Tina Fink, Vida Forstnerič, Iva Hafner-Bratkovič, Sara Orehek, Žiga Strmšek, Mateja Manček-Keber, Peter Pečan, Hana Esih, Špela Malenšek, Jana Aupič, Petra Dekleva, Tjaša Plaper, Sara Vidmar, Lucija Kadunc, Mojca Benčina, Neža Omersa, Gregor Anderluh, Florence Pojer, Kelvin Lau, David Hacker, Bruno Correia, David Peterhoff, Ralf Wagner, Roman Jerala
Effective and safe vaccines against SARS-CoV-2 are highly desirable to prevent casualties and societal cost caused by Covid-19 pandemic. The receptor binding domain (RBD) of the surface-exposed spike protein of SARS-CoV-2 represents a suitable target for the induction of neutralizing antibodies upon vaccination. Small protein antigens typically induce weak immune response while particles measuring tens of nanometers are efficiently presented to B cell follicles and subsequently to follicular germinal center B cells in draining lymph nodes, where B cell proliferation and affinity maturation occurs. Here we prepared and analyzed the response to several DNA vaccines based on genetic fusions of RBD to four different scaffolding domains, namely to the foldon peptide, ferritin, lumazine synthase and β-annulus peptide, presenting from 6 to 60 copies of the RBD on each particle. Scaffolding strongly augmented the immune response with production of neutralizing antibodies and T cell response including cytotoxic lymphocytes in mice upon immunization with DNA plasmids. The most potent response was observed for the 24-residue β-annulus peptide scaffold that forms large soluble assemblies, that has the advantage of low immunogenicity in comparison to larger scaffolds. Our results support the advancement of this vaccine platform towards clinical trials.
Covid-19 is a pandemic viral disease caused by SARS-CoV-2 that emerged in 2019 and infected >23 millions of people across the world while the number of casualties is approaching 1 million. Since we currently lack an effective treatment of the disease and containment of the virus without imposing high cost for the society, vaccination seems to be the best hope to stop the waves of infection that continue to spread throughout the world. An effective vaccine should trigger formation of a protective humoral and cell mediated immune response against the viral components that will either inhibit viral entry and replication or kill virus-infected cells. Different vaccination platforms for the presentation of viral components have been used, including inactivated or attenuated viral particles, purified proteins, mRNA, plasmid DNA, nonreplicating viruses; each of them with particular features. The advantages of DNA plasmid delivery including the speed of adaptation to new targets, cost effective production, stability at ambient temperature without the need for a cold chain make it a potentially attractive vaccination platform, although no DNA plasmid vaccine has been approved for humans so far. Antibodies triggered by a vaccine should preferentially be focused to the domains and epitopes that can prevent viral recognition of the receptor, block viral fusion with cell membrane or interfere with viral replication in other ways. The majority of vaccines in current clinical trials are based on trimeric full length spike protein or its stabilized devivatives, through which the virus attaches to the host cell receptor ACE2. In this case antibodies against diverse surface exposed epitopes of the Spike protein are generated, where some of them may not prevent recognition of the ACE2 receptor or fusion and may even facilitate viral entry through an antibody dependent enhancement mechanism (ADE), as suggested before for SARS CoV and MERS CoV, as a highly undesirable property of a vaccine. Therefore focusing immune response to the receptor binding domain (RBD) of the Spike protein may increase the probability of inducing neutralizing antibodies. Tertiary structure of the RBD in the complex with the ACE2 receptor has been determined. Masking the receptor binding domain by the antibodies can prevent viral binding to the receptor and subsequent infection of target cells. Indeed it has been shown that RBD induces formation of neutralizing antibodies, similar as demonstrated for the related pathogenic coronaviruses and monoclonal antibodies targeting RBD have demonstrated effectiveness in preclinical studies.
The ability of viral proteins or its domains to induce formation of antibodies depends on the structure and size of antigen. Viral surface proteins are typically presented to the immune system of the host in the form or nanoparticles tens of nm in diameter that present tens of copies of viral proteins. It has been shown that cells respond better to particles that present multiple copies of the antigen in comparison to the monomeric proteins, due to clustering of B cell receptors on B-lymphocytes, increased avidity of multimeric proteins and facilitated entry of particles above 20 nm to the lymph nodes Affinity maturation, class switching and memory cell formation takes place in germinal centers inside lymph node follicles, therefore transport of protein antigens to lymph nodes is preferable in comparison to the insoluble protein antigen aggregates. It has been shown that nanoparticles above 50 nm are retained longer inside lymph node follicles and presented on the dendrites of follicular dendritic cells. Presentation of the target antigen domain in the multimeric form may be accomplished by chemical conjugation of a viral protein domain to the nanoparticle or by a genetic fusion to the scaffold-forming polypeptide domain. Examples of the attachment of immunogenic domains to the virus-like capsule are capsid proteins of bacterial, plant or animal viruses such as Qβ, HPV, JCV, HBcAg, cowpea chlorotic mottle virus capsids and many others, nonviral protein compartments such as ferritin, lumazine synthase and encapsulin and de novo designed protein or DNA cages.
Comparison of different scaffolding platforms could unravel which of them triggers the strongest response. An important issue of this strategy is that antibodies or cellular response may also be targeted against the scaffolding domain or a delivery vector, which could impair the efficiency of subsequent immunizations with the same vaccine type eliminate cells that produce the components of the scaffold proteins. Implementation of the smallest scaffolding domain could therefore represent an advantage. Scaffolds used so far that form nanoparticles typically comprise 100-200 amino acid residues (e.g. ferritin, Qβ, lumazine synthase). On the other hand fusion to peptide tags with high aggregation propensity has also been used for vaccines. In this case the amyloid or helical assembly-promoting peptide tag induced formation of large fibers that were decorated with peptide epitopes. For the induction of most efficient antibody response it is important that the antigens readily reach the lymph nodes where affinity maturation of the adaptive immune response takes place. Particles with sizes below ~500 nm can traffic through the lymphatic system into the germinal centers and can be taken up by antigen presenting cells such as dendritic cells, therefore large insoluble aggregates are likely suboptimal.
Here we compared several strategies to present the RBD of the spike protein of SARS-CoV-2 on nanoscale particles, with stoichiometry in the range of 1 to >60 copies. Those assemblies were genetically encoded and implemented in animal studies as a DNA vaccine encoding fusion proteins for secretion from mammalian cells. We show that RBD fused to the scaffolding domains strongly increased the titer of RBD-specific antibodies that also recognized the Spike protein, neutralized interaction between the Spike and ACE2 receptor and provided protection in a surrogate infection assay. The proportion of the amino acid residues of the scaffold in those constructs ranged from as small as 8% to ~50% and those representing smallest fraction generated very low level of scaffold-directed antibodies even in 1 or 2 booster applications. Interestingly the response was best with fusion to a β-annulus peptide (RBD-bann) that had the propensity to generate large soluble oligomers rather than for other large structurally well defined scaffolds. Titers of antibodies against RBD after single immunization was most potent for larger RBD assemblies (RBD-bann and RBD-AaLS) and for all scaffolds substantially larger than for the monomeric RBD. This DNA vaccine also induced cytotoxic lymphocytes and γIFN producing lymphocytes. Our results warrant the advance of this type of vaccines into clinical trials. This type of nanoscaffolded assemblies could be implemented in either DNA, mRNA, viral or isolated protein-based platforms or combinations thereof in well considered prime-boost regimens.
Molecular model structures of designed nanovaccines were prepared by performing homology modelling with Modeller (version 9.23). Structures PDB ID: 6VW and PDBID: 6VSB were used as templates for the receptor binding domain (RBD). For the RBD-ferritin construct, composed of 24 domains, PDB ID: 3EGM was used as template for the ferritin cage. RBD-foldon-RBD was modelled as a trimer based on the template PDB ID: 1RFO. For RBD-bann, homology modelling was used to first construct a model of the trimeric subunit based on the structure of the tomato bushy stunt virus (TBSV) (PDB ID: 2TBV). Twenty trimeric subunits were then assembled into a RBD-decorated protein cage by employing icosahedral symmetry characteristic for TBSV, resulting in a 60-mer with a diameter of approximately 26 nm. Model of the design RBD-AaLS, composed from 60 subunits, was built by employing PDBID: 1HQK as template for the lumazine synthase scaffold.
DNA constructs were prepared using conventional methods based either on synthetic DNA (Twist, IDT) or purchased commercially (Genscript) or clones from plasmids with viral proteins generously provided by Prof. Nevan Krogan, UCSF. The construct encoding prefusion ectodomain of the SARS-CoV-2 Spike protein was a generous gift from Prof. Jason McLellan, University of Texas, Austin. The RBD domain of Spike (residues R319-S591) was codon-optimized for H. sapiens and synthesized (Genscript) into pcDNA3.1 (+) with a human pregnancy specific glycoprotein 1 signal peptide at the N-terminus and a 3C-protease cleavage site followed by a His-tag at the C-terminus. ACE-2 (hsACE-2, residues S19 to D615) was codon optimized for H. sapiens and cloned into the pTwist_CMV_BetaGlobin_WPRE_Neo (Twist Biosciences). It is preceded with a human pregnancy specific glycoprotein 1 signal peptide, and C-terminally tagged with a 3C protease cleavage site, twin-strep tag and 10x His tag.RBD domain of the Spike protein encompassing residues P330 to K521 was fused with polypeptide scaffolds and inserted in pcDNA3.1 (+) vector with a CD45 signal peptide at the N-terminus. β -annulus scaffold peptide from the tom
