key: cord-0813408-h980fqkg
authors: Zabaleta, Nerea; Bhatt, Urja; Hérate, Cécile; Maisonnasse, Pauline; Sanmiguel, Julio; Diop, Cheikh; Castore, Sofia; Estelien, Reynette; Li, Dan; Dereuddre-Bosquet, Nathalie; Cavarelli, Mariangela; Gallouët, Anne-Sophie; Pascal, Quentin; Naninck, Thibaut; Kahlaoui, Nidhal; Lemaitre, Julien; Relouzat, Francis; Ronzitti, Giuseppe; Thibaut, Hendrik Jan; Montomoli, Emanuele; Wilson, James M.; Le Grand, Roger; Vandenberghe, Luk H.
title: Durable immunogenicity, adaptation to emerging variants and low dose efficacy of AAV-based COVID19 platform in macaques
date: 2022-05-10
journal: Mol Ther
DOI: 10.1016/j.ymthe.2022.05.007
sha: 9b4685aa202db1b631b27f742367ecbfb9d44b02
doc_id: 813408
cord_uid: h980fqkg

The COVID19 pandemic continues to have devastating consequences on health and economy, even after the approval of safe and effective vaccines. Waning immunity, the emergence of variants of concern, breakthrough infections, and lack of global vaccine access and acceptance perpetuate the epidemic. Here, we demonstrate that a single injection of an AAV-based COVID19 vaccine elicits at least 17-month-long neutralizing antibody responses in non-human primates at levels that were previously shown to protect from viral challenge. To improve the scalability of this durable vaccine candidate, we further optimized the vector design for greater potency at a reduced dose in mice and nonhuman primates. Finally, we show that the platform can be rapidly adapted to other variants of concern to robustly maintain immunogenicity and protect from challenge. In summary, we demonstrate this class of AAV can provide durable immunogenicity, provide protection at dose that is low and scalable, and be adapted readily to novel emerging vaccine antigens thus may provide potent tool in the ongoing fight against SARS-CoV-2.

with SARS-CoV-2 after week 9 and shown to have near-sterilizing protective immunity 33 . 126 Importantly, all 4 animals in Figure 1A presented neutralizing antibody titers in range with the 127 titers observed in protective immunity ( Figure 1B ) at all timepoints measured from week 8 to 128 week 70. This study is ongoing and intended for long-term follow-up of Spike neutralizing 129 responses. Additionally, cross-reactivity with the better escape VOC variants (Beta, Delta and although detectable, while subgenomic RNA was undetectable in all except one AC1 NHP 159 (Figures 2F and 2G) . This observation was confirmed by the analysis of lung lymph nodes by 160 PET scan ( Figure 2H ). Vaccinated animals did not show an activation of lymph nodes after 161 challenge, which was observed in control animals, due to an active SARS-CoV-2 infection in the 162 lungs ( Figure 2H ). CT scan did not reveal a significant difference in lung lesions due to the mild 163 phenotype of SARS-CoV-2 infection in NHPs ( Figure S1E ). Lung histology analysis of 164 vaccinated animals 30 to 35 days after challenge suggests less lesions due to COVID19 infection 165 in AC1 vaccinated animals while no significant difference was observed between the scores of 166 controls and AC3 vaccinated animals ( Figure 2I ). 167 Antibody responses after challenge increased in all the animals, including controls (Figures 2A, 168 2B, S1A and S1B). Figure 2A illustrates that 2 of the animals treated with AC1 were non-169 responders, since the antibody levels after challenge followed the same trend as the unvaccinated 170 and challenged controls. All AC3 animals however did seroconvert prior to the challenge, 171 indicating that at the 10 11 gc level the AAVCOVID platform can perform reliably.

Biodistribution was assessed for AC1 and AC3 at all doses tested ( Figure S2A ). Results show 173 that AAVCOVID primarily biodistributes to the injected muscle, the regional lymph node and 174 spleen, while only minimal systemic biodistribution is observed in tissues like liver; at a dose of 175 10 11 gc approximately 1 vector genome per 10,000 diploid genomes is detected in any of the 4 176 liver lobes. 177 In summary, the AC1 and AC3 dose reduction challenge studies indicated (a) that AC3 at the 178 10 11 gc dose led to 100% seroconversion and a strong T-cell response, yet was unable to achieve 179 the previously demonstrated level of protection in the upper and lower airway as AC1 at the 10x 180 higher dose 33 and (b) that AC1 at the 10 11 gc dose was unable to achieve full seroconversion, notwithstanding use of an identical viral vector capsid to AC3 carrying a superior antigen (full 182 length prefusion stable spike compared to S1). The only remaining variable in the constructs 183 between AC1 and AC3 were the regulatory regions of the promotor (SV40 in AC1 and CMV in 184 AC3) and the polyadenylation sequences (SV40 in AC1 and a bovine growth hormone (bGH) in 185 AC3).

Second generation AAVCOVID platform is optimized for capsid and promoter 187 Based on the experience with AC1 and AC3 in the above studies and prior experiment 33 , we 188 sought to further optimize the various characteristics of a broadly applicable vaccine platform: 189 manufacturing, seroconversion, and potency of immunogenicity and protection at the lowest 190 dose possible. We next explore optimizations of both vector capsid (mainly toward optimized 191 and consistency of production) and potency (mainly toward dose reduction).

First, we evaluated the AAV11 serotype, a close homolog of AAVrh32.33. AAV11 is a natural 193 serotype that was isolated from the liver of a cynomolgus monkey 35 as opposed to the 194 AAVrh32.33 which is man-made capsid and therefore more likely to suffer from structural 195 deficits that hamper production and reduce yields 36 . From structural comparison with other 196 known AAV serotypes, AAVrh32.33, AAV4 and AAV12 are the closest related serotypes to 197 AAV11 37 . The VP1 sequence of AAV11 and AAVrh32.33 are 99.7% homologous with 2 amino 198 acid difference (K167R and T259S in AAV11).

To ensure vaccine properties of AAVrh32.33 were retained, AAV11 vectors containing the same 200 cassette as AC1 (SV40 promoter expressing Spp) were produced and tested in mouse 201 immunogenicity studies. 6-8 weeks male and female C57BL/6 mice with 10 11 and 10 10 gc dose 202 of AAV11-Spp vaccine and compared to AAVrh32.33-based AC1 candidate. Spike binding and 203 J o u r n a l P r e -p r o o f neutralizing responses were similar between mice vaccinated with AC1 and AAV11-Spp across 204 doses and genders (Figures 3A and 3B) . Cellular responses to the transgene were also preserved 205 for the AAV11-based candidate, with robust IFN-γ responses against Spike peptides, mainly 206 subunit 1 (S1) peptides and very low IL-4 secretion ( Figures 3C and 3D ). The biodistribution 207 pattern of the vectors was analyzed on day 7 after IM administration, same distribution profiles 208 were observed for AAVrh32.33 and AAV11 with most vector copies in the injected muscle 209 (right gastrocnemius) ( Figure 3E ). Same results were observed in BALB/c mice injected with 210 these vectors ( Figure S3 ). AAV11 was the serotype used for all subsequent of studies.

Based on the observations in the NHP dose reduction studies in Figure 2 , we hypothesized that 212 increasing promoter strength would further optimize the immunogenicity of the AAVCOVID 213 platform. This was further supported by expression data in C57BL/6 that previously 214 demonstrated the CMV driven antigen expression from AC3 was far greater than the SV40 215 expression in AC1 33 . We thus designed AAV expression cassettes to improve the expression of 216 Spp. Spp was chosen as an antigen over S1 as prior studies in mice clearly indicated its 217 superiority for generating neutralizing responses to SARS-CoV-2 and similar antigen designs in 218 the currently FDA-approved vaccines have been highly efficacious and safe in large populations create ACE1, ACM1 and ACC1 vectors, respectively ( Figure 4A and S4A). ACC1 promoter, 227 due to the long size of the promoter, resulted in an oversized recombinant genome, which could 228 lead to fragmented genome packaging and lower vector yields at scale 38,39 . In vitro expression 229 studies revealed improved expression of Spike protein in cells infected with ACM1 and ACC1 230 compared to AC1 ( Figure S4B ). This was confirmed in C57BL/6 female animals that received 231 these candidates by measuring Spike mRNA levels in the injected muscle 7 days after a 10 11 gc 232 IM injection ( Figure 4B and S4C). Higher expression resulted in significantly higher RBD-233 binding antibody levels in animals vaccinated with ACM1 compared to AC1-SPA and ACE1 at 234 3 doses ranging from 2x10 9 gc to 10 11 gc. Interestingly, ACM1 achieved full seroconversion with 235 a single dose as low as 2x10 9 gc per mouse, while 20% of AC1-SPA animals at the same dose Figure 4D ). ACC1 also showed increased transduction in the injected muscle and increased 239 antibody responses, in line with ACM1 ( Figures S4C and S4D ).

To further validate the efficacy of ACM compared to AC at the low 10 11 gc dose, we performed a 242 cynomolgus study in which animals were challenged with SARS-CoV-2. An ACM vector was 243 generated expressing the Beta strain of SARS-CoV-2.

with modestly delayed kinetics, in line with the experience with AC1 or AC3 33 ( Figure 5B and 250 5C). IFN-γ-mediated cellular responses as measured by ELISPOT on peripheral blood 251 mononuclear cells (PBMCs) were elevated by week 4 ( Figure 5D ). Cross-neutralization was 252 measured by RBD-binding, ACE2 inhibition and pseudovirus assay ( Figure S5 ). Binding 253 antibody levels were very similar for different VOC RBDs ( Figure S5A ), but ACE2 inhibition 254 and pseudovirus neutralization were superior for Beta and Gamma variants compared to for 255 Wuhan, Alpha and Delta ( Figures S5B and S5C ). Regarding active replication of the virus, only one animal presented sgRNA detectable above the 264 limit of quantification on day 3 ( Figure 5F ). sgRNA was not detectable in BAL samples on day 3 265 ( Figure 5F ). These data demonstrated a protective effect from infection of ACM-Beta from 266 SARS-CoV-2 Beta infection.

Biodistribution of the ACM-Beta vector was found to be consistent with AC1 at the same dose, 268 primarily directed to the injected muscle, draining lymph node and spleen. Systemic 

Gene-based vaccines can be designed and developed more quickly to respond to epidemic threats 284 or the emergence of novel pathogenic strains, e.g. VOCs in the case of COVID19. The 285 responsiveness of the gene-based platforms such as mRNA is primarily due to the DNA-based 286 template (e.g. plasmid DNA) as a substrate for the production process and the generic nature of 287 the production and purification process independent of the encoded antigen. This is in contrast to 288 other vaccine approaches that require viral or recombinant protein production which is slower 289 and specific to even subtle changes of the antigen.

AAV-based vaccines indeed rely on a plasmid-based substrate to initiate production that can be 291 generated within days following the emergence and sequencing of a novel pathogen. Its 292 production and purification are dependent on the viral capsid which is kept consistent using the 293 vectors specific to each VOC were developed and tested in vivo for immunogenicity as 295 illustrated in Figure 7 . First, the SARS-CoV-2 Beta VOC is reported to be highly antigenically 296 distinct to other variants, and hence is significantly less neutralized in individuals exposed to or 297 immunized with the ancestral Wuhan Spike. Interestingly however, individuals infected with 298 Beta may develop stronger cross-reactivity to Wuhan and most of the other VOCs 40 . Indeed, 299 C57/BL6 mice also developed high titers of neutralizing antibodies against Wuhan, Alpha and 300 Gamma VOCs following immunization with ACM-Beta compared to the neutralization potency 301 to the Beta VOC itself ( Figure 7A ). In line with prior observations, cross-neutralization was 302 lower for the Delta VOC 41 .

Next, we sought to evaluate the consistency of performance in terms of immunogenicity of the 304 ACM platform in the context of Wuhan, Beta, and Delta Spike antigens. Figure 7B and 7C 305 illustrates that both binding and neutralizing antibody titers are analogous for each of these 306 vaccine candidates. VOC cross-reactivity of each of these vaccine responses was interrogated 307 and illustrates their unique antigenic profile ( Figure 7C ). A separate more recent study included 308 ACM-Omicron encoding the Omicron Spp similarly demonstrating similar potency to ACM-1 309 and ACM-Delta in mice 12 days after vaccination ( Figure 7D ). Interestingly, cross-reactivity is 310 between Wuhan and Omicron ( Figure 7D ) is greatly reduced compared to Wuhan and Beta 311 binding IgG antibody ( Figure S7 ). Same trend was observed in cross-neutralizing antibody titers 312 ( Figure 7E ).

The constantly evolving COVID19 pandemic requires vaccines and vaccine regimens to adapt to 317 the rapidly changing threat. Past experience demonstrates that vaccines are indeed a key tool in 318 managing the ongoing crisis, however for vaccines to eventually suppress the epidemic that tool 319 may need to be sharpened; rapid global deployment is needed to prevent the emergence of new 320 variants, vaccines need to have breadth and/or adaptability to be effective against current and 321 future VOCs, protection from disease needs to be durable, and ideally also prevent transmission.

Here, we evaluate and optimize an AAV-based COVID19 vaccine platform in its potential to 323 address some of the limitations that have been exposed.

Previously, we demonstrated proof-of-concept data that a first generation AAVCOVID candidate convalescence of an ICU cohort in humans. We further demonstrated previously that this AAV-331 based vaccine product is high yielding in production and was adapted to a scalable 332 manufacturing process. The vaccine product was found to be stable when stored for 1 month at 333 room temperature and at least 12 weeks at 4ºC in a simple modified saline buffer.

These preclinical data, if recapitulated in human subjects, suggest that the profile of 335 AAVCOVID may overcome some of the limitations of current approved COVID-19 vaccine 336 (e.g. durable immunogenicity from a single dose, improved storage stability, potential for strong 337 upper airway protection). As articulated by Dr. Fauci and colleagues most recently 42 , there is a continued need to fight epidemics and specifically future coronavirus outbreaks by accelerating 339 the development of improved vaccine technologies specifically on attributes AAVCOVID may 340 hold based on the presented data.

However, for this technology to be further consider toward clinical translation, several 342 outstanding concerns warrant addressing that speak to safety, efficacy in humans, and feasibility.

The studies presented here specifically sought to improve on potency for a lower dose to be 344 sufficiently robust in terms of seroconversion and level of immunogenicity. A target of 10 11 gc 345 was established based on models to attain feasibility for scaled production and sufficiently low 346 production cost in line with vaccine applications. immunogenicity and efficacy data, we were able to redesign the vaccine platform. By correlating 353 AC1 and AC3 relative performance vis à vis their distinct design features, we hypothesized that 354 increasing antigen expression would permit a potency increase and a dose reduction. However, 355 due to size constraints the CMV promoter used in AC3 could not be transferred to AC1.

Therefore, we designed a construct with a minimal CMV promoter to achieve higher expression 357 within the packaging limitation of AAV to drive the pre-fusion stable full-length SARS-CoV-2 (d.p.e.) while blood was taken at 2, 3, 4, 7, 10 and 14 days , Bronchoalveolar lavages (BAL) 433 were performed using 50 mL sterile saline at 3 and 11 d.p.e. CT scans were performed at D3 and 434 D7 to quantify lung lesions.

Blood cell counts, hemoglobin and hematocrit were determined from EDTA blood using a 436 DXH800 analyzer (Beckman Coulter). of-sars-cov-2-institut-pasteur-paris.pdf?sfvrsn=3662fcb6_2). The limit of detection was 582 estimated to be 2.87 log10 copies of SARS-CoV-2 sgRNA per mL and the limit of quantification 583 was estimated to be 3.87 log10 copies per mL. For imaging sessions, animals were first anesthetized with Ketamine (10mg/kg) + Metedomidine 588 (0.05mg/kg) and then maintained under isofluorane 2% in a supine position on a patient warming blanket (Bear Hugger, 3M) on the machine bed with cardiac rate, oxygen saturation and 590 temperature monitoring.

CT was performed under breath-hold 5 minutes prior to PET scan for attenuation correction and 592 anatomical localization. The CT detector collimation used was 64 × 0.6 mm, the tube voltage 593 was 120 kV and intensity of about 150mAs. Automatic dose optimization tools (Dose Right, Z-mounted on coated glass slides (Superfrost+, Thermo) and stained with haematoxylin and eosin 612 (H&E) with automated staining processor (Autostainer ST5020, Leica).

Each slide was scored in 20 different spots at X40 magnification (Plan Apo 40X, 0.95 614 Numerical aperture, 0.86 mm² per Field of View). On each spot, 5 different parameters were 615 assessed: Septal cellularity, Septal fibrosis, Type II pneumocytes hyperplasia and alveolar 616 neutrophils. A systematic histopathology scoring was used and described in Table S1 . Each score 617 were then cumulated for each assessed field of view for cranial and caudal lobes. separately to minimize competitive PCR multiplexing issues prior to analysis and delta delta Ct 648 normalization 51 . The limit of detection of the assay was 10 copies/reaction, therefore, wells with 649 less than 10 copies were considered negative. 

Efficacy and Safety of the 729 mRNA-1273 SARS-CoV-2 Vaccine

Safety and Efficacy of 733 the BNT162b2 mRNA Covid-19 Vaccine

Safety and Efficacy of 737 NVX-CoV2373 Covid-19 Vaccine

Safety and 741 Efficacy of Single-Dose Ad26.COV2.S Vaccine against Covid-19. The New England journal 742 of medicine

Duration of Protection 745 against Mild and Severe Disease by Covid-19 Vaccines

Durability of mRNA-1273 749 vaccine-induced antibodies against SARS-CoV-2 variants

Differential Kinetics of

Immune Responses Elicited by Covid-19 Vaccines. The New England journal of medicine

SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the 768 COVID-19 Virus

COVID-19-neutralizing antibodies predict disease severity and survival

A longitudinal study of SARS-CoV-2-775 infected patients reveals a high correlation between neutralizing antibodies and COVID-776 19 severity

Assessing transmissibility 780 of SARS-CoV-2 lineage B.1.1.7 in England

Detection of a 784 SARS-CoV-2 variant of concern in South Africa

Genomics 788 and epidemiology of the P.1 SARS-CoV-2 lineage in Manaus

was driven by the persistence of endemic lineages and P.1 793 emergence

SARS-CoV-2 Spike Mutations

the Second Wave of COVID-19 in

Microorganisms 9

Multiple SARS-CoV-2 variants escape neutralization by vaccine-induced humoral 801 immunity

Omicron boost in vaccinated macaques elicits comparable B cell expansion, neutralizing 856 antibodies and protection against Omicron. bioRxiv

An AAV-based, 860 room-temperature-stable, single-dose COVID-19 vaccine provides durable 861 immunogenicity and protection in non-human primates

Clinical 865 and virological data of the first cases of COVID-19 in Europe: a case series. The Lancet

Two novel adeno-associated 868 viruses from cynomolgus monkey: pseudotyping characterization of capsid protein

Naturally occurring 872 singleton residues in AAV capsid impact vector performance and illustrate structural 873 constraints

Capsid Structures Reveals Clade-Specific Features

Packaging capacity of adeno-associated virus 878 serotypes: impact of larger genomes on infectivity and postentry steps

Effect of genome size on AAV vector packaging

Reactive Neutralizing Antibody Responses Elicited by SARS-CoV-2 501Y.V2 (B.1.351). The 886

Reduced 889 neutralization of SARS-CoV-2 B.1.617 by vaccine and convalescent serum

Universal Coronavirus Vaccines 892 -An Urgent Need

Evaluation of a novel multiplexed assay for 896 determining IgG levels and functional activity to SARS-CoV-2

A trivalent HCV vaccine elicits broad and synergistic polyclonal antibody 901 response in mice and rhesus monkey

Detection of 2019 904 novel coronavirus (2019-nCoV) by real-time RT-PCR

Europeen sur les maladies transmissibles = European communicable disease bulletin 25

Virological 909 assessment of hospitalized patients with COVID-2019

Digital Photon Counting PET/CT: The Physics 913

Powering Precision Nuclear Medicine

Radiological findings from 81 patients with COVID-19 pneumonia in Wuhan, China: a 916 descriptive study. The Lancet. Infectious diseases

Time Course of Lung Changes at Chest CT during Recovery from 920

COVID-19)

Hydroxychloroquine 924 use against SARS-CoV-2 infection in non-human primates

Analysis of relative gene expression data using real-927 time quantitative PCR and the 2(-Delta Delta C(T)) Method