Serum specimens from 12 healthy recipients of COVID-19 Ad26/Ad5 vaccines were collected at 1-, 2-, 3-, and 6-month(s) post-second vaccine dose (MPV). The samples were titrated to measure total Ig antibody responses to SARS-CoV-2 spike, RBD, S1, S2, nucleoprotein antigens and bovine serum albumin (BSA) as a negative control ( Supplementary Figure S1 ). Areas under the curve (AUC) were calculated and plotted to examine changes in antibody levels over time ( Figure 1 ). Spike-, RBD-, and S1- specific responses were detectable in all 12 vaccinees at 1 MPV and, as expected, declined over time. From 1 MPV to 6 MPV, the median values significantly decreased by 2.4 to 5.3-fold for the three different antigens ( Figure 1 ), consistent with previous observations (15). The anti-S2 responses were detected only in a fraction of vaccinees at low levels ( Supplementary Figure S1 ) and also dropped over time (median 3.6-fold decrease from 1 MPV to 6 MPV) ( Figure 1 ). The responses to nucleoprotein were below or close to the BSA control and did not change over time, except for one individual (subject 6) who demonstrated elevated levels of anti-nucleoprotein antibodies at 3 MPV and 6 MPV ( Figure 1 , Supplementary Figure S1 ), indicating a breakthrough infection prior to the 3 MPV time point.

We then examined Ig isotypes (IgM, IgG1-4 and IgA1, 2) elicited against SARS-CoV-2 spike ( Figure 2 ) and RBD ( Supplementary Figure S2 ) in sera from the 12 vaccinated subjects as compared to four samples from non-vaccinated COVID-19-negative controls. Total Ig responses were tested in parallel as positive controls. Spike-specific IgG1 and IgG4 were detected in almost all vaccinated subjects at 1 MPV (75 – 100% of responders) while no spike-specific IgM, IgG2, IgG3, IgA1 and IgA2 were detected ( Figure 2A ). At 2 MPV and 3 MPV, the spike-specific IgG1 responses were still detected in 100% and 75% of responders, respectively, but IgG4 responses dropped below background as soon as 2 MPV and remained undetected at 3 MPV. Interestingly, at 2 MPV very low IgG3 responses to spike were seen in 75% of vaccinees and also in 58% at 3 MPV, including in subject 6 with a breakthrough infection ( Figure 1 ). When we compared the levels of spike-specific IgG1 and IgG4 responses over time, a decrease was observed between 1 MPV and 3 MPV (median 1.2- and 1.3-fold decline for IgG1 and IgG4, respectively) ( Figure 2B ).

The results were similar for RBD-specific isotypes. IgG1 and IgG4 were detectable at 1 MPV and the levels declined over time ( Supplementary Figures S2A, B ). IgG4 dropped below background starting from 2 MPV while IgG1 remained detectable in 92 – 100% of the vaccinees. However, no RBD-specific IgG3 responses were detected at any time point ( Supplementary Figure S2A ).

The neutralization capacities of the vaccinees’ sera were evaluated against recombinant vesicular stomatitis virus (VSV) expressing spike glycoprotein of WT (Wuhan Hu-1), beta, alpha, and delta strains ( Figure 3 ). There was a trend of lower neutralizing titers against the beta and delta strains as compared to WT and alpha, similar to the pattern previously reported (16) and also seen with neutralizing activities induced by LNP-mRNA vaccines and after infection (25). The same trend was maintained from 1 MPV to 6 MPV. With a limited sample size (n = 12), a significantly lower titer was observed only against the beta vs. WT and alpha strains at 1 MPV ( Figure 3A ). When we examined the neutralization titers over time against each of the four different strains, we observed no significant decline in titers between 1 MPV and the subsequent time points ( Figure 3B ). These results indicate the stability of neutralization titers against the four variants over time in spite of declining spike-binding Ab levels during the same duration ( Figure 1 ), in agreement with previously reported data (26). When we normalized the neutralization titers to the spike-binding Ig levels as ratios of ID₅₀ to anti-spike total Ig AUC at 1 MPV vs. 6 MPV, a trend of increased ratios was observed that varied depending on the viral strains (from 2 to 5-fold increase) and did not reach statistical significance with this limited sample size ( Figure 3C ). Overall, the data indicate the maintenance of neutralizing potency against the early virus strains over six months post-vaccination.

In addition to virus neutralizing activities, we assessed the Fc-mediated capacities of spike- and RBD-specific antibodies in sera from the Ad26/Ad5 vaccine recipients. To this end, we measured C1q and FcγR binding using C1q and FcγRI, IIa and IIIa proteins in the multiplex bead assay ( Figure 4 , Supplementary Figures S3 , S4 ). Titrating amounts of C1q and FcγR binding to both spike- and RBD-specific antibodies were observed, and at 1 MPV the binding levels were above background for all 12 vaccine recipients ( Supplementary Figures S3 , S4 ). To show changes in the C1q and FcγR binding activity over time, AUC was calculated from each of the titration curves.

The C1q binding levels to spike- ( Figure 4A ) and RBD-specific ( Figure 4B ) antibodies varied among samples from the 12 vaccinees but all declined between 1 MPV and 6 MPV, with median fold decreases of 221 for spike and 6.1 for RBD. At 6 MPV, very low or no C1q binding activities were detectable ( Supplementary Figures S3 , S4 ). The declines were also noted from 1 MPV and 3 MPV at 7.1-fold for spike and 2.5-fold for RBD. The magnitude of the decrease was faster than that of the total Ig levels for spike but was comparable for RBD ( Supplementary Figure S5 ).

Similarly, the binding of FcγRI, IIa and IIIa receptors to spike- and RBD-specific antibodies were detected in all samples at 1 MPV, albeit at variable levels ( Figures 4C, D , Supplementary Figures S3 , S4 ). The levels of FcγR binding decreased significantly from 1 MPV to 3 MPV, with median fold decreases of 1.7 – 2.8 ( Figures 4C, D ). These rates of decline were not as steep as those of C1q binding ( Figures 4A, B ). The changes were slower than or consistent with the declining antibody levels observed over the same duration after vaccination ( Supplementary Figure S5 ).

We subsequently asked whether SARS-CoV-2 spike-specific antibodies elicited by Ad26/Ad5 vaccination (1 MPV) displayed differential properties and functions as compared to the antibodies studied in our previous study from LNP-mRNA vaccinees (1 MPV after 2 doses) and from convalescent COVID-19 patients (>189 days post-symptom onset) (25, 27). First, we compared the antibody isotypes in sera of these three groups. Comparing the percentage of responders for each of the antibody isotypes tested (IgM, IgG1-4 and IgA1-2), we observed distinct profiles among the groups ( Figure 5 ). Although all three groups generated IgG1 as the dominant IgG subtype, the Ad26/Ad5 vaccinees produced only IgG1 and IgG4 antibodies. In contrast, LNP-mRNA vaccination generated all four IgG subtypes as well as IgM, IgA1, and IgA2, although the frequencies varied among isotypes and for anti-spike vs. RBD antibodies. Likewise, the Ad26/Ad5 vaccinees differed from COVID-19-convalescent individuals, all of whom produced IgM and IgG1, and a fraction also had IgG2, IgG3, IgG4, and IgA1 ( Figure 5 ). The same isotype profile was apparent at earlier time points after COVID-19 (<37 days post-symptom onset) (28).

Next, we assessed the functional potencies of antibodies elicited by Ad26/Ad5 vaccine vs. LNP-mRNA vaccines vs. virus infection. We compared serum neutralization ID₅₀ titers against WT, alpha, beta and delta strains observed in the three groups ( Figure 6 ). Except for lower neutralization titers against the delta strain in Ad26/Ad5- vs. LNP-mRNA-vaccinated individuals, the overall ID₅₀ titers against all four strains were similar among the Ad26/Ad5, LNP-mRNA, and COVID-19 groups ( Figure 6A ). To account for the variable levels of spike-specific antibodies in the individual samples, we also calculated the ratios of neutralization ID₅₀ over spike-specific total Ig levels and observed no significant difference ( Figure 6B ), indicating comparable neutralizing potencies induced in the three groups.

Finally, we compared the potencies of Fc-mediated activities mediated by spike and RBD-specific serum antibodies elicited in the Ad26/Ad5, LNP-mRNA, and COVID-19 groups ( Figures 7 , 8 ). To account for the different antibody levels in the sera, we calculated the ratios of Fc activities over the level of total Ig binding the respective antigens. Comparison of C1q deposition revealed that LNP-mRNA-induced antibodies had a significantly higher potency than antibodies from COVID-19 convalescent patients ( Figure 7A ), consistent with our past findings (25). Ad26/Ad5-induced antibodies also exhibited higher potencies than COVID-19-induced antibodies, although a significant difference was observed with RBD-specific and not spike-specific antibodies ( Figure 7A ). The C1q binding potencies of antibodies elicited by Ad26/Ad5 and LNP-mRNA vaccines were comparable.

For FcγR binding, a similar pattern was observed for all three activating receptors studied (FcγRI, Iia and IIIa): the Ad26/Ad5 and LNP-mRNA groups had spike- and RBD-specific antibodies with higher binding capacities for FcγRs than the COVID-19 group ( Figure 7B ). The Ad26/Ad5-induced antibodies displayed a consistent trend of higher FcγR binding capacity than the LNP-mRNA-induced antibodies, although a significant difference was achieved only for the FcγRIIIa binding of anti-spike antibodies ( Figure 7B ). The greater Fc-mediated activity induced by vaccination vs. infection was similarly observed when we measured the FcγRIIIa intracellular signaling in a cell-based assay ( Supplementary Figure S6 ).

Correlation analysis of complement and FcγR binding potencies revealed distinct patterns among the three groups ( Figure 7C ). The LNP-mRNA group showed uniformly strong correlations among C1q and FcγR binding activities of spike-and RBD-specific antibodies, whereas the COVID-19 group exhibited significant but weaker correlations. In the Ad26/Ad5 group, on the other hand, poor or no correlations were observed for most of the studied parameters.

ADCP was also measured in titrated sera from the three groups ( Figures 8A, B ). The Ad26/Ad5 had lower spike-specific ADCP and higher RBD-specific ADCP than the other two groups ( Figure 8C ). However, when the serum antibody levels were accounted for by calculating the ADCP/total Ig ratios, a trend of higher ADCP potencies was observed in the Ad26/Ad5 group vs. the other two groups, albeit a significant difference was attained only for RBD-specific ADCP between the Ad26/Ad5 and LNP-mRNA groups ( Figure 8D ). Notably, whereas the levels of spike-specific total Ig correlated with spike-specific ADCP ( Figure 8E left panel), correlation was absent between RBD-specific total Ig and RBD-specific ADCP for all three groups ( Figure 8E right panel), suggesting potential differences in Fc functions of antibodies targeting different regions or epitopes of spike. Altogether, the examination of Fc functions revealed the distinctive qualities of antibodies elicited by Ad26/Ad5 vaccines, LNP-mRNA vaccines, and SARS-CoV-2 infection.

Sera were collected from 12 Sputnik V Ad26/Ad5-vaccinated individuals recruited for this study at the Administración Nacional de Laboratorios e Institutos de Salud “Dr. Carlos Malbrán” (ANLIS MALBRAN) in Buenos Aires, Argentina (A.E.V, A.E, C.P.). Volunteers received their first dose either on 29 December or 31 December 2020, and their second dose either on 20 January 2021 or 2 February 2021. Data presented in this study were from sera collected at 1, 2, 3 and 6 months post-second dose. Serum samples were obtained from human volunteers of both sexes and de-identified prior to use in experiments. Sex was not considered as a biological variable in this study. Serum collection at ANLIS MALBRAN was approved by the Research Ethics Committee of its Unidad Operativa Centro de Contención Biológica (UOCCB) on 9 February 2021. Written informed consent was obtained as per institutional policy.

For comparison, a total of 27 specimens from the recipients of Moderna mRNA-1273 or Pfizer/BioNtech BNT162b2 LNP-mRNA vaccines (1 MPV after 2 doses), 15 specimens from convalescent COVID-19 patients (>189 days post-symptom onset), and four specimens from contemporaneous non-vaccinated COVID-19-negative subjects were also evaluated. Eighteen of these specimens were obtained from volunteers enrolled in IRB-approved protocols at the Icahn School of Medicine at Mount Sinai (IRB#17-00060, IRB#19-01243) and the James J. Peter Veterans Affairs Medical Center (IRB#BAN-1604): seven [RN#1, RN#4 and RV#1-5] after LNP-mRNA immunization; seven [RP#2-5, 7, 12, 13] after infection; and four negative controls. Eight additional convalescent plasma samples (CVAP samples) were obtained from 199-243 days after symptom onset under the JJPVAMC Quality Improvement project “Evaluation of the clinical significance of two COVID-19 serologic assays”. Twenty post-LNP mRNA vaccination plasma samples were also collected from participants in the longitudinal observational PARIS (Protection Associated with Rapid Immunity to SARS-CoV-2, IRB#20-03374) study. All participants signed written consent forms prior to sample and data collection. All participants provided permission for sample banking and sharing. All samples were heat-inactivated before use.

SARS-CoV-2 spike and RBD proteins were produced as described previously (49, 50). S1 (amino acids 16-685), S2 (amino acids 686-1213), and nucleoprotein (amino acids 1-419) antigens were purchased from ProSci Inc, CA (#97-087, #97-079 and #11-184, respectively). All antigens were of the parental SARS-CoV-2 Wuhan-Hu-1 or WA1 strain.

Measurement of total Ig and Ig isotypes to SARS-CoV-2 antigen-coupled beads was performed as described (25, 27, 28). The quantification was based on median fluorescent intensity (MFI) values at the designated sample dilutions. For total Ig responses, specimens were diluted 10-fold from 1:100 to 1:100,000, reacted with antigen-coated beads, and treated sequentially with biotinylated anti-human total Ig antibodies and PE-streptavidin. The isotyping assays were performed at a single 1:200 dilution using human Ig isotype or subclasses antibodies and the MFI values were shown.

For the C1q assay, beads with spike-Ab or RBD-Ab complexes were incubated with C1q Component from Human Serum (Sigma, #C1740) for 1 hour at room temperature, followed by an anti-C1q-PE antibody (Santa Cruz, #sc-53544 PE) for 30 minutes at room temperature (25, 27).

For the FcγR assays, beads with spike-Ab or RBD-Ab complexes were incubated with His-tagged recombinant FcγRI/CD64 (R&D Systems™, #1257FC050), FcγRIIa/CD32a (R&D Systems™, #1330CD050/CF) or FcγRIIIa/CD16a (R&D Systems™, #4325FC050) proteins for 1 hour at room temperature, followed by an anti-His-PE antibody (R&D Systems™, #IC050P) for 30 minutes at room temperature. FcγRIIIa signaling was also assessed using a reporter cell co-culture system according to the protocol published previously (51).

The relative levels of binding were obtained as MFI, from which titration curves were plotted and areas-under the curves (AUC) were calculated.

This assay used 293T-hACE2-TMPRSS2 target cells which were seeded at a density of 4 × 10⁴ cells per well in flat-bottom 96-well plates (Fisher Scientific, #08-772-3) after coating with collagen (Millipore, #08-115). The cells were incubated at 37°C/5% CO₂ overnight (~20 hours). On the following day, recombinant VSV virions expressing SARS-CoV Spike and a GFP reporter were pre-incubated with 4-fold serially diluted serum (starting from 1:5 to 1:20,480) in DMEM with 10%FBS and 1%PenStrep for a minimum of 10 minutes at room temperature, and then added to the target cells. At 13 hours post-infection, GFP counts were acquired by the Celigo imaging cytometer (Nexcelom Biosciences, version 4.1.3.0). Each assay was performed in triplicates.

To calculate ID₅₀, GFP counts from “no serum” conditions were set to 100%; GFP counts of each serum-treated condition were normalized to no serum control well. Inhibition curves were generated using Prism 9.1.2 (225) (GraphPad Software) with “log (inhibitor) vs. normalized response-variable slope” settings.

Assays to measure spike-specific ADCP were performed using a protocol reported previously (25, 27). Briefly, FluoSpheres carboxylate-modified yellow-green fluorescent microspheres (Thermo Fisher, #F8823) were coupled with SARS-CoV-2 spike or RBD proteins using the xMAP Antibody Coupling Kit (5 µg protein/~36.4x10⁹ beads, Luminex #40-50016). Spike-conjugated microspheres were incubated with diluted plasma for 2 hours at 37°C in the dark. After washing and centrifugation (2,000 g, 10 min), the beads (~3x10⁸ beads, 10 µL/well) were incubated with THP-1 cells (0.25x10⁵ cells, 200 µL/well) for 16 hours. The samples were analyzed on an Attune NxT flow cytometer (Thermo Fisher, #A24858). Data analysis was performed using FCS Express 7 Research Edition (De Novo Software). ADCP scores were calculated as follows: (% microsphere positive cells) x (geometric mean fluorescent intensity of the microsphere positive cells)/1000.

Statistical analyses were performed as designated in the figure legends using GraphPad Prism 9 (GraphPad Software, San Diego, CA). Correlation matrices were generated using R version 4.1.0 (The R Foundation for Statistical Computing) and corrplot package.