Construction and rescue of B/HPIV3 expressing the S-6P version of the B.1.617.2/Delta and B.1.1.529/Omicron variants

Previously, we generated and evaluated in hamsters and rhesus macaques the B/HPIV3 vector vaccine candidate B/HPIV3/S-6P, expressing the SARS-CoV-2 S protein of the ancestral Wuhan-Hu-1 strain (GenBank MN908947), prefusion-stabilized by six proline substitutions [23,25,26]. In the present study, we generated B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P expressing the full-length S proteins derived from SARS-CoV-2 B.1.617.2/Delta or B.1.1.529/Omicron variants, respectively, following the same strategy (Fig 1A). The S-Delta-6P and S-Omicron-6P open reading frames (ORFs) (aa 1–1,273) were codon-optimized for human expression and include six proline substitutions to stabilize S in the prefusion form [22,23]. Each S ORF was framed by nucleotide adapters containing the BPIV3 gene start and gene end signal sequences and placed as an additional gene between the N and P genes in the B/HPIV3 vector [25,26] (Fig 1A). B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P were recovered from cDNA by reverse genetics and passaged once on Vero cells to generate passage 2 (P2) working stocks. Working stocks were titrated by dual-staining immunoplaque assay to confirm titers [6.2 and 6.3 log₁₀ plaque-forming units (PFU) per ml for B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P]. Ninety-five and 85% of the plaques generated by the B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P stocks, respectively, were positive for HPIV3 and S antigen by dual-staining immunoplaque assay. Sanger sequencing of overlapping PCR fragments covering the entire genome (excluding regions of the genome ends complementary to the terminal sequencing primers) did not reveal any adventitious mutations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Replication of B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P in hamsters.

(A) Genome map of the B/HPIV3 empty vector, B/HPIV3/S-6P expressing the S-6P stabilized version of the spike protein (S) of the ancestral SARS-CoV-2 strain (blue, [25,26], and of the newly-generated versions B/HPIV3/S-Delta-6P (green) and B/HPIV3/S-Omicron-6P (orange). BPIV3  N, P, M and L genes are in white and HPIV3 F and HN genes in pink. Sequences corresponding to S ORFs (aa 1-1,273) from Wuhan-Hu-1, Delta or B.1.1.529 strains were codon-optimized for expression in humans, the furin cleavage site “RRAR” was removed and replaced by “GSAS” [22], and six proline substitutions were introduced to stabilize the S antigen in the prefusion form [23]. An additional gene encoding the respective S ORF, flanked by a BPIV3 gene start (light grey) and gene end (dark grey) signal sequence, was inserted between the BPIV3 N and P genes [25,26]. (B) Vero cells seeded in 6-well plates were mock-infected or infected with the indicated virus using an MOI of 1 PFU/cell and incubated at 32°C for 48 h . Cell lysates were harvested and analyzed by SDS-PAGE under denaturing and reducing conditions and by western blotting. The SARS-CoV-2 S protein, the BPIV3 N protein and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; included as loading control) were detected using primary antibodies (see Methods) followed by immunostaining with infrared fluorophore labeled secondary antibodies and infrared imaging. Images were acquired and analyzed using Image Studio software (LiCor). The approximate molecular weight in kDa is shown on the left. (C) Timeline of the hamster experiment. Five to six-week-old hamsters in groups of 46 were intranasally immunized with 5 log₁₀ PFU of B/HPIV3 empty vector or B/HPIV3 expressing the S-6P, S-Delta-6P or S-Omicron-6P. (D) On days 3 and 5 post-immunization (pi), five hamsters per group each day were euthanized, and vaccine titers were determined from nasal turbinates (D, left panel) and lungs (D, right panel) by dual-staining immunoplaque assay. The limit of detection (dotted line) is 50 PFU/g of tissue. Geometric mean titers (GMTs) for each group are indicated above x axes. Two-way ANOVA with Sidak post-test; exact p values are indicated for levels of significance p < 0.05. (E) Percentages of S-positive plaques in nasal turbinates on days 5 pi. Due to low level of replication on day 3 pi in nasal turbinatesand on day 3 and 5 pi in lungs, these samples were not investigated. Each animal is represented by a circle and GMTs with geometric standard deviations are shown.

https://doi.org/10.1371/journal.ppat.1012585.g001

We also confirmed that the B/HPIV3 S-expressing vectors expressed the full-length S antigen by western blot. Monolayers of Vero cells in 6-well plates were inoculated with the P2 working stocks of B/HPIV3 and S-expressing versions at an MOI of 1 PFU per cell. After 48 h incubation at 32°C, cell lysates were harvested, and expression of SARS-CoV-2 S and BPIV3 N was evaluated by western blot (Fig 1B). GAPDH was included as a control. Using an S-specific monoclonal antibody, a single strong band with an apparent molecular weight of approximately 200 kDa corresponding to the uncleaved form of SARS-CoV-2 S (S0) was detected in all cases in lysates of cells infected with the B/HPIV3 S-expressing viruses. This confirmed that each B/HPIV3 S-expressing vector efficiently expressed the full-length S antigen.

We also evaluated the genetic stability of B/HPIV3 S-expressing vectors during serial passages on Vero cells. Monolayers of Vero cells in 25 cm² flasks were inoculated with P2 working stocks of B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P or B/HPIV3/S-Omicron-6P using an MOI of 0.01 PFU/cell, and incubated at 32°C. On day 6 post-infection, corresponding to the peak of virus replication [25], supernatants were harvested. One ml of supernatant was used to infect new monolayers of Vero cells in 25 cm² flasks, and this procedure was repeated one more time for a total of three serial passages (P3 through P5). At the end of the experiment, virus titers after each passage were determined by immunoplaque assay (S1 Fig). All viruses replicated efficiently and reached titers between 6.7 to 7.7 log₁₀ PFU/ml at the end of the last passage (P5). Viral RNA was extracted from the virus-containing supernatants collected after the last passage, and a 5.7 kb fragment of each genome was amplified by long-range PCR, containing the complete N gene, the added gene encoding the respective S-6P antigen, and the P gene start signal. PCR products were sequenced by nanopore long-range sequencing. No mutation in the gene start or gene end consensus sequences or in the S ORF was detected in any of the passaged B/HPIV3 S-expressing vectors, suggesting that under these experimental conditions, the S ORF and the surrounding regions are genetically stable.

B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P replicate efficiently in the respiratory tract of hamsters

The replication, immunogenicity, and protective efficacy of B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P were evaluated in five- to six-week-old golden Syrian hamsters, and B/HPIV3/S-6P and the B/HPIV3 empty vector were included as controls (see Fig 1C for timeline of the experiment). On day 0, four groups of 46 hamsters each were immunized intranasally with 5 log₁₀ PFU of B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P, or B/HPIV3/S-Omicron-6P. On days 3 and 5 post-immunization (pi), five hamsters per group were euthanized and nasal turbinates (NTs) and lungs were harvested. Tissue homogenates were prepared, and vaccine virus titers were determined by immunoplaque assay.

Similarly to previous studies, the B/HPIV3 empty vector replicated to high titers on day 3 pi (geometric mean titers [GMTs] of 6.9 and 5.4 log₁₀ PFU/g in NTs and lungs, respectively, Fig 1D, left and right panel). On day 5 pi, B/HPIV3 titers were significantly lower than on day 3 in NTs (6.0 log₁₀ PFU/g) but remained at a level similar to day 3 in lungs (5.6 log₁₀ PFU/g). On day 3, the GMTs of all S-expressing versions (B/HPIV3/S-6P, B/HPIV3/S-Delta-6P, and B/HPIV3/S-Omicron-6P) were significantly lower than those of the empty B/HPIV3 vector [between 32- and 1,995-fold (NTs) and 794- and 2,512-fold (lungs), p < 0.0001]. By day 5 pi, titers in NTs of all three S-expressing vectors had increased to levels comparable to those in B/HPIV3-infected hamsters on day 3 (Fig 1D, left panel). While the pattern of increases in titers of the S-expressing vectors from day 3 to day 5 in the NTs seemed similar, titers of B/HPIV3/S-Omicron-6P were significantly lower compared to those of B/HPIV3/S-Delta-6P on day 3 and compared to those of B/HPIV3/S-6P on day 5. In the lungs, titers of all S-expressing B/HPIV3 vectors had also increased by day 5 (GMTs ranging from 3.8 to 4.5 log₁₀ PFU/g), but peak titers remained lower than those of B/HPIV3 (13-fold for B/HPIV3/S-6P; 79-fold for B/HPIV3-S-Delta-6P and B/HPIV3/S-Omicron-6P). Thus, replication of B/HPIV3-S expressing vectors was delayed in the NTs and delayed and reduced in the lungs due to the insertion of the foreign S gene into the virus genome.

The stability of the S expression by the B/HPIV3 vectors in NTs at day 5 pi was also evaluated using a dual-staining immunoplaque assay (Fig 1E). Between 96 and 98% of the plaques from NT-derived B/HPIV3/S-Delta-6P, B/HPIV3/S-Omicron-6P or B/HPIV3/S-6P expressed S. This suggests substantial stability of S expression in the upper airways. Due to the low titers of B/HPIV3 S-expressing vectors in the NT at day 3 pi and in the lungs at day 3 and 5 pi, stability of S expression was not evaluated in those samples.

B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P induced mucosal and serum anti-S antibody responses in hamsters

We next evaluated the mucosal antibody responses induced by the B/HPIV3 S-expressing vectors using nasal washes (NW) collected on day 21 pi from 18 of the 36 remaining hamsters per immunized group that were selected randomly (Fig 2A). S-specific IgG and IgA titers were determined by IgG/IgA ELISA based on S-6P antigen derived from the ancestral Wuhan-Hu-1 strain. This dual ELISA allows detection of IgA and IgG in the same sample by sequential reads (see Materials and Methods); the IgA detection relies on a highly-sensitive dissociation-enhanced lanthanide fluoride immune assay (DELFIA), providing for a higher sensitivity of IgA detection compared to IgG detection. Thus, IgG and IgA titers cannot be directly compared.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Mucosal and serum antibody responses in immunized hamsters.

(A) IgG (top and bottom left panels), IgA (top and bottom middle panels) and secretory IgA (sIgA) (top right panel) anti-S mucosal antibody titers were evaluated against Wuhan-Hu-1 S (top panels) or B.1.1.529/Omicron S (bottow panels) by ELISA (IgG) or DELFIA (IgA and sIgA) from nasal washes (NW) collected on day 21 post-immunization (pi, n = 18 hamsters per group picked at random). The fraction of animals with titers above the limit of detection is indicated above x axes. (B-E) On day 24 or 25 pi, serum was collected from n = 36 hamsters per group. (B) The IgG (left panels) and IgA (right panels) anti-S antibody titers were evaluated by ELISA (IgG) or DELFIA (IgA) using purified preparations of S antigen from the Wuhan-Hu-1 strain (top row), B.1.617.2/Delta (middle row) or B.1.1.529/Omicron variants (bottom row). ELISA results using purified RBD antigen preparations from the Wuhan-Hu-1 strain are shown in S2 Fig. The limit of detection of reciprocal antibody titers (dotted line) is 1 log₁₀ (A) or 2 (B) log₁₀. (C) Sera collected on day 24 or 25 pi were also analyzed for SARS-CoV-2 neutralizing antibody titers. Live virus neutralization assays were performed using the WA1/2020 strain or B.1.617.2/Delta or BA.1/Omicron variants. Titers were expressed as 50% neutralizing doses. (A-C) Non-parametric Kruskal Wallis test with Dunn post hoc test (Fig 2A and C) and One-way ANOVA with Tukey or Sidak post-test for other data sets; exact p values are indicated for levels of significance p < 0.05. Each hamster is represented by a symbol and medians with interquartile ranges are shown. (B-C) GMTs are indicated above x axes. (D-E) Binding inhibition of soluble ACE2 protein to SARS-CoV-2 S proteins from 20 different variants by serum antibodies from immunized hamsters. Data are represented as a heatmap with the median percent inhibition (from n = 36 hamsters per group) of ACE2 binding relative to a non-serum control indicated for each variant (D), or as a radar plot, with each segment representing a variant, and percent inhibition indicated by concentric circles (E).

https://doi.org/10.1371/journal.ppat.1012585.g002

Using the ancestral S antigen, we detected mucosal anti-S IgG in nasal washes from 8/18 and 7/18 hamsters immunized with B/HPIV3/S-6P or B/HPIV3/S-Delta-6P, respectively (GMT ELISA titer of 1.3 and 1.2 log₁₀, respectively). No hamsters immunized with B/HPIV3/S-Omicron-6P had IgG antibodies to the ancestral version of S detectable in nasal washes (Fig 2A, top left panel). However, 17/18, 12/18 and 4/18 hamsters immunized with B/HPIV3/S-6P, B/HPIV3/S-Delta-6P or B/HPIV3/S-Omicron-6P, respectively, had mucosal anti-S IgA detectable in nasal washes (GMT ELISA titer from 1.04 to 1.7 log₁₀), indicating that all S-expressing vectors induced a mucosal anti-S IgA response (Fig 2A, top middle panel), even though B/HPIV3/S-Omicron-6P was significantly less efficient in inducing a mucosal anti-S antibody response to the ancestral version of S, with a high animal-to-animal variability. We also tested nasal washes for secretory IgA (sIgA) to the Wuhan-Hu-1 S antigen (Fig 2A, top right panel; note that the assay requires undiluted samples; volume limitations restricted the number of analytes in this study). Previous studies found that local mucosal IgA was not induced by parenterally administered vaccines [27]. Thus, detection of S-specific sIgA would unequivocally confirm the induction of mucosal antibody responses by the S-expressing vectors. Secretory IgA has been shown to have higher neutralizing potency compared to monomeric IgA or even IgG [28,29]. Indeed, we found that a subset of hamsters immunized with the B/HPIV3/S-6P (4/15) or B/HPIV3/S-Delta-6P (2/16) exhibited low levels of sIgA in the upper airways.

Since the B/HPIV3/S-Omicron-6P immunized hamsters had low or no NW IgG and only low IgA titers to Wuhan-Hu-1 S antigen, detectable in 4/18 animals, we also evaluated the levels of anti-B.1.1.529/Omicron S IgG and IgA in the upper airways (Fig 2A, lower left and right panels). By B.1.1.529/Omicron S ELISA, only a subset of B/HPIV3/S-6P immunized hamsters had NW IgA or IgG detectable, and none of the B/HPIV3/S-Omicron-6P immunized hamsters had mucosal antibodies detectable in this vaccine-matched assay, suggesting that the low level of mucosal anti-S IgG and IgA detected by Wuhan-1 S ELISA in B/HPIV3/S-Omicron-6P immunized animals might reflect the low level of anti-S IgG and IgA induced by this vaccine in the upper airways, rather than a mismatch of vaccine and ELISA antigens. As expected, no mucosal anti-S IgG, IgA or sIgA antibody response was detectable in hamsters immunized with the B/HPIV3 empty-vector control.

We also determined the serum IgG and IgA titers elicited by the B/HPIV3 vectors to the S proteins of the ancestral strain as well as B.1.617.2/Delta and B.1.1.529/Omicron variants at day 24 or 25 pi using dual IgG/IgA ELISAs. Purified S proteins from the indicated strains were used as antigens (Fig 2B). An additional ELISA was performed to evaluate antibodies binding the receptor binding domain (RBD, derived from the ancestral strain; S2 Fig).

All three S-expressing B/HPIV3 vectors elicited robust serum IgG and IgA responses to both Wuhan-Hu-1 S and RBD (IgG and IgA ELISA GMT between 5.0-6.2 log₁₀, and 3.5-5.8 for S and RBD, respectively, Fig 2B top panels and S2 Fig). However, serum IgG and IgA GMTs to the ancestral version of the S protein elicited by B/HPIV3/S-Omicron-6P were 6- to 16-fold lower than those elicited by the matched vaccine B/HPIV3/S-6P- or by B/HPIV3/S-Delta-6P (p < 0.0001, Fig 2B, top panels), and anti-RBD IgG and IgA GMTs elicited by B/HPIV3/S-Omicron-6P were 31–88-fold lower (p < 0.0001, S2 Fig), reflecting antigenic differences between S and RBD proteins of the Wuhan-Hu-1, Delta, and B.1.1.529/Omicron variants [30].

The homologous and heterologous antibody responses induced by each B/HPIV3-vectored S antigen to the S protein from Delta or B.1.1.529/Omicron variants were also determined (Fig 2B, middle and bottom panels), showing that B/HPIV3/S-6P and B/HPIV3/S-Delta-6P-elicited antibodies bound significantly stronger to the Delta S protein than those elicited by B/HPIV3/S-Omicron-6P. Overall lower IgG titers were detected when S-2P of the B.1.1.529/Omicron variant was used as the antigen; B.1.1.529/Omicron anti-S IgG GMTs were comparable in all groups, while B.1.1.529/Omicron anti-S IgA GMTs elicited by B/HPIV3/S-6P and B/HPIV3/S-Delta-6P were significantly lower than those in B/HPIV3/S-Omicron-6P immunized hamsters (Fig 2B, bottom panel). As expected, no anti-S or anti-RBD IgG/IgA antibodies were induced in hamsters immunized by the B/HPIV3 empty-vector control (Figs 2B, S2).

The 50% serum neutralizing antibody titers (ND₅₀) to SARS-CoV-2 variants corresponding to the three different B/HPIV3-expressed S antigens (WA1/2020, B.1.617.2/Delta, and B.1.1.529/Omicron) were determined at BSL3 by ND₅₀ assay for nine randomly selected hamsters from each group (Fig 2C). As expected, B/HPIV3/S-6P induced strong WA1/2020-neutralizing titers, similarly to B/HPIV3/S-Delta-6P (GMTs of 2.25 log₁₀ in both groups, Fig 2C, top panel), while only one of the nine B/HPIV3/S-Omicron-6P immunized hamsters had WA1/2020-neutralizing antibodies detectable. Against SARS-CoV-2 of the B1.617.2/Delta variant, the matched vaccine candidate B/HPIV3/S-Delta-6P induced higher neutralizing titers than B/HPIV3/S-6P (GMT 2.45 vs 2.01), while only two of nine B/HPIV3/S-Omicron-6P immunized animals had very low B.1.617.2/Delta neutralizing serum antibody titers detectable (Fig 2C, middle panel). However, using SARS-CoV-2 BA.1/Omicron, exclusively the sera from B/HPIV3/S-Omicron-6P immunized hamsters had robust neutralizing antibody titers detectable, while only a single B/HPIV3/S-6P- and two B/HPIV3/S-Delta-6P-immunized hamsters had very low BA.1/Omicron ND₅₀ titers detectable (Fig 2C, bottom panel). Thus, each vector induced strong neutralizing serum antibodies against its best-matching SARS-CoV-2 variant of a similar magnitude (ND₅₀ GMTs of 2.25 log₁₀, 2.45 log₁₀, and 2.47 log₁₀, for WA1/2020, B.1.617/Delta, and BA.1/Omicron, respectively). There was substantial cross-neutralization of antibodies elicited by the ancestral and the Delta antigens between the ancestral strain and Delta variant, but this was not true for the BA.1/Omicron variant. No SARS-CoV-2 neutralizing antibodies were detected in sera from B/HPIV3 control immunized hamsters.

To evaluate further the antigenic breadth of the vaccine candidates, we evaluated the sera in an ACE2 binding inhibition assay as a surrogate to live virus neutralization assays. This assay evaluates the ability of serum antibodies to inhibit binding of soluble ACE2 receptor to S proteins derived from 20 different SARS-CoV-2 variants, including the recently circulating XBB variants (Fig 2D and E). Results from this assay are expressed as % inhibition of ACE2 binding relative to a non-serum buffer-only negative control and are shown as a heat map (Fig 2D, n = 36 per immunized group). The assay allows evaluation of the breadth of the serum antibody responses using relatively small volumes of serum. We found that sera from B/HPIV3/S-6P- and B/HPIV3/S-Delta-6P immunized hamsters highly efficiently inhibited ACE2 binding to S from the vaccine-matched strains (Wuhan-Hu-1 and B.1.617.2/Delta), but also of early variants, B.1.1.7/Alpha and B.1.351/Beta (96–100 median % inhibition; Fig 2D and E). Interestingly, the binding inhibition of S proteins from BA.1 and BA.1+R346K variants by these sera was robust (58–65 median % inhibition), while their ability to inhibit binding of ACE2 to S from BA.1+L452R, BA.2, BA.3, BA.4/BA.5, BQ.1, BF.7, BN.1 variants and derivatives as well as the recently circulating XBB variants was stronger (62–86% median inhibition). The results for each vaccine candidate were combined in a radar plot (Fig 2E), showing that the breadth of serum antibodies elicited by B/HPIV3/S-6P and B/HPIV3/S-Delta-6P was comparable, with the strongest reactivity to variants of the B.1.1.7/Alpha, B.1.351/Beta, and B.1.617.2/Delta lineages.

In contrast, serum antibodies in hamsters immunized with B/HPIV3/S-Omicron-6P showed robust inhibition against these pre-Omicron strains (60–70 median % inhibition, Fig 2D and E), but they efficiently inhibited ACE2 binding of S proteins from the BA.1, BA.2, BA.3 Omicron sublineages and derivatives (94–100 median % inhibition). However, their ACE2 binding inhibition of S proteins from more recent SARS-CoV-2 variants (BA.4/5, BA.2.75, BF7, BN.1, BQ.1, XBB.1 and derivatives) was lower than that of hamsters immunized with B/HPIV3/S-6P and B/HPIV3/S-Delta-6P, with median % inhibition ranging between 42% and 66%. As expected, sera from B/HPIV3 empty vector immunized hamsters did not inhibit binding of ACE2 to S from the 20 different variants analyzed.

Weight change and lung inflammatory response of immunized hamsters following challenge with SARS-CoV-2 WA1/2020, B.1.617.2/Delta, or BA.1/Omicron variants

We next evaluated the protective efficacy of the B/HPIV3-S expressing vectors against challenge with homologous or heterologous SARS-CoV-2 variants (see Fig 1C for timeline). To do so, the remaining 36 hamsters from each immunization group (B/HPIV3 empty-vector control, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P, or B/HPIV3/S-Omicron-6P) were transferred to a BSL3 facility. Hamsters from each group were randomly distributed into three subgroups (n = 12 per subgroup), and, on day 32 pi, challenged with 4.5 log₁₀ 50% tissue culture infectious doses (TCID₅₀) of SARS-CoV-2 of either the WA1/2020 or B.1.617.2/Delta variant, or the BA.1/Omicron variant, representing the major circulating variant at the time the study was designed.

After SARS-CoV-2 challenge, hamsters were weighed daily for 12 days to monitor for weight loss (Fig 3A). Animals immunized with the B/HPIV3 empty-vector control and challenged with WA1/2020 or B.1.617.2/Delta exhibited moderate weight loss from day 2–8 post-challenge (pc; 14 and 7 median % weight loss on day 8, respectively) before re-gaining weight from day 8–12 pc (Fig 3A, left and middle panels). Animals immunized with the S-expressing B/HPIV3 vectors were all protected from weight loss following challenge with WA1/2020 (p < 0.001 from day 2–7 pc, p < 0.01 on day 8 pc and p < 0.01 on day 11 and 12 pc compared to the B/HPIV3 empty vector) or B.1.617.2/Delta (p < 0.001 on day 2 and 3 pc compared to the B/HPIV3 empty vector) (Fig 3A, left and middle panels). Following challenge with BA.1/Omicron, no weight loss was observed in the empty-vector control group, nor in any hamsters immunized with an S-expressing B/HPIV3 vector (Fig 3A, right panel). This was expected, as previous studies have shown that BA.1/Omicron does not induce weight loss in hamsters [31].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Weight change and lung inflammatory responses of immunized hamsters upon SARS-CoV-2 challenge.

On day 32 pi, the 36 remaining hamsters per immunized group were sub-divided into three subgroups of 12 hamsters each. Each subgroup was challenged with 4.5 log₁₀ TCID₅₀ per animal of SARS-CoV-2 WA1/2020, B.1.617.2/Delta or BA.1/Omicron (see Fig 1C for timeline). (A) Body weights were monitored daily from day 0 to 12 post-challenge (pc). Data are shown as mean percent body weight relative to the day 0 weight (n = 12 hamsters/subgroup from day 0 to 3, n = 6 hamsters/subgroup from day 4 to 12 except for B/HPIV3-immunized and WA1-challenged subgroup with 4 or 5 animals from day 8 to 12). Mixed-effects analysis with Dunnett post-test; exact p values are indicated for levels of significance p < 0.05. (B) Expression of inflammatory/antiviral genes in lungs on day 3 pc. On day 3 pc, six hamsters per subgroup were euthanized and lungs were harvested and homogenized. Total RNA was extracted from lung homogenates and the expression of 10 inflammatory/antiviral genes was evaluated by RT-qPCR using TaqMan assays [25]. Results were analyzed using ΔΔCt method and normalized to beta actin. Relative expression was expressed as log₂ fold change over the expression level in five unimmunized, unchallenged hamsters [25] and represented as heatmaps with one heatmap per challenge virus. Non-detected genes are crossed.

https://doi.org/10.1371/journal.ppat.1012585.g003

Six hamsters per subgroup were euthanized on day 3 pc, and lungs and nasal turbinates were collected. To assess effects of SARS-CoV-2 challenge further, we determined lung inflammatory cytokine responses by measuring the expression of 10 inflammatory/antiviral genes by RT-qPCR in their lungs. Data represented as heatmaps show the fold increase of gene expression (in log₂) over the mean expression in five mock-immunized, mock-challenged hamsters that were derived from a separate previous study [25] (Fig 3B). Upon challenge with WA1/2020, B.1.617.2/Delta or BA.1/Omicron, a strong increase in expression of inflammatory/antiviral genes was detected in the lungs of B/HPIV3-immunized hamsters (1–12 log₂ fold-increase in expression, with CCL5, IFN-L, IRF7 and MX2 being the most-upregulated genes) (Fig 3B, B/HPIV3 columns).

No or low increases in expression of inflammatory/antiviral genes were detected in B/HPIV3/S-6P- or B/HPIV3/S-Delta-6P-immunized hamsters after challenge with either homologous or heterologous SARS-CoV-2 strains WA1/2020 or B.1.617.2/Delta. However, following heterologous challenge of these B/HPIV3/S-6P- or B/HPIV3/S-Delta-6P immunized animals with the more antigenically distant BA.1/Omicron variant, moderate expression of some inflammatory/antiviral genes were observed in 4/6 and 2/6 hamsters, respectively (Fig 3B, B/HPIV3/S-6P and B/HPIV3/S-Delta-6P columns, bottom row). This suggested that immunization with B/HPIV3/S-6P or B/HPIV3/S-Delta-6P was less effective against lung inflammatory responses after BA.1/Omicron challenge.

On the other hand, no or low increases in expression of inflammatory/antiviral genes were detected in B/HPIV3/S-Omicron-6P immunized hamsters after homologous BA.1/Omicron challenge (Fig 3B, B/HPIV3/S-Omicron-6P column, bottom row), while a moderate to strong increase in expression of inflammatory/antiviral genes was detected in the lungs of B/HPIV3/S-Omicron-6P-immunized hamsters following heterologous WA1/2020 or B.1.617.2/Delta virus challenge (Fig 3B, B/HPIV3/S-Omicron-6P column, top and middle rows), suggesting that immunization with B/HPIV3/S-Omicron-6P was less effective against challenge with the heterologous WA1/2020 or B.1.617.2/Delta variant. These results are consistent with the greater antigenic distance between BA.1/Omicron and the ancestral WA1/2020 strain or the B.1.617.2/Delta variant.

Replication of SARS-CoV-2 WA1/2020, B.1.617.2/Delta or BA.1/Omicron challenge virus in immunized hamsters

We also evaluated SARS-CoV-2 challenge virus replication in nasal turbinates (NTs) and lungs from these six hamsters per challenge subgroup that had been euthanized on day 3 pc. Challenge viral loads were first determined from lung tissue homogenates by RT-qPCR using previously-described TaqMan assays targeting SARS-CoV-2 subgenomic E (sgE) or N (sgN) mRNA as indicators for active challenge virus replication. As expected, WA1/2020 and B.1.617.2/Delta challenge virus replicated efficiently in the lungs of hamsters previously immunized with the B/HPIV3 empty vector (GMTs from 8.6 to 10.2 log₁₀ copies of sgE or sgN/g, respectively, of lung tissues) (Fig 4A, left and middle panels). BA.1/Omicron also replicated efficiently in the empty-vector control immunized hamsters, albeit to a lower level (GMT of 6.9 and 8.8 log₁₀ copies of sgE and sgN/g of tissues, respectively) (Fig 4A, right panel).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Protection of immunized hamsters from SARS-CoV-2 challenge virus replication.

On day 3 pc, 6 of 12 hamsters per subgroup were euthanized and nasal turbinates and lungs were harvested, and homogenized. Aliquots were used to evaluate SARS-CoV-2 challenge virus replication by RT-qPCR (A) or viral culture (B). (A) Total RNA was extracted from aliquots of lung homogenates and SARS-CoV-2 viral loads were determined by RT-qPCR using TaqMan assays to quantify subgenomic E or N mRNA (Limit of detection: 5.1 log₁₀ copies per g, dotted line). (B) Titers of SARS-CoV-2 challenge viruses, WA1/2020, B.1.617.2/Delta or BA.1/Omicron were determined from aliquots of nasal turbinate and lung homogenates and expressed in log₁₀ TCID₅₀ per g of tissue (Limit of detection: 1.5 log₁₀ TCID₅₀ per g, dotted line). Each hamster is represented by a symbol and GMTs with standard deviations are shown. GMTs are also indicated above x axes. Non-parametric Kruskal Wallis test with the Dunn post hoc test; exact p values are indicated for levels of significance p < 0.05.

https://doi.org/10.1371/journal.ppat.1012585.g004

After challenge with WA1/2020 or B.1.617.2/Delta, hamsters immunized with B/HPIV3/S-6P or B/HPIV3/S-Delta-6P had no detectable virus replication in the lungs, indicating complete protection, except for one hamster immunized with B/HPIV3/S-6P that exhibited a low level of B.1.617.2/Delta sgN RNA (Fig 4A, left and middle panel). However, in 2/6 B/HPIV3/S-Omicron-6P-immunized hamsters challenged with WA1/2020, sgE/N RNA was detected, albeit at a lower level than in empty-vector immunized animals (GMT of 5.8 log₁₀ vs 10.2 log₁₀ copies/g of sgN, Fig 4A, left panel). In addition, all six B/HPIV3/S-Omicron-6P-immunized hamsters in the B.1.617.2/Delta challenge subgroup had sgN RNA detectable in the lungs (GMT of 6.8 log₁₀ copies/g of sgN), albeit greatly reduced compared to the B.1.617.2/Delta viral load in empty-vector immunized hamsters (Fig 4A, middle panel). This indicates that immunization with B/HPIV3/S-Omicron-6P conferred less-than-complete protection against WA1/2020 or B.1.617.2/Delta challenge virus replication in the lungs, reflecting the antigenic distance between BA.1/Omicron and WA1/2020 or B.1.617.2/Delta variants.

In BA.1/Omicron-challenged subgroups of B/HPIV3/S-6P and B/HPIV3/S-Delta-6P immunized hamsters, sgE or sgN was detected in 4/6 immunized hamsters per subgroup, albeit at a low level (GMT of 6.0 and 5.6 log₁₀ copies/g of sgN, for B/HPIV3/S-6P and B/HPIV3/S-Delta-6P immunized hamsters, respectively), indicating less-than-complete protection against heterologous BA.1/Omicron challenge. On the other hand, animals in BA.1/Omicron-challenged subgroups immunized with the matched vector B/HPIV3/S-Omicron-6P had no sgE or sgN detectable in lungs and appeared fully protected from challenge virus replication (Fig 4A, right panel).

We also evaluated SARS-CoV-2 replication in NT and lungs on day 3 pc by titration of clarified tissue homogenates on Vero cells. As expected, in NT and lung homogenates from B/HPIV3 empty-vector immunized animals, WA1/2020, B.1.617.2/Delta and BA.1/Omicron challenge viruses were detectable at substantial titers (between 4.8-6.1 log₁₀ and 4.6-6.3 log₁₀ TCID₅₀/g in NT and lungs, respectively) (Fig 4B), indicating that these variants replicated efficiently in B/HPIV3 empty vector immunized hamsters.

Following homologous or heterologous challenge with WA1/2020 or B.1.617.2/Delta variants (Fig 4B, left and middle panels), hamsters from the B/HPIV3/S-6P- and B/HPIV3/S-Delta-6P-immunized groups had no or relatively low levels of challenge virus detectable in the NTs (undetectable to 2,500-fold reduced titers compared to hamsters in B/HPIV3 empty-vector control group, p < 0.0001); in their lungs, challenge virus was undetectable, with the exception of one hamster in the B/HPIV3/S-6P immunized group challenged with B.1.617.2/Delta (p < 0.0001 compared to B/HPIV3 empty vector, Fig 4B, left and middle panels). Following BA.1/Omicron challenge (Fig 4B, right panel), virus titers in NT of the B/HPIV3/S-6P- and B/HPIV3/S-Delta-6P-immunized hamsters were 250-fold lower than in the B/HPIV3 empty-vector immunized group (Fig 4B, right panel, p < 0.0001). In lungs collected from animals in these two immunization groups, BA.1/Omicron was detectable only in two hamsters per group at low levels close to the limit of detection (Fig 4B, right panel, p < 0.0001 compared to B/HPIV3 empty vector). Thus, hamsters immunized with B/HPIV3/S-6P- and B/HPIV3/S-Delta-6P were robustly but nevertheless less effectively protected against BA.1/Omicron than against WA1/2020 or B.1.617.2/Delta.

On the other hand, all hamsters immunized with B/HPIV3/S-Omicron-6P and challenged with the ancestral WA1/2020 had challenge virus detectable in NT (GMT 100-fold reduced compared to B/HPIV3 empty-vector immunized hamsters, p < 0.0001, Fig 4B, left panel). However, in their lungs, with the exception of one hamster that had low titers of WA1/2020 detectable, no WA1/2020 virus was detectable, indicating substantial protection by B/HPIV3/S-Omicron-6P against the ancestral WA1/2020. In addition, low levels of B.1.617.2/Delta and BA.1/Omicron challenge virus were detected in the NT and lungs of the B/HPIV3/S-Omicron-6P-immunized hamsters (Fig 4B, middle and right panels).

Magnitude and breadth of anamnestic mucosal and serum antibody responses upon SARS-CoV-2 challenge

In the remaining six hamsters per challenge subgroup, we evaluated mucosal and serum anamnestic immune responses following challenge. To do so, nasal washes (NW) were collected on day 21/22 pc from the same hamsters that had NW collected on day 21 pi. On day 23/24 pc, animals were necropsied and bronchoalveolar lavage and sera were collected (see Fig 1C for timeline). Anti-S and anti-RBD IgG and IgA titers were then determined by ELISA using S antigen from the Wuhan-Hu-1 strain (Figs 5A-C, S3). The level of anti-S sIgA in BAL was also determined. The breadth of the anamnestic serum antibody response was evaluated further in an assay that measures binding inhibition of soluble ACE2 receptor to S antigens from 20 variants (Fig 5D and 5E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Mucosal and serum antibody responses in immunized hamsters upon SARS-CoV-2 challenge.

On day 23 or 24 pc, the six remaining hamsters per subgroup were euthanized, and bronchoalveolar lavage (BAL) and sera were collected (see Fig 1C for timeline of the experiment) for evaluation of the antibody response by ELISA (IgG) or DELFIA (IgA and sIgA) (A-C) or ACE2 binding inhibition assay (D-E). (A) Anti-S IgG (left panel), IgA (middle panel) and sIgA (right panel) antibody titers in BAL (limit of detection: 1 log₁₀, dotted line for IgG and IgA). N = 6 hamsters per subgroup with the exception of n = 5 for B/HPIV3 immunized/WA1/2020 challenged, B/HPIV3/S-6P-immunized/BA.1 challenged, B/HPIV3/S-Delta-6P-immunized/BA.1 challenged, B/HPIV3/S-Omicron-6P-immunized/WA1/2020 challenged and n = 4 for B/HPIV3/S-6P-immunized/Delta challenged. The anti-S sIgA DELFIA titers were determined for hamsters with sufficient volumes of remaining BAL samples, and the animal numbers and fractions of hamsters with detectable levels are indicated on top of the graph. (B-C) Serum anti-S IgG (B) and IgA (C) titers after challenge (left panels) and fold changes (in log₁₀) of post-challenge titers over post-immunization titers (right panels). N = 6 with the exception of n = 5 for B/HPIV3-immunized/WA1/2020 challenged. Purified preparations of S from the Wuhan-Hu-1 strain were used as antigens in the ELISA. Each hamster is represented by a symbol and medians with interquartile ranges are shown. One-way ANOVA with Tukey post-test; exact p values are indicated for levels of significance p < 0.05. (D-E) Binding inhibition of soluble ACE2 protein to SARS-CoV-2 S proteins from 20 different variants by serum antibodies from immunized and challenged hamsters. Data are represented as a heatmap with one heatmap per challenge virus and with the median percent inhibition of ACE2 binding relative to a non-serum control indicated (D) or as radar plots (E). N = 6 with the exception of n = 5 for B/HPIV3-immunized/WA1/2020 challenged.

https://doi.org/10.1371/journal.ppat.1012585.g005

Using the ancestral Wuhan-Hu-1 S antigen, we detected only low to background levels of nasal anti-S IgG and IgA by ELISA in B/HPIV3 empty-vector control immunized hamsters 21/22 days after challenge with the three SARS-CoV-2 strains (S3 Fig). The strongest responses were detected by IgA ELISA in three of five animals after WA1/2020 challenge, in the same order of magnitude as those on day 21 after immunization with B/HPIV3/S-6P (S3B Fig, top row). Overall, after SARS-CoV-2 challenge of B/HPIV3 empty-vector control immunized animals, primary S-specific nasal antibody responses were detected in fewer animals than on day 21 after immunization with B/HPIV3/S-6P and B/HPIV3/S-Delta-6P. In NW from the B/HPIV3/S-6P and B/HPIV3/S-Delta-6P-immunized hamsters, no or very weak anamnestic antibody responses were detected after SARS-CoV-2 challenge (S3A and S3B Fig, middle panels), with small post-challenge anti-S IgG and IgA increases only in a few animals that had low or background levels of anti-S IgG or IgA three weeks after immunization. B/HPIV3/S-Omicron-6P-immunized hamsters had no to low anti-Wuhan-Hu-1 S IgG and IgA detectable in the upper airways three weeks following immunization; most animals in the B/HPIV3/S-Omicron-6P immunized groups exhibited anamnestic increases in nasal anti-S IgG and IgA after WA1, Delta, or BA.1 challenge, detectable by Wuhan Hu-1 S ELISA (S3A and S3B Fig, right panels). The absence of post-challenge increases in most animals with nasal IgG and IgA detectable after immunization with B/HPIV3/S-6P or B/HPIV3/S-Delta-6P suggested that protection against SARS-CoV-2 challenge (Fig 4) was restrictive with respect to anamnestic nasal antibody responses.

We also evaluated the mucosal antibody levels in the lower airways using BAL samples collected at necropsy (day 23/24 pc, Fig 5A). Even though there was variability within the groups, anti-S IgG and IgA antibodies were detected in the lower airways of most of the B/HPIV3 empty vector-immunized hamsters and sIgA antibodies detected from some hamsters 3 weeks after SARS-CoV-2 challenge, revealing primary mucosal antibody responses following high levels of challenge virus replication in these control animals (Fig 4). Since it is not possible to obtain sequential BAL samples from the same animals, the kinetics of mucosal responses in the lower airways could not be assessed directly. Compared to the B/HPIV3 empty-vector control groups, post-challenge anti-S IgG and IgA titers in B/HPIV3/S-6P immunized animals were significantly higher in all challenge groups, reflecting a strong and durable primary response in the lower airways to the B/HPIV3-vectored S-6P antigen, and/or a boost of mucosal antibodies after challenge. Compared to the post-challenge titers in B/HPIV3-empty vector controls, BAL anti-S IgG and IgA titers were higher in seven of nine challenge groups that had previously been immunized with S-expressing B/HPIV3, and anti-S sIgA titers were higher in six of nine of these challenge groups. Even though we were not able to directly compare pre- and post-challenge antibody levels in the BAL from the same animals, these results indirectly suggest that anamnestic mucosal antibody responses occurred in the lower airways in most of the groups immunized with the S-expressing B/HPIV3 vectors.

In addition, we evaluated the serum IgG and IgA responses to the ancestral Wuhan-Hu-1 S and to the RBD on day 23/24 pc by ELISA (Fig 5B and 5C, left panels; anti-RBD titers are shown in S4 Fig). We detected high anti-S and anti-RBD serum IgG and IgA titers in all challenged hamsters. After BA.1/Omicron challenge, B/HPIV3 empty-vector immunized control animals had about 10-fold lower Wuhan-Hu-1 S-specific IgG and IgA antibody titers than animals primed with any of the S-expressing versions of B/HPIV3, including the B/HPIV3/S-Omicron-6P primed animals (Fig 5B and 5C, left panels). Comparison of paired post-immunization and post-challenge sera (Fig 5B and C, right panels) revealed that the anamnestic responses to the S antigen after challenge in all groups that had previously been immunized with S-expressing versions of B/HPIV3 were limited (medians <16-fold), suggesting strong protection of the S expressing B/HPIV3 vectors against matched or heterologous challenge with these three variants. We noted that matched challenge of B/HPIV3/S-Omicron-6P primed animals induced the strongest anamnestic responses, detectable by Wuhan-Hu-1 S ELISA, possibly reflecting affinity maturation and increase of antigenic breadth after challenge. The comparison of the RBD-specific responses of the paired post-immunization and post-challenge sera showed that the anamnestic IgG and IgA responses to the RBD region after challenge in animals previously immunized with S-expressing B/HPIV3 vaccines were generally stronger than the anamnestic responses to the whole S protein, especially in the B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P immunized groups, irrespective of the challenge virus (medians up to 79-fold) (S4 Fig).

To evaluate further the breadth of antibody responses post-challenge, sera were also evaluated for their ability to inhibit the binding of soluble ACE2 protein to S of 20 SARS-CoV-2 variants (Fig 5D and 5E). As shown in the heat maps (Fig 5D) and radar plots (Fig 5E), sera from B/HPIV3 empty-vector immunized/WA1/2020 or B.1.617.2/Delta challenged hamsters efficiently inhibited the binding of ACE2 to S of Wuhan-Hu-1 or B.1.617.2/Delta as well as early variants such as B.1.1.7 or B.1.351 (95–100% inhibition) (Fig 5D, top and middle panel, top rows and Fig 5E, left and middle panel, red lines). However, using sera from these animals, ACE2 binding inhibition to S from more recent variants was overall lower (44–85% inhibition). Sera from B/HPIV3 empty vector immunized/BA.1/Omicron challenged hamsters exhibited a different breadth with strong ability to inhibit ACE2 binding to S of BA.1, BA.2 and BA.3 variants (83–96% inhibition) but reduced ability to inhibit ACE2 binding to S from early or more recent SARS-CoV-2 variants including XBB variants (31–54% inhibition) (Fig 5D, bottom panel, top row and Fig 5E, right panel).

Sera from B/HPIV3/S-6P or B/HPIV3/S-Delta-6P immunized hamsters exhibited increased breadth after challenge with WA1/2020, Delta or BA.1/Omicron (Figs 2D, 2E, 5D and 5E, blue and green lines) with almost complete ACE2 binding inhibition to all evaluated S antigens (77–100% inhibition). Finally, sera from B/HPIV3/S-Omicron-6P immunized hamsters also exhibited increased breadth after challenge with WA1/2020, Delta or BA.1, with nevertheless a lower ability to inhibit ACE2 binding to S proteins from the more recent SARS-CoV-2 isolates (BF.7, BN.1, BQ, XBB and derivatives; 62–85% inhibition, Figs 2D, 2E, 5D and 5E, orange lines).

Ethics statement

Hamster studies were approved by the Animal Care and Use Committee of the National Institutes of Allergy and Infectious Diseases. The animal experiments were performed following the Guide for the Care and Use of Laboratory Animals by the NIH.

SARS-CoV-2 viruses

The SARS-CoV-2 USA-WA1/2020 virus was obtained through BEI Resources (cat # NR-52281) and was passaged twice on Vero TMPRSS2 cells. SARS-CoV-2 isolate hVoV-19/USA/MD-HP05285/2021 (Lineage B.1.617.2; Delta Variant) was contributed by Andrew S. Pekosz and was obtained through BEI Resources (cat# 55673). The SARS-CoV-2 isolate hCoV-19/USA/HI-CDC-4359259-001/2021 (lineage B.1.1.529; Omicron variant) was deposited by the Centers for Disease Control and Prevention (CDC) and obtained through BEI Resources (cat# NR-56486) and was passaged once on Vero TMPRSS2 cells. All experiments with SARS-CoV-2 were conducted in Biosafety Level (BSL)-3 containment laboratories approved for use by the US Department of Agriculture and CDC.

Generation of recombinant B/HPIV3 expressing SARS-CoV-2 spike protein

A cDNA clone encoding the B/HPIV3 antigenome expressing the 1,273 aa full-length version of the 6P-stabilized SARS-CoV-2 S protein derived from the ancestral Wuhan-Hu-1 strain (GenBank MN908947), prefusion-stabilized by six proline substitutions [23] to generate B/HPIV3/S-6P, was generated previously [25,26]. For the present study, we generated versions of B/HPIV3 that express the S-6P stabilized full-length versions of S proteins derived from SARS-CoV-2 B.1.617.2/Delta (GISAID EPI_ISL_3066877, sequenced by Long J, Renzette N, Adams M, Omerza G, Kelly K, Li L, The Jackson Laboratory, Farmington, CT) and B.1.1.529/Omicron variants [GISAID EPI_ISL_6795833, sequenced by Strydom A. et al., ZARV/NHLS, Department of Medical Virology, University of Pretoria [41]]. The ORF encoding the S-Delta-6P and S-Omicron-6P open reading frames (ORFs) (aa 1–1,273) were codon-optimized for human expression and include six proline substitutions to stabilize S in the prefusion form [23]. In addition, the S1/S2 polybasic furin cleavage motif “RRAR” was ablated and substituted by the “GSAS” motif [22]. Each ORF was framed by nucleotide adapters containing the BPIV3 gene start and gene end signal sequences. The resulting sequences were synthetized de novo and inserted into the B/HPIV3 antigenome cDNA using a singular Asc I restriction site in the 3’ noncoding region of the N gene, placing the additional gene between the N and P genes in the B/HPIV3 antigenome plasmid. The sequences of the antigenomic cDNAs were confirmed completely by Sanger sequencing, and plasmids were used to transfect BHK21 cells, clone BSR T7/5 [42], to produce the recombinant B/HPIV3 vectors expressing the S-6P antigens derived from B.1.617.2/Delta and B.1.1.529/Omicron isolates. Virus stocks were grown in Vero cells, and viral genomes purified from recovered virus were sequenced in their entirety by Sanger sequencing from overlapping uncloned RT-PCR fragments, confirming the absence of any adventitious mutations.

Western blotting

Viral protein expression by the B/HPIV3 S-expressing vectors or the B/HPIV3 empty vector was evaluated as described previously [25]. Briefly, Vero cells were seeded in 6-well plates and, on the next day, inoculated with P2 working stocks of B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P or B/HPIV3/S-Omicron-6P at an MOI of 1 PFU per cell. After incubation for 48 h at 32°C, cells were washed with DPBS and lysed using 300 μl 1X NuPAGE LDS Sample Buffer (NP0007, Thermo Fisher). Lysates were clarified by QIAshredder (79656, Qiagen). NuPAGE Sample Reducing Agent (NP0009, Thermo Fisher) was added, and lysates were denatured for 10 min at 90°C. The lysates were separated on NuPAGE Bis-Tris Mini Protein Gels, 4–12% (NP0335BOX, Thermo Fisher). The resolved proteins were transferred to polyvinylidene difluoride membranes included in iBlot 3 Transfer Stacks (IB34002, Thermo Fisher) using an iBlot 3 Western Blot Transfer Device (Thermo Fisher). Membranes were blocked with Intercept Blocking Buffer (LiCor) at room temperature for 1 h and incubated with primary antibodies; a goat antiserum raised against a recombinantly expressed secreted form (amino acids 1–1208) of the S-2P stabilized SARS-CoV-2 S protein [22,24]; a rabbit polyclonal hyperimmune serum against purified B/HPIV3 [17]; a mouse monoclonal antibody to GAPDH (G8795, Sigma) (all at 1:5,000) in Intercept Blocking Buffer containing 0.05% Tween 20 overnight at 4°C. Membranes were further incubated with infrared dye-labeled secondary antibodies from LiCor (IRDye 800CW donkey anti-goat IgG (926-32214); IRDye 680RD donkey anti-rabbit IgG (926-68073); IRDye 680RD donkey anti-mouse IgG) (925-68072) (all at 1:10,000) in Intercept Blocking Buffer containing 0.05% Tween 20 for 1 h at room temperature and scanned using an Odyssey CLx Imager (LiCor).

Virus passaging and long-range nucleotide sequencing

Working stocks of B/HPIV3/S-6P, B/HPIV3/S-Delta-6P, B/HPIV3/S-Omicron-6P and B/HPIV3 empty vector control (passage 2) were serially passaged on Vero cells to evaluate the genetic stability of S gene. Monolayers of Vero cells in 25 cm² flasks were inoculated with B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P or B/HPIV3/S-Omicron-6P at an MOI of 0.01 PFU/cell and incubated at 32°C. On day 6 pi, which corresponds to the peak of replication [25], supernatant from each flask was harvested, aliquoted and snap-frozen in dry ice and stored at -80°C. Then, 1 ml of virus supernatant was used to infect a new monolayer of Vero cells in 25 cm² flasks and this procedure was repeated one more time for a total of three serial passages. At the end of the experiment, virus titers of the clarified supernatants collected after each passage were determined by immunoplaque assay. Viral RNA was extracted from P5 stocks of B/HPIV3/S-6P, B/HPIV3/S-Delta-6P and B/HPIV3/S-Omicron-6P using QIAamp Viral RNA Mini Kit (52904, Qiagen), and long-range cDNA was synthesized with Maxima H Minus First Strand cDNA Synthesis Kit (K1652, Thermo Fisher) according to the manufacturers’ instructions. Long-range PCR was performed to amplify a 5.7 kb fragment using the SequalPrep Long PCR Kit with dNTPs (A10498, Thermo Fisher) with the forward primer annealing to nt position 10–39 (17 nucleotides upstream of the N gene start signal sequence) and the reverse primer annealing to nt position 5692–5715 (33 nucleotides downstream of the P gene coding sequence), containing the complete N gene, the added gene encoding the respective S-6P antigen, and P gene start signal. PCR products were purified by a standard gel extraction method and subjected to long-range nanopore sequencing (Quintara Biosciences, Frederick, MD).

Replication, immunogenicity and protective efficacy of B/HPIV3 S expressing vectors in hamsters

Five-to six-week-old male golden Syrian hamsters (Mesocricetus auratus) (n = 184) were obtained from Envigo Laboratories (Frederick, MD). The experiments were conducted in BSL2 and BSL3 facilities approved by the CDC. The timeline of the procedures and sampling is described in Fig 1C. Two days before immunization, serum was collected from each hamster. On day 0, 46 hamsters per group (four groups total) were immunized under isoflurane anesthesia with 5 log₁₀ PFU/hamster of B/HPIV3, B/HPIV3/S-6P, B/HPIV3/S-Delta-6P or B/HPIV3/S-Omicron-6P. On days 3 and 5 pi, five hamsters per immunized group were necropsied and nasal turbinates and lungs were harvested. Tissues were homogenized, clarified by centrifugation, and aliquots were snap-frozen in dry ice. Vaccine replication was evaluated by titration of clarified supernatants using an immunoplaque assay. On day 21 pi, 18 of the 36 remaining hamsters per immunized group were randomly picked and nasal washes were performed under isoflurane anesthesia using 200 μl of 1X PBS. Aliquots were snap-frozen in dry ice for evaluation of the mucosal antibody response by ELISA. On days 24 and 25 pi, serum was collected from the remaining 36 animals per immunized group.

Between days 26 and 28 pi, hamsters were transferred to a BSL3 facility for challenge with SARS-CoV-2. On day 32 pi, the 36 hamsters per immunized group were subdivided into three subgroups of 12 hamsters each and intranasally inoculated under isoflurane anesthesia with 4.5 log₁₀ TCID₅₀/hamster of SARS-CoV-2 WA1/2020, B.1.617.2/Delta or BA.1/Omicron challenge virus. Hamsters were monitored for weight loss and clinical signs of SARS-CoV-2 infection for 12 days after challenge. On day 35 pi (day 3 pc), six of the 12 hamsters per subgroup were necropsied and nasal turbinates and lungs were harvested. Then, tissues were homogenized and aliquots were snap-frozen in dry ice for subsequent titration of SARS-CoV-2 challenge virus by determination of the 50% tissue culture infectious dose (TCID₅₀) in Vero E6 cells and evaluation of the inflammatory response by RT-qPCR. On days 53 and 54 pi (days 21 and 22 pc), nasal washes of the six remaining hamsters per subgroup were performed under isoflurane anesthesia using 200 μl of 1X PBS. Aliquots were snap-frozen in dry ice for evaluation of the mucosal antibody response by ELISA. On days 55 and 56 pi (days 23 and 24 pc), the remaining animals were necropsied and serum was collected. Bronchoalveolar lavage was also done on each animal using 1 ml of 1X PBS and aliquots were snap frozen in dry ice for evaluation of the mucosal antibody response by ELISA.

Dual-staining immunoplaque assay

Replication of B/HPIV3 and derivatives was evaluated from nasal turbinates and lung homogenates by a dual-staining immunoplaque assay as previously described [24,25]. Briefly, sub-confluent monolayers of Vero cells in 24-well plates were inoculated with 10-fold serially-diluted samples and incubated for 2 h at 32°C. Then, cells were overlaid with 1 ml per well of Opti-MEM (Thermo Fisher) containing 0.8% methylcellulose (Sigma), 1% L-glutamine (Thermo Fisher), 2.5% penicillin-streptomycin (Thermo Fisher), 0.5% amphotericin B (Thermo Fisher) and 0.1% gentamicin (Thermo Fisher). After incubation for 7 days at 32°C, cell monolayers were fixed overnight with ice-cold 80% methanol followed by 1 h incubation with Odyssey Blocking Buffer. Then, the HPIV3 antigens and SARS-CoV-2 S protein were detected by dual-immunostaining of plaques using a rabbit anti-HPIV3 serum and a human anti-SARS-CoV-2 S monoclonal primary antibody (CR3022) and the IRDye 680RD donkey anti-rabbit IgG and IRDye 800CW goat anti-human IgG secondary antibodies (LiCor). Plates were scanned using an Odyssey Infrared Imager (LiCor) and staining for PIV3 proteins and SARS-CoV-2 S was visualized in green and red, respectively, generating yellow plaque staining when merged.

IgG and IgA or secretory IgA dual ELISA

Expression and purification of the SARS-CoV-2 S-6P and RBD antigens from the Wuhan-Hu-1 strain was previously described [24]. Plasmids encoding SARS-CoV-2 S-2P from B.1.617.2/Delta or B.1.1.529/Omicron were a kind gift from Dr. Peter Kwong and Dr. I-Ting Teng (VRC, NIAID, NIH). S proteins were expressed and purified as previously described [43]. IgG, IgA and secretory IgA (sIgA) antibody titers against SARS-CoV-2 S or its receptor binding domain (RBD) in sera, nasal washes or BAL samples were evaluated by a previously described ELISA assay designed to detect IgG and IgA by sequential reads in the same sample. Briefly, black 96-well plates (MaxiSorp, Thermo Fisher, cat #437111) were coated with 75 ng per well of purified S or RBD (100 ng per well of S for sIgA) in 50 mM carbonate coating buffer, and incubated overnight at 4°C. Plates were washed three times and blocked with 250 µl DPBS containing 5% dry milk (W/V). Samples were serially diluted in DPBS with 5% dry milk and 0.2% IGEPAL CA-630 by eleven 3-fold dilutions, starting from 1:100 (sera), 1:10 (nasal washes and BAL) or from the original undiluted samples for sIgA. After a 1-hour incubation, plates were washed, and 100 µl per well of secondary antibodies [goat anti-hamster IgG(H+L)-conjugated with HRP (Thermo Fisher, cat# PA1-29626, 1:10,000, rabbit anti-hamster IgA conjugated with biotin (Brookwood Biomedical, sab3002a, 1:1,000) or mouse anti-rhesus J chain-biotin (Nonhuman Primate Reagent Resource, PR3316, 1:1,000)] in dilution buffer were added. Plates were washed and 100 µl per well of diluted Streptavidin-Europium (Perkin Elmer, cat# 1244-360), diluted 1:2,000 in DPBS+ 0.2% IGEPAL CA-630, was added. Plates were incubated for 1 h and washed. Fifty µl of Pierce ECL Substrate (Thermo Fisher, cat# 32106) per well were added, and plates were read using a Synergy Neo2 (BioTek) plate reader to collect IgG luminescence signals. Plates were washed, and 100 µl per well of Enhancement Solution (Perkin Elmer: 4001-0010) was added. Plates were read again using a program for time-resolved fluorescence (TRF; excitation 360/40; emission 620/40) to collect IgA (or sIgA) data. Results were determined by (i) calculating the average reading from duplicate wells, (ii) subtraction of the average reading from blank samples, (iv) setting the cut-off value to the blank average plus three standard deviations, and IgG and IgA (or sIgA) titers of each sample were determined in sequential reads by interpolating the sigmoid standard curve generated on Prism 9.0 as previously described [44,45].

SARS-CoV-2 neutralization assay

SARS-CoV-2 neutralizing antibody titers in hamster sera were determined by live SARS-CoV-2 virus neutralization assays, performed in a BSL3 facility. Heat-inactivated sera were 2-fold serially diluted in Opti-MEM and mixed with 100 TCID₅₀ of SARS-CoV-2 WA1/2020, B.1.617.2/Delta or BA.1/Omicron. After incubation at 37°C for 1 h, the mixtures were added to quadruplicate wells of Vero E6 cells in 96-well plates and incubated for four days. The 50% neutralizing dose (ND₅₀) was defined as the highest dilution of serum that completely prevented cytopathic effects in 50% of the wells and was expressed as a log₁₀ reciprocal value.

ACE2 binding inhibition assay

The ability of hamster sera to inhibit binding of ACE2 to SARS-CoV-2 S was evaluated in an ACE2 binding inhibition assay (Meso Scale Diagnotics, MSD). V-PLEX SARS-CoV-2 Panel 25 (K15586U), Panel 27 (K15609U) and Panel 34 (K15693U) Kits (MSD) were used. Each kit contains 96-well, 10-spot plates coated with soluble ectodomains of S proteins from the wild-type SARS-CoV-2 (Wuhan-Hu-1) and variants (Alpha, Beta, Delta and Omicron sublineages) representing a total of 20 different S antigens from the three different kits. Each assay was performed according to the manufacturer’s instructions and as previously described [25]. Heat-inactivated hamster sera were diluted at 1:20 in MSD diluent and evaluated in duplicate. Plates were read using a MESO QuickPlex SQ 120MM. The average electrochemiluminescence signals from duplicate wells for each serum and in wells with diluent only were calculated. The ACE2 binding inhibition by each serum is expressed as percent inhibition relative to the diluent.

Quantification of SARS-CoV-2 viral genomes and host gene expression in lungs

One hundred μl of lung homogenate from each of six hamsters per subgroup from the day 3 post-SARS-CoV-2 challenge time point was subjected to total RNA extraction using TRIzol LS Reagent and Phasemaker Tubes Complete System (Thermo Fisher) in combination with PureLink RNA Mini Kit (Thermo Fisher). The extracted RNA was used to evaluate the level of expression of inflammatory/antiviral host genes as well as to quantify SARS-CoV-2 viral genomes by RT-qPCR as previously described [25].

Briefly, total RNA was used to synthesize cDNA using High-Capacity RNA-to-cDNA Kit (Thermo Fisher). Then, the level of expression of 10 inflammatory/antiviral host genes (CCL2, CCL3, CCL5, IFN-G, IFN-L, IL-1B, IP-10, IRF7, MX2, TNF-A) was evaluated in triplicate by RT-qPCR using TaqMan Fast Advanced Master Mix (Thermo Fisher) and previously-described TaqMan assays [46–48] on the QuantStudio 7 Pro. Beta-actin was also included as a housekeeping gene. The relative level of expression of each evaluated gene was expressed as log₂ fold changes over the expression determined from five non-immunized, non-challenged hamsters [25].

Quantification of SARS-CoV-2 viral genomes from total lung RNA was performed as previously described [25]. Briefly, SARS-CoV-2 subgenomic E and N mRNA was quantified in triplicate by RT-qPCR using TaqMan RNA-to-Ct 1-Step Kit (Thermo Fisher) and previously-reported TaqMan primers and probes [49–52] on QuantStudio 7 Pro Real-Time PCR System (Thermo Fisher). Standard curves were generated by serially diluting pcDNA3.1 plasmids containing each target gene sequence. The limit of detection was 5.1 log₁₀ copies per g of tissue.

Statistical analyses

Data sets were assessed for significance using non-parametric Kruskal Wallis test with the Dunn post hoc test, one-way ANOVA with Tukey’s or Sidak’s post-test or two-way ANOVA with Sidak’s post-test using Prism 9.0 (GraphPad Software). Time course data sets were assessed by mixed-effects analysis with Dunnett post-test; exact p values are indicated for levels of significance p < 0.05. Data were only considered significant at P ≤ 0.05. Details on the statistical comparisons can be found in the figures, figure legends, and results.

The numerical data used in all figures are included in S1 Data.