Coronavirus spike protein-specific antibodies indicate frequent infections and reinfections in infancy and among BNT162b2-vaccinated healthcare workers

Based on the overall lower amino acid sequence identity (Supplementary Table 1) and on our previous work with SARS-CoV-210, spike protein subunit 1 (S1) was selected as the antigen to set up an enzyme immunoassay (EIA) for the detection of HCoV S binding IgG antibodies. S1 subunits were produced as mFc-fusion (S1-mFc) proteins in HEK-293F cells, purified (Supplementary Fig. 1) and used in EIA.

Seropositivity and HCoV spike-specific IgG antibody levels in children of 1 to 3 years of age

Serum specimens collected in 2009–2013 from 140 healthy children (74 male and 66 female, Table 1) at 1, 2, and 3 years of age were analyzed by EIA for antibodies against S1 proteins of four seasonal HCoVs (229E, HKU1, NL63, OC43), MERS and SARS-CoV-2. The absorbance values were converted to EIA units enabling reliable identification of seropositive samples and to compare antibody levels against different coronavirus species (Fig. 1). An increase in the geometric mean antibody levels between 1- and 2-year samples was significant for all four seasonal HCoVs (p < 0.0001) while the change between 2 and 3 years was significant only for NL63 and OC43 (p < 0.0001 and p = 0.0004, respectively). One participant had low levels of MERS and SARS-CoV-2 S1-binding IgG antibodies at 2 and 3 years of age.

Table 1 Characteristics of the study cohorts and serum sampling intervals.
Figure 1
figure 1

HCoV S1-specific IgG antibody responses in children at 1, 2 and 3 years of age. HCoV S1-specific IgG antibody levels were measured with EIA from sera collected from 140 children at 1, 2, and 3 years of age, during the years 2009 to 2013. Geometric means and geometric standard deviations of antibody levels are shown. Differences between the groups were analyzed using Wilcoxon matched pairs test. Two-tailed P-values < 0.05 were considered statistically significant. P-values are < 0.0001 (1111) and 0.0004 (111). Dashed line indicates the cutoff value for seropositivity.

The rate of antibody positive children increased by age for the four seasonal HCoVs and the antibody levels remained elevated in most of the seropositive children (Table 2). In all age groups the seropositivity was the highest for OC43 with a seropositivity of 31% in 1-year-olds and 81% in 3-year-olds. At 3 years of age the seropositivity rate for 229E (37%) was lower than for the other seasonal HCoVs (59% for HKU1, 76% for NL63, and 81% for OC43). It is noteworthy that the cumulative seropositivity for HKU1 declined between 2 and 3 years of age, while in 229E, NL63 and OC43 the annual and cumulative seropositivity rates increased almost at the same rate (Table 2). A decline in anti-HKU1 S1 antibodies was observed in previously seropositive children whose sequential samples showed a decrease in geometric mean IgG antibody levels (GMALs) from 2 to 3 years (44 to 33 EIA units). For 229E, NL63, and OC43 the mean IgG antibody levels in seropositive children remained relatively stable during the 3-year follow-up. Gender had no significant effect on the HCoV S1 antibodies in the studied children at any time point (Supplementary Fig. 2).

Table 2 Seroprevalence of IgG antibodies against HCoV S1 in 140 children at ages of 1, 2, and 3 years.

A subset of seropositive 1-year-old or 2-years-old children showed a diagnostic increase in HCoV S1-specific antibody levels in subsequent samples at 2 or 3 years of age, respectively, indicating a likely reinfection or re-exposure (below referred to as reinfection). Altogether 31% (27/88) of the children seropositive for any seasonal HCoV at 1 year showed a diagnostic/significant increase (more than 20 EIA units between sequential samples) in anti-S1 HCoV antibodies by the age of 2 years. A similar rate of likely reinfection for any HCoV (34%) was observed between 2 and 3 years for children who were seropositive at the age of 2 years.

Correlation of anti-HCoV S1 and anti-HCoV N antibody levels

HCoV infections have been shown to induce antibodies for both HCoV S and N proteins9,10. We have previously analyzed anti-HCoV N IgG antibody levels for the same set of serum samples from children (n = 420)18 allowing us to use the existing data to analyze the correlation of N and S1 protein-specific antibody responses. We calculated the correlation coefficients for S1 and N EIA data of each HCoV (Fig. 2). The correlation coefficients of S1 and N IgG antibody levels for 229E and NL63 were relatively high (r = 0.66 and r = 0.65, respectively, p < 0.0001). The rates of correlation for anti-HKU1 S1 and N, and anti-OC43 S1 and N IgG antibodies were somewhat lower, yet significant (r = 0.44 and r = 0.55, respectively, p < 0.0001). However, the negative samples may contribute considerably to the higher rates of correlation since the removal of samples that were negative in both assays weakened the correlation of all assay pairs (Supplementary Fig. 3).

Figure 2
figure 2

Correlation of HCoV N and S1 antibody responses. IgG antibody levels for 229E, HKU1, NL63, OC43, MERS, and SARS-CoV-2 S1 and N proteins were measured for 420 serum samples from 1 to 3-year-old children. The correlation of anti-S1 and anti-N antibody levels for each HCoV was analyzed using Spearman’s matched pairs test. The correlation coefficient (r) and two-tailed p-value for each pair is shown.

COVID-19 vaccination has no effect on HCoV S1 IgG antibody levels in adults

To investigate whether COVID-19 vaccination induces the production of HCoV S1 cross-reactive antibodies, we analyzed sera of COVID-19 vaccinees with HCoV S1 protein-specific EIAs. Sera were collected from 113 HCWs (25–65 years old, Table 1) before vaccination with BNT162b2 (0D; August 2020 to January 2021), three weeks after the second BNT162b2 dose (2D; January to March 2021), and three weeks after the third vaccination dose with BNT162b2 or mRNA-1273 (3D; October 2021 to January 2022). Geometric mean IgG antibody levels (GMALs) for the seasonal HCoV S1 proteins decreased from 0 to 2D but increased from 2 to 3D (44 EIA units at 0D vs. 41 EIA units at 2D vs. 44 EIA units at 3D for 229E, both p < 0.0001; 35 vs. 33 vs. 36 for HKU1, p = 0.0007 and p = 0.0008; 50 vs. 46 vs. 50 for NL63, p < 0.0001; 43 vs. 41 vs. 46, for OC43 p < 0.0001; Fig. 3). GMALs for MERS S1 remained practically negative while two doses of BNT162b2 increased the SARS-CoV-2 S1 GMALs from negative to high levels (GMAL of 1.6 EIA units at 0D vs. 105 at 2D, p < 0.0001). The third mRNA vaccine dose (BNT162b2 or mRNA-1273) further boosted the SARS-CoV-2 antibody levels (GMAL of 105 EIA units at 2D vs. 113 at 3D, p < 0.0001).

Figure 3
figure 3

HCoV S1-specific antibody responses of COVID19 mRNA vaccinated health care workers (HCWs). (af) HCoV S1-specific IgG antibody levels were measured with EIA from 339 serum specimens collected before vaccination (0D, Jan/21), and three weeks after second (2D, Feb/21) and third vaccination (3D, Nov/21) from 113 HCWs who received two doses of BNT162b2 with a three-week interval in December 2020 to February 2021 and a third dose of BNT162b2 or mRNA-1273 in September to December 2021. Geometric means and geometric standard deviations of antibody levels are shown. Differences between the groups were analyzed with Wilcoxon matched pairs test. Dashed line indicates the cutoff value for seropositivity. Two-tailed p-values < 0.05 were considered statistically significant. P-values 1111 < 0.0001 and 111 (b) from left to right are 0.0007 and 0.0008.

Comparison of HCoV S1-specific antibody responses between genders showed similar antibody levels at each time point with the exception of anti-NL63 S1 antibodies, which were slightly, although not significantly higher in female than in male HCWs (GMAL of 37 EIA units for male vs 51 for female at 0D, 33 vs 48 at 2D, and 43 vs. 50 at 3D; Supplementary Fig. 4). Unequal number of the HCWs in the genders (9 males and 104 females) limited the strength of the comparisons between the groups.

Prevalence of HCoV infections in 2020–2021

To investigate how the circulation of different HCoVs associated with the changes in HCoV antibody levels in the study population, we analyzed the monthly number and prevalence of HCoV PCR-positive samples in hospitalized patients in Southwest Finland health district in 2020–2021 (Fig. 4). The multiplex RT-qPCR with Allplex Respiratory Panel 3 (Seegene Inc.), which detected 229E, NL63, and OC43 (but not HKU1) RNA, showed the circulation of these viruses in early 2020 (January–March) with a peak of 12% of positive HCoV samples in March 2020 (7 229E-positive, 15 NL63-positive, and 19 OC43-positive out of 349 tested samples). This was followed by 12 months (April 2020 to March 2021) of very low number of positive samples (only 4 NL63-positive samples out of 2356 tested samples). Starting from April 2021 the circulation of OC43 was observed monthly until the end of the study period. NL63 was detected from May to July 2021 but no circulation was seen later in the year. 229E was detected sporadically after May 2021 and occasionally between October and December 2021. Detection of SARS-CoV-2 with laboratory developed RT-qPCR showed higher number of positive cases starting from autumn 2020 with peaks in March and December 2021.

Figure 4
figure 4

Monthly prevalence of HCoV RT-qPCR-positive samples in Southwest Finland hospital district in 2020–2021. Multiplex RT-qPCR test (with Allplex Respiratory panel 3, Seegene) including the detection of 229E, NL63, and OC43, or an in-house RT-qPCR for SARS-CoV-2 was used to analyze respiratory samples at the Clinical microbiology unit of Turku University Hospital in 2020 to 2021. Monthly numbers of positive samples for 229E, NL63, OC43 (a), SARS-CoV-2 (c) as well as the ratio of positive samples of tested samples (b and d) are presented. Horizontal orange lines indicate the timelines for serum specimens collected before (0D), three weeks after two doses (2D) and three weeks after the third (3D) mRNA vaccine dose. The number of tested samples for multiplex RT-qPCR varied from 104 to 377 (228 on average) per month and for SARS-CoV-2 RT-qPCR 2280 to 41,300 (22,983 on average) per month.

Congruence of seasonal HCoV occurrence and S1 IgG antibody levels

The absence of serologically observable seasonal HCoV reinfections (defined as > 20 EIA unit increase in antibody levels of sequential samples) between 0 and 2D (Fig. 3a–d) correlated with the low number of seasonal HCoV-positive samples in Autumn 2020 to March 2021 in Southwest Finland health district. The increases in seasonal HCoV S1 IgG antibody levels between the second and third vaccine doses (8 months apart) on the other hand suggested potential circulation of HCoVs and reinfection amongst HCWs (Fig. 5). The PCR-data on the rates of 229E, NL63, and OC43 detections from May 2021 onwards matched well with the serological changes among some of the HCWs on the same period. Diagnostic rises in antibody levels from 2 to 3D indicated a reinfection rate of 5.3% (6/113) for 229E, 6.2% (7/113) for HKU1, 3.5% (4/113) for NL63, and 14.2% (16/113) for OC43. Interestingly, each participant (n = 7) showing an increase in HKU1 S1 antibodies had also an increase in OC43 S1 antibodies (Fig. 5).

Figure 5
figure 5

Increases in HCoV S1 binding IgG antibody levels indicate reinfection. HCWs received two doses of BNT162b2 and a third dose of BNT162b2 or mRNA-1273 in 2020–2021. Serum samples were collected before vaccination (0D), 3 weeks after the second (2D), and 3 weeks after the third dose (3D) and analyzed for 229E (a), HKU1 (b), NL63 (c), and OC43 (d) S1 binding IgG antibodies with EIA. An increase of > 20 EIA units between sequential samples was considered as an indication for reinfection or re-exposure. Sequential serum samples of different individuals are connected with lines. In figure (d) purple-marking indicates the antibody levels of individuals who have an increase in antibodies for both OC43 S1 and HKU1 S1 and green-marking those with an increase only in OC43 S1 binding antibodies. Statistical differences between the time points were analyzed using Wilcoxon matched pairs test. Two-tailed p-values < 0.05 were considered statistically significant. P-values from left to right marked with 1 (panels a and b) are 0.031, 0.031, 0.016, and 0.016. 1111 < 0.0001.

Correlation of HCoV S1-binding antibody levels

To evaluate the presence of homologous and potentially cross-reactive anti-S1 antibodies, we used the data from 339 serum samples of COVID-19 vaccinated HCWs to analyze the correlation of seasonal HCoV S1-binding IgG antibodies. The highest rates of correlation were observed for anti-229E and anti-NL63 S1 antibodies (r = 0.524, p < 0.0001), and anti-HKU1 and anti-OC43 S1 antibodies (r = 0.631, p < 0.0001) but also other assay pairs showed moderate correlation coefficients (r > 0.334, p < 0.0001 for other anti-HCoV S1 antibody level pairs) (Fig. 6).

Figure 6
figure 6

Correlation of seasonal HCoV S1 binding antibody responses. Pairwise correlation of 229E, HKU1, NL63, and OC43 S1 binding IgG antibody levels were analyzed from the data of 113 HCWs (n = 339 serum samples) with Spearman’s matched pairs test. The correlation coefficient (r) and two-tailed p-value for each pair is shown.

In children (420 serum samples, Supplementary Fig. 5), a moderate correlation was observed for anti-HKU1 S1 and anti-OC43 S1 binding antibodies (r = 0.589, p < 0.0001) and for anti-229E S1 and anti-NL63 S1 binding antibodies (r = 0.512, p < 0.0001) while the correlation coefficient values of other anti-HCoV S1 pairs was lower (r < 0.26). Pairwise comparison of S1 amino acid sequences (Supplementary Table 1) showed that the pairs of 229E and NL63, as well as HKU1 and OC43 shared more identical amino acids (50% and 59%, respectively) in comparison to other sequence pairs (10–22%), which may contribute to sequence identity-related immunological correlation.

The correlation coefficient values for anti-SARS-CoV-2 S1 antibodies and other anti-HCoV S1 antibodies after COVID-19-vaccination were low at 2D and 3D time points (Supplementary Fig. 6). Despite the low correlation coefficient values, likely due to mainly seronegative specimens, the correlation of anti-SARS-CoV-2 S1 antibodies was statistically significant (p < 0.05) with anti-OC43 antibodies at 2D (r = 0.3033, p = 0.0011), and with anti-MERS antibodies both at 2D and 3D (r = 0.2462, p = 0.0086 and r = 0.2098, p = 0.0257, respectively).

Homologous and cross-reactive S1 antibodies in immunized animals

The data from HCoV reinfections in HCWs as well as the correlations between anti-S1 antibody levels suggested potential cross-reactivity between HCoV S1-specific antibodies. To validate further the antigens chosen for EIA, we immunized rabbits and guinea pigs with the HCoV S1 proteins without mFc-fusion (except for OC43 for which S1 with mFc was used due to unsuccessful production of OC43 S1 without mFc, Supplementary Fig. 1). Each animal was immunized with one antigen. Antigens without mFc-fusion were used in immunization to avoid formation of antibodies against the mFc domain. The specificity of immune sera was examined by immunofluorescence (IF) assay for homologous and cross-reactive antibodies using 229E, OC43, and NL63 virus-infected cells (Huh7 or LLC-MK2). Virus-infected cells were detected with high IF titers (1:1000 to 1:5000) by the corresponding anti-S1 sera of immunized rabbits and guinea pigs (Table 3, Supplementary Fig. 7). In addition to homologous reactivity of sera, cross-reactivity was also observed. Cells infected with 229E were recognized by anti-HKU1 S1 rabbit serum (titer 50) and by both rabbit and guinea pig anti-NL63 S1 sera (titer 200; Table 3, Supplementary Fig. 8). OC43-infected cells were recognized by rabbit and guinea pig anti-HKU1 S1 sera (titer 200, Supplementary Fig. 9) and NL63-infected cells were recognized by rabbit anti-OC43 S1-mFc sera (titer 200, Supplementary Fig. 10).

Table 3 Cross-reactive and homologous titers of serum from HCoV S1 immunized rabbit and guinea pig.

Source link

girlfriends having fun with their favorite toy.amateur girls double dong full insertion.
sex tube my golden pussy is not beautiful and.