To understand if mucosal vaccination with ancestral or variant-specific spike antigens could protect hamsters from disease caused by omicron and delta variants of concern (VOCs), we vaccinated hamsters with a non-replicating recombinant adenovirus type 5 (rAd5) that expresses a transgene of interest in the same gene cassette as a molecular adjuvant, a short nucleotide sequence that forms a hairpin RNA which acts as an innate immune antagonist within the same cell ( Figure 1A ) (23, 24). Hamsters (n=6/group) were immunized with 1x10⁹ infectious units (IU) of rAd5-S-omicron or rAd5-S-delta vaccines and compared to hamsters immunized with rAd5-S-Wuhan via the intranasal route in preparation for viral challenge ( Figures 1B, C ). Additionally, one group from the cohort with rAd5-S-delta received an oral administration of rAd5-S-Wuhan prior to challenge with delta virus ( Figure 1C ). Serum samples were collected at D-1 prior to primary vaccination, D27 (one day prior to boost), and D41 (two weeks post boost) and IgG and IgA were measured. Additionally, nasal washes and samples from oropharyngeal swabs were collected at these timepoints for the assessment of mucosal IgA ( Figure 1B ). Timepoints were collected before and after administration of boosting dose so that the effect of multiple vaccine doses on binding antibody levels and on antibody avidity could be assessed. At D56, one month post boost, animals were challenged with the homologous VOCs. The challenge doses were chosen based on previous Syrian hamster titration studies performed by Bioqual, Inc (Rockville, MD) where detectible levels of viral replication and/or disease could be observed. Samples from oropharyngeal swabs were collected each day after challenge and lung tissue was collected at the end of challenge for pathology and measurement of infectious viral load by TCID50. Infection with the omicron BA.1 variant causes only mild disease in hamsters and thus the challenge was ended after six days, as opposed to the 10 days observed with the delta variant challenge, so that pathology could be more easily assessed before the animals fully recovered.

Serum was collected at indicated days post vaccination and IgG levels were quantified using the Meso Scale Discovery (MSD) platform utilizing electrochemiluminescent detection. As no spike-specific hamster antibodies exist for use as standards, values are reported as relative light units (RLU) and compared to either placebo-vaccinated or baseline (D-1) samples, rather than normalized to a standard curve. Vaccination with both rAd5-S-Wuhan and rAd5-S-omicron elicited cross reactive IgG to Wuhan, omicron and other VOC spike proteins after both prime and boost ( Figure 2A ). Further, a significant increase in binding antibodies to both Wuhan and omicron spike proteins was observed after boost vaccination with rAd5-S-Wuhan, but not with rAd5-S-omicron (p = 0.0119 and 0.0026, respectively), which reached its maximum level of binding by D27 ( Figure 2A ). Next, we tested whether the serum IgG at D41 were cross-reactive to other spike and RBD proteins of VOCs in addition to omicron. rAd5-S-Wuhan elicited cross-reactive IgG to variant spike and RBD proteins. In comparison, vaccination with rAd5-S-omicron still elicited cross-reactive IgG to Wuhan alpha, beta, gamma, and delta variant antigens tested, albeit at significantly reduced levels of binding antibody compared to rAd5-S-Wuhan ( Figures 2B, C ) (p < 0.005 – p <0.0001). Vaccination with rAd5-S-omicron and rAd5-S-Wuhan generated similar levels of binding IgG to omicron antigens ( Figures 2B, C ).

To determine if intranasal administration of rAd5-S vaccines elicited mucosal IgA, oral swab samples from D-1, D27, and D41 were analyzed for RBD and spike-specific IgA. Because experimental variability may be observed between swab samplings, specific IgA was first normalized to total IgA and expressed as fold rise over D-1 so that a trend could be observed. Both Wuhan and omicron vaccinations elicited increased RBD and spike-specific IgA compared to samples from pre-immune and placebo animals ( Figure 2D ). Similar to the serum samples, there were increases in binding following boost vaccination with rAd5-S-Wuhan, whereas boosting with rAd5-S-omicron did not increase the level of binding observed after prime vaccination ( Figure 2D ). Although not normalized to total IgA, raw values of spike-specific IgA mirrored these results ( Figure S1A ). As a second measure of mucosal IgA, nasal washes were collected at D-1, D27, and D41 of the study and evaluated for RBD and spike-specific IgA. Again, relative levels of spike specific IgA were determined by first normalizing to total IgA, then calculating the fold rise over D-1. Similar trends were seen when comparing antibody responses obtained from nasal washes of immunized animals, with all groups having increased omicron RBD and spike-specific IgA by D41 ( Figure 2E , Figure S1B ).

In human phase I clinical trials using the same vaccine platform delivered to the ileum via enterically coated tablets, long lasting secretory IgA was observed in mucosal secretions (22). As a proxy for oral tablet delivery in humans, the second cohort of hamsters included a group that were vaccinated via oral gavage in parallel to groups with intranasal administration in preparation for challenge with the delta variant ( Figure 1C ). We sought to understand if the delivery of rAd5-S-Wuhan by oral gavage could generate cross-reactive antibodies in serum, nasal, and oral samples. Additionally, we wanted to see if immune responses from oral administration compared similarly to that of intranasal administration.

Serum from vaccinated animals was analyzed for cross-reactive IgG against spike and RBD proteins. As this cohort of hamsters was vaccinated in preparation for challenge with the delta variant, Wuhan and delta antigens were assayed. Vaccination by oral gavage elicited significantly higher binding IgG to Wuhan and delta antigens over D-1 (p < 0.0001 at D27 and D41) ( Figure 3A ). In addition, intranasal vaccination with rAd5-S-Wuhan and rAd5-S-delta were able to generate cross-reactive IgG compared to baseline levels to both Wuhan and delta spike proteins (p < 0.001) ( Figure 3A ). Next, we tested whether the serum IgG at D41 from this cohort were cross-reactive to other spike and RBD proteins of VOCs in addition to delta. Intranasal administration of rAd5-S-Wuhan gave higher IgG responses to RBD and spike proteins than oral administration, although this difference was only significant in the case of Wuhan and delta spike proteins (p = 0.0472 and p = 0.0497) ( Figures 3B, C ). Intranasal administrations of rAd5-S-Wuhan and rAd5-S-delta elicited similar levels of serum IgG against spike and RBD, although vaccination with rAd5-S-delta performed significantly better against beta (p = 0.0423), delta (p = 0.0012), and omicron (p 0.0043) spike proteins, and beta (p = 0.0374), gamma (p = 0.0481), and delta RBD proteins (p = 0.0059) ( Figures 3B, C ). Oral administration of rAd5-S-Wuhan was not statistically compared to intranasal administration of rAd5-S-delta as both the antigen and route differed between the two groups.

In addition to quantifying the presence of serum IgG, we also observed IgA in mucosal secretions. Oropharyngeal swabs were tested for the presence of specific IgA as described in Figure 2D . At the timepoints tested, all vaccinated groups elicited higher IgA when compared to the pre-immune and placebo samples ( Figure 3D ). Although not normalized to total IgA, raw values of spike-specific IgA mirrored these results ( Figure S1A ). Lastly, we looked at nasal wash IgA as we were particularly interested to see if oral administration of rAd5 could elicit nasal IgA. By D41, all vaccinated groups tested showed cross-reactive antibodies in the nasal secretions ( Figure 3E ; Figure S1D ). No specific or total nasal IgA was observed from the hamsters vaccinated intranasally with rAd5-S-Wuhan at D27, likely due to a technical error. Thus, this data point was excluded from analysis ( Figure 3E ; Figure S1D ).

To determine if the antibodies generated by boost vaccination increased antibody specificity, we tested whether antibody avidity increased following boost vaccination by measuring the avidity index (25). In all iterations, vaccination with the matched antigen elicited the highest affinity antibodies against the homologous variant protein ( Figure 4 ). The avidity index of both serum IgG and IgA and oral IgA to spike and/or RBD generally increased with boost. Interestingly, the IgA avidity index to omicron spike elicited by rAd5-S-omicron vaccination started high and stayed high ( Figure 4C ), whereas serum IgG increased in avidity with boost ( Figure 4A ). This phenomenon was only observed in the serum with respect to the spike protein as there was no superiority observed by antibody avidity towards omicron RBD with antibodies generated by both vaccinations giving equivalent avidity indexes ( Figure 4B ) and equal levels of binding antibody to omicron RBD ( Figure 2C ). Avidity of IgA from oral swabs also increased with boost, but this increase in avidity index was very slight ( Figure 4C ). Additionally, the avidity indexes of oral IgA were higher at D27 compared to the avidity index of serum IgG at D27 ( Figure 4 ). This observation is in line with previous findings that secretory IgA is of higher avidity due to its valency. Lastly, as a comparator to other VOC proteins, the avidity index of antibodies generated after rAd5-S-Wuhan and rAd5-S-omicron vaccination were measured against delta spike and RBD proteins. In line with what was seen before, rAd5-S-Wuhan generated antibodies of the higher avidity when compared with the omicron vaccination, again indicating that Wuhan-based vaccination may elicit broadly cross-reactive antibodies across many VOCs, particularly in comparison to rAd5-S-omicron.

One month after vaccination series (D56), the omicron cohort of hamsters was challenged with 4.84x10⁴ TCID50 of the omicron BA.1 variant of SARS-CoV-2 and infection proceeded for six days. Both rAd5-S-Wuhan and rAd5-S-omicron protected hamsters from weight loss due to infection (p = 0.0186 and 0.0013 respectively) ( Figures 5A, B ). In addition, viral shedding, as measured by qRT-PCR of genomic RNA (gRNA) from oropharyngeal swab, was significantly reduced by both rAd5-S-Wuhan and rAd5-S-omicron compared to placebo (p = 0.0001 and p < 0.0001, respectively) ( Figures 5C, D ). At the end of the infection period, the lungs of hamsters were collected and TCID50 was determined. Three of the six hamsters in the placebo group had detectable levels of infectious virus in their lungs while no detectable virus was recovered from the lungs of vaccinated hamsters ( Figure 5E ).

The second cohort of hamsters was vaccinated with rAd5-S-Wuhan and rAd5-S-delta in preparation for challenge with the delta variant of SARS-CoV-2. Hamsters were infected with 3.48x10³ TCID50/animal and monitored for 10 days post challenge for weight loss and viral load. Animals vaccinated intranasally with rAd5-S-Wuhan and rAd5-S-delta experienced no weight loss compared to the placebo control. There was a small dip in body weight for hamsters vaccinated orally with rAd5-S-Wuhan, however all animals recovered quickly and all vaccine groups were significantly protected from weight loss compared to placebo (p=<0.0001 for all groups) ( Figures 5F, G ). Viral loads from oropharyngeal swabs were analyzed by qRT-PCR for subgenomic (sgRNA) and gRNA. Viral loads and viral replication were significantly reduced in the rAd5-S-Wuhan oral and intranasal groups with the greatest reduction of viral RNA levels in the homologous rAd5-S-delta vaccination group (p=0.0007, <0.0001, and <0.0001 respectively) ( Figures 5H–K ).

In addition to monitoring clinical symptoms of disease above, the lungs of infected animals were observed for signs of inflammation, as measured by bronchiolo-alveolar hyperplasia, and bronchoalveolar/interstitial and vascular/perivascular inflammation. Fewer animals from all vaccinated groups, as compared to placebo, showed signs of inflammation ( Figures 6A–D ). rAd5-S-delta was the most efficacious with the greatest decrease in severity of lung pathology compared to placebo animals in the cohort challenged with the delta variant. Bronchiolo-alveolar hyperplasia was observed in almost all animals challenged with either omicron or delta VOCs; however, severity was decreased compared to placebo treated animals. In the omicron cohort, there were no significant differences in efficaciousness between rAd5-Wuhan and r-Ad5-omicron. Overall, there were fewer animals with lung inflammation from the cohort infected with the delta variant than that of omicron. However, this is likely due to the timing of infection as by day 10 post infection with the delta variant, most of the animals had recovered in body weight ( Figures 5F, G ) whereas the cohort infected with the omicron variant had not yet recovered ( Figures 5A, B ).

Male Golden Syrian hamsters (n=6) aged 6-8 weeks, were vaccinated according to the groupings in Figure 1B with a dose of 1x10⁹ infectious units (IU) diluted in 1xPBS in a total volume of 100 µL for intranasal vaccination (50 µL/nostril) and 1 mL total volume for oral gavage. Vaccinations occurred four weeks apart. Serum, nasal washes, and oral swabs were collected on day -1 (D1), day 27 (D27), and day 41 (D41) as described below. At D52, the animals were anesthetized by injecting 80 mg/kg ketamine and 5 mg/kg xylazine via intramuscular route in preparation for challenge. Animals were challenged with an appropriate dose via intranasal administration using a total volume of 100 μL per animal (50 μL/nostril), administered dropwise. SARS-CoV-2 delta variant (ATCC NR-56116 LOT#: 70047614) was dosed at 3.48x10³ TCID₅₀ per hamster and SARS-CoV-2 omicron variant (ATCC NR-56486 LOT#: 70049695 was dosed at 4.8x10⁴ TCID₅₀ per hamster as pre-determined in titration studies performed by Bioqual, Inc. Animals were monitored daily for any abnormal clinical observation and body weights were recorded. All in-life animal handling occurred at Bioqual, Inc (Rockville, MD).

Adenovirus type 5 (rAd5) vaccines were generated based on the published spike sequence of SARS-CoV-2 Wuhan (Genbank Accession No. MN908947.3), delta (GISAID Accession No. EPI-ISL-2570775), and omicron (BA.1) (GISAID Accession No. EPI_ISL_6699769) variants. These sequences were inserted into a recombinant plasmid containing rAd5 sequence that is deleted in E1 and E3 genes. The respective transgenes were cloned into the E1 region that additionally contains a downstream molecular dsRNA adjuvant in the gene cassette and is expressed together with the transgene in the target cell. rAd5 particles were generated by transfection of transgene containing DNA into Expi293F cells (ThermoFisher Scientific) to generate rAd5 virions which were purified by CsCl density centrifugation (38, 51).

Blood collection was performed one day prior to each vaccination (D-1, D27) and two weeks post final vaccination (D41). Prior to collection, animals were anesthetized with up to 5% isoflurane and blood was collected via the retro-orbital sinus vein. Samples were allowed to clot for 30 minutes – 1 hr at room temperature and centrifuged at 9300 x g for 10 minutes. Serum was stored at -80˚C.

Hamster nasal wash samples were collected following anesthesia using isoflurane; a soft tipped catheter was used to flush 400 µL of 1X PBS into the nasal cavity and a collection device was placed under the opposite nostril to collect the fluid. The recovery yield was documented to be approximately 200 to 250 µL. Hamster oral swab samples for antibodies were collected using a sterile flocked swab (Copan Diagnostics FLOQSwabs 501CS01) by inserting into the mouth and swabbing both cheeks for at least 30 seconds. The swab was snap-frozen until elution in 200 µL 1X PBS, 0.25 M NaCl (Corning cat#21-031-CV, Acros Organics cat#327300025) and collected by centrifugation. Post-challenge swab samples were collected in cryovials containing 1 mL 1X PBS for qPCR viral load testing.

Vaccine-induced serum IgG specific to SARS-CoV-2 spike and RBD variants was measured with MSD V-PLEX COVID-19 Serology kit panel 22 and 23 (MSD cat#K15563U, K15571U). Plates were coated, blocked, washed, and incubated with sample and detection antibody according to the manufacturer’s instructions. Serum samples were diluted at 1:4000 in Diluent-100 (MSD cat#R50AA). A goat anti-hamster IgG antibody (Invitrogen cat#31115) was sulfo-tagged with the MSD GOLD SULFO-TAG NHS-Ester Conjugation Pack (MSD cat#R31AA) and diluted to 1 µg/mL in Diluent-100. Plates were read on a Meso QuickPlex SQ 120 instrument (MSD) and sample values were reported as relative light units (RLUs). Data analysis was performed in GraphPad Prism (Version 9.4.1).

Vaccine-induced mucosal IgA specific to SARS-CoV-2 variant spike and RBD were measured with the Mesoscale Discovery (MSD) 4-spot U-PLEX Development Pack (MSD cat#K15229N). Spots were linked with anti-Syrian hamster IgA antibody (Brookwood Biomedical cat#sab3001a) biotinylated with the EZ-Link Sulfo-NHS-LC-Biotin kit (Thermo Fisher Scientific cat#A39257) and either biotinylated Wuhan, delta, or omicron variant SARS-CoV-2 spike trimer and RBD proteins (ACROBiosystems cat#SPD-C82E9, SPN-C82Ec, SPD-C82Ed, SPN-C82Ee, SPD-C82E4). Plates were coated, blocked, washed, and incubated with sample and detection antibody according to the manufacturer’s recommendations. Nasal samples were diluted at 1:15 and oral samples were diluted at 1:5 in Diluent-100 (MSD cat#R50AA). The anti-Syrian hamster IgA antibody was sulfo-tagged with the MSD GOLD SULFO-TAG NHS-Ester Conjugation Pack (MSD cat#R31AA) and diluted with Diluent-100 to 2 µg/mL for the nasal samples and 1 µg/mL for the oral samples. Plates were read on a Meso QuickPlex SQ 120 instrument (MSD). Samples were reported as relative light units (RLUs). Due to the variability in mucosal sampling, samples were normalized by total IgA and expressed as fold change was reported as vaccinated over unvaccinated sample. Data analysis was performed in GraphPad Prism (Version 9.4.1).

Serum and mucosal antibody avidity was determined by MSD using either MSD V-PLEX COVID-19 Serology kit panel 25 (MSD cat# K15583U-2) or U-PLEX Development Pack (MSD cat#K15229N). U-plex spots were linked with biotinylated Wuhan, delta, or omicron variant SARS-CoV-2 spike trimer and RBD proteins (ACROBiosystems cat#SPD-C82E9, SPN-C82Ec, SPN-C82Ee, SPD-C82E4). Plates were coated, blocked, washed, and incubated with sample and detection antibody according to the manufacturer’s recommendations and as described above with the exception that after incubation of sample, an additional 10-minute incubation of 3M urea in 1X PBS + 0.05% Tween-20 was performed. Avidity was calculated as (RLU_urea)/(RLU_PBST)*100.

Tissue sections were homogenized in 0.5 mL cold medium (DMEM + 10% FBS + 0.05 mg/mL gentamicin) for and centrifuged (2000 xg, 4°C, 10 minutes) to remove debris and supernatants collected. Clear flat-bottom 96-well culture microplates (BD Falcon Cat. #: 353072) were seeded with Vero TMPRSS2 cells at 2.5x10⁴ cells per well in growth media (DMEM + 10% FBS + 1% Penicillin/Streptomycin + 10 µg/mL Puromycin) and incubated at 37°C, 5% CO2 until 80-100% confluent. A 10-fold dilution series of processed tissue sample was plated in growth media (DMEM + 2% FBS + 1% Penicillin/Streptomycin + 10 µg/mL Puromycin). Plates were incubated at 37°C, 5% CO₂ for 4 days. After incubation, the presence of cytopathic effects (CPE) was plated, and the TCID50 value per mL was calculated using the Read-Muench formula based on the tissue section weight, homogenization volume, and sample volume used in the assay assigned a calculated TCID₅₀ titer per gram of tissue. Data analysis was performed in GraphPad Prism (Version 9.4.1).

Post-challenge oral swabs were analyzed at the DUKE Human Vaccine Center IVQAC. The assay for SARS-CoV-2 quantitative Polymerase Chain Reaction (qPCR) detects total RNA using the WHO primer/probe set E_Sarbeco (Charité/Berlin). A QIAsymphony DSP, automated sample preparation platform along with a virus/pathogen DSP midi kit, were used to extract viral RNA from 800 μL of sample. A reverse primer specific to the envelope gene of SARS-CoV-2 (5′-ATA TTG CAG CAG TAC GCA CAC A-3′) was annealed and then reverse transcribed into cDNA using SuperScript™ III Reverse Transcriptase with RNase Out. The resulting cDNA was treated with RNase H (Thermo Fisher Scientific) and then added to a custom 4x TaqMan™ Gene Expression Master Mix containing primers and a fluorescently labeled hydrolysis probe specific for the envelope gene of SARS-CoV-2 (forward primer 5′-ACA GGT ACG TTA ATA GTT AAT AGC GT-3′, reverse primer 5′-ATA TTG CAG CAG TAC GCA CAC A-3′, probe 5′-6FAM/AC ACT AGC C/ZENA TCC TTA CTG CGC TTC G/IABkFQ-3′). SARS-CoV-2 RNA copies per reaction were interpolated using quantification cycle data and a serial dilution of a highly characterized custom DNA plasmid containing the SARS-CoV-2 envelope gene sequence. Mean RNA copies per milliliter were then calculated by applying the assay dilution factor with a limit of detection (LOD) approximately 62 RNA copies per mL of sample.

SARS-CoV-2 N gene subgenomic mRNA was measured by a one-step RT-qPCR. To generate standard curves, a SARS-CoV-2 E gene sgRNA sequence, including the 5′UTR leader sequence, transcriptional regulatory sequence, and the first 228 bp of E gene, was cloned into a pcDNA3.1 plasmid. For generating SARS-CoV-2 N gene sgRNA, the E gene was replaced with the first 227 bp of N gene. The respectively pcDNA3.1 plasmids were linearized, transcribed using MEGAscript T7 Transcription Kit, and purified with MEGAclear Transcription Clean-Up Kit. The purified RNA products were quantified on Nanodrop, serial diluted, and aliquoted as E sgRNA or N sgRNA standards. RNA extracted from samples or standards were then measured in Taqman custom gene expression assays using TaqMan Fast Virus 1-Step Master Mix and custom primers/probes targeting the E gene sgRNA (F primer: 5′ CGATCTCTTGTAGATCTGTTCTCE 3′; R primer: 5′ ATATTGCAGCAGTACGCACACA 3′; probe: 5′ FAM-ACACTAGCCATCCTTACTGCGCTTCG-BHQ1 3′) or the N gene sgRNA (F primer: 5′ CGATCTCTTGTAGATCTGTTCTC 3′; R primer: 5′ GGTGAACCAAGACGCAGTAT 3′; probe: 5′ FAM-TAACCAGAATGGAGAACGCAGTG GG-BHQ1 3′). Standard curves were used to calculate E sgRNA in copies per mL; the limit of detections (LOD) for N sgRNA assays were approximately 31 copies per mL of sample.

At necropsy, the left lung of each animal was collected and placed in 10% neutral buffered formalin. Tissue sections were trimmed and processed to hematoxylin and eosin (H&E) stained slides and examined by a board-certified pathologist at Experimental Pathology laboratories, Inc. (EPL) (Sterling, VA). Findings were graded from one to five (1=minimal, 2=mild, 3=moderate, 4=marked, 5=severe).