A replication-incompetent adenovirus type 55 vaccine induces broad and durable protective immunity against pathogenic adenoviruses

Abstract

Human adenovirus types 55, 11, and 14 (HAdV-55, -11, and -14) are pathogenic respiratory viruses for which no drugs or vaccines are currently available. We report the generation of a replication-incompetent rAd55-5E4 with deleted E1 and E3 genes, which only replicates in cells that provide E1 proteins in trans. In mice and non-human primates, vaccination with live non-replicating rAd55-5E4 elicited robust and durable neutralizing antibody (nAb) and cell-mediated immune responses against HAdV-55, as well as cross-reactivity against HAdV-11 and HAdV-14. Furthermore, vaccination with the live non-replicating rAd55-5E4 elicited much stronger immune responses than inactivated rAd55-5E4. In transgenic mice that express human desmoglein-2, the cellular receptor for HAdV-55, -11 and -14, vaccination with rAd55-5E4 or passive transfer of macaque immune sera collected at 66 weeks post vaccination effectively protected against challenges with HAdV-55, HAdV-11, and HAdV-14. Epitope profiling revealed that nAbs mainly recognize epitopes on hexon hypervariable regions 1, 2, 5, and 7, as well as the fiber knob. This study supports the feasibility of developing replication-incompetent HAdVs as vaccines against pathogenic HAdVs.

Graphical abstract

Ling Chen and colleagues engineered a replication-incompetent HAdV-55 (rAd55-5E4) that can only be propagated in cells that provide E1 proteins in trans, ensuring its safety for application. Vaccination in mice and macaques elicited protective immunity against HAdV-55, HAdV-11, and HAdV-14. rAd55-5E4 is a promising vaccine candidate and vector for clinical development.

Introduction

Human adenoviruses (HAdVs) are non-enveloped, double-stranded DNA viruses that belong to the family Adenoviridae, genus Mastadenovirus.¹ To date, more than 116 HAdV genotypes (http://hadvwg.gmu.edu/) have been identified. Several HAdVs, such as HAdV-1, -2, -3, -4, -5, -6, -7, -11, -14, and -55 are common pathogens linked to acute respiratory diseases (ARDs).^2,3 Among these, HAdV-55, previously known as HAdV-11a, was first reported to cause severe ARDs and death in China in 2006, has re-emerged as one of the major HAdVs circulating globally, especially in eastern Asia.^4,5,6,7 Compared with HAdV-3, -4, and -7, which usually cause ARDs in young children and immunocompromised individuals, HAdV-55 exhibits somewhat distinct pathogenic properties.^8,9,10 HAdV-55 infection tends to develop into pneumonia with a poorer prognosis.^10,11 Severe pneumonia has been recorded in both pediatric patients and immuno-competent young adults.^10,12 A proportion of critical cases died despite receiving medical intervention.^9,10 The majority of people are seronegative for HAdV-55 antibodies,¹³ indicating a high susceptibility to HAdV-55 infection in the population. There are no specific drugs approved for treating HAdV infections. Several broad-spectrum antiviral drugs, such as ribavirin, ganciclovir, and cidofovir, have been used off-label to treat severe HAdV infections in immunocompromised patients; however, they are not commonly recommended due to limited efficacy and side effects.^11,14 The only approved live oral HAdV vaccine, comprising wild-type HAdV-4 and -7 in enteric-coated capsules, is only effective against HAdV-4 and -7.¹⁵ Its use is restricted within the US military base, but it is not suitable for the general population due to the risk of viral shedding and infection that may pose a threat to the general population, especially to immune-compromised individuals.¹⁶ Therefore, research and development of safe and effective vaccines for pathogenic adenoviruses, including HAdV-55, is necessary.

HAdV-55 evolves from an intertypic recombination between HAdV-11 and HAdV-14, two pathogenic HAdV-B associated with renal and respiratory diseases, respectively.¹⁷ Most of the HAdV-55 genome is identical to that of HAdV-14, including fiber, except that the hexon gene is evolutionarily closer to HAdV-11. HAdV-14 has been reported to cause ARDs in military and civilian populations. Of the adenovirus infections in the United States from 2003 to 2016, 5.9% were caused by HAdV-14p1, which also caused at least 10 reported deaths in the United States from 2006 to 2007,^18,19 with subsequent outbreaks in Europe and China. HAdV-11 also causes ARDs and is the main pathogen in acute hemorrhagic cystitis and kidney transplant recipients.^20,21 Similar to other HAdVs, the icosahedral capsid of HAdV-55 consists of three major proteins (hexon, penton base, and fiber) and at least four minor proteins (IIIa, VI, VIII, and IX). Trimeric hexon proteins form each facet and recruit dynein for transport to the nucleus at the post-endosomal stage. Penton bases cap the 12 vertices and trigger endocytosis upon binding to the co-receptor, integrin. Trimeric fiber proteins protrude from penton bases and mediate viral binding to the primary cellular receptor, human desmoglein-2 (hDSG2).²² It has been reported that hexon and fiber proteins, in particular the hyper-variable regions (HVRs) of hexon and the knob domain of fiber, are the major targets recognized by neutralizing antibodies (nAbs).²³ Vaccination using full-length hexon proteins encoded by DNA plasmids or hexon-chimeric adenoviruses can elicit HAdV nAbs in mice.^23,24 Vaccination using fiber protein from HAdV-11 can elicit nAbs against multiple HAdV-Bs.²⁵ However, the proper folding and trimerization of hexons rely on the assistance of the HAdV L4-100K chaperon protein.²⁶ In the context of ectopic expression, the post-translational modifications of hexon might be inefficient without L4-100K.²⁷ In addition, fiber-targeting nAbs have been reported to be less potent than hexon-targeting nAbs in inhibiting HAdV-5 infection.²⁸ Given that either natural HAdV infection or the live, oral HAdV vaccine can induce durable antibody and cell-mediated immune responses to confer effective protection in humans,^28,29 we propose that a replication-incompetent HAdV-55, in which natural infection can be mimicked, may be the best strategy for provoking effective protective immunity while avoiding safety concerns.

To achieve a replication-incompetent phenotype, the E1 and E3 genes are usually deleted. HAdV E1 proteins play crucial roles in initiating the viral life cycle and modulating the host cell cycle, whereas the E3 proteins mainly antagonize the host immune defenses and are dispensable for viral replication in cell cultures.^30,31 Replication-incompetent HAdV-5 can be propagated in HAdV E1 gene-transformed human embryonic kidney 293 (HEK293) cells, which express HAdV-5 E1 proteins in trans. However, replication-incompetent HAdV-Bs, with E1 and E3 deleted, were unable to propagate efficiently in HEK293 cells. This is mainly because the E1B 55K protein of HAdV-5 cannot form complex with the E4 34K protein (encoded by E4 ORF6) of HAdV-B, and thus cannot suppress the DNA repair responses of the host cells and inhibit the exporting of host mRNAs.³² E4 34K forms a complex with the E1B 55K protein, which is essential for efficient adenoviral DNA replication and for the activation of viral late genes. It has been reported that replacing the E4 34K of HAdV-B with that of HAdV-5 can restore the propagation capacity of replication-incompetent HAdV-Bs in HEK293 cells.³³

In this study, we generated a replication-incompetent HAdV-55, in which the E1 and E3 genes were deleted, and the native E4 region was replaced with that of HAdV-5 (termed as rAd55-5E4). We tested the propagation capacity of rAd55-5E4 in human-derived cell lines, and assessed the immunogenicity in mice and in non-human primates (rhesus macaques). By employing hDSG2 transgenic mice that express the major cellular receptor for the infection of HAdV-Bs,²² we investigated the protective effects of rAd55-5E4 vaccination and immune serum transfer against infections of HAdV-55 and two related HAdV-11 and HAdV-14. We also assessed the epitope recognition profile of rAdV55-5E4-induced nAbs in non-human primates to understand the neutralizing mechanism better.

Results

Generation of a replication-incompetent rAd55-5E4 vaccine candidate

To construct a replication-incompetent HAdV-55 vaccine candidate, we deleted the E1and E3 regions of the wild-type HAdV-55 Shanxi-Y16 strain (GeneBank: NM_KF911353.1), and replaced the region covering E4 ORF2 to E4 ORF6 with the corresponding region of HAdV-5 (Figure 1A). The resultant rAd55-5E4, as well as a reporter virus rAd55-5E4-EGFP that harbored an expression cassette encoding an enhanced green fluorescent protein (EGFP) in the original E1 region, could be successfully rescued. In HEK293 cells, the genome copies and plaque size of rAd55-5E4 was comparable with that of replication-competent rAd55-EGFP, of which the E1 region was retained but the E3 region was deleted and inserted with an expression cassette encoding EGFP. Modifications of the early genes showed no significant impacts on the morphology of HAdV-55 virions (Figure 1B). SDS-PAGE analysis of the major components of rAd55-5E4 capsids, including hexon, penton base, and fiber, showed no deviation from rAd55-EGFP (Figure 1C).

Figure 1.

Generation and characterization of rAd55-5E4

(A) Genome structure of rAd55-5E4, rAd55-5E4-EGFP, and rAd55-EGFP. (B) Morphology of rAd55-5E4 and rAd55-EGFP viral particles (VPs). One representative image for each HAdV taken under transmission electron microscopy is shown. Scale bar, 100 nm (left) and 20 nm (right). (C) SDS-PAGE analysis of viral lysates. Purified rAd55-EGFP, rAd55-5E4, and rAd5 VPs were treated with SDS in the presence of β-mercaptoethanol and separated by SDS-PAGE. The protein bands were stained with Coomassie Brilliant Blue. (D) Plaques formed by rAd55-EGFP and rAd55-5E4-EGFP in HEK293 or A549 cells. Plaques were observed under a fluorescence microscope. One representative image for each virus is shown. Scale bar, 200 μm. (E and F) Growth curves (E) and cytopathic effects (F) of rAd55-EGFP and rAd55-5E4-EGFP in HEK293 cells. HEK293 cells in 96-well plates were infected with rAd55-5E4-EGFP or rAd55-EGFP at 5 × 10⁶ vp per well. At the indicated time points, viral genome copies in cells and cell viabilities were measured by qPCR and CCK8, respectively. (G and H) Growth curves (G) and cytopathic effects (H) of rAd55-EGFP and rAd55-5E4-EGFP in A549 cells. All tests were performed in duplicate. Shown are the results of one of two independent experiments. The Pearson correlation coefficient was used to measure the strength of the correlation between the growth curves or cytopathic effects of rAd55-EGFP and rAd55-5E4-EGFP. Correlation coefficients were calculated using Student’s t test, and p values of <0.05 were considered statistically significant.

We first assessed the infectivity of rAd55-5E4 in human-derived cell lines using plaque-forming assays. In HEK293 cells that express HAdV-5 E1 proteins, both rAd55-5E4-EGFP and rAd55-EGFP could form plaques of similar sizes without observable differences. In human lung carcinoma A549 cells that lack HAdV-5 E1 genes, rAd55-5E4-EGFP could not form plaques as rAd55-EGFP did (Figure 1D). Accordingly, in HEK293 cells, rAd55-5E4-EGFP and rAd55-EGFP replicated efficiently, with the genome copy numbers increased by 903 and 1,337-folds (peak value versus baseline) within 48 h of infection (Figure 1E), which resulted in obvious cytopathic effects (Figure 1F). In A549 cells that do not express HAdV-5 E1 proteins, only replication-competent rAd55-EGFP, but not replication-incompetent rAd55-5E4-EGFP, could replicate and caused cytopathic effects (Figures 1G and 1H). Therefore, rAd55-5E4 can only replicate in cells that provide HAdV-5 E1 gene products in trans, thus providing safety assurance for in vivo application.

rAd55-5E4 elicits nAb and cell-mediated immune responses against HAdV-55, -11, and -14 in mice

To assess the immunogenicity of rAd55-5E4, we vaccinated BALB/c mice intramuscularly (i.m.) with rAd55-5E4 at 2 × 10⁸ or 2 × 10⁹ viral particles (vp), or 2 × 10⁹ vp of inactivated rAd55-5E4 (iAd55) (which was inactivated with β-propiolactone) twice at a 4-week interval (Figure 2A). Mice receiving PBS were used as no vaccine control. At 2 weeks after the first dose (prime), rAd55-5E4 elicited significant HAdV-55 nAb response in a dose-dependent manner (Figure 2B). Notably, the nAb titers against HAdV-55 in mice receiving 2 × 10⁸ vp rAd55-5E4 were comparable with mice receiving a 10-fold higher dose of iAd55 (Figure 2B). A booster dose further enhanced the nAb titers (Figure 2B). In mice receiving the high dose (2 × 10⁹ vp) of rAd55-5E4, the nAb titers reached 1:65,282 (range, 1:37,419–1:83,732), more than 22-fold greater than those in mice receiving 2 × 10⁹ vp of iAd55 with nAb titers at 1:2,867 (range, 1:1,301–1:4,660). The low dose of rAd55-5E4 was comparable with a 10-fold higher dose of iAd55 in inducing nAb responses (Figure 2B). Given that HAdV-55 shares a similar hexon with HAdV-11 and similar fiber with HAdV-14,^34,35 we tested whether the rAd55-5E4-induced antibody responses had cross-neutralizing activities against HAdV-11 and -14. After the first dose, all vaccinated mice elicited an HAdV-11 nAb response, and the titers were elevated by the booster dose (Figure 2C). However, the first dose elicited HAdV-14 nAb response only in mice receiving the high dose of rAd55-5E4 but not in mice receiving either the low dose of rAd55-5E4 or the high dose of iAd55. After the booster dose, HAdV-14 nAbs were detected in all vaccinated mice, with the highest nAb titers observed in mice receiving the high dose of rAd55-5E4 (Figure 2D). Taken together, rAd55-5E4 is more effective than iAd55 in eliciting nAbs against HAdV-55, -11, and -14 in mice.

Figure 2.

Immunogenicity of rAd55-5E4 in mice

(A) Schematic diagram of rAd55-5E4 vaccination in mice. Six-week-old BALB/c mice (n = 5 per group) were vaccinated with rAd55-5E4 using the indicated dosage through at a 4-week interval. Two weeks after each vaccination, nAb and CMI responses were assessed. (B) HAdV-55 nAb titers in mice sera at 2 weeks after prime or boost vaccination. (C) HAdV-11 nAb titers in mice sera at 2 weeks after prime or boost vaccination. (D) HAdV-14 nAb titers in mice sera at 2 weeks after prime or boost vaccination. Each data point represents an individual mouse. (E) Location of each peptide pool in hexon protein. Peptides corresponding with HAdV-55 hexon were divided into six subpools, P1 (1–130 aa), P3 (121–320 aa, HVR1–6), P4 (310–410 aa), P6 (400–460 aa, HVR7), P7 (450–710 aa), and P8 (700–947 aa). Peptides corresponding with the HVRs of HAdV-14 hexon were divided into P2 (121–320 aa, HVR1–6) and P5 (400–460 aa, HVR7). The percentages shown in the dark gray, yellow, and green areas indicate the similarity of the protein sequence in that region with the corresponding region of HAdV-55. The sequence covered by each peptide pool has been marked. (F) IFN-γ⁺ SFCs in mice at 2 weeks after prime or boost vaccination. Splenic lymphocytes from each group of mice (n = 5) were pooled together and stimulated with the whole peptide pool P9 (F) or individual subpools (G). The numbers of SFCs in 1 million splenic lymphocytes are shown. Each data point represents one replicate. Error bars represent means ± SDs. Data are one representative result of two independent experiments. rAd55^lo, mice vaccinated with 2 × 10⁸ vp of rAd55-5E4; rAd55^hi, mice vaccinated with 2 × 10⁹ vp of rAd55-5E4; iAd55, mice vaccinated with 2 × 10⁹ vp of inactivated HAdV-55; PBS, mice receiving PBS. Comparisons were performed by one-way ANOVA. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ns, not significant.

To determine whether rAd55-5E4 vaccination elicited any cell-mediated immune (CMI) responses, mouse splenocytes were obtained and stimulated with a hexon peptide pool (Figure 2E) to measure the interferon-γ (IFN-γ)-secreting cells (spot-forming cells [SFCs]) by an enzyme-linked immunospot (ELISpot) assay. After the first dose, IFN-γ⁺ SFCs became detectable in mice receiving rAd55-5E4 but were nearly undetectable in those receiving iAd55 (Figure 2F). A booster vaccination with rAd55-5E4 significantly enhanced the frequencies of IFN-γ⁺ SFCs. The high dose of rAd55-5E4 elicited the highest CMI response (Figure 2F). In contrast, two doses of iAd55 elicited a CMI response lower than the 10-fold lower dosage of rAd55-5E4 (Figure 2F). We also dissected the hexon region containing dominant T cell epitopes using hexon peptide subpools. Each peptide corresponds with the amino acid (aa) sequence on HAdV-55 hexon as P1 (1–130 aa), P3 (121–320 aa, HVR1-6), P4 (310–410 aa), P6 (400–460 aa, HVR7), P7 (450–710 aa), and P8 (700–947 aa), where P3 represents to HVR 1–6 and P6 represents to HVR7. Given that HAdV-55 hexon is similar to the HAdV-11 hexon but differs from the HAdV-14 hexon, the most differentiated P2 and P5 peptide pools were constructed, representing the HVR1–6 and HVR7 regions of HAdV-14 hexon, respectively (Figure 2E). The C-terminal region of hexon P7 and P8 (covering aa 450–947) stimulated the highest frequency of IFN-γ⁺ SFCs (Figure 2G). P5 and P6, corresponding with only HVR7, showed no significant CMI response (Figure 2G). The other regions also stimulated IFN-γ⁺ SFCs, but the frequencies were not as high as the C-terminal region (Figure 2G). Hence, the C-terminal region of hexon contains dominant T cell epitopes for CMI response. Since this region is conserved among HAdV-55, -11, and -14, the rAd55-5E4-elicited CMI response is cross-react with HAdV-11 and -14.

Vaccination with rAd55-5E4 protects against HAdV-55, -11, and -14 in hDSG2 transgenic mice

We next employed hDSG2 transgenic mice to assess the protective effects of rAd55-5E4. This mouse line constitutively expresses hDSG2, the primary cellular receptor for HAdV-55, -11 and -14, and thus supports HAdV-55 infection.^22,36 hDSG2 mice that were 6–8 weeks of age received an i.m. injection of rAd55-5E4 at 2 × 10⁸ vp, 2 × 10⁹ vp, or iAd55 at 2 × 10⁹ vp per mouse (Figure 3A). Mice receiving PBS were used as a non-vaccine control group. Six weeks later, we challenged the mice via tail vein injection of a large bolus of 2 × 10¹⁰ vp of rAd55-SEAP, a replication-competent HAdV-55 expressing secreted alkaline phosphatase (SEAP). In the non-vaccine group (mice receiving PBS), rAd55-SEAP infection led to significant SEAP activity in the sera after the challenge, along with the highest level of viral genomic DNA. In mice receiving rAd55-5E4 or iAd55, the SEAP activity in the sera was significantly lower (Figure 3B). The high dose of rAd55-5E4 was more effective in reducing SEAP activity than the low dose of rAd55-5E4 or the high dose of iAd55 (Figure 3B). The genomic copy numbers of rAd55-SEAP in the liver were also significantly lower, and a high dose of rAd55-5E4 achieved the most significant reduction (Figure 3C). Therefore, rAd55-5E4 is significantly much more effective than iAd55 in protecting against HAdV-55 infection.

Figure 3.

rAd55-5E4 vaccination protects against HAdV-55, -11, and -14 infection in hDSG2^+/+ mice

(A) Schematic diagram of rAd55-5E4 vaccination and HAdV-55, -11, and -14 challenge in hDSG2^+/+ transgenic mice. Six-week-old female hDSG2^+/+ mice (n = 8 per group for HAdV-55 and -11 and n = 10 per group for HAdV-14) were vaccinated i.m. with 2 × 10⁹ vp of rAd55-5E4 or iAd55. The mice were challenged 42 days later with 2 × 10¹⁰ vp of rAd55-SEAP, rAd11-SEAP, and rAd14-SEAP via tail vein injection, respectively. (B) Serum SEAP activities in mice at 3 and 7 days after challenge with rAd55-SEAP. (C) rAd55-SEAP genome copies in liver tissues at 7 days after challenge. (D) Serum SEAP activities in mice at 3 and 7 days after challenged with rAd11-SEAP. (E) rAd11-SEAP genome copies in liver tissues at 7 days after challenge. (F) Serum SEAP activities in mice at 3 and 7 days after challenged with rAd14-SEAP. (G) rAd14-SEAP genome copies in liver tissues at 7 days after challenge. Each data point in (B–G) represents an individual mouse. Error bars represent means ± SDs. Data are one representative result of two independent experiments. Statistical analysis is performed by one-way ANOVA. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. ns, not significant.

To determine whether rAd55-5E4 vaccination could confer cross-protection against HAdV-11 and -14 infection, we challenged the mice with a large bolus of 2 × 10¹⁰ vp of rAd11-SEAP or rAd14-SEAP via tail vein injection, which resulted in significant elevation of serum SEAP activity in non-vaccinated mice. rAd55-5E4 vaccination reduced the serum SEAP activities caused by rAd11-SEAP infection. A low dose of rAd55-5E4 also showed such effects, albeit with less potency (Figure 3D). iAd55, however, only mildly reduced SEAP activity. Nevertheless, both rAd55-5E4 and iAd55 reduced the copy numbers of rAd11-SEAP genomes in the liver (Figure 3E). Notably, rAd55-5E4, regardless of the dose, protected against HAdV-14 infection, as evidenced by the reduced SEAP activity and lower viral genome copy numbers compared with the mice receiving PBS (Figures 3F and 3G). iAd55 only conferred mild protection against HAdV-14 infection. Together, rAd55-5E4 is much more effective than iAd55 in protecting against infection with HAdV-11 and -14.

rAd55-5E4 elicits nAb and CMI responses against HAdV-55, -11, and -14 in rhesus macaques

We next assessed the immunogenicity of rAd55-5E4 in non-human primates to obtain data that may be more relevant to humans. A total of 12 rhesus macaques were equally divided into 4 groups for vaccination (Table S2). Macaques in group 1 (No. 1, 2, and 3) and group 2 (No. 4, 5, and 6) were vaccinated i.m. twice with 2 × 10¹⁰ or 1 × 10¹¹ vp of rAd55-5E4, respectively. Macaques in group 3 (No. 7, 8, and 9) received 1 × 10¹¹ vp of iAd55 at the same time points. Macaques in group 4 (No. 10, 11, and 12) received PBS and served as non-vaccine controls (Figure 4A). No significant adverse effects were observed after vaccination, including changes in body weight, temperature, and the complete blood count (Figure S1). A single low dose of rAd55-5E4 elicited a nAb response against HAdV-55. The titers reached peak values within 2–4 weeks after vaccination, declined slowly, and persisted at more than 1:2,000 for at least 32 weeks (Figure 4B). A single high dose of rAd55-5E4 elicited a significantly higher nAb response than the low dose. A single high dose of iAd55 also elicited a nAb response, but it was the lowest among three vaccinated groups. Macaques received a booster vaccination at 32 weeks after the first dose. One macaque (no. 9) in group 3 (iAd55) died accidentally during anesthesia. nAb titers against HAdV-55 elevated to higher levels after the booster vaccination (Figures 4B–4D and S2A). We calculated the areas under the curve (AUCs) representing total antibodies over a sustained period after priming or booster vaccination. The AUCs were significantly higher in the high-dose rAd55-5E4-vaccinated macaques than the low dose of rAd55-5E4-vaccinated macaques (Figures 4F and 4G). After the booster vaccination, the nAb titers further increased by more than 300-fold in all vaccinated macaques (Figures 4B–4D). Although showing a gradual decline, the titers remained over 1:100,000 for at least 34 weeks after booster vaccination with rAd55-5E4. iAd55 vaccination also elicited HAdV-55 nAb response, but the AUCs were significantly lower than those elicited by rAd55-5E4, and the durability seemed to be somewhat limited (Figures 4D and 4F). Thus, rAd55-5E4 can elicit robust and long-lasting HAdV-55 nAb responses against HAdV-55 in non-human primates.

Figure 4.

The nAb responses elicited by rAd55-5E4 in Chinese macaques

(A) Schematic diagram of vaccination assays in Chinese macaques. Four-year-old Chinese macaques (n = 3 per group) were vaccinated with rAd55-5E4 or iAd55 using the indicated dosage through an i.m. route. (B) Titers of nAbs against HAdV-55, -11, and -14 in macaques vaccinated with 2 × 10¹⁰ vp of rAd55-5E4. (C) Titers of nAbs against HAdV-55, -11, and -14 in macaques vaccinated with 1 × 10¹¹ vp of rAd55-5E4. (D) Titers of nAbs against HAdV-55, -11, and -14 in macaques vaccinated with 1 × 10¹¹ vp of iAd55. (E) Titers of nAbs against HAdV-55, -11, and -14 in macaques receiving PBS. Each graph shows the kinetics of nAb titers for one macaque. Gray background indicates the time points after the booster vaccination. The values for each time point reflect the mean of three technical replicates. (F and G) Area under the curve (AUC) of nAb titers after the prime (F) or booster (G) vaccination. Each data point represents the AUC for one macaque. Error bars represent means ± SDs. Statistical analysis is performed by two-way ANOVA. ∗p < 0.05; ∗∗p < 0.01. ns, not significant.

We next examined the cross-reactive nAb response against HAdV-11 and -14 elicited by rAd55-5E4. The HAdV-11 nAb response showed similar kinetics as the HAdV-55 nAb response (Figures S2A and S2B), but the titers and AUCs were relatively lower (Figures 4B, 4C, and 4F). After the booster vaccination, the titers increased by about 300-fold, and remained over 1:50,000 for at least 34 weeks (Figures 4B and 4C). In macaques receiving iAd55, however, only a low level of HAdV-11 nAb response was elicited after booster vaccination (Figure 4D). The titers declined more rapidly and remained around 1:1,000 at 34 weeks after the booster vaccination (Figures 4F and 4G). After each vaccination, the AUCs were much lower in macaques receiving iAd55 than in those receiving rAd55-5E4 (Figures 4F and 4G). Compared with the HAdV-55 and HAdV-11 nAb responses, the HAdV-14 nAb response, which is mainly directed against the fiber protein, showed somewhat different kinetics (Figure S2C). After the priming vaccination, both rAd55-5E4 and iAd55 elicited very low HAdV-14 nAb responses (Figures 4B–4D). After the booster vaccination, only the high dose of rAd55-5E4 increased the HAdV-14 nAb titers to be comparable with the HAdV-11 nAb titers (Figures 4B and 4C). In contrast, the high dose of iAd55 elicited a very weak HAdV-14 nAb response, even after the booster vaccination (Figure 4D). The AUCs were significantly lower in macaques receiving iAd55 than in those receiving an equal dose of rAd55-5E4 (Figures 4F and 4G). Thus, rAd55-5E4 can provoke nAb responses against both HAdV-11 and HAdV-14 in non-human primates, and a booster dose could confer a greater antibody response.

In order to understand the epitopes recognized by nAbs, we investigated the neutralizing activities of immune sera against a series of chimeric HAdV-14 viruses, in which each one or all seven HVRs of the hexon protein has been replaced with the corresponding HAdV-55 HVRs (Figure 5A) (designated as HAdV-14/55H1 to HAdV-14/55H7, or HAdV-14/55H1-7).²³ The sera containing nAbs to a particular HAdV-55 HVR would neutralize the infection of the chimeric HAdV-14 that contains this HAdV-55 HVR. After a single low-dose rAd55-5E4 vaccination, macaque no. 1 showed inhibiting activities against the infections of HAdV-55, HAdV-14/55H1-7, and HAdV-14/55H1 (Figure 5B), suggesting that this macaque mounted nAb response primarily targeting HVR1. Macaques no. 2 and 3 showed inhibiting activities against the infections of HAdV-55, HAdV-14/55H1-7, HAdV-14/55H1, and HAdV-14/55H2 (Figure 5B), suggesting that these macaques mounted an nAb response primarily targeting HVR1 and HVR2. Therefore, a single low dose of rAd55-5E4 elicited mainly nAbs against hexon HVR1 and to a lesser extent HVR2 (Figures 5B and 5D). These sera showed lower neutralizing activities against HAdV-14, indicating that a single low dose of rAd55-5E4 elicits weak nAbs against the fiber (Figures 5B and 5F). The sera of high dose rAd55-5E4-vaccinated macaques (group 2) showed neutralizing activities against HAdV-55, HAdV-14/55H1-7, HAdV-14/55H1, HAdV-14/55H2, HAdV-14/55H5, and to a lesser extend HAdV-14/55H4 and HAdV-14/55H7 (Figure 5B). This result demonstrated that a single high dose of rAd55-5E4 elicits nAbs against HVR1, -2, and -5 of HAdV-55 hexon, and to a lesser extent against HVR4 or -7 (Figures 5B and 5D). After the booster vaccination, the sera of macaques receiving rAd55-5E4, regardless of low or high doses, showed significant neutralizing activities against HAdV-14/55H1, HAdV-14/55H2, HAdV-14/55H5, HAdV-14/55H7, and HAdV-14/55H1-7, as well as HAdV-55 (Figures 5C and 5E), suggesting that a booster vaccination expands the epitopes recognition profiles of nAb response. The sera from two-dose-vaccinated macaques (groups 1 and 2) showed moderate inhibitions against HAdV-14, suggesting that the booster vaccination increased the fiber-specific nAb response, but the potency was less than hexon-specific nAbs (Figures 5C, 5E and 5F). In contrast, iAd55 mainly elicited nAbs against hexon HVR1 but minimal or no nAbs against other hexon HVRs and fiber, even after the booster vaccination (Figures 5C, 5E, and 5F). Taken together, a single high dose of rAd55-5E4 induced nAb epitope recognition spectrum was broader than that of a low-dose vaccination. Regardless of the vaccination dose and times, rAd55-5E4 induced nAb epitope recognition that was superior to iAd55, including hexon HVR and fiber.

Figure 5.

Epitope recognition profiling for macaque antibody responses

(A) Genome structure of HAdV14/55HVR chimeric reporter viruses. (B–E) Mapping the neutralizing epitopes recognized by macaque nAbs. Serum from macaques 2 weeks after primary (B) or booster (C) vaccination was serially diluted (from 1:100 to 1:312,500) and incubated with HAdV14, HAdV55, or HAdV14/55HVR chimeric reporter viruses for testing. The nAb titers were calculated as the reciprocal of the dilutions at which 50% of EGFP expression was inhibited. The heatmap shows the means of nAb titers of the three recapitulations. (D) HVR epitopes analysis for primary vaccination serum. (E) HVR epitopes analysis for boost vaccination serum. The values show the percentage of nAb titers contributed by each HVR epitope in different groups to HAdV-55 nAb. (F) Western blotting verified the recognition of HAdV-55 fiber knob protein by the sera of primed (top) and boosted (bottom) vaccinated macaques. Purified HAdV-55 fiber knob was electrophoresed on the gel. After transfer, all the macaque sera were incubated at a dilution of 1:2,000 for binding antibody test.

We also assessed the CMI response elicited by rAd55-5E4 in macaques. After the priming vaccination, IFN-γ⁺ SFCs were detectable in high-dose rAd55-E4-vaccinated macaques (Figures 6A–6C). After the booster vaccination, the frequencies of IFN-γ⁺ SFCs were significantly elevated in rAd55-E4 vaccinated macaques, the values reached a peak at 4 weeks and declined thereafter (Figures 6A and 6B). The frequencies of IFN-γ⁺ SFCs of iAd55-vaccinated macaques were only mildly elevated. This result suggests that rAd55-5E4 is more effective than iAd55 in inducing a CMI response.

Figure 6.

CMI response elicited by rAd55-5E4 in Chinese macaques

(A) Frequencies of IFN-γ⁺ SFCs in the PBMCs of macaques receiving 2 × 10¹⁰ vp of rAd55-5E4. (B) Frequencies of IFN-γ⁺ SFCs in the PBMCs of macaques receiving 1 × 10¹¹ vp of rAd55-5E4. (C) Frequencies of IFN-γ⁺ SFCs in the PBMCs of macaques receiving 1 × 10¹¹ vp of iAd55. (D) Frequencies of IFN-γ⁺ SFCs in the PBMCs of macaques receiving PBS. PBMCs were stimulated with HAdV-55 hexon peptide pool. Each graph shows the frequencies of IFN-γ⁺ SFCs at the indicated time points. Gray background indicates the time frame after the booster vaccination. The values for each time point reflect the mean of three technical replicates. Error bars represent mean ± SD.

Passive transfer of macaque immune sera protected mice against HAdV-55, -11, and -14 infection

We first tested whether serum antibodies from vaccinated macaques can confer protection against HAdV-55 infection in hDSG2 transgenic mice. Mice were injected intraperitoneally with 40 μL of immune macaque sera from each vaccination group 1 day before the rAdV55-SEAP challenge (Figure 7A). Serum samples were collected at 10 months after the booster vaccination and pooled with equal portions from each macaque in that group. Mice that received PBS or the sera of non-vaccinated macaques were used as controls. Passive transfer of rAd55-5E4 immune sera could reduce the recipients’ serum SEAP activities caused by rAd55-SEAP infection, and the sera of high-dose-vaccinated macaques showed a greater protection (Figure 7B). Passive transfer of iAd55 immune sera could modestly reduce the recipients’ serum SEAP activities but to a much lesser extent than rAd55-5E4 immune sera (Figure 7B). Compared with the non-vaccine macaque sera, the immune sera from low-dose rAd55-5E4-vaccinated macaques or high-dose iAd55-vaccinated macaques reduced the viral genome copy numbers in the liver by 2-fold, whereas the immune sera from high-dose rAd55-5E4-vaccinated macaques reduced the viral genome copy numbers in the liver by 11-fold (Figure 7C), demonstrating that rAd55-5E4 elicits more robust protective nAb response against HAdV-55 than did iAd55. Notably, the immune sera of from high dose rAd55-5E4 vaccinated macaques also protected against infection of rAd11-SEAP and rAd14-SEAP, as evidenced by the reduction in serum SEAP activities and viral genome copy numbers in the liver (Figures 7D and 7E). The immune sera from low-dose rAd55-5E4-vaccinated macaques or high-dose iAd55-vaccinated macaques showed comparable protective efficacy, but both were less effective than the immune sera from high-dose rAd55-5E4-vaccinated macaques (Figures 7D–7G). Taken together, rAd55-5E4 vaccination in non-human primates can elicit protective nAbs against infections of HAdV-55, HAdV-11, and HAdV-14.

Figure 7.

Protective effects of macaque immune sera against HAdV-55, -11, and -14 infection in hDSG2^+/+ transgenic mice

(A) Schematic diagram of serum transfer and HAdV challenge assays in hDSG2^+/+ mice. Six-week-old hDSG2^+/+ mice (n = 5 per group) were i.p. administered with macaque sera at the indicated dosage. One day later, mice were i.v. challenged with rAd55-SEAP, rAd11-SEAP, or rAd14-SEAP at the indicated dosage. At 5 days after challenge, mice sera were collected and the SEAP activities were measured. At 5 days after challenge, mice were sacrificed and the liver tissues were harvested and subjected to viral load examination. (B) SEAP activities in mice sera after rAd55-SEAP challenge. (C) HAdV-55 genome copies in mice liver tissues. (D) SEAP activities in mice sera after rAd11-SEAP challenge. (E) HAdV-11 genome copies in mice liver tissues. (F) SEAP activities in mice sera after rAd14-SEAP challenge. (G) HAdV-14 genome copies in mice liver tissues. rAd55^lo, transfer with the sera of macaques receiving 2 × 10¹⁰ vp of rAd55-5E4; rAd55^hi, transfer with the sera of macaques receiving 1 × 10¹¹ vp of rAd55-5E4; iAd55, transfer with the sera of macaques receiving 1 × 10¹¹ vp of iAd55; no vac, transfer with the sera of macaques receiving PBS. PBS, transfer with an equal volume of PBS. Each data point represents the mean value of triple technical replicates for one mouse. Error bars indicate means ± SDs. Statistical analysis is performed by one-way ANOVA. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. ns, not significant.

Discussion

The sporadic outbreaks of HAdV-55 poses a public health challenge. The replication-incompetent rAd55-5E4 described here presents a promising vaccine candidate for protection against HAdV-55-, HAdV-11-, and HAdV-14-related infections and diseases. The oral administration of wild-type HAdV-4 and -7 have demonstrated protective efficacy, resulting in a nearly 100-fold decrease in HAdV cases in the US military within 2 years of reintroduction.³⁷ However, this vaccine is limited in its use among healthy military recruits at military bases, as infectious viruses can be shed in the stool for up to 4 weeks following vaccination,¹⁶ leading to the risk of transmission among close contacts and vulnerable populations such as immunocompromised individuals, infants, and elderly people.³⁸ In this context, rAd55-5E4 can be propagated only in cell lines that provide the HAdV-5 E1 protein in trans (e.g., HEK293) and is unlikely to produce replication-competent HAdV due to the low sequence homology of the left arm and the absence of continuously repeated similar sequences, thus providing greater safety. Indeed, vaccines based on replication-incompetent HAdV vectors, such as E1-and E3-deleted HAdV-5, -26, and -35, have shown good safety records in clinical trials and real-world applications.^39,40,41 Unlike the inactivated iAd55 vaccine, rAd55-5E4 elicited more robust nAb and CMI responses in mice and rhesus macaques. The rAd55-5E4-elicited nAb response also showed good durability and passive transfer of immune sera collected nearly 1 year after vaccination effectively protected hDSG2 mice against HAdV-55, HAdV-11, and HAdV-14 infection.

Various technologies have been employed to develop vaccines against pathogenic HAdVs, including live-attenuated HAdVs, hexon-chimeric HAdVs, genetic vectors, virus-like particles (VLPs), and subunit vaccines.^{24,42,43,44,45} Deletion of the E3 genes reduces the cytopathic effects of HAdV-4 and -7 but does not affect the immunogenicity and replication capabilities.⁴² However, safety concerns still exist with replication-competent HAdVs, especially in immunocompromised individuals.^46,47 Displaying the hexon HVRs of a pathogenic HAdVs on the capsid of less virulent HAdVs are able to elicit nAb response against both types of HAdVs, but the titers are usually lower than those elicited by the parental HAdVs.^23,48 A recombinant HAdV-5 carrying an expression cassette for HAdV-3 hexon elicits HAdV-3 nAb response,⁴⁴ but improper folding of HAdV-3 hexon in the absence of homologous L4-100K may dampen the immunogenicity.²⁷ HAdV-7 VLPs, which resemble wild-type HAdV-7 but lack genetic material, can induce an HAdV-7 nAb response but lack a CMI response.⁴⁵ Recombinant hexon or fiber proteins have also been explored as subunit vaccines, which can induce a nAb response.^25,43 However, the production of VLPs can be expensive, and the immunogenicity of subunit vaccine, especially the CMI response, may be limited.^49,50 Compared with these approaches, rAd55-5E4 has significant advantages. Its replication-incompetent nature enhances safety compared with replication-competent HAdVs, which may expand the use of this vaccine in immunocompromised and pregnant individuals.^51,52 During the coronavirus disease 2019 (COVID-19) pandemic, vaccines using ChAdOx1 and HAdV-26 vectors caused rare but serious side effect of thrombosis with thrombocytopenia syndrome,^53,54 but no such side effect was observed with COVID-19 vaccines using the HAd-5 vector. Whether this is related to adenovirus serotype is unclear. Whether rAd55-5E4 candidate vaccine carries a potentially risk is also unknown. This requires further clinical trials to investigate.

rAd55-5E4 mimics the binding and entry of HAdV-55 and retains all the native neutralizing epitopes, providing superior immunogenicity over hexon-chimeric HAdVs, VLPs, and subunit vaccines. In addition, rAd55-5E4 can be produced in HEK293 cells at lower cost than other technologies, which has been demonstrated by AstraZeneca’s production and global distribution of billion doses of adenovirus vectored COVID-19 vaccine during severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. These merits help to improve vaccine accessibility and expand the scope of target population. An analysis of published reports revealed that the seroprevalence of HAdV-55 in healthy people is lower than commonly used HAdV-5, ranging from 13.2% to 28.2% and 54.9% to 84.5%, respectively.⁵⁵ Combined with the safety and high delivery efficiency of the replication-incompetent HAdV-55, rAd55-5E4 has the potential to be developed as a new vaccine vector.

The development of HAdV vaccines has been hindered by the lack of animal models that enable human HAdV infection, partly due to the variations in proteins receptors between humans and animals that facilitate virus infection. Rodent cells are not permissive for HAdV-Bs infection, partially due to differences in cellular receptors, such as CD46 and DSG2, which share only 46.6% and 80.7% aa identity between mouse and humans.²² The use of hDSG2 transgenic mice enables the assessment of the protective effects conferred by rAd55-5E4. Using this model, we have demonstrated that vaccination with rAd55-5E4 protected against HAdV-55 infection. The efficacy correlated well with the titers of HAdV-55 nAbs, in line with observations from clinical trials of the live, oral HAdV-4 and -7 vaccines.⁵⁶ However, rodent cells cannot fully support HAdV genome replication and late mRNA translation,^57,58 indicating that, even with hDSG2 expression, mice may not accurately recapitulate the ARDs caused by HAdVs in humans.⁵⁹ We recently showed that monoclonal nAbs from rAd55-5E4-vaccinated rhesus macaques effectively prevented HAdV-55-casued pneumonia in tree shrews, a type of small mammals closely related to primates.⁶⁰ This study demonstrated that the nAb response, if high enough, can be sufficient to prevent HAdV-55 infection and diseases. Nevertheless, a comprehensive assessment of the protective efficacy of rAd55-5E4 should be performed in future studies using more appropriate animal models, such as tree shrews or, ideally, non-human primates.

Notably, rAd55-5E4 elicited a more robust nAb response than iAd55 in C57BL/6 mice and rhesus macaques, indicating the advantage of using a live virus over an inactivated whole virus. After inoculation, rAd55-5E4 could be internalized by target cells, especially antigen-presenting cells such as macrophages and dendritic cells.⁶¹ The genome of rAd55-5E4, containing multiple CpG motifs, may trigger innate immunity by activating a variety of intracellular sensors and cytosolic DNA sensors.^62,63 The resultant cytokine milieu may play a role in B cell development and IgG production.⁶⁴ As for iAd55, treatment with β-propiolactone destroys viral genome and maybe also damage the antigenicity of vps,⁶⁵ which impacts the induction of the nAb response. In rhesus macaques, both receptor-dependent entry and phagocytosis may occur after rAd55-5E4 inoculation.²² In this context, antigenic components of rAd55-5E4 might be presented to host immunity more efficiently than iAd55. It has been reported that E1-deleted HAdVs still express the early and late genes, but at much lower levels.⁶⁶ The expression of viral antigens, together with the adjuvant effects of viral genome, contribute to the induction of a higher nAb response than iAd55.

Another merit of rAd55-5E4, as a live replication-incompetent vaccine, is that it is able to provoke significant CMI response. Although nAb responses have been recognized as a benchmark for protection against HAdV infection, T cell responses are also involved in controlling HAdV diseases.^67,68 CD4⁺ T cells provide essential signals to B cells for affinity maturation and memory formation, thereby maximizing the nAb response.^69,70 CD4⁺ T cells also optimize CD8⁺ T cell response in magnitude and quality.^71,72 For example, CD8⁺ T cells facilitate the clearance of HAdV-infected host cells and inhibit disease progression.^73,74 Indeed, reconstitution of HAdV-specific CD4⁺ and CD8⁺ T cells ameliorates HAdV infection in stem cell transplant recipients.⁷⁴ In addition, rAd55-5E4 could elicit a CMI response against the conserved region of the hexon protein. These responses not only add to the protection against HAdV-55 infection but may also provide cross-protection against other pathogenic HAdVs containing such epitopes. A recent paper reported that intranasal administration of a SARS-CoV-2 vaccine can induce respiratory mucosal secretory IgA antibodies that possess broader and stronger neutralizing activity,⁷⁵ suggesting that intranasal administration may be a vaccination route suitable for preventing respiratory viral infections. In the current study, we focused on evaluation of the immunogenicity of rAd55-5E4 via i.m. injection. Future studies will be carried out to investigate the effects of intranasal or oral immunization routes on inducing mucosal and systemic immune responses.

In summary, we demonstrate that replication-incompetent rAd55-5E4, with the E1 and E3 genes deleted and the E4 genes replaced with those of HAdV-5, can only be produced in HEK293 cells or cell lines that provide E1 proteins in trans, can elicit durable protective immunity in both mice and non-human primates. These results support rAd55-5E4 as a promising vaccine candidate that warrants further development.

Materials and methods

Ethics statement and animals

All animal experiments were carried out at the Laboratory of Animal Center of Guangzhou Institute of Biomedicine and Health (GIBH), Chinese Academy of Sciences. The experimental protocols have been approved by the Institutional Animal Care and Use Committee of GIBH (Nos. 2015014, 2018027, for mice experiments and 2018042 for Chinese macaque experiments). The 6-week-female BALB/c mice were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd. The hDSG2 transgenic mice were constructed as previously described²² and kept in the animal center. The Chinese macaques (4 years old, female, 3–5 kg) were purchased from Guangdong Landau Biotechnology Co. Ltd. All the Chinese macaques have been confirmed to be negative for nAbs against HAdV-55, HAdV-11, and HAdV-14.

Cells and adenoviruses

The HEK293 (No. CRL-1573) and A549 (No. CCL-185) cells were purchased from the American Type Culture Collection (ATCC) and kept in the laboratory which was cultured in DMEM with 10% fetal bovine serum (FBS). The adenoviruses used in this study, HAdV-55 (GenBank: NM_KF911353.1) and HAdV-14 (GenBank: NM_JQ824845.1) have been mentioned in the previous article.¹³ HAdV-11 (GenBank: NM_AF532578.1) was purchased from ATCC (No. VR-12). Replication-competent HAdV-11, HAdV-14, and HAdV-55, which carried the expression cassettes for EGFP or SEAP in the original E3 region, as well as the HVR chimeric HAdV-14 viruses expressing the EGFP reporter, have been described previously.^13,60,23 All the HAdVs used in the research were propagated in HEK293 cells and purified by cesium chloride gradient centrifugation. The virus particle (VP) assay is based on a spectrophotometric estimation of the amount of protein and nucleic acid in a purified rAd55 preparation. The number of VPs is calculated using the formula: VP count = absorbance at 260 nm × dilution factor × 3.5 × 10⁹. Virus stocks were stored at −80°C.

Peptides and proteins

Peptides of 20 aa in length, with 10-aa overlaps between two consecutive peptides, were chemically synthesized. Given that HAdV-55 hexon is similar to HAdV-11 hexon but distinct from HAdV-14 hexon in the seven HVRs, the peptides corresponding with the HAdV-55 hexon were divided into six peptide pools according to the location of the HVRs (Figure 2E): P1 (1–130 aa), P3 (121–320 aa, HVR1–6), P4 (310–410 aa), P6 (400–460 aa, HVR7), P7 (450–710 aa), and P8 (700–947 aa). In addition, we synthesized two peptide pools covering HAdV-14 HVR1 to HVR6 as well as HVR7, designated as P2 (121–320 aa, HVR1–6) and P5 (400–460 aa, HVR7), respectively (Figure 2E). In this context, IFN-γ-secreting cells stimulated by P1, P3, P4, P6, P7, and P8 represent the CMI responses against HAdV-55, HAdV-11, and HAdV-14, whereas those stimulated by P2 and P5 mostly represent the CMI responses against HAdV-14.

HAdV-55 fiber knob proteins (aa 118–325) were produced in E. coli as described previously.²³

Constructions of rAd55-5E4

The genomic plasmid for rAd55-5E4 was constructed based on our previously reported pAd55ΔE3.¹³ In brief, the left and right homologous recombination fragments corresponding with the HAdV-55 E1 region were amplified by PCR using primer 55E1L F and 55E1L R (for the left fragment) and primer pair 55E1R F and 55E1R R (for the right fragment) (Table S1). The fragments were cloned into a pVAX1 plasmid (Thermo Fisher Scientific) to obtain the pGK551 shuttle plasmid. The HAdV-55 E1 region was deleted by homologous recombination between linearized pGK551 and pAd55ΔE3 to obtain pAdV55ΔE1ΔE3 in competent E. coli BJ5183 cells (Agilent). Subsequently, the HAdV-55 E4 region (nt 32,036–33,954) was amplified by PCR using primer 55E4 F and 55E4 R (Table S1) and cloned into a pMD18T (TaKaRa) vector to obtain the p55E4 shuttle plasmid. The HAdV-5 E4 region (nt 33,188–35,106, covering ORF2 to ORF6) was also amplified using primer pair 5E4 F and 5E4 R (Table S1). HAdV-5 E4 region was then inserted into p55E4 to replace the corresponding E4 ORFs of HAdV-55 by homologous recombination. The resultant p55-5E4 was linearized by enzymatic restriction digestion with MluI and PmeI (New England Biolabs [NEB]) and subjected to homologous recombination with linearized pAdV55ΔE1ΔE3 to obtain the genomic plasmid pAdV55ΔE1ΔE3-5E4. Finally, the genomic sequence of rAd55-5E4 was released from pAdV55ΔE1ΔE3-5E4 by enzymatic restriction digestion with AsiSI (NEB) and transfected into HEK293 cells to rescue rAd55-5E4. rAd55-5E4 was propagated and titrated similarly as described above.

iAd55 was obtained by treating rAd55-5E4 with 0.1% β-propiolactone at 4°C overnight.

SDS-PAGE

We incubated 1 × 10¹⁰ vp of purified rAd55-5E4, rAd55-EGFP, and rAd5 VPs with SDS and β-mercaptoethanol at 100°C for 10 min. Subsequently, the lysates were subjected to SDS-PAGE. The gel was stained with Coomassie Brilliant Blue staining solution (Genestar). The images were captured using the Gel Doc XR Imaging System (Bio-Rad).

Plaque formation assays

The HEK293 and A549 cells were seeded into six-well plates and subjected viral infection after 24 h, with doses of rAd55-5E4-EGFP and rAd55-EGFP at 1 × 10⁵ vp per well. At 2 h after infection, the culture medium was replaced with MEM containing 2% FBS and 0.7% agarose gel. The plates were then incubated at 37°C with 5% CO₂ in the atmosphere for approximately 10 days. The formation of fluorescent plaques by rAd55-5E4-EGFP and rAd55-EGFP was observed under an inverted fluorescence microscope (Nikon) and representative images were collected.

Growth curves

The HEK293 and A549 cells were seeded into 96-well plates. The HEK293 cells were infected 24 h later with rAd55-5E4-EGFP and rAd55-EGFP at 5 × 10⁶ vp per wells. The A549 cells were infected with rAd55-5E4-EGFP and rAd55-EGFP at 4 × 10⁷ vp per wells. At 3 h after infection, the culture media were replaced with DMEM containing 5% FBS. At 12, 24, 36, 48, 60, 72, and 96 h after infection, cells were harvested, lysed by three freeze-thaw cycles, and subjected to quantitative PCR (qPCR) to measure the viral genome copies.

Cytopathic effect detection

HEK293 and A549 cells were infected with rAd55-5E4-EGFP or rAd55-EGFP as described. At 12, 24, 36, 48, 72, 96, 120, and 144 h after infection, the cell cultures were treated with CCK-8 detection reagent (Vazyme) at 10 μL per well and incubated for another 2 h. Finally, the optical density value at 450 nm were recorded with a Microplate reader (Bio Tek).

Animal experiments

To test the immunogenicity of rAd55-5E4 in mice, 6-week-old female BALB/c mice were i.m. immunized with rAd55-5E4 at 2 × 10⁸ or 2 × 10⁹ vp per mouse, or with iAd55 at 2 × 10⁹ vp per mouse (termed as iAd55 hereafter), twice at a 4-week interval. Control mice were inoculated with an equal volume of PBS. At 2 weeks after each immunization, one-half of the mice in each group were euthanized. Mice sera were collected and the nAb titers were examined by neutralization assays. Mice spleens were harvested and splenic lymphocytes were isolated and subjected to ELISpot assay.

To test the protective efficacy of rAd55-5E4 in mice, hDSG2 transgenic C57BL/6 mice (male and female) were i.m. immunized once with rAd55-5E4 or iAd55 at the same doses as mentioned above. At 6 weeks after immunization, mice were intravenously (i.v.) challenged with 2 × 10¹⁰ vp of rAd55-SEAP, rAd11-SEAP, or rAd14-SEAP. At 3 and 7 days after infection, sera from mice challenged with the SEAP reporter viruses were collected and the SEAP activities was measured using the Phospha-Light System (Thermo Fisher Scientific). At 7 days after infection, mice were sacrificed. The liver tissues were harvested and homogenized, and the viral genomes were extracted using QIAamp DNA Mini Kit (Qiagen) and quantified by qPCR using iTaq Universal SYBR green Supermix (Bio-Rad) according to the manufacturer’s instructions.

To test the immunogenicity of rAd55-5E4 in non-human primates, twelve 4-year-old female Chinese rhesus macaques randomly divided into 4 groups (Table S2). Macaques in groups 1–3 were i.m. immunized with 2 × 10¹⁰ vp of rAd55-5E4, 1 × 10¹¹ vp of rAd55-5E4, and 1×10¹¹ vp of iAd55, respectively. Macaques in group 4 receiving PBS were used as controls. The macaques received a booster immunization 32 weeks later with each vaccine at similar doses as the first immunization. After each immunization, macaque sera were collected at the indicated time points as shown in Figure 4 and subjected to neutralization assays. PBMCs were isolated at 2, 4, or 6 weeks after each immunization and subjected to ELISpot assays.

To test the protective effects of immunization-elicited macaque nAbs, macaque sera from the same group collected at 66 weeks after the first immunization were equally pooled together and diluted 5-fold with PBS. The 6- to 8-week-old hDSG2 transgenic mice were adoptively transferred with 200 μL diluted macaque sera by an intraperitoneal injection. One day later, mice were i.v. challenged with 2 × 10¹⁰ vp of rAd55-SEAP, rAd11-SEAP, or rAd14-SEAP. At 5 days after the challenge, mice sera were collected and SEAP activities were measured. Mice were sacrificed, and viral genomes were extracted from mouse liver tissues and subjected to qPCR.

Neutralization assay

The titers of anti-HAdV nAbs in mouse or macaque sera were tested by neutralization assays as described previously. In brief, HEK293 cells were seeded in 96-well plates at 5 × 10⁵ cells per well and cultured for 2 days. Each serum sample was serially diluted in phenol red-free DMEM medium and incubated with rAd55-SEAP, rAd11-SEAP, rAd14-SEAP, or chimeric HAdV-14 reporter viruses at 2 × 10⁶ vp per well. One hour later, 200 μL of the mixtures was added onto HEK293 cells and incubated at 37°C for 24 h. For the SEAP reporter HAdVs, culture supernatants were harvested to measure the SEAP activities using the Phospha-Light System (Thermo Fisher Scientific) according to the manufacturer’s protocol. Relative light units were recorded by GloMax Discover Microplate Reader (Promega). For the chimeric HAdV-14 reporter viruses, EGFP signals were measured by ImageXpress Micro Confocal (Molecular Devices). The nAb titers were calculated as the reciprocal of the dilutions, at which 50% of SEAP or EGFP expression was inhibited.

ELISpot

The CMI responses induced by rAd55-5E4 were measured by IFN-γ ELISpot assays. In brief, sterile 96-well microtiter plates (Merck Millipore) were coated with anti-mouse or anti-monkey IFN-γ coating antibody (U-CyTech) at 4°C overnight. Mouse splenic lymphocytes were isolated using mouse 1× lymphocytes separation medium (Dakewe Biotech Co.). Macaque PBMCs were isolated using a density gradient medium Lymphoprep (Stemcell). The cells were seeded onto the ELISpot plates at 3 × 10⁵ cells per well and stimulated with hexon peptide pools (4 μg/mL each peptide) at 37°C overnight. Subsequently, cell cultures were discarded, and the plates were washed with PBST five times. The plates were then added with biotinylated detection antibody (U-CyTech) and incubated at 37°C for another 16 h. Finally, the plates were developed with alkaline phosphatase conjugated streptavidin (U-CyTech) and NBT/BCIP reagent (Pierce). The spots were counted with an ELISpot reader (Bioreader 4000).

qPCR

The growth curves of rAd55-5E4 in cell lines and HAdV genomic copies in mouse liver tissues were determined by qPCR. In brief, viral genomes were extracted from cell lysates or tissue homogenates using QIAamp DNA Mini Kit (Qiagen) and subjected to qPCR using iTaq Universal SYBR green Supermix (Bio-Rad). The primers have been listed in Table S1. The procedures were set up as 95°C for 5 min; 40 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s. Melting curves were established at 65°C–95°C with an increment of 0.5°C per cycle for 5 s. Standard curves were established using serially diluted genomic plasmids for HAdV-55, HAdV-11, and HAdV-14 to calculate the genome copies.

Western blot analysis

To test the fiber-specific antibodies elicited by rAd55-5E4, 1 μg of purified HAdV-55 fiber knob protein was subjected to SDS-PAGE in the presence or absence of β-mercaptoethanol. After transfer, the membranes were blocked using 5% milk in PBST at room temperature for 2 h, and then incubated with each macaque serum (1:2,000 dilution) for 1 h. Finally, the membranes were incubated with HRP-conjugated goat anti-human secondary antibodies (SeraCare Life Sciences) and developed with chemiluminescent HRP substrate (Merck).

Statistical analysis

Statistical analyses and graphical presentations were conducted with GraphPad Prism version 8.0 (GraphPad Software). Differences among groups were determined by one-way ANOVA. Throughout the text, figures, and legends, the following terminology is used to show statistical significance: ∗p < 0.05; ∗∗p < 0.01; and ∗∗∗p < 0.001.

Data and code availability

The data supporting the findings of this study are available within the paper (and its supplemental information files).

Acknowledgments

This study was partially supported by the National Natural Science Foundation of China (32200759 and 32370998), the National Key Research and Development Program of China (2024YFA10920001), the Youth Innovation Promotion Association of CAS (2022361), the Guangdong Basic and Applied Basic Research Foundation (2022A1515012002 and 2021A1515110549), the China Postdoctoral Science Foundation (2020M682942), the Open Fund from the State Key Laboratory of Respiratory Disease (SKLRD-Z-202109, SKLRD-Z-202319, and SKLRD-Z-202208), and the Guangzhou Basic Research Program-City-University(Hospital)-Enterprise Joint Funding Project (2025A03J3621). We thank Prof Qiwei Zhang (Jinan University, China) and Prof Chengfeng Qin (Beijing Institute of Microbiology and Epidemiology, China) for help in providing research material. We thank Rui Mao, Zhi Wang, Cuie Li, Xiangjie Feng, and Weimin Zhang for their assistance in macaque and mouse experiments.

Author contributions

Y.F. conceived and designed this study, performed the experiments, analyzed the data and wrote the manuscript. T.S., L.L., X.S., and C.Y. were involved in the development of rAd55-5E4, conducted the virological assays and mouse studies. W.L., L.H., J.L., S.Z., W.Z., C.Y., C.W., and S.C. provided reagents. X.Z., X.Y., X.S., Y.L., and Z.L. were involved in the macaque experiments and virological and immunological assays. X.N., R.Z., and P.L. revised the manuscript. L.F. and L.C. conceived the study, interpreted the data, and wrote and revised the manuscript. All authors edited and approved the final manuscript.

Declaration of interests

The Guangzhou nBiomed Ltd. has filed patents on a replication-incompetent adenovirus type 55 vaccine naming L.C. and L.F. as inventors. C.W. and L.C. are employees of Guangzhou nBiomed Ltd. L.C. serves as Chief Scientific Advisor for Guangzhou nBiomed Ltd.