key: cord-0327991-d9egxof5
authors: Bellier, Bertrand; Saura, Alicia; Luján, Lucas A.; Molina, Cecilia R.; Lujan, Hugo D.; Klatzmann, David
title: A thermostable oral SARS-CoV-2 vaccine induces mucosal and protective immunity
date: 2021-09-09
journal: bioRxiv
DOI: 10.1101/2021.09.09.459634
sha: b541f951e751a3bcdfc7c1f60c562d90f6eda589
doc_id: 327991
cord_uid: d9egxof5

An ideal protective vaccine against SARS-CoV-2 should not only be effective in preventing disease, but also in preventing virus transmission. It should also be well accepted by the population and have a simple logistic chain. To fulfill these criteria, we developed a thermostable, orally administered vaccine that can induce a robust mucosal neutralizing immune response. We used our platform based on retrovirus-derived enveloped virus-like particles (e-VLPs) harnessed with variable surface proteins (VSPs) from the intestinal parasite Giardia lamblia, affording them resistance to degradation and the triggering of robust mucosal cellular and antibody immune responses after oral administration. We made e-VLPs expressing various forms of the SARS-CoV-2 Spike protein (S), with or without membrane protein (M) expression. We found that prime-boost administration of VSP-decorated e-VLPs expressing a pre-fusion stabilized form of S and M triggers robust mucosal responses against SARS-CoV-2 in mice and hamsters, which translate into complete protection from a viral challenge. Moreover, they dramatically boosted the IgA mucosal response of intramuscularly injected vaccines. We conclude that our thermostable orally administered e-VLP vaccine could be a valuable addition to the current arsenal against SARS-CoV-2, in a stand-alone prime-boost vaccination strategy or as a boost for existing vaccines.

Current field observations show that a protective vaccine against SARS-CoV-2 is likely the 20 only means of controlling the pandemic 1-3 . To fulfill this promise, these vaccines should 21 ideally be effective in preventing infection and virus transmission and, importantly, well 22 accepted by the population. In underdeveloped countries, vaccines should also have a 23 simple logistic chain 4,5 . 24

Regarding efficacy, years of vaccine research have demonstrated that vaccine protective 25 effects rely in large part on systemic neutralizing antibodies, while local cytotoxic T cell 26 responses are for the most part responsible for virus eradication after a productive 27 infection 6 . Moreover, for upper respiratory tract infection, a robust mucosal immunity is 28 likely required to minimize virus transmission 7-9 . Regarding vaccine hesitancy, oral 29 administration would favor acceptance and minimize the risk of adverse events 10 . 30

Regarding logistics, oral administration would also ease mass vaccination, and a 31 thermostable vaccine would ensure a much-simplified logistic chain. Likewise, an optimal 32 vaccine against SARS-CoV-2 should be thermostable, orally administered and able to 33 induce a robust mucosal neutralizing immune response. 34

To tackle the challenge of producing such a vaccine, we used our platform based on 35 retrovirus-derived enveloped virus-like particles (e-VLPs) that has been developed to 36 generate neutralizing antibody (NAb) 11 . Indeed, these e-VLPs have the same lipid 37 membrane as the cell they derive from. Likewise, virus envelope proteins that VLPs 38 express have the same conformation as they have on the lipid membrane of an infected 39 cell, and on the virus itself. As NAbs are mostly targeted to conformational structures, e-40 Fluid collection. Blood was collected weekly from the retro-orbital sinus of hamsters and 104 serum was separated and stored at −80 °C. BAL was collected through the trachea by 105 injection-aspiration of 1 mL of PBS with protease inhibitors. 

Designing and selecting the immunogens. 122

Initially, the spike protein S of SARS-CoV-2 was evaluated and several variants for 123 stabilization of the receptor-binding domain (RBD) and stabilization in the pre-fusion state 124 were designed (Fig.1A) . Point mutations, Cys-molecular clamps, furin-cleavage site 125 elimination, and Proline (Pro) substitutions 23,24 were generated and cloned. Spike protein 126 variants that conserved their own cytoplasmic tail (CT), although it can be advantageous 127 to swap it for the CT of VSV-G to improve pseudotyping onto eVLPs, or that were modified 128 to delete its ER retention signal were also designed. Then, those VLPs were produced and 129 validated for the correct composition as described 25 . VSP-pseudotyped VLPs were orally 130 administered to BALB/c mice and the level of serum IgGs was determined by ELISA (Fig. 2) . The levels of serum IgG after immunization with Sst1 were higher than with the wild-type 162 spike, and the addition of M further increased the IgG response. The levels of serum IgGs were also higher than those of IgA were, as expected. However, these values are 164 comparable to if not higher than those obtained after immunization with other vaccine 165 formulations 36,37 or those found in plasma from convalescent patients 38, 39 . 166

Altogether, these results clearly show that plain e-VLPs are already good immunogens 167 when administered by injection in the absence of any adjuvant, highlighting that VLPs are 168 structures well recognized by the immune system. These responses are strongly increased 169 when VSPs were present on e-VLPs according to the VSPs' intrinsic TLR4-dependent 170 adjuvant properties 25, 35 . 171

Oral administration of the same validated immunogens showed that the absence of the 173

Giardia VSP decorating the different e-VLPs led to no detectable immune response, most 174 likely due to destruction of the VLPs in the upper small intestine (Fig. 4) . Additionally, the 175 modification of the CT of S appeared detrimental in inducing either serum IgG (Fig. 4 left)  176 or IgA (Fig. 4 right) as compared with S having the wild-type CT. However, serum IgG and 177

IgA titers were augmented when M was incorporated into the VSP-e-VLPs (Fig. 4) . 178

Noteworthy, the serum IgA induced by the VSP-decorated e-VLPs were higher after oral 179 administration (Fig. 4 right) than after i.m. injections (Fig. 3 right) . 

We then compared the best e-VLPs expressing the Sst1 spike and M proteins with the e-195

VLPs expressing a wild-type spike for the generation of neutralizing antibodies. 196

Interestingly, there was no NAb generated after i.m. injection when the VSP was not 197 present on the VLPs (Fig. 6) , highlighting its adjuvant effect. With the VSPs present, both 198 wild-type and stabilized S was able to generate NAbs, as observed for the different 199 commercial vaccines already being administered to humans 1,40-42 . Remarkably, the titer of 200 NAbs generated after oral administration are equivalent to those generated after i.m. 201 injections (Fig. 6) , highlighting the efficiency of VSP-e-VLPs as immunogens. 202

Given the incomplete level of protection afforded by some vaccines and the constant 204 emergence of new viral variants, these orally administered immunogens could possibly be 205 good boosts for existing vaccines. For these reasons, an oral boost was applied to animals 206 previously vaccinated by i.m. injections. In these animals, a third dose of the oral 207 formulation containing VSP, stabilized S and M induced a major increase in the levels of 208

IgA in BAL as compared to those that only received two doses intramuscularly (Fig. 7) . 209

Animals immunized with e-VLPs in which stabilized S and M were present on the particles 211 with or without VSP pseudotyping were challenged with SARS-CoV-2 and the clinical 212 response of the hamsters was determined by monitoring their weight 21 . Control animals 213 lost weight during the two weeks following the viral challenge (Fig. 8) and then recovered, 214 as reported for experimental infections in hamsters 20,21 . Hamsters that were only 215 immunized intramuscularly were not fully protected as they had only a slightly lower 216 weight loss. In contrast, oral immunization with VSP-e-VLPs fully prevented weight loss, 217 whether or not the M protein was present, and similarly to animals that were immunized 218 by injection first and then boosted orally (Fig. 8) . 219 220 Discussion 221

The differences observed between the same formulations administered either orally or 222 intramuscularly in these animals suggest that although the oral route is expected to show 223 a higher degree of variation among animals, this was not the case. This could be explained Besides its ease, oral administration is known for also having the advantage of triggering 245 better mucosal immunity. This is indeed the case here, with high levels of plasma but also 246 bronchoalveolar lavage IgA detectable only after oral administration. This is an obvious 247 advantage for a vaccine against SARS-CoV-2, as it should reduce viral transmission. In this 248 line, SARS-CoV-2 was still detected in BAL of i.m. vaccinated macaques that otherwise 249 appeared protected from infection. Whether a better mucosal response, as afforded by 250

VSP-e-VLPs, will completely sterilize challenged macaques requires further investigation. 251

We have not tested the specific T cell response in this study. However, it is known that e-252

VLPs do induce robust cellular responses; indeed, using VSP-HA-VLPs, a strong cytotoxic T 253 lymphocyte response was generated that was able to kill HA-expressing tumor cells 25 . 254

Moreover, the IgG and IgA responses here are notoriously T cell-dependent and the good 255 antibody response thus attests to a good T cell response 44 . In this line, we previously 256 showed that the fusion of a viral peptide to Gag, the retroviral protein precursor that 257 drives the formation and release of the viral particle/VLPs, produces additional strong T 258 cell responses against this peptide 12 . The fusion to Gag of large fragments or the SARS-259

CoV-2 N structural protein, or a stretch of immunodominant and/or conserved peptides, 260 would be a mean to further enhance the immunogenicity of VSP-e-VLPs and enhance 261 protection against variants. 262 SARS-CoV-2 e-VLPs and VSP-e-VLPs could be used as a stand-alone vaccine, likely with a 263 prime-boost scheme of administration. VSP-e-VLPs are thermostable 19 , retaining their 264 properties at room temperature and tolerating several freeze-thaw cycles, and could thus 265 be particularly advantageous for vaccination in countries where refrigeration of vaccine 266 supplies is problematic. VSP-e-VLPs could also be used as a boost for other vaccine 267 designs. In this regard, it is still unknown how long the protection afforded by the currently used vaccines will last. The follow-up of infected patients indicates that, at least 269 for some patients, the persistence of NAbs and the duration of protection might last a few 270 months 45 . These findings, plus the advent of viral variants, make it likely that the global 271 population will need to boost the immune response of vaccinees regularly. For some 272 vaccine designs, and particularly those based on adenoviral vectors, the re-administration 273 of the same vector might not be very efficient due to the immune response generated 274 against the vector. For these, a boost with VSP-e-VLPs might be particularly interesting. 275

For other vaccine designs, and especially if repeated administrations are needed over the 276 years, an orally administered vaccine might be more acceptable. 277

The SARS-CoV-2 pandemic calls for vaccination of very large groups of people. This 278 requires a suitable production of vaccine with an excellent safety profile. Noteworthy, we 279 contributed to the design of an anti-CMV e-VLP vaccine based on our e-VLP platform that 280 has already been used in patients, demonstrating scalable GMP production and an 281 excellent safety profile 15,17,46 . 282 Altogether, given the specific issues of each vaccine design (thermostability, side effects, 283 lack of mucosal immunity induction, immunogenicity against the vector, among other 284 benefits), the availability of multiple vaccines against SARS-CoV-2 improves our chances of 285 controlling the pandemic. In this regard, a thermostable orally administered e-VLP vaccine 286 will be a valuable addition to the current arsenal against this virus.

Immune response to SARS-CoV-2 variants of concern in vaccinated 289 individuals

Immune life history, vaccination, and the dynamics of SARS-291

CoV-2 over the next 5 years

Potential effects of vaccinations on the prevention of COVID-19: 293 rationale, clinical evidence, risks, and public health considerations

Vaccine hesitancy, vaccine refusal and the 296 anti-vaccine movement: influence, impact and implications

Vaccine Hesitancy: Causes

Consequences, and a Call to Action

A guide to vaccinology: from basic principles to new 301 developments

Induction of secretory immunity and memory at mucosal surfaces. 303 Vaccine

Mucosal immunity and vaccines

Mucosal vaccines -fortifying the frontiers

Intranasal and oral vaccination with protein-309 based antigens: advantages, challenges and formulation strategies

Enveloped virus-like particle platforms: 312 Vaccines of the future?

Virus-like particle-based vaccines against hepatitis C 315 virus infection

DNA vaccines expressing retrovirus-like particles are efficient 317 immunogens to induce neutralizing antibodies

Replication-competent vectors 319 and empty virus-like particles: new retroviral vector designs for cancer gene therapy or 320 vaccines

An enveloped virus-like particle vaccine expressing a stabilized 322 prefusion form of the SARS-CoV-2 spike protein elicits potent immunity after a single dose

A prime-boost strategy using virus-like particles pseudotyped for 326 HCV proteins triggers broadly neutralizing antibodies in macaques

Enveloped virus-like particle expression of human 329 cytomegalovirus glycoprotein B antigen induces antibodies with potent and broad 330 neutralizing activity

Using Protozoan Surface Proteins for 332 Effective Oral Vaccination

Efficient oral vaccination by bioengineering virus-like particles 334 with protozoan surface proteins

Syrian hamsters as a small animal model for SARS-CoV-2 infection 336 and countermeasure development

Oral SARS-CoV-2 Inoculation Establishes Subclinical Respiratory 339 Infection with Virus Shedding in Golden Syrian Hamsters

In vitro susceptibility of 10 clinical isolates of SARS coronavirus to 342 selected antiviral compounds

Structural analysis of full-length SARS-CoV-2 spike protein from 344 an advanced vaccine candidate

Structure-based design of prefusion-stabilized SARS-CoV-2 spikes

Efficient oral vaccination by bioengineering virus-like particles 348 with protozoan surface proteins

Cryo-EM structure of the 2019-nCoV spike in the prefusion 350 conformation

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G 352 Increases Infectivity of the COVID-19 Virus

Spike mutation D614G alters SARS-CoV-2 fitness

Analysis of the SARS-CoV-2 356 spike protein glycan shield reveals implications for immune recognition

Site-specific 359 glycan analysis of the SARS-CoV-2 spike

The SARS-CoV-2 envelope and membrane proteins modulate 361 maturation and retention of the spike protein, allowing assembly of virus-like particles

SARS-CoV-2-reactive T cells in healthy donors and patients with 364 COVID-19

Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans 366 with COVID-19 Disease and Unexposed Individuals

Selective and cross-reactive SARS-CoV-2 T cell epitopes in 368 unexposed humans

Using Protozoan Surface Proteins for 370 Effective Oral Vaccination

A potent neutralizing human antibody reveals the N-terminal domain 372 of the Spike protein of SARS-CoV-2 as a site of vulnerability

Neutralizing antibody titres in SARS-CoV-2 infections

Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-378

Throughput Single-Cell Sequencing of Convalescent Patients' B Cells

Robust T Cell Immunity in Convalescent Individuals with 381 Asymptomatic or Mild COVID-19

A systematic review of SARS-CoV-2 vaccine candidates

A Single-Dose Intranasal ChAd Vaccine Protects Upper and 385 Lower Respiratory Tracts against SARS-CoV-2

A promising inactivated whole-virion SARS-CoV-2 387 vaccine. The Lancet Infectious Diseases

Retrovirus-Based Virus-Like Particle Immunogenicity and Its 389 Modulation by Toll-Like Receptor Activation

Enhanced SARS-CoV-2 Neutralization by Secretory IgA in vitro

What defines an efficacious COVID-19 vaccine? A review of 394 the challenges assessing the clinical efficacy of vaccines against SARS-CoV-2

The Status of Vaccine Development Against the Human 397 Cytomegalovirus

Acknowledgments: This work was supported by grants from FONCYT (PICT-E 0234, and 401 PICT-2116), CONICET (D4408), and UCC (80020150200144CC) of Argentina

Forster Award of the Alexander von Humboldt Foundation of Germany to H.D.L., and 403 recurrent funding by Sorbonne University and INSERM to DK

Author contributions: HDL & DK conceived the project. BB, HDL & DK designed the 405 experiments, and supervised students and technicians. AS, LL, CM

proteins used as immunogens. Structural elements include the S1 and S2 ectodomains 415 derived from the original Wuhan variant (Swt) in which specific mutation were inserted. 416The native furin cleveage site was mutated (RRAR à QQAQ) in all variants (Sst 1 -Sst 5 ) to be 417 protease resistant. Specific subsitution (in red) and respective position were indicated. 418The spike variants with a CT modified to abolish ER retention (modCT) were also 419 generated.