mice immunized using a homologous prime-boost regimen or a heterologous prime-boost regimen with the HAdV-5-vectored Ebola vaccine were assessed using ELISA and ELISPOT assays. The ELISA and ELISPOT results showed that mice primed with rLS/EB-GP and boosted with HAdV-5-MakGP (NDV+HAdV-5) or, reversed, primed with HAdV-5-MakGP and boosted with rLS/EB-GP (HAdV-5+NDV) exhibited more-potent EBOV GP-specific antibody and cellular immune responses than those receiving the same vaccine twice. The most robust EBOV GP-specific antibody and T-cell responses were detected in the HAdV-5-MakGP-primed and rLS/EB-GP-boosted (HAdV-5+NDV) mice. These results suggest that the HAdV-5 prime-NDV boost regimen is more effective in stimulating EBOV-specific immunity than homologous regimens alone, indicating the potential boosting ability of the NDV vector in human vaccine use.

The 2013-16 Ebola outbreak in West Africa was the most devastating human infectious disease witnessed in recent times, causing over 28,000 reported cases and more than 11,000 deaths [7] . Recent studies have revealed that the Ebola virus (EBOV) spreads by human-to-human transmission through contact with body fluids of infected people; even the recovered can still transmit the virus via their semen for up to 7 weeks after recovery from illness [5] . Due to the ease of transmission and lack of approved antiviral therapy, further small outbreaks could occur at any time. This threat highlights the need for a vaccine to prevent infection and avoid repetition.

In 2017, the China FDA approved an Ebola vaccine based on a human adenovirus serotype 5 (HAdV-5) vector. A phase I clinical trial of the HAdV-5-vectored Ebola vaccine in China has shown that two doses of the homologous HAdV-5 Ebola vaccine with a six-month interval elicit a robust humoral response but only a weak cellular immune response [17] . Unlike the induction of antibodies, repeated homologous boosting with the same vaccine does not increase the magnitude of the cellular immune response, as pre-existing anti-vector immunity inhibits antigen processing [13] . The suppression of immune responses caused by HAdV-5 pre-existing immunity (PEI) to the homologous vaccine boost has also been extensively reported in animal models and clinical trials [3, 27] . To address this concern, researchers have attempted to develop a heterologous prime-boost immunization strategy. Several pre-clinical and clinical studies have shown that heterologous prime-boost regimens generate stronger and broader immune responses and provide significant durability benefits when compared to homologous boosting [21, 24, 25] . These advantages of heterologous prime-boost regimens over homologous regimens might be due to their ability to avoid the consequences of pre-existing anti-vector immunity. The heterologous prime-boost strategy delivers the same or similar antigens using different vaccine vectors, which redirects the immune response to the antigens rather than the vectors. Because of the highlighted potential mentioned above, such a strategy represents a new way of immunization and has been applied to the development of vaccines against a wide range of pathogens [13] .

Newcastle disease virus (NDV) is a non-segmented, single-stranded negative-sense RNA virus belonging to the genus Orthoavulavirus, family Paramyxoviridae [23] . Several characteristics make NDV a promising vaccine vector against human and veterinary pathogens. NDV is safe for administration to humans, due to a natural host range restriction [8] . In clinical trials for cancer treatment, NDV as an oncolytic agent was well tolerated without causing adverse effects [11] . Since NDV is an avian virus that is antigenically distinct from human pathogens, there is no preexisting immunity in humans that would limit the utility of an NDV vaccine vector [1] . Furthermore, NDV replicates in the respiratory tract and induces strong humoral and cellular immune responses, both at the mucosal and systemic levels, in murine and nonhuman primate models, thus making this vector an attractive candidate for vaccine development [1, 15] .

In the present study, we generated an infectious clone of the NDV LaSota strain as a vaccine vector, inserted the surface glycoprotein (GP) gene of EBOV into the infectious clone, and rescued a recombinant NDV that expresses the GP protein of the EBOV variant Makona. Evaluation of the NDV-vectored Ebola vaccine in mice through different immunization regimens with the HAdV-5-vectored Ebola vaccine allowed us to better understand the immune responses elicited after heterologous and homologous prime-boost regimens.

HEp-2, DF-1, and Vero cell lines, obtained from the American Type Culture Collection (Manassas, VA), were maintained in Dulbecco's minimal essential medium (DMEM) supplemented with 10% fetal bovine serum (FBS). The modified vaccinia virus vTF7-3, expressing the T7 RNA polymerase, was obtained from ATCC (VR-2153; ATCC) and was grown in primary chicken embryo fibroblast cells. The NDV LaSota strain obtained from the China Veterinary Culture Collection Center was propagated in 9-day-old embryonated specific-pathogen-free (SPF) chicken eggs as a stock. Total RNA was extracted from the allantoic fluid of the infected eggs at 72 h post-inoculation using TRIzol Reagent (Invitrogen).

The construction of a plasmid carrying the full-length antigenomic cDNA of the NDV LaSota strain has been described previously [31] . In brief, RNA was extracted and subjected to RT-PCR with four pairs of specific primers using a Superscript TM III One Step RT-PCR System with Platinum Taq High Fidelity (Invitrogen) to generate four overlapping PCR fragments of the entire viral genome (Fig. 1 ). These four cDNA fragments were sequentially cloned into a modified pBluescript vector using an In-Fusion® PCR Cloning Kit (Clontech) to construct a complete genomic cDNA of the LaSota virus. The final plasmid containing the full-length LaSota strain genome, designated as pFLC-LS, was amplified in Stbl2 cells at 30°C and purified using a QIAprep Spin Miniprep Kit (QIAGEN, Valencia, CA).

Viral rescue from infectious clones for members of the paramyxovirus family requires the co-expression of the viral nucleoprotein (NP), phosphoprotein (P), and polymerase (L) genes [9] . Amplicons containing the open reading frames (ORFs) of the NP, P, and L genes were generated by RT-PCR from the virus genome and cloned into a modified pBluescript vector using an In-Fusion PCR Cloning Kit (Clontech). The resulting plasmid constructs were designated as p-NP, p-P, and p-L, respectively.

The plasmid containing the GP gene of the EBOV Makona strain was a kind gift from Dr. Xu, Beijing Institute of Biotechnology [28] . The procedure for construction of the recombinant cDNA clone containing the GP gene of EBOV is illustrated in Fig. 1 . Briefly, a cDNA fragment representing the ORF encoding the EBOV GP was amplified using paired specific primers into which gene end-intergenic-gene start motifs sequences were incorporated. An optimal Kozak translation sequence was inserted upstream of the GP gene translation start codon to enhance foreign gene expression. Subsequently, the PCR product containing the GP ORF along with the NDV trans-acting signal sequences was introduced into the P-M intergenic region of the pFLC-LS vector using an In-Fusion ® PCR Cloning Kit (Clontech). The resulting recombinant cDNA clone was designated as pLS/EB-GP.

The procedures for transfection of HEp-2 cells with the full-length cDNA clone, pFLC-LS or pLS/EB-GP, and the supporting plasmids, p-NP, p-P, and p-L, were described previously [32] . HEp-2 cells grown in a six-well plate were infected with vTF7-3 at a multiplicity of infection (MOI) of 3. Subsequently, the cells were cotransfected with 2 μg of pFLC-LS or pLS/EB-GP together with 1 μg of p-NP, 0.5 μg of p-P, and 0.1 μg of p-L using 6 μl of Lipofectamine TM 2000 (Invitrogen). After 6 h of incubation at 37°C, the medium was replaced with DMEM containing 2% FBS, and the cells were incubated for three days at 37°C. The supernatants were then inoculated into the allantoic cavities of 10-dayold embryonated SPF eggs. After 4 days of incubation, the allantoic fluid was harvested, and the rescued viruses (rLS and rLS/EB-GP) were screened by hemagglutination assay (HA). The nucleotide sequences of the rescued viruses were determined by sequencing the RT-PCR products amplified from the viral genomes as described previously [12] .

The procedure for IFA was described previously [31] . Vero cells in 6-well plates were infected with rLS or rLS/EB-GP at an MOI of 0.01. The primary antibody was rabbit anti-EBOV GP protein antibody (Sino Biological Inc, Beijing, China). The secondary antibody was fluorescein (FITC)conjugated goat anti-rabbit IgG (H+L) (Thermo Fisher). Fluorescence images were observed under an inverted fluorescence microscope at 100× magnification with a matching wavelength filter for FITC.

The replication-defective HAdV-5-MakGP vaccine was developed and prepared by the Beijing Institute of Biotechnology and Tianjin CanSino Biotechnology as described previously [33] . The open reading frame (ORF) of the EBOV GP gene along with the NDV trans-acting signal sequences was amplified and inserted into the pFLC-LS vector in the non-coding region downstream of the P gene, resulting in the pLS/EB-GP clone. The direction of the T7 promoter is indicated by a bold black arrow. HDVRz and T7Φ represent the site of the hepatitis delta virus ribozyme and the T7 terminator sequence, respectively. The NDV gene start and gene end sequences, Kozak sequence, and the EBOV GP ORF are boxed A group of 4-to 6-week old female BALB/c mice was obtained from Beijing Vital River Laboratory Animal Co. Ltd (Beijing, China) and housed in an animal biosafety level 2 facility. The mice were confirmed by ELISA to be free of EBOV antibodies prior to immunization and then randomly divided into six groups of six mice each and inoculated with the initial dose of vaccine on day 0 and the booster on day 21, as illustrated in Fig. 2 At 21 and 49 days post-immunization (PI), blood was collected from each mouse, and serum was obtained by centrifugation of whole blood. Isolated sera were transferred to labeled tubes and stored at -20°C until analysis. Eighteen mice from each group were humanely euthanized at day 49, and splenocytes were isolated for ELISPOT assay.

96-well polystyrene microtiter plates (Costar, Bethesda, MD) were coated overnight at 2-8°C with 2 μg of EBOV GP protein (subtype Zaire, strain H. sapiens-wt/GIN/2014/ Kissidougou-C15) (Sino Biological Inc, Beijing, China) per ml. Plates were washed with PBS containing 0.05% Tween (PBST) and blocked with 5% goat serum for 1 h at 37°C. After blocking, twofold serial dilutions of sera were added to EBOV-GP-protein-coated plates in duplicate and incubated for 1 h at 37°C. The bound antibodies were detected using horseradish-peroxidase-conjugated goat anti-Mouse IgG antibody (Thermo Fisher). After a 1-h incubation, color was developed by adding TMB (Solarbio, Beijing, China) peroxidase substrate, and absorbance was read at 450 nm in a Bio-Rad 550 ELISA plate reader (Bio-Rad, Hercules, CA). The antibody titer was expressed as the reciprocal of the highest antibody dilution that yielded greater than or equal to 2.1 times the background absorbance of PBS (the negative control) [19] .

A commercial kit (mouse IFN-γ precoated ELISPOT kit, Dakewe Biotech, Shenzhen, China) was used for the ELIS-POT assay. Lymphocytes from the spleen were harvested from each mouse, and a single cell suspension (5 × 10 5 cells/well) was prepared and stimulated with 20 μg of the EBOV GP protein described above per ml. After incubation at 37°C for 24 h, the cells were removed and the plates were processed according to the manufacturer's instructions. The number of spots was determined automatically using an automatic CTL Immunospot® Reader (Cellular Technology, Shaker Heights, Ohio).

Statistical analysis was performed using GraphPad Prism. All data were shown as the mean and standard deviation. Data were statistically analyzed using the unpaired, twotailed Student's t-test. p-values of <0.05 were considered significant.

A full-length cDNA clone encoding the complete antisense genome of the NDV LaSota strain was constructed by assembling four overlapping PCR fragments through PCR and In-Fusion PCR cloning, as illustrated in Fig. 1 . The GP gene of EBOV flanked by the NDV trans-acting elements was inserted into the P-M intergenic region of the LaSota full-length clone as an additional transcription unit, which increased the length of the recombinant clone by 2232 nucleotides (nt). Thus, the total length of the cDNA clone in the pLS/EB-GP plasmid was 17,418 nt and therefore divisible by 6, conforming to the "Rule of Six" [16] . Then, the NDV LaSota virus and the recombinant virus, designated as rLS BALB/c mice were divided randomly into six groups and were immunized at day 0 (prime) and day 21 (boost). Serum samples were collected at day 21 and 49 PI for ELISA assay. Mice were sacrificed at day 49 PI, and splenocytes were isolated for ELISPOT assay and rLS/EB-GP, respectively, were rescued from the cDNAs as described in Materials and methods. Sequencing analysis confirmed the nucleotide sequence fidelity of the rescued viruses (Fig. 3) .

To examine the expression of EBOV GP protein by rLS/ EB-GP, Vero cells were infected with rLS/EB-GP at an MOI of 0.01. After 48 h, the cells were fixed and incubated with a rabbit antiserum directed against the EBOV GP protein, followed by immunostaining with FITC-conjugated goat anti-rabbit IgG. Positive staining was observed in rLS/EB-GP-infected Vero cells using an anti-EBOV GP antibody and an FITC-labeled secondary antibody (Fig. 4a) . In contrast, no fluorescence was detected in NDV-LaSota-infected cells after staining with the same antiserum and FITC conjugate (Fig. 4b) . This result demonstrated that the EBOV GP protein was expressed from the rLS/EB-GP-recombinantinfected cells.

To assess the effectiveness of heterologous prime-boost vaccination regimens with rLS/EB-GP and HAdV-5-MakGP for inducing immune responses, BALB/c mice were immunized on day 0 (prime) and day 21 (boost) with these vectored EBOV vaccines as illustrated in Fig. 2 . EBOV-GP-specific antibodies in the collected sera were detected by ELISA. As shown in Fig. 5 , although EBOV-GP-specific IgG titers were relatively low in all groups after the first immunization (prime), they increased dramatically after the second immunization (boost). The heterologous prime-boost with HAdV-5 and NDV induced a significantly higher geometric mean titer (GMT) of GP-specific IgG at day 49 post-immunization (PI) than the homologous prime-boost regimens (HAdV-5+NDV group vs. HAdV-5 group, p < 0.05; HAdV-5+NDV group vs. NDV group, p < 0.05). The mice in the NDV+HAdV-5 group produced approximately 1.5-fold and 3.1-fold higher GP-specific IgG titers than those receiving the NDV and HAdV-5 homologous regimens, respectively. These results demonstrated that the heterologous primeboost regimens could elicit more-robust EBOV-GP-specific antibody responses than the homologous regimens.

Splenocytes isolated from the immunized mice at day 49 PI were assayed by IFN-γ-specific ELISPOT to analyze the antigen-specified T-cell responses induced by the different Fig. 3 Sequencing analysis of the rescued viruses. The RT-PCR products generated from the rescued viruses rLS (A) and rLS/EB-GP (B) were sequenced directly with 32 and 37 pairs of gene-specific prim-ers, respectively. Nucleotide sequence editing, assembly, and comparison analysis were carried out using the DNASTAR Lasergene software (DNASTAR, Madison, WI, USA) immunization regimens. Interestingly, as shown in Fig. 6 , the magnitude of the IFN-γ-specific ELISPOT responses correlated positively with the antibody responses in all immunization groups. Specifically, splenocytes obtained from mice in the heterologous HAdV-5+NDV and NDV+HAdV-5 regimen groups exhibited significantly higher IFN-γ expression than those from mice in the homologous HAdV-5 regimen (p < 0.05) and rLS/EB-GP regimen (p < 0.05) groups. The IFN-γ-specific ELIS-POT data, together with the antibody response results, demonstrate that mice immunized with the heterologous prime-boost regimens not only elicited stronger humoral responses but also induced stronger cellular responses than those receiving the homologous regimens.

In the present study, we successfully generated an NDV LaSota strain-based recombinant virus, rLS/EB-GP, expressing the GP protein of the EBOV variant Makona, as a vaccine candidate, using reverse genetics techniques. The immune responses of mice to this rLS/EB-GP recombinant vaccine and an HAdV-5-based EBOV experimental vaccine were evaluated by prime-boost regimens. The ELISA and ELISPOT results indicated that heterologous prime-boost regimens elicit more-robust immune responses than homologous regimens. The HAdV-5 prime and rLS/EB-GP boost immunization strategy was found to be the most effective approach for stimulating EBOVspecific immunities among the different regimens. Human adenovirus has emerged as an ideal vaccine vector because of its relative safety, high transgene expression levels, and its capacity to elicit robust humoral and cellular immune responses to the encoded antigens. [30] . However, serosurveillance has shown that PEI to HadV-2 and HAdV-5 existed in up to 50% of the American population and up to 95% in people from other countries. This high seroprevalence limits the efficacy of adenovirus (Ad)-vectored vaccines [4] . To circumvent the problem of PEI interference and improve the effectiveness of Ad-based vaccines, researchers have developed non-human primate (e.g., chimpanzee and gorilla) adenovirus vectors [6] , explored different immunization routes [22] , and, notably, investigated prime-boost regimens with heterologous vectors [18] . Regarding the last of these, studies have demonstrated that heterologous primeboost regimens employing a DNA prime and an HAdV-5 or ??modified vaccinia virus Ankara?? (MVA) boost could effectively induce humoral and cellular immune responses, and more importantly, improve vaccine efficacy, regardless of pre-existing Ad or MVA immunity [14, 29] . Three EBOV vaccine candidates adopting the heterologous prime-boost strategy have undergone different phases of clinical trials [20] . Results from these investigations indicated that both heterologous vector vaccine regimens of ChAd3-EBOV and Ad26.ZEBOV vaccines boosted with MVA elicited stronger B-cell and T-cell immune responses to EBOV than those induced by single-dose prime vaccination alone [2, 10] .

EBOV transmission occurs exclusively through contact with mucosal surfaces. It might therefore be advantageous to use a vaccine vector with a natural tropism for the respiratory tract, which could induce the first line of defense at mucosal points of entry and robust systemic immune responses in the respiratory tract. Numerous researches have shown that NDV, an avian paramyxovirus, is an attractive vaccine vector for human pathogens, inducing humoral and cellular immune responses at both the mucosal and systemic levels in murine and non-human primate models [15] . Therefore, in this study, we adopted NDV as a vaccine vector to express the EBOV GP protein as a heterologous boost vaccine with the hope of avoiding the PEI interference associated with Ad-based primed vaccines. The results presented in this report are in agreement with the findings from the EBOV vaccine trials using the heterologous prime-boost immunization strategy [20] , confirming that heterologous prime-boost regimens can elicit more-robust humoral and cellular responses to EBOV than homologous regimens [10, 26] .

We observed with the homologous regimens that although two doses of rLS/EB-GP or HAdV-5 induced detectable EBOV-GP-specific antibody and T-cell responses in mice, the immune responses elicited by HAdV-5 were twofold higher than those induced by NDV. This difference might be due to differences in the amount of antigen expressed/ presented by these two vectors or the different immunization routes used to stimulate the immune system. Furthermore, the DNA vector HAdV-5 and the RNA vector NDV undoubtedly stimulate innate pathways differently. Therefore, we adopted the non-uniform but normal dosage of HAdV-5 (1 × 10 7 IFU) and NDV (1 × 10 6 PFU) in this study. Further studies may be needed to determine whether a higher dose of NDV could enhance vaccine effectiveness.

In contrast to the homologous regimens, the heterologous prime-boost with HAdV-5 and rLS/EB-GP induced stronger immune responses than the homologous regimens. The most robust EBOV-GP-specific antibody and T-cell responses were detected by ELISA and ELISPOT in the HAdV-5 prime and NDV boost group, while the reversed NDV prime-HAdV-5 boost ranked second. The reason for this difference is still unclear. A possible explanation may be differences in the way stimulating innate immunity is stimulated by an RNA virus (NDV) and a DNA virus (Ad), which would affect the direction and strength of the subsequent adaptive immune response. This hypothesis needs to be tested in future studies by analyzing markers of immune responses.

Newcastle disease and other avian paramyxoviruses

Safety and immunogenicity of a 2-dose heterologous vaccination regimen with Ad26.ZEBOV and MVA-BN-Filo Ebola vaccines: 12-month data from a phase 1 randomized clinical trial in Uganda and Tanzania

Immunogenicity of recombinant adenovirus serotype 35 vaccine in the presence of pre-existing anti-Ad5 immunity

Novel adenovirus vector-based vaccines for HIV-1

Assessment of the risk of Ebola virus transmission from bodily fluids and fomites

Development of chimpanzee adenoviruses as vaccine vectors: challenges and successes emerging from clinical trials

The Ebola outbreak, 2013-2016: old lessons for new epidemics

Newcastle disease virus, a host range-restricted virus, as a vaccine vector for intranasal immunization against emerging pathogens

Evaluation of Newcastle disease virus chimeras expressing the Hemagglutinin-Neuraminidase protein of velogenic strains in the context of a mesogenic recombinant virus backbone

A Monovalent Chimpanzee Adenovirus Ebola Vaccine Boosted with MVA

Phase I/II trial of intravenous NDV-HUJ oncolytic virus in recurrent glioblastoma multiforme

Generation and evaluation of a recombinant Newcastle disease virus expressing the glycoprotein (G) of avian metapneumovirus subgroup C as a bivalent vaccine in turkeys

Prime-boost vaccine strategy against viral infections: Mechanisms and benefits

A phase 1/2 study of a multiclade HIV-1 DNA plasmid prime and recombinant adenovirus serotype 5 boost vaccine in HIV-Uninfected East Africans (RV 172)

Newcastle disease virus as a vaccine vector for development of human and veterinary vaccines

Paramyxovirus mRNA editing, the "rule of six" and error catastrophe: a hypothesis

Immunity duration of a recombinant adenovirus type-5 vector-based Ebola vaccine and a homologous prime-boost immunisation in healthy adults in China: final report of a randomised, double-blind, placebo-controlled, phase 1 trial

Safety and immunogenicity of an rAd26 and rAd5 vector-based heterologous prime-boost COVID-19 vaccine in two formulations: two open, non-randomised phase 1/2 studies from Russia

Standardization of immunoglobulin M capture enzyme-linked immunosorbent assays for routine diagnosis of arboviral infections

Ebola vaccine trials: progress in vaccine safety and immunogenicity

Team SS (2017) First-inhuman evaluation of the safety and immunogenicity of an intranasally administered replication-competent sendai virus-vectored HIV type 1 gag vaccine: induction of potent T-cell or antibody responses in prime-boost regimens

Airway delivery of an adenovirus-based Ebola virus vaccine bypasses existing immunity to homologous adenovirus in nonhuman primates

Newcastle disease and related avian paramyxoviruses

Chimpanzee adenovirus vaccine generates acute and durable protective immunity against ebolavirus challenge

Development of a preventive vaccine for Ebola virus infection in primates

Use of ChAd3-EBO-Z Ebola virus vaccine in Malian and US adults, and boosting of Malian adults with MVA-BN-Filo: a phase 1, single-blind, randomised trial, a phase 1b, open-label and double-blind, dose-escalation trial

Strategies to overcome host immunity to adenovirus vectors in vaccine development

An Adenovirus vaccine expressing Ebola virus variant Makona glycoprotein is efficacious in guinea pigs and nonhuman primates

Overcoming immunity to a viral vaccine by DNA priming before vector boosting

Adenoviral vector-based strategies against infectious disease and cancer

Newcastle disease virus (NDV) recombinants expressing infectious laryngotracheitis virus (ILTV) glycoproteins gB and gD protect chickens against ILTV and NDV challenges

P and M gene junction is the optimal insertion site in Newcastle disease virus vaccine vector for foreign gene expression

Safety and immunogenicity of a novel recombinant adenovirus type-5 vectorbased Ebola vaccine in healthy adults in China: preliminary report of a randomised, double-blind