Efficacy of ChAdOx1 vaccines against SARS-CoV-2 Variants of Concern Beta, Delta and Omicron in the Syrian hamster model

ChAdOx1 nCoV-19 (AZD1222) is a replication-deficient simian adenovirus–vectored vaccine encoding the spike (S) protein of SARS-CoV-2, based on the first published full-length sequence (Wuhan-1). AZD1222 was shown to have 74% vaccine efficacy (VE) against symptomatic disease in clinical trials and over 2.5 billion doses of vaccine have been released for worldwide use. However, SARS-CoV-2 continues to circulate and consequently, variants of concern (VoCs) have been detected, with substitutions in the S protein that are associated with a reduction in virus neutralizing antibody titer. Updating vaccines to include S proteins of VoCs may be beneficial over boosting with vaccines encoding the ancestral S protein, even though current real-world data is suggesting good efficacy against hospitalization and death following boosting with vaccines encoding the ancestral S protein. Using the Syrian hamster model, we evaluated the effect of a single dose of AZD2816, encoding the S protein of the Beta VoC, and efficacy of AZD1222/AZD2816 as a heterologous primary series against challenge with the Beta or Delta variant. We then investigated the efficacy of a single dose of AZD2816 or AZD1222 against the Omicron VoC. As seen previously, minimal to no viral sgRNA could be detected in lungs of vaccinated animals obtained at 5 days post inoculation, in contrast to lungs of control animals. Thus, these vaccination regimens are protective against the Beta, Delta, and Omicron VoCs in the hamster model.

Introduction At the end of 2019, the causative agent of COVID-19, severe acute respiratory syndrome 36 coronavirus 2 (SARS-CoV-2), was first detected in Wuhan, China 1,2 . As of February 2 nd 2022, SARS-37 CoV-2 has infected an estimated 380 million people, causing more than 5 million deaths 3 . Its emergence 38 prompted the rapid development of vaccines based on the viral receptor binding protein, spike (S) [4][5][6] . 39 Several vaccines demonstrated efficacy through clinical trials in less than a year 7-11 and were approved 40 for emergency use by different regulatory bodies worldwide. Over 4.4 billion people are estimated to 41 have received at least one dose of COVID-19 vaccination 3 . One of those vaccines is AZD1222 42 (ChAdOx1 nCoV-19), developed by Oxford University and produced by AstraZeneca. AZD1222 is a 43 replication-deficient simian adenovirus-vectored vaccine encoding the non-stabilized S protein of 44 Wuhan-1, one of the first published full-length SARS-CoV-2 sequences 12 . AZD1222 was shown to be 45 highly effective in clinical trials, demonstrating 74% vaccine efficacy against symptomatic disease 7 . A 46 two dose primary series of AZD1222 is approved for usage in more than 170 countries, and more than 47 between binding antibodies or pseudotype VN titer against any S variant and sgRNA in swabs (Extended 113 Data Figure 3). 114 Lung pathology was scored by a board-certified veterinary pathologist blinded to study groups ( Figure 3). 115 SARS-CoV-2-related pathology was observed in all animals of the control group. On day 3 post Beta 116 challenge, minimal-to-moderate acute bronchiolitis was observed affecting less than 1% of the lung. 117 Histological lesions consisted of a moderate subacute broncho-interstitial pneumonia affecting between 118 30-50% of pulmonary tissue. Lesions were characterized by broncho-interstitial pneumonia centered on 119 terminal bronchioles and extending into the adjacent alveoli. Alveolar septa were expanded by edema 120 fluid and leucocytes. SARS-CoV-2 antigen staining was numerous within bronchiolar epithelium on day 121 3, whereas this had mostly moved to type I and II pneumocytes on day 5. In contrast, antigen staining in 122 the vaccinated groups was relatively low; samples obtained from the prime only group were mostly 123 negative, whereas antigen staining in samples obtained from the prime-boost group was between none to 124 moderate ( Figure 3, Extended Data Figure 4). 125 To investigate the efficacy of a vaccine optimized for the Beta S against the Delta variant, new groups of 126 hamsters vaccinated as described above were challenged with 10 4 TCID50 of the Delta variant via the 127 intranasal route. As observed upon challenge with the Beta variant, vaccinated animals did not show any 128 weight loss throughout the experiment, whereas control animals did lose weight ( Figure 4A). Indeed, 129 differences in weight loss between the control and vaccinated groups were significant on day 7 ( Figure  130 4B). High levels of sgRNA could be detected on both day 3 and 5 in lung tissue of control animals (12/12 131 samples positive, median of 5.6 x 10 9 and 5.0 x 10 9 copies/gram, respectively). In contrast, the majority of 132 lung tissue obtained from vaccinated animals was negative for sgRNA, viral sgRNA was detected in 1/6 133 samples on each day for the prime boost group, versus 2/6 on day 3 and 0/6 on day 5 for the prime only 134 group ( Figure 4C). Significant differences in sgRNA detected in oropharyngeal swabs were limited to 135 day 3 (both vaccinated groups) and day 5 (prime only group) compared to controls ( Figure 4D). In 136 contrast to what we observed in animals inoculated with the Beta variant, we did not see a decrease in the 137 window of shedding of vaccinated animals compared to control animals. Area-under-the-curve analysis (as a measurement of total amount of sgRNA detected in oropharyngeal swabs throughout the 139 experiment) showed that animals that received a prime only vaccination shed significantly less than 140 control animals ( Figure 4E). 141 In sera collected on day 5 post challenge, a significantly higher binding antibody titer against Delta S 142 compared to ancestral S was detected in the prime only group (Extended Data Figure 2). In the live VN 143 assay, higher VN titers were found against the Beta VoC compared to ancestral virus in the prime only 144 group (Extended Data Figure 2). Significant differences in the pseudotype VN assay were found between 145 Omicron and Beta as well as Delta VoCs (Extended Data Figure 2), as well as between ancestral and 146 E484K S in the prime only group, and ancestral and K417N S in the prime boost group. (Extended Data 147 Figure 1C). A linear correlation was found between sgRNA in swabs and correlating binding antibodies 148 against ancestral, Beta and Delta S. For live VN titers, a significant correlation was found with the Beta 149 and Delta, but not ancestral, VoCs (Extended Data Figure 5). prime only group, a minimal-to-moderate bronchiolitis was observed in some animals on day 3. In the 153 prime boost group, bronchiolitis was either absent or minimal. This was combined with reduced antigen 154 staining in the bronchiolar epithelium in both groups compared to controls. No pathology or antigen 155 staining was observed on day 5, except for one animal in the prime boost group, which is the only animal 156 that was positive for sgRNA at this time point. 157 We then investigated the efficacy of AZD1222 and AZD2816 against the Omicron VoC. Here, groups of 158 hamsters were vaccinated with a single dose of AZD1222, AZD2816, or ChAdOx1 GFP and serum was 159 collected 14 days post-vaccination ( Figure 6A). Binding antibodies against different S proteins were 160 detected using the Mesoscale V-PLEX SARS-CoV-2 panel 23, in-house optimized for hamster sera. 161 Upon vaccination with AZD1222, binding antibodies were highest for ancestral and Alpha S and lowest 162 for Omicron. Upon vaccination with AZD2816, antibody levels were similar for ancestral, Alpha, Beta, 163 and Gamma S, but dropped for Delta and Omicron S ( Figure 6B). Live VN titers against the Omicron VoC were significantly lower compared to the ancestral variant ( Figure 6C). 28 days after vaccination, 165 animals were challenged with the ancestral variant or Omicron VoC. As previously reported 25 , we did not 166 see weight loss in control hamsters challenged with the Omicron VoC, whereas this weight loss was 167 present in control hamsters challenged with ancestral virus ( Figure 6D). Four animals per group were 168 euthanized on day 3 and day 5. No significant differences in lung:body weight ratio were observed, 169 although lung:body weight ratio was relatively high on day 5 for control hamsters inoculated with 170 ancestral virus ( Figure 6E). As expected, AZD1222 vaccination resulted in significantly reduced viral 171 genome copies in lung tissue ( Figure 6F). However, replication of the Omicron VoC in hamster lung 172 tissue was low and although a reduction in genome copies in lung tissue of vaccinated hamsters was 173 observed, particularly in the animals which received AZD2816, no significance was reached. 174 Oropharyngeal swabs were obtained on day 1-5 and analyzed for sgRNA. Shedding of the control animals 175 infected with the Omicron VoC was similar to control and vaccinated animals that were infected with the 176 ancestral variant, albeit lower on day 1. However, vaccinated animals inoculated with the Omicron VoC 177 had significantly lower shedding on day 3 (both groups), day 4, and day 5 (AZD2816 vaccinated group 178 only) compared to controls ( Figure 6G). Area-under-the-curve analysis was performed on the four 179 animals per group that were euthanized on day 5 as a measure of total amount of virus shed throughout 180 the experiment. Significantly lower shedding was detected in the vaccinated animals infected with the 181 Omicron VoC, but not in those infected with the ancestral virus ( Figure 6H). 182  Thus, we investigated the protective efficacy of the vaccine AZD2816, which encodes the S protein of the 213 Beta VoC, in the hamster model. In contrast to control animals, upon challenge with either the Beta, 214 Delta, or Omicron VoC, little-to-no viral RNA was found at 5 days post challenge in the lower respiratory dose AZD2816, are protective against all three VoCs in the hamster model. 217 Vaccine and variant-specific differences were observed in the different experiments. In the Beta VoC 218 study, lung tissue from 2/6 prime boost vaccinated animals were positive for sgRNA at day 3, combined 219 with higher antigen staining in this group compared to the prime only group. Furthermore, whereas total 220 shedding was reduced in the prime only group compared to controls, this was not the case for the prime 221 boost group. This suggests that initial priming with one VoC S may shape the immune response to 222 subsequent vaccinations. Indeed, our humoral immune response analysis showed higher titers for the Beta 223 S and E484K mutation compared to ancestral or Delta S in the prime only group, but not the prime boost 224 group. In contrast, in the Delta VoC study, the prime boost group appeared to be slightly better protected 225 than the prime only group, mostly evident in pathology scoring. This may be due to the higher quantity of 226 antibodies in the prime boost group compared to the prime only group.  Interestingly, whereas we previously reported on the lack of reduction in virus detected in oropharyngeal 239 swabs when vaccines were given via the intramuscular route 33,34 , vaccinated hamsters inoculated with the 240 Omicron VoC displayed significantly reduced shedding compared to controls, even though shedding of 241 control animals was at levels equal to animals inoculated with the ancestral virus. This difference was 242 particularly evident in the AZD2816 group. Omicron has an E484A mutation in the S protein, whereas the 243 Beta VoC has an E484K mutation. In our pseudovirus VN assays, we show a higher neutralization of 244 pseudotypes with the E484K mutation compared to ancestral S in serum obtained from hamsters that only 245 received the AZD2816 vaccination. It is possible that this also translates to the E484A mutation. 246 However, we did not see an increase neutralization in live virus assays against the Omicron VoC in serum 247 obtained from hamsters vaccinated with AZD2816 compared to those vaccinated with AZD1222. Further 248 research is needed to determine whether the small difference observed between the two vaccines against 249 the Omicron VoC is relevant, and why shedding is reduced. Our study confirms that AZD2816 is immunogenic in the hamster model and protects against infection of 256 the lower respiratory tract against the Omicron, Beta, and Delta VoC. Likewise, a single dose of 257 AZD1222 protects against the Omicron VoC. Furthermore, initial immunization with AZD1222 followed 258 by immunization with AZD2816 results in full protection against the Beta and Delta VoCs, and we 259 predict it will also protect against Omicron. This confirms previous reports that a full antigenic match 260 between the vaccine and the challenged virus is not required for protection of the lower respiratory tract. 261

Acknowledgments 262
We would like to thank Mehul Suthar, Kathleen Cordova, Brian Smith, Jade Riopelle, Lara Myers,  All other authors declare no competing interests. 291

Ethics Statement 296
Animal experiments were conducted in an AAALAC International-accredited facility and were approved 297 by the Rocky Mountain Laboratories Institutional Care and Use Committee following the guidelines put
Body weights were recorded daily. Oropharyngeal swabs were collected in 1 mL of DMEM2. On day 3 321 and 5, 4-6 animals from each group were euthanized and lung samples were taken for qRT-PCR analysis, 322 virus titrations and histopathology. The remaining six animals in each group were monitored daily until 323 day 21. 324

RNA extraction and quantitative reverse-transcription polymerase chain reaction 325
RNA was extracted from DMEM2 containing oropharyngeal swabs using the QiaAmp Viral RNA kit 326 (Qiagen), and lung samples were homogenized and extracted using the RNeasy kit (Qiagen) according to 327 the manufacturer's instructions and following high-containment laboratory protocols. Five μL of 328 extracted RNA was tested with the Quantstudio 3 system (Thermofisher) according to the manufacturer's 329 instructions using viral RNA specific assays 38,39 . A standard curve was generated during each run using 330 SARS-CoV-2 standards containing a known number of genome copies. 331

Virus neutralization 332
Sera were heat-inactivated (30 min, 56 °C). After an initial 1:10 dilution of the sera, two-fold serial 333 dilutions were prepared in DMEM2. 100 TCID50 of SARS-CoV-2 was added to the diluted sera. After a 334 60 min incubation at 37°C and 5% CO2, the virus-serum mixture was added to VeroE6 cells and cells 335 were further incubated for 6 days before assessment of CPE. The virus neutralization titer was expressed 336 as the reciprocal value of the highest dilution of the serum that still inhibited virus replication. Three 337 different positive serum controls were done next to NIBSC sera sample 20/130 by three different 338 technicians, to determine IU/mL equivalent. NIBSC sera readout was 640-1066, compared to reported 339 value at 1300 (1.5x higher). All serum samples were subsequently accompanied by positive controls on 340 the plate. Assays were only approved if positive controls fell within the range previously determined by 341 three technicians. Values were then multiplied by 1.5 to determine IU/mL. 342

Generating lentiviral based pseudotypes bearing the SARS-CoV-2 S protein 343
Lentiviral-based SARS-CoV-2 pseudotyped viruses were generated in HEK293T cells incubated at 37°C, 344 5% CO2 as previously described 40  Vaccinated hamster serum samples were diluted 10,000x, and ChAdOx1 GFP-vaccinated hamster serum 387 samples were diluted 1,000x. The plate was sealed with shaking at room temperature for 2 hours, 388 followed by 3 washes with 1X MSD Wash buffer. An in-house MSD GOLD SULFO-TAG NHS-Ester 389 (MSD, R31AA-2) conjugated goat anti-hamster IgG secondary antibody (Thermo Fischer, SA5-10284) 390 was diluted 10,000x in diluent 100 and 50 µL was applied to each well of the plate. The plate was sealed 391 with shaking at room temperature for 1 hour. After incubation, the plate was washed with 1X MSD Wash 392 buffer as before, and 150 µL of MSD Gold Read Buffer B was added per well. The plate was read 393 immediately by the MSD instrument. Arbitrary units (AU) were assigned to the standard curve of pooled 394 SARS-CoV-2-positive hamster sera, which was used on each plate. AU/mL were calculated using the 395 Lungs were perfused with 10% neutral-buffered formalin and fixed for at least 8 days. Tissue was 398 embedded in paraffin, processed using a VIP 6 Tissue-Tek (Sakura Finetek)   protein (brown, 2 nd and 4 th column), 100x, scale bar = 100 µm; Limited bronchiolitis with epithelial cell 575 necrosis observed on day 3, which was resolved on day 5, in animals that received an AZD2816 576 vaccination. No pathology observed in animals which received an AZD1222 + AZD2816 vaccination. 577 Compared to controls, limited staining of bronchiolar epithelium observed on day 3, which was resolved 578 on day 5 in both vaccine groups. Control animals show progression from bronchiolitis on day 3 to 579 bronchointerstitial pneumonia on day 5, at which point alveolar septa are expanded by edema fluid and 580 leucocytes. Staining of bronchiolar epithelial cells, type I&II pneumocytes, and rare macrophages on both 581 days. Images are representative of observations within 100% of a complete lung section containing all 582 lobes. 583