iScience ll OPEN ACCESS Article Preclinical evaluation of a COVID-19 vaccine candidate based on a recombinant RBD fusion heterodimer of SARS-CoV-2 Antonio Barreiro, Antoni Prenafeta, Gregori Bech- Sabat, ..., Paula Cebollada Rica, Andreas Meyerhans, Laura Ferrer antonio.barreiro@hipra.com (A.B.) antoni.prenafeta@hipra.com (A.P.) Highlights An RBD-based recombinant protein vaccine (PHH-1V) against COVID-19 has been generated PHH-1V elicits humoral and cellular responses against different SARS- CoV-2 variants PHH-1V vaccination prevents mortality in SARS-CoV-2 infected K18- hACE2 mice PHH-1V vaccination reduces Beta, Delta and Omicron infection in lower respiratory airways Barreiro et al., iScience 26, 106126 March 17, 2023 ª 2023 The Author(s). https://doi.org/10.1016/ j.isci.2023.106126 iScience ll OPEN ACCESS Article Preclinical evaluation of a COVID-19 vaccine candidate based on a recombinant RBD fusion heterodimer of SARS-CoV-2 Antonio Barreiro,1,11,* Antoni Prenafeta,1,11,12,* Gregori Bech-Sabat,1 Mercè Roca,1 Eva Perozo Mur,1 Ricard March,1 Luis González-González,1 Laia Madrenas,1 Júlia Corominas,1 Alex Fernández,1 Alexandra Moros,1 Manuel Cañete,1 Mercè Molas,1 Thais Pentinat-Pelegrin,1 Clara Panosa,1 Alberto Moreno,1 Ester Puigvert Molas,1 Eva Pol Vilarrassa,1 Jordi Palmada,1 Carme Garriga,1 Teresa Prat Cabañas,1 Javier Iglesias-Fernández,2 Júlia Vergara-Alert,3 Cristina Lorca-Oró,3 Núria Roca,3 Leira Fernández-Bastit,3 Jordi Rodon,3 Mònica Pérez,3 Joaquim Segalés,4,5 Edwards Pradenas,6 Silvia Marfil,6 Benjamin Trinité,6 Raquel Ortiz,6 Bonaventura Clotet,6,7 Julià Blanco,6,7 Jorge Dı́az Pedroza,8 Rosa Ampudia Carrasco,8 Yaiza Rosales Salgado,8 Jordina Loubat-Casanovas,8 Sara Capdevila Larripa,8 Julia Garcia Prado,6,8 Jordi Barretina,8 Marta Sisteré-Oró,9 Paula Cebollada Rica,9 Andreas Meyerhans,9,10 and Laura Ferrer1 SUMMARY 1HIPRA, Avda. La Selva, 135, Current COVID-19 vaccines have been associated with a decline in infection rates, Amer, 17170 Girona, Spain prevention of severe disease, and a decrease in mortality rates. However, SARS- 2Nostrum Biodiscovery, Avenue de Josep Tarradellas, CoV-2 variants are continuously evolving, and development of new accessible 8-10, 3-2, 08029 Barcelona, COVID-19 vaccines is essential to mitigate the pandemic. Here, we present Spain data on preclinical studies in mice of a receptor-binding domain (RBD)-based re- 3Institutde Recerca i Tecnologia Agroalimentàries combinant protein vaccine (PHH-1V) consisting of an RBD fusion heterodimer (IRTA), Centre de Recerca en comprising the B.1.351 and B.1.1.7 SARS-CoV-2 variants formulated in SQBA Sanitat Animal (CReSA, IRTA- UAB), Campus de la UAB, adjuvant, an oil-in-water emulsion. A prime-boost immunisation with PHH-1V in 08193 Cerdanyola del Vallès, BALB/c and K18-hACE2 mice induced a CD4+ and CD8+ T cell response and Spain RBD-binding antibodies with neutralizing activity against several variants, and 4UniversitatAutònoma de Barcelona, CReSA (IRTA- also showed a good tolerability profile. Significantly, RBD fusion heterodimer UAB), Campus de la UAB, vaccination conferred 100% efficacy, preventing mortality in SARS-CoV-2 in- 08193 Cerdanyola del Vallès, Spain fected K18-hACE2 mice, but also reducing Beta, Delta and Omicron infection in 5Departament de Sanitat i lower respiratory airways. These findings demonstrate the feasibility of this re- Anatomia Animals, Facultat combinant vaccine strategy. de Veterinària, UAB, 08193 Cerdanyola del Vallès, Spain 6IrsiCaixa. AIDS Research Institute, Germans Trias i INTRODUCTION Pujol Research Institute In December 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was identified as the (IGTP), Can Ruti Campus, UAB, 08916 Badalona, Spain etiological agent of the coronavirus disease 2019 (COVID-19). Soon afterward, the scientific community 7University of Vic–Central and the pharmaceutical industry began focusing on the development of effective COVID-19 vaccines to University of Catalonia (UVic- mitigate the health emergency. Thanks to these efforts, several vaccines are currently available, and UCC), Vic, 08500 Catalonia, Spain more than 12.9 billion doses have been administered worldwide (November 2022).1 The decline in new 8Comparative Medicine and infection rates in many countries coincides with the introduction of vaccines. However, COVID-19 cases Bioimage Centre of continue to emerge, probably because of the appearance and evolution of SARS-CoV-2 variants, the Catalonia, Germans Trias i decline of immunological protection provided by the current vaccines, and especially the lack of homoge- Pujol Research Institute (CMCiB-IGTP), 08916 neous distribution of COVID-19 vaccines, with only 22.5% of people in low-income countries having Badalona, Spain received at least one dose (December 2021).2,3 As the global outbreak continues, the pandemic is far 9InfectionBiology from being over and it is not clear if the available vaccines will be sufficient to revert the situation. Thus, Laboratory, Department of it is still critical to develop second-generation vaccines using different platforms that are effective against Experimental and Health Sciences (DCEXS), Universitat new variants and that could be further used as a booster, particularly to maintain or even enhance immunity Pompeu Fabra (UPF), 08003 against SARS-CoV-2.4,5 Moreover, it is of relevance that these recent vaccines can be stored in refrigerated Barcelona, Spain conditions, making them easier to distribute and avoiding more expensive and less available ultra-low tem- perature storage and transport conditions to ensure their global supply. Currently authorized vaccines, Continued iScience 26, 106126, March 17, 2023 ª 2023 The Author(s). 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ll iScience OPEN ACCESS Article whether approved under emergency use or fully licensed, are based on viral vectors, inactivated viruses, nucleic acid-based vaccines, and protein subunit vaccines.6 SARS-CoV-2 is a betacoronavirus belonging to the subfamily Coronovirinae, within the family Coronaviri- dae and the order Nidovirales. The SARS-CoV-2 genome is a positive-sense single-stranded RNA (+ssRNA) molecule. The genome size ranges between 27 and 32 kbp, one of the largest known RNA viruses. The genomic structure of SARS-CoV-2 contains at least six open reading frames (ORFs), encoding for at least four structural proteins, namely envelope or spike (S) glycoprotein; membrane (M) proteins, responsible for the shaping of the virions; envelope (E) proteins, responsible for the virions assembly and release; and nucleocapsid (N) proteins, involved in the RNA genome packaging.7 The trimeric S glycoprotein of SARS-CoV-2 is the primary target of viral neutralizing antibodies and has been the main protein candidate for vaccine development.8 Consistent with SARS-CoV, angiotensin-converting enzyme 2 (ACE2) binding of the S protein allows cellular entry of SARS-CoV-2 viral particles.9 This protein consists of two domains, S1 and S2, allowing the binding of the viral particles and cellular entry by fusing with the host cell membrane.10 The receptor-binding domain (RBD) (Thr333-Gly526) is found in the S1 domain and it contains a highly immunogenic receptor-binding motif (RBM) that directly interacts with ACE2 and neutralizing antibodies.11 Therefore, most key mutations are found in the RBM, allowing the virus to adapt and escape the previously developed immunity.12,13 To date, several SARS-CoV-2 VoCs with key mutations in the S protein have emerged: Alpha (B.1.1.7), Beta (B.1.351), Gamma (P.1), Delta (B.1.617.2) and Omicron (B.1.1.529).14,15 The S protein is the primary target in vaccine development against betacoronaviruses because of its acces- sibility for immune recognition.16 It has been reported that two proline substitutions in the original S pro- tein sequence (S-2P) of MERS-CoV, SARS-CoV and HKU1 coronavirus maintain the antigenic conformation but retain S proteins in the prototypical prefusion conformation.17 Thus, learning from these previous re- sults, this S-2P design is used in the licensed mRNA-based vaccines Comirnaty (Pfizer-BioNTech)18 and Spi- kevax (Moderna),19 which substitute the K986 and V987 residues for prolines from the original S protein variant. Likewise, the adenoviral vector-based vaccine Jcovden (Johnson & Johnson) also contains DNA- encoding for the S-2P protein of SARS-CoV-2.20 Adjuvanted protein-based subunit vaccines represent an important type of vaccine, yet their development has lagged compared to other platforms because of the need to optimize the manufacturing process for each protein antigen. The most advanced subunit vaccine program against COVID-19 is the Novavax vac- cine (NVX-CoV-2373, marketed as Nuvaxovid), which is produced in insect cells in combination with a saponin-based adjuvant (Matrix-M).21 This vaccine consists of the stable full-length S protein in the antigen- ically optimal prefusion conformation. In addition, the Sanofi-GSK vaccine, known as VidPrevtyn Beta, con- sists of soluble prefusion-stabilized S trimers from SARS-CoV-2 produced in insect cells combined with the AS03 adjuvant.22 Both vaccines have been tested in human clinical trials and have recently received an EU marketing authorisation.21,23 Notably, recombinant proteins are competitive vaccine candidates with an adequate safety profile, no risk of genome integration, no live components, and suitable for people with compromised immune systems,24 showing high productivity yields and good stability profiles.24–26 Most cloned neutralizing antibodies target the RBD in the S1 domain,27 although there are additional immunogenic epitopes outside this domain.28 It has been reported that more than 90% of neutralizing an- tibodies isolated from convalescent patients target the RBD in one-third of cases.29 Moreover, depleted sera and plasma samples from individuals vaccinated with a 250-mg dose of the mRNA-1273 vaccine 10Catalan Institution for showed that up to 99% of neutralising antibodies target the RBD, even though the antigen is based on Research and Advanced the whole prefusion spike conformation.30 Given that the RBD domain of the S protein directly interacts Studies (ICREA), Pg. Lluı́s with the ACE2 receptor, RBD-targeting antibodies are not expected to cause antibody-dependent Companys 23, 08010 Barcelona, Spain enhancement (ADE), unlike non-neutralizing or sub-neutralizing antibodies,31 highlighting the importance 11These authors contributed of the RBD in the immune response against SARS-CoV-2 and emphasizing its relevance as a powerful and equally efficient immunogen in vaccine design. 12Lead contact *Correspondence: In view of the inherent particularities of the S protein, and especially the RBD domain, our team developed antonio.barreiro@hipra.com a vaccine-candidate platform based on this immunogen. From amongst the tested preliminary vaccine can- (A.B.), antoni.prenafeta@hipra.com didates, combined with one or several adjuvants, we finally proceeded with a protein-based subunit vac- (A.P.) cine candidate, namely PHH-1V, consisting of a recombinant RBD fusion heterodimer of the B.1.351 and https://doi.org/10.1016/j.isci. B.1.1.7 variants of SARS-CoV-2 expressed in Chinese hamster ovary (CHO) cells and formulated with the 2023.106126 2 iScience 26, 106126, March 17, 2023 iScience ll Article OPEN ACCESS squalene-based adjuvant (SQBA). Specifically, the SQBA adjuvant is an oil-in-water emulsion comprising well-known components which are used as adjuvants in human medicine.32 Hence, the main aims of this study were to assess the safety and efficacy of the PHH-1V vaccine in BALB/c and K18-hACE2-transgenic mice models, and to characterize the RBD fusion heterodimer antigen and its immunogenicity. RESULTS Recombinant RBD fusion heterodimer expression and characterization The antigen of the PHH-1V vaccine candidate is a recombinant RBD fusion heterodimer based on the B.1.351 (Beta) and B.1.1.7 (Alpha) SARS-CoV-2 variants (Figure 1A). The N-terminal monomer contains the amino acid sequence of the SARS-CoV-2 RBD protein from the B.1.351 variant, whereas the C-terminal monomer contains the amino acid sequence of the SARS-CoV-2 RBD protein from the B.1.1.7 variant. The rationale behind this antigenic construct was based on maximization of the affinity constant toward its target receptor, allowing the accommodation of each RBD variant bound to a single hACE monomer within the same or a different receptor. Three-dimensional structural models generated with AlphaFold233 high- lighted the coexistence of two different conformations of the PHH-1V construct. More specifically, one of the conformations is characterized by a protein-protein interaction between both RBD variants, whereas the other presents separated RBD domains stabilized by interactions of the N-/C-terminal regions (Fig- ure 2B). RBD monomer binding toward individual human ACE2 (hACE2) units requires the preferential adoption of a separated RBD stabilized conformation, and thus construct generation followed this requirement. The protein-protein interaction energies of two construct variant candidates, B.1.351-B.1.1.7 and B.1.1.7-B.1.351, were estimated by means of molecular mechanics-generalized Born surface area (MM- GBSA) calculations. MM-GBSA results highlighted the B.1.351-B.1.1.7 variant as the preferred construct candidate with a protein-protein interaction energy of 78.23 kcal mol1 as compared to the value of 98.06 kcal mol1 for the inverse construct. This suggests that the N-/C-terminal stabilized conformation is energetically more favourable in the B.1.351-B.1.1.7 construct than in the B.1.1.7-B.1.351 construct. The interaction energies of both studied constructs, in a stabilized N-/C-terminal conformation, and the hACE2 receptor were also computed by means of MM-GBSA simulations (Figure 2B). Although both models showed similar binding affinities to hACE2, the Beta N-terminus plus Alpha C-terminus configuration clearly exposed those mutations involved in the higher affinity toward the human ACE2 receptor on the protein surface and the potential immune evasion by both variants. Hence, the selection of the B.1.351-B.1.1.7 fusion heterodimer as the PHH-1V vaccine antigen was based on the lower free energy required for the formation of the stabilized N-/C-terminal conformation. The heterodimer is expressed in mammalian CHO cells and is formulated with the SQBA adjuvant. After expressing the antigen in a bioreactor fed-batch cultivation, it is purified by a downstream process consist- ing of sequential stages, including depth and tangential filtration, chromatography steps, and sterile filtra- tions. The final product is a highly purified antigen, as determined by SDS-PAGE (Figure 1C) and SEC-HPLC (Figure 1D), suitable for vaccine formulation. Moreover, surface plasmon resonance (SPR) analysis showed an affinity constant of 0.099 nM for hACE2 (Figure 1E). Recombinant RBD fusion heterodimer antigen immunogenicity in BALB/c mice RBD-specific binding and neutralizing antibody titers on PHH-1V vaccination BALB/c mice (Environ, IN, USA) were immunized with different doses of the recombinant RBD fusion het- erodimer antigen (group B: 0.04 mg, group C: 0.2 mg, group D: 1 mg, group E: 5 mg; and group F: 20 mg) on days (D) 0 and 21. Mice were also immunized with PBS as a control group (group A). A schematic view of the immunization protocol is depicted in Figure 2. The prime immunization of BALB/c mice with the PHH-1V candidate induced higher titers of RBD binding antibodies in groups C to F compared to the control (group A, immunized with PBS) on day 21 post-first immunization (D21) (p<0.01) (Figure 3A). After the prime-boost immunization, all vaccinated groups (B to F) reached higher IgG titers than the control group on D35/D37 (14/16 days after the boost; p<0.01). On D35/D37, specific SARS-CoV-2 RBD-binding antibodies were detected in groups B to D in a dose- dependent manner, with significant differences between these groups (p<0.01). However, no significant differences were observed between the groups immunized with more than 1 mg of recombinant RBD fusion iScience 26, 106126, March 17, 2023 3 ll iScience OPEN ACCESS Article A B C D E Figure 1. Structure and characterization of the B.1.351 (Beta) - B.1.1.7 (Alpha) receptor-binding domain (RBD) heterodimer, immunogen of PHH-1V (A) Structural representation of the RBD heterodimer. Top: sequence diagram. Bottom left: front view of the RBD heterodimer cartoon structure. Bottom right: top view of the antigen surface structure. Mutations are highlighted in green (K417N), cyan (E484K) and yellow (N501Y). (B) Computation modeling for PHH-1V vaccine. Top: AlphaFold2 results for the B.1.351-B.1.1.7 construct. This highlights the presence of two different construct conformations: (1) Stabilized N-/C-terminal conformation and (2) adopting protein-protein interactions. Bottom: hACE2 receptor-construct model derived from MD calculations of the B.1.351-B.1.1.7 construct. RBD residues 1 to 219 and 220 to 439 are shown in gray and orange, respectively, whereas ACE2 monomers are shown as a transparent surface and cartoon representation in violet and purple. (C) SDS-PAGE. The reduced and non-reduced purified antigens were loaded at three serial dilutions: 1/10, 1/20 and 1/40. M: molecular weight ladder. C: BSA control. (D) SEC-HPLC chromatogram of the purified antigen. (E) Surface plasmon resonance (SPR) for the quantitative evaluation of the affinity between the antigen and its natural ligand, the human ACE2 receptor. RU: resonance units. heterodimer antigen (groups D to F). Therefore, the IgG response was saturated from 1-mg immunization. Likewise, the IgG2a/IgG1 ratios were calculated as a surrogate of the Th1/Th2 cellular immune response to estimate the type of cellular immune response elicited by the vaccine. The IgG2a/IgG1 ratio of groups E and F was 0.74 and 0.75, respectively, which suggests a balanced Th1/Th2 immunogenic response on PHH-1V vaccination in mice (Figure 3B). 4 iScience 26, 106126, March 17, 2023 iScience ll Article OPEN ACCESS Figure 2. Schematic representation of the experimental protocol in BALB/c and K18-hACE2 mice for the safety, immunogenicity, and efficacy assessment For safety and immunogenicity assays (in the top side), 48 five-week-old female BALB/c mice were allocated to 6 groups (n = 8) and were injected intramuscularly with two doses of 0.1 mL of the PHH1-1V vaccine on days 0 (prime) and 21 (boost). Then, animals were monitored daily for clinical signs and bodyweight was recorded weekly until D35/D37, when they were euthanized and both spleens and blood were collected. For safety, immunogenicity and efficacy assays (in the bottom side), K18-hACE2 mice were allocated to 4 groups (efficacy against SARS-CoV-2 D614G) or 3 groups (efficacy against different VoCs), and were injected intramuscularly with two doses of 0.1 mL of the PHH1-1V vaccine on days 0 (prime) and 21 (boost). On D35 animals were challenged with 103 TCID50 of the SARS-CoV-2 or different VoCs, blood samples were collected to analyze neutralizing activity, and they were monitored daily for clinical signs and mortality. Then, challenged animals were chronologically euthanized on D37, D39 and D42/D43; and several tissue samples were collected for several analyses. Schematic artwork used in this figure is provided by Servier Medical Art under a Creative Commons Attribution 3.0 Unported License. SARS-CoV-2 neutralizing antibodies titers in sera from BALB/c mice were determined by a pseudovirus- based neutralisation assay (PBNA) against the S protein of different variants on D35/D37 (14/16 days after the boost). Prime-boost immunization of groups C to F induced higher neutralizing antibody titers against the S protein of the Alpha variant compared to the control group A (p<0.01) (Figure 3C). No neutralizing antibody response was observed in group B, although IgG binding antibodies were detected on D35/ D37. The mean neutralizing antibody titers observed in groups C and D remained the same because no statistically significant differences were observed. However, vaccination with 5 mg (group E) and 20 mg (group F) of RBD fusion heterodimer antigen induced higher neutralizing titers than group C and groups C and D, respectively. Of interest, high neutralizing titers against all the tested variants (Alpha, Beta, Delta and Omicron BA.1) were detected in sera from group F compared to control group A (p<0.01) (Figure 3D). RBD-specific cellular immune response on PHH-1V vaccination The characterization of the antigen-specific response of splenic T cells 14/16 days after the boost immuni- zation was performed by intracellular cytokine staining (ICS) and enzyme-linked immunospot (ELISpot) as- says in female BALB/c mice from groups A (control), E and F (vaccinated with 5 mg or 20 mg of recombinant protein RBD fusion heterodimer, respectively). The ICS data indicate that on stimulation with an RBD pep- tide pool, splenocytes from group F displayed significant activation of CD4+ T cells expressing IFN-g (p<0.01), IL-2 (p<0.05) and Th1-like cytokines (IFN-g and/or TNF-a and/or IL-2; p<0.05) compared to the control group (Figure 4A). No significant antigen-specific response of CD4+ T cells expressing TNF-a or IL-4 was observed in group F when compared to the control group. Notably, immunization of mice with a lower RBD dose (group E) did not induce CD4+ T cells secreting IFN-g, TNF-a, IL-2 and/or IL-4 after the in vitro restimulation. Furthermore, splenocytes from group F showed activation of CD8+ T cells ex- pressing IFN-g (0.05