A specific HAdV7-Luc recombinant virus was constructed by replacing the E1A region with a luciferase expression cassette and used for the neutralization assay. To determine the optimal conditions for HAdV7-Luc infection of target A549 cells, infections with HAdV7-Luc were performed using different concentrations of viral particles. A linear relationship between the fluorescence value and multiplicity of infection (MOI) was obtained at MOI = 0.0625–1 (Figure 1A), and no significant difference was observed when using different infectious doses (MOI = 0.25, 0.50, 0.75, or 1.00) to detect the nAb values of ten HAdV7-positive serum samples (one-way ANOVA test, n = 10, p > 0.05) (Figure 1B). Therefore, an infectious dose with a MOI value of 0.75 was chosen as the optimal viral dose for this assay. To further verify the reliability of the assay, we measured the same serum samples using the traditional CPE (cytopathic effect) method, and the results showed good consistency (R² = 0.8612, p < 0.0001) (Figure 1C). These results demonstrated the stability and repeatability of the HAdV7-Luc-based neutralization assay.

We first evaluated the overall prevalence of nAbs against HAd7-Luc in sera collected from 2,350 participants compared with that of nAbs against HAd5-Luc. HAdV nAb titers were divided into four subgroups (< 12, negative; 12–200, low; 200–1,000, moderate; >1,000, high), similar to those reported in other HAdV seroprevalence studies (26). Intriguingly, most infected people had low nAb titers (47.1%; 95% CI: 45.0–49.1%) against HAdV7, while approximately even distribution of acquired nAb titers were observed for HAdV5 (Figure 2A). The overall seroprevalence of anti-HAdV was higher for HAdV5 (74.7%; 95% CI: 73.0–76.5%) than for HAdV7 (58.6%; 95% CI: 56.6–60.6%) (Pearson chi-squareare test, p < 0.001) (Figure 2B). It was found that 11.5% (95% CI: 10.2–12.8%) of the participants had moderate or high HAdV7 nAb titers (>200), and 1.4% (95% CI: 0.9–1.8%) had very high titers (>1,000); these proportions were much lower than the corresponding rates of 50.7% (95% CI: 48.7–52.7%) and 23.1% (95% CI: 21.4–24.8%) of participants with HAdV5 nAb titers (Pearson chi-squareare test, p < 0.001) (Figure 2C). The overall geometric mean titer of HAdV7 nAb among HAdV7-seropositive samples was 74.4 (95% CI: 69.9–79.3), which was much lower than that for HAdV5 (408.0; 95% CI: 379.0–439.3) (Mann–Whitney test, p < 0.001) (Figure 2D). These results suggested that nAbs against HAdV7 were less prevalent and were present at lower titers than nAbs against HAdV5.

We further analyzed the samples collected from China and Sierra Leone respectively. As shown in Figure 3, the samples from Sierra Leone had higher seroprevalence rates (65.7%; 95% CI: 61.5–69.8%) but lower nAb titers (49.7; 95% CI: 44.6–55.3) and the proportion (5.4%; 95% CI: 3.4–7.4%) of HAdV7 nAbs at a moderate to high level (>200) were lower than the corresponding values of 13.2% (95% CI: 11.6–14.7%) in China (p < 0.001) (Figures 3A–D). Besides, among all regions and adenovirus types, participants from Xizang and Gansu had the highest proportions of high preexisting titers (>1,000) (Supplementary Figure S1A). For HAdV7, the highest seroprevalence and nAb titers were observed in Xizang (Supplementary Figures S1B, C). In addition, the proportion of moderate to high nAb titers (>200) appeared to be relatively higher in Xizang, Xinjiang, and Beijing (Supplementary Figure S1D). These results indicated that there were geographic variations in the prevalence of natural HAdV7 infection.

To illustrate the impact of age on HAdV7 nAb seroprevalence, serum nAb levels were stratified into four age groups, as shown in in Figure 4. We analyzed 1,664 serum samples according to age and sex because the remainder of the data were incomplete. The overall HAdV7 seropositive rates were 42.0% (95% CI: 38.0–46.0%), 60.0% (95% CI: 55.5–64.6%), 66.1% (95% CI: 61.3–71.0%), and 68.8% (95% CI: 63.2–74.3%) in the 18–30, 31–40, 41–50, and >50 years age groups, respectively. The seroprevalence significantly increased with age (Cochran–Armitage trend test, p < 0.001) (Figures 4A, B), although HAdV7 nAb levels were consistent among the different age groups (Jonckheere-Terpstra test, p = 0.093) (Figure 4C). The same trend was also observed among samples from Sierra Leone (Supplementary Figures S2A, B) and China (Supplementary Figures S2C, D). In addition, the proportions of individuals with moderate to high anti-HAdV5 antibodies (titers >200) were 5.3% (95% CI: 3.5–7.1%), 10.4% (95% CI: 7.5–13.2%), 10.3% (95% CI: 7.2–13.4%), and 13.4% (95% CI: 9.3–17.5%) in the corresponding age groups, which also increased with increasing age (Cochran–Armitage trend test, p < 0.001) (Figure 4D). In contrast, for HAdV5, a slightly different trend was observed; although increasing age was associated with a higher seroprevalence rate, younger age (< 30 years) appeared to be inversely associated with a higher titer (Figures 4C, D). These results indicated the correlation between age and naturally occurring nAbs against HAdV7, indicating a continuously increasing infection rate of HAdV7 among healthy adults with increasing age.

As with HAdV5, no statistically significant difference in anti-HAdV7 nAb levels between males and females was observed. The overall HAdV7 seropositive rates in males and females were 56.3% (95% CI: 52.9–59.6%) and 54.7% (95% CI: 51.3–58.1%), respectively, showing no significant differences (Pearson chi-square test) (Figures 5A, B), similar to those observed individually in China and in Sierra Leone (Supplementary Figures S3A, B). The overall HAdV7 nAb titers were also comparable between males and females (Supplementary Figures 5C, D), as well as in Sierra Leone (Supplementary Figure S3C), whereas higher HAdV7 nAb titers were detected in males than in females in China (Mann–Whitney test) (Supplementary Figure S3D). These results demonstrated the indiscriminate infection of HAdV7 between genders.

As shown in Table 1, the HAdV7 seroprevalence among HAdV5-seropositive samples was 59.1% (95% CI: 56.8–61.4%), with no significant difference from that among HAdV5-seronegative samples (57.2%; 95% CI: 53.2–61.2%) (Pearson chi-square test). Likewise, HAdV5 seroprevalence among HAdV7-seropositive samples (75.3%; 95% CI: 73.0–77.6%) was similar to that among HAdV7-seronegative samples (73.9%; 95% CI: 71.1–76.7%) (Pearson chi-square test). In addition, HAdV7 nAb titers in HAdV5-seropositive samples were comparable to those in HAdV5-seronegative samples, and HAdV5 nAb titers were unaffected by HAdV7 seroprevalence (Mann–Whitney test) (Figures 6A, B). These results demonstrated that antibodies against HAdV7 clearly not participated in effective barrier to infections against HAdV5, so as to antibodies against HAdV7 upon HAdV5 infection. Drawing on previous research (24, 27), these data indicated that antibodies against HAdV7 and HAdV5 provided hardly any cross protection.

The 5' untranslated region (UTR) and 3' UTR of HAdV7 were polymerase chain reaction (PCR)-amplified from the HAdV7-GZ08 (GQ478341) strain using the primers Ad7-re-Left-F/R and Ad7-re-Right-F/R, respectively, and the kanamycin coding sequence and PBR322 replication origin site from our previously generated pAd4-dE3-luc plasmid were amplified using primers Ad7-re-V1-F/R. Then, a UTR_kanamycin_ PBR322_UTR shuttle plasmid was produced using a Gibson Assembly Cloning Kit and designated pAd7-re. Next, pAd7-V0 was formed during homologous recombination of the HAdV7 genome with pAd7-re in Escherichia coli BJ5183 competent cells (Agilent Technologies, CA, USA). A replication-incompetent pAd7ΔE3 mutant was produced using the restriction enzyme EcoRI (New England Biolabs, MA, USA). For the construction of pAd7ΔE3-Luc, we first amplified the luciferase expression cassette from our previously generated pAd4ΔE3-Luc by PCR using the primers Ad7-Luc1-F and Ad7-Luc2-R. The E1 region was spliced from pAd7ΔE3 via digestion with AatII and BsiWI and replaced with the ligated luciferase amplicon. The origin shuttle plasmid pAd7-re (GenBank accession number: OP902171) and the ultimately used plasmid pAd7ΔE3-Luc (GenBank accession number: OP902172) were well confirmed by sequencing, demonstrating the correctness of the recombination. The oligonucleotide primers used in this study are summarized in Supplementary Table S1.

To produce recombinant HAdV7-Luc viruses, HEK293 cells were transfected with the linearized pAd7ΔE3-Luc plasmid. Cells were washed with PBS once and refreshed with DMEM the day after transfection. The proliferative viruses were purified by ion-exchange chromatography on a Source 30Q column (GE Healthcare, Beijing, China). Then, HAdV7 viral titers were determined in HEK293 cells using an Adeno-XTM Rapid Titer Kit (Clontech, CA, USA), as previously described (24). In brief, HEK293 cells [5 × 10 (4) cells/mL] in 24-well plates were infected with HAdV7-Luc after 10-fold serial dilution. Specific anti-adenovirus hexon antibodies were used to label the infected viruses, and the visualized infected cells were quantified using a microscope. The infectious units (IFU)/mL was calculated as [infected cells/field) × (field/well)/(volume virus (mL) × (dilution factor)].

A total of 2,350 serum samples from individuals in China and Sierra Leone in West Africa were included in this seroepidemiological study. Among them, 734 samples were from Hubei obtained in March 2020, 43 from Gansu obtained in August 2019, 169 from Sichuan obtained in August 2019, 273 from Beijing obtained in March 2017, 120 from Jiangsu obtained throughout the year 2015, 206 from Xinjiang obtained in August 2018, 307 from Xizang obtained in August 2020, and 498 from Freetown, the capital of Sierra Leone, obtained in October 2015. The blood donors were all healthy adults between 18 and 65 years old. A total of 1,664 serum samples had complete age and sex information, while the rest of the samples had incomplete data. Samples collected in Sierra Leone (34), Hubei (35–37), and Jiangsu (38) were involved in the clinical trial of an HAdV-5-based vaccine. The demographic data of the enrolled participants are summarized in Table 2, and no statistically significant differences were identified with regard to sex and age. The sera obtained after centrifugation at 2,000 rpm for 15 min were kept in a deep freezer at −20°C and inactivated for 30 min at 56°C before use.

The study protocol was approved by the Scientific Review Committee and the relevant institutional review boards. The samples collected in Sierra Leone were involved in the clinical trial of a HAdV5-based vaccine, which was approved by the Sierra Leone Ethics Committee and the Pharmacy Board of Sierra Leone and was performed in accordance with the Declaration of Helsinki.

The adenovirus neutralization assay was performed as previously described (24), except a newly synthesized HAdV7-Luc virus was used. In brief, HAdV7-Luc was incubated with 3-fold serial dilutions of heat-inactivated sera (starting at 1:12 and ending at 1:8,748) for 1 h at 37°C before 2 × 10 (4) A549 cells were added to 96-well plates. Simultaneously, diluent without serum was used as a negative control. After incubation for 24 h at 37°C, the inoculum was removed, and luciferase expression was detected using a Luciferase Assay System (Promega, Madison, WI). The reciprocal of the dilution multiple corresponding to a 90% reduction in the fluorescence value (IC₉₀) compared to that of the negative control was designated as the nAb titer, using non-linear regression with four parameters in GraphPad.

In addition, a traditional cytopathic effect (CPE) assay was conducted to validate the stability, reliability, and accuracy of the HAdV7-Luc-based neutralization assay. A mixture of HAdV7-Luc and serum was incubated with HEK293 cells for 6 days before cell viability measurement using a Cell Proliferation Assay kit (Promega, Madison, WI). The correlation of these two methods was then analyzed using Spearman correlation analysis.

The recombinant HAdV-5 vector-based luciferase expressing virus was constructed using the AdMax system (Microbix Biosystems, ON, Canada) as detailedly described previously (24, 39). Briefly, E1/E3 region of Ad5-based vector was substituted by a luciferase expression cassette, then the recombinant adenovirus was generated in the adenoviral packaging cells and propagated on HEK293 cells. HAdV-5-specific nAbs were assessed with the same method described above instead that HAdV5-Luc reporter virus was used.

Descriptive statistics were generated. The software PASW Statistics 18 and GraphPad Prism 7.0 were used for data analysis. Seropositivity ratios and nAb titers were evaluated using the Pearson chi-square test and Mann–Whitney test, respectively. The trend of the association of seroprevalence with age was analyzed by the Cochran–Armitage trend test, and the trends in nAb titers among age groups were analyzed by the Jonckheere–Terpstra test. n.s., nonsignificant, p > 0.05; ^*p, between 0.01 and 0.05; ^**p, between 0.01 and 0.001; ^***p < 0.001.