Broad blockade antibody responses in human volunteers after immunization with a multivalent norovirus VLP candidate vaccine: immunological analyses from a phase I clinical trial

Abstract

Background: Human noroviruses (NoVs) are the primary cause of acute gastroenteritis and are characterized by antigenic variation between genogroups and genotypes and antigenic drift of strains within the predominant GII.4 genotype. In the context of this diversity, an effective NoV vaccine must elicit broadly protective immunity. We used an antibody (Ab) binding blockade assay to measure the potential cross-strain protection provided by a multivalent NoV virus-like particle (VLP) candidate vaccine in human volunteers.

Methods and findings: Sera from ten human volunteers immunized with a multivalent NoV VLP vaccine (genotypes GI.1/GII.4) were analyzed for IgG and Ab blockade of VLP interaction with carbohydrate ligand, a potential correlate of protective immunity to NoV infection and illness. Immunization resulted in rapid rises in IgG and blockade Ab titers against both vaccine components and additional VLPs representing diverse strains and genotypes not represented in the vaccine. Importantly, vaccination induced blockade Ab to two novel GII.4 strains not in circulation at the time of vaccination or sample collection. GII.4 cross-reactive blockade Ab titers were more potent than responses against non-GII.4 VLPs, suggesting that previous exposure history to this dominant circulating genotype may impact the vaccine Ab response. Further, antigenic cartography indicated that vaccination preferentially activated preexisting Ab responses to epitopes associated with GII.4.1997. Study interpretations may be limited by the relevance of the surrogate neutralization assay and the number of immunized participants evaluated.

Conclusions: Vaccination with a multivalent NoV VLP vaccine induces a broadly blocking Ab response to multiple epitopes within vaccine and non-vaccine NoV strains and to novel antigenic variants not yet circulating at the time of vaccination. These data reveal new information about complex NoV immune responses to both natural exposure and to vaccination, and support the potential feasibility of an efficacious multivalent NoV VLP vaccine for future use in human populations.

Trial registration: ClinicalTrials.gov NCT01168401.

Conflict of interest statement

MTF, CWM, KD, and JS do not have any competing interests. LCL and RSB have received royalties from a licensing agreement with Ligocyte (now Takeda). FB is an employee of Takeda Pharmaceuticals International. CR, RRG, RFB, and PMM are employees of Takeda Vaccines. RRG has stock in Takeda Vaccines, Inc. CR holds stock or options in Takeda Pharmaceuticals. RFB holds patents on Takeda's norovirus vaccine drug substance, owns Takeda stock, is PI of a research contract from the Department of Defense for norovirus vaccine development, is a board member of the Montana BioScience Alliance, and is a registered lobbyist. JF is an employee of EMMES and is contracted through Takeda Vaccines.

Figures

Fig 1. Temporal relationship between epidemiologically important GII.4 strains, relative to the virus-like particles and sera used in this study.

The vaccine GII.4 component (GII.4C) is a consensus VLP composed of GII.4.2002, GII.4.2006a, and GII.4.2006b sequences. GII.4.2006b.P.D302 represents a strain that evolved in vivo and was isolated from an immune-compromised person.

Fig 2. Characteristics of virus-like particles.

Ab responses to a diverse panel of GI (blue), GII.4 (grey), and non-GII.4 GII (green) VLPs were compared in this study.

Fig 3. Mean EC50 IgG titer in vaccinated participants.

Serum samples collected from participants who received the 50/50-μg VLP dose were assayed for IgG reactivity to a panel of GI (blue), GII.4 (grey), and non-GII.4 GII (green) VLPs. The seroresponse rate is the ratio of the number of participants with a ≥4-fold titer increase above day 0 titer compared to the total number of samples tested at day 0 for each VLP. Bolded values denote significant increases in GMFR above baseline.

Fig 4. Mean EC50 blockade antibody titer in vaccinated participants.

Serum samples collected from participants who received the 50/50-μg VLP dose were assayed for blockade Ab to a panel of GI (blue), GII.4 (grey), and non-GII.4 GII (green) VLPs. The seroresponse rate is the ratio of the number of participants with a ≥4-fold titer increase above day 0 titer compared to the total number of samples tested at day 0 for each VLP. Bolded values denote significant increases in GMFR above baseline.

Fig 5. Day 0 blockade antibody titers below the assay limit of detection for any norovirus virus-like particle are predictive of a ≥4-fold increase, but not overall blockade Ab titer, at day 7.

Day 0 blockade Ab titers for all of the NoV VLPs studied were compared to the corresponding day 7 fold increase (A) and titer of blockade Ab (B) using GEEs (n = 10 participants, 82 samples). The likelihood of responding to vaccination with a ≥4-fold increase in blockade Ab titer to any NoV VLP was 3-fold greater if the day 0 titer was below the assay limit of detection. Day 0 blockade Ab titer did not correspond with day 7 titer. The dotted lines mark the lower limit of detection of the blockade Ab assay. The solid grey line marks a 4-fold increase in blockade titer at day 7. RR, relative risk.

Fig 6. Mean EC50 IgG titers to novel GII.4 strain virus-like particles.

Serum samples collected from participants who received the 50/50-μg VLP dose were assayed for IgG reactivity to the vaccine components and to two novel GII.4 VLPs. GI VLP is shaded blue; GII.4 VLPs are shaded grey. The seroresponse rate is the ratio of the number of participants with a ≥4-fold titer increase above day 0 titer compared to the total number of samples tested at day 0 for each VLP. Bolded values denote significant increases in GMFR above baseline.

Fig 7. Mean EC50 blockade antibody titers to novel GII.4 strain virus-like particles.

Serum samples collected from participants who received the 50/50-μg VLP dose were assayed for blockade Ab to the vaccine components and to two novel GII.4 VLPs. GI VLP is shaded blue; GII.4 VLPs are shaded grey. The seroresponse rate is the ratio of the number of participants with a ≥4-fold titer increase above day 0 titer compared to the total number of samples tested at day 0 for each VLP. Bolded values denote significant increases in GMFR above baseline.

Fig 8. Amino acid sequence of identified GII.4 blockade antibody epitopes (A, D, and E) and the regulating domain of epitope F (NERK motif) in GII.4 virus-like particles relevant to this study.

Color indicates antigenic groupings based on epitope A sequence. *The amino acid coordinates of epitope F are unknown. The NERK motif is a temperature-sensitive regulator of Ab access to epitope F [32].

Fig 9. Evolving IgG responses to virus-like particles throughout the time course of the study.

Individual points represent VLPs, and the distances between points represent the overall differences in the magnitude of IgG responses from all ten participants against these NoV VLPs on day 0 (A–C), day 7 (D–F), and day 180 (G–I). Specifically, the distance between VLPs (with each unit in any dimension [D] relating to a 3.4-fold difference in total IgG responses) shows how similar total IgG responses across the ten participants were to each pair of VLPs tested within this study and how vaccine components (noted in each panel) cluster with other NoV strain VLPs. Despite similarities across these responses, we are able to show that GI VLPs (orange) cluster with the GI.1 vaccine component, early GII.4 VLPs (blue) cluster with the GII.4C vaccine component, and late GII.4 VLPs (green) and the other GII VLPs (pink) cluster together and away from the other VLPs. For each time point, the x-axis is that showing the most variation between all VLPs, then the y-axis, then the z-axis. Therefore, down each column, we can see how immune responses change and track through the time course of the study. Of note are the clusterings of each virus subtype through these responses.

Fig 10. Evolving blockade antibody responses to virus-like particles throughout the time course of the study.

Individual points represent VLPs, and the distances between points represent the overall differences in the magnitude of blockade Ab responses from all ten participants against these NoV VLPs on day 0 (A–C), day 7 (D–F), and day 180 (G–I). Specifically, the distance between VLPs (with each unit in any dimension [D] relating to a 3.2-fold difference in blockade Ab, or EC50 response) shows how similar total blockade Ab responses across the ten participants were between each pair of VLPs tested within this study, and how vaccine components (noted in each panel) cluster with other NoV strain VLPs. Across these responses, we are able to show that GI VLPs (orange) cluster with the GI.1 vaccine component, early GII.4 VLPs (blue) cluster with the GII.4C vaccine component, and the late GII.4 VLPs (green) and the other GII VLPs (pink) cluster together and away from the other VLPs. For each time point, the x-axis is that showing the most variation between all VLPs, then the y-axis, then the z-axis. Therefore, down each column, we can see how immune responses change and track through the time course of the study. Of note are the clusterings of each virus subtype through time, although with clear distinction of the vaccine components at day 7. At day 180, as titers have fallen, the blockade Ab distinctions between VLPs are diminished across the panels (G–I), with the exception of responses to GII.4.1997, suggesting that a memory Ab response to this strain may be driving the GII.4-reactive vaccine response.

Fig 11. Vaccination results in a rapid but transient increase in antibody titer to multiple blockade epitopes in multiple GII.4 strains.

Serum samples were evaluated for ability to block binding of mouse mAbs to epitope (Epi) A or F in GII.4.1997 and GII.4.2006b using a BOB assay. Sigmoidal curves were fit to the mean percent control binding (percent of mouse mAb bound to VLP in the presence of serum pretreatment compared to the amount of mouse mAb bound in the absence of serum pretreatment), and the mean EC50 titer (1/serum dilution) for BOB calculated. Dotted line marks 0.5 times the assay limit of detection. Error bars represent 95% confidence intervals. An asterisk indicates that the EC50 titer is significantly different from that of day 0.

Fig 12. Proposed mechanisms for antibody responses induced by GI.1/GII.4C multivalent vaccine.

At the time of vaccination, adult participants have a lifetime of NoV exposure history and a pool of NoV-reactive memory B cells (Bmem cells), both strain-specific and strain cross-reactive clones. Vaccination activates memory B cells to undergo SHM of the variable region of the Ab gene, to proliferate, and, for some cells, to differentiate into plasma cells secreting high-affinity Ab by day 7 post-vaccination. The GI.1 vaccine component elicits activation of both GI.1-specific memory B cells and memory B cells with specificity for shared GI epitopes, resulting in increased Ab to the panel of GI VLPs, but the strongest response to the homotypic GI.1 VLP because more blockade epitopes are unique to GI.1 than are shared across the GI VLPs. By day 180, low levels of GI.1-specific Ab persist (A). In comparison, the GII.4C vaccine component does not elicit a strong strain-specific response but could in theory activate memory B cells with GII.4.2002, GII.4.2006a, or GII.4.2006b specificity. However, the uniformity in the GII.4 VLP response across an antigenically diverse panel suggests that GII.4C preferentially activates memory B cells for conserved GII.4 epitopes and a smaller subset of memory B cells for a conserved GII epitope, resulting in more potent GII.4 and less potent GII blockade Ab production at day 7. The GII.4C Ab response continues to track with GII.4.2002 and GII.4.1997 through day 35, but by day 180, only GII.4.1997 Ab responses remain distinct, suggesting that the common GII.4 blockade epitopes recognized by the vaccine-induced Abs are most similar to sequences found in GII.4.1997, possibly because of extensive long-term immune focusing for this strain. Ab responses to at least two epitopes are maintained. Epitope F is a conserved GII.4 blockade epitope located sub-surface on the particle, and proposed epitope A′ is likely a surface-exposed blockade epitope physically near, overlapping, or within epitope A (B).