Maternal immune response and placental antibody transfer after COVID-19 vaccination across trimester and platforms

Abstract

The availability of three COVID-19 vaccines in the United States provides an unprecedented opportunity to examine how vaccine platforms and timing of vaccination in pregnancy impact maternal and neonatal immunity. Here, we characterize the antibody profile after Ad26.COV2.S, mRNA-1273 or BNT162b2 vaccination in 158 pregnant individuals and evaluate transplacental antibody transfer by profiling maternal and umbilical cord blood in 175 maternal-neonatal dyads. These analyses reveal lower vaccine-induced functions and Fc receptor-binding after Ad26.COV2.S compared to mRNA vaccination and subtle advantages in titer and function with mRNA-1273 versus BN162b2. mRNA vaccines have higher titers and functions against SARS-CoV-2 variants of concern. First and third trimester vaccination results in enhanced maternal antibody-dependent NK-cell activation, cellular and neutrophil phagocytosis, and complement deposition relative to second trimester. Higher transplacental transfer ratios following first and second trimester vaccination may reflect placental compensation for waning maternal titers. These results provide novel insight into the impact of platform and trimester of vaccination on maternal humoral immune response and transplacental antibody transfer.

Fig. 1. Maternal vaccine-induced titers are comparable between mRNA-1273 and BNT162b2 vaccination, but lower after Ad26.COV2.S vaccination.

A Spike-specific IgG1 and Fc-receptor (FcR) binding were measured by Luminex. The dot plots show the titer for pregnant individuals who received Ad26.COV2.S (red), mRNA-1273 (yellow) or BNT162b2 (blue). The black diamond represents the group median. Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S. IgG1: p = 0.00035; p = 0.00048. FcR2a: p = 0.00018; p = 0.00019. FcR2b: p = 0.00018; p = 0.00019. FcR3a: p = 0.00018; p = 0.00019. B The dot plots show the Spike-specific antibody-dependent cellular phagocytosis (ADCP), antibody-dependent neutrophil phagocytosis (ADNP), antibody-dependent complement deposition (ADCD), and antibody-dependent NK cell degranulation, as measured by % CD107a + NK cells, in maternal samples. The black diamond represents the group median. Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S. ADCP: p = 0.00018 for both; ADNP: p = 0.0001 both; ADCD: p = 0.0001 both; % CD107a+: p = 0.00019 for both. C The dot plots show the S1- (top) or S2- (bottom) specific IgG1 or FcR-binding in maternal samples. Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The black diamond represents the group median. The exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S. FcR2a S2: p = 0.00018, p = 0.0006. FcR3a S2: p = 0.000045, p = 0.000036. D A partial-least squares discriminant model (PLSDA) was built using least absolute shrinkage and selection operator (LASSO)-selected SARS-CoV-2 specific antibody features in maternal samples, using vaccine type as the outcome variable. Each dot represents a sample, with the color representing the vaccine type. The ellipses represent the 95% confidence interval for the vaccine. E The barplot shows the latent variable (LV) 1 for the LASSO-selected features for the PLSDA in (D). Features with a positive loading along LV1 are enriched in mothers who received an mRNA vaccination, and features with a negative loading are enriched in mothers who received Ad26.COV2.S. Nearly all features are enriched in mRNA vaccine recipients. Source data are presented in Source Data File 1.

Fig. 2. Trimester of vaccination affects vaccine-induced antibody titer in maternal samples.

A The dot plots show the Spike-directed IgG1 and FcR-binding in maternal samples by trimester of vaccination. Significance was determined by Kruskal–Wallis test. No significant differences were found. The black diamond represents the group median. B The dot plots show the Spike-directed antibody-dependent cellular phagocytosis (ADCP), antibody-dependent neutrophil phagocytosis (ADNP), antibody-dependent complement deposition (ADCD) and antibody-dependent NK cell degranulation, as measured by % CD107a + NK cells, in maternal samples by trimester of vaccination. Significance was determined by Kruskal–Wallis test. No significant differences were found. The black diamond represents the group median. C The polar plots show the mean percentile rank for Spike-specific features in the first, second, and third trimesters of vaccination. D A PLSDA was built using LASSO-selected antibody features in maternal plasma for mothers who received Ad26.COV2.S, mRNA-1273, or BNT162b2 using the trimester of vaccination as the outcome variable. Each dot represents a sample, with the color representing the trimester. The ellipses represent the 95% confidence interval for the trimester. Source data are presented in Source Data File 1.

Fig. 3. Vaccine-induced antibody titer in cord samples is comparable between mRNA-1273 and BNT162b2 vaccination but lower after Ad26.COV2.S vaccination.

A Spike-specific IgG1 and Fc-receptor (FcR) binding were measured by Luminex. The dot plots show the titer for cords whose mothers who received Ad26.COV2.S (red), mRNA-1273 (yellow) or BNT162b2 (blue). Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The black diamond represents the group median. Exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S. IgG1: p = 0.001, p = 0.0031. FcR2a: p = 0.00036, p = 0.00018. FcR2b: p = 0.00012, p = 0.00009. FcR3a: p = 0.000072, p = 0.00006. B The dot plots show the Spike-specific antibody-dependent cellular phagocytosis (ADCP), antibody-dependent neutrophil phagocytosis (ADNP), antibody-dependent complement deposition (ADCD) and antibody-dependent NK cell degranulation, as measured by % CD107a + NK cells, in cord samples. Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The black diamond represents the group median. Exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S, then mRNA-1273 vs BNT162b2 (if relevant). ADCP: p = 0.000036, p = 0.000033, p = 0.02. ADNP: p = 0.00003, p = 0.000028. ADCD: p = 0.000026, p = 0.005, p = 0.0025. CD107a+: p = 0.00085, p = 0.006. C The dot plots show the S1 (top) or S2 (bottom) specific IgG1 or FcR-binding in cord samples. Significance was determined by Kruskal–Wallis test followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. The black diamond represents the group median. Exact p-values stated in the following order for all subpanels: mRNA-1273 vs Ad26.COV2.S, then BNT162b2 vs Ad26COV2.S. FcR2a S2: p = 0.00045 for both. FcR2b S2: p = 0.00045, p = 0.00072. FcR3a S2: p = 0.00045, p = 0.0027. D A partial-least squares discriminant model (PLSDA) was built using LASSO-selected SARS-CoV-2 specific antibody features in cord samples, using vaccine type as the outcome variable. Each dot represents a sample, with the color representing the vaccine type. The ellipses represent the 95% confidence interval for the vaccine. E The barplot shows the latent variable (LV) 1 for the least absolute shrinkage and selection operator (LASSO)-selected features for the PLSDA in (D). Features that with a positive LV1 loading were enriched in the cords whose mothers received an mRNA vaccine. Source data are presented in Source Data File 1.

Fig. 4. Efficient transfer of vaccine-induced antibodies to cord blood.

A The dot plots show the Spike-specific IgG1 titer or FcR3a binding for maternal plasma (M) and cord blood (C). Lines connect maternal-cord dyads and the color represents vaccine type, Ad26.COV2.S (red), mRNA-1273 (yellow) or BNT162b2 (blue). Significance was determined by Wilcoxon signed-rank test (2-sided) followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. Exact p-value for FcR3a: mRNA-1273 p = 0.0000052, BNT162b2 p = 0.042. B The dot plots show the Spike-specific antibody-dependent cellular phagocytosis (ADCP), antibody-dependent neutrophil phagocytosis (ADNP), antibody-dependent complement deposition (ADCD), and antibody-dependent NK cell degranulation, as measured by % CD107a + NK cells, for maternal plasma (M) and cord blood (C). Lines connect maternal-cord dyads and the color represents vaccine type, Ad26.COV2.S (red), mRNA-1273 (yellow) or BNT162b2 (blue). Significance was determined by Wilcoxon signed-rank test (2-sided) followed by posthoc Benjamini–Hochberg correction adjustment, *p < 0.05, **p < 0.01, ***p < 0.001. Exact p-values stated in the following order for all subpanels: mRNA-1273 M-C, then BNT162b2 M-C, unless otherwise noted. ADCP: p = 0.00011, p = 0.0066. ADNP: p = 0.00007, p = 0.0099. ADCD: BNT162b2 p = 0.011. CD107a+: p = 0.00021, p = 0.047. C–E A multilevel PLSDA (mPLSDA) was built for Ad26.COV2.S (C), mRNA-1273 (D) and BNT162b2 (E) using sample type, maternal blood (M, light purple) or cord blood (CB, dark purple) as the outcome variable. Features were selected using LASSO prior to building the models. The dot plots (top) show the scores plots for the mPLSDA. Each dot represents a sample, with the color representing the sample type. The ellipses represent the 95% confidence interval for the sample type. The bar plots show the LV1 for the mPLSDA built in each respective subfigure. Source data are presented in Source Data File 1.

Fig. 5. Transfer efficiency differs by trimester of vaccination.

A The dot plots show the Spike-specific IgG titer in maternal (M) or cord (C) plasma. Lines connect paired dyads. The color indicates the trimester of vaccination, first (orange), second (blue) or third (pink). Significance was determined by a Wilcoxon signed-rank test (2-sided), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p-values presented first trimester: p < 0.0001 (GraphPad Prism does not provide a more exact p-value for this comparison), second trimester: p < 0.0001. B The dot plot shows the transfer ratio (cord titer/maternal titer) of Spike-specific IgG to the cord. Color indicates the trimester of vaccination. Significance was determined by Kruskal–Wallis test. ****p < 0.0001. The horizontal line represents the group median. Exact p-value 1st vs 3rd TR: p < 0.0001 (GraphPad does not provide any more exact p-value); 2nd vs 3rd p < 0.0001. C The dot plot shows the Spike-specific IgG titer in cord blood by trimester that the mother received COVID-19 vaccination. Significance was determined by a one-way ANOVA followed by correction for multiple comparisons, *p < 0.05. The horizontal line represents the group median. Exact p-value 1st vs 2nd p = 0.017. D The dot plots show the IgG Spike titer in maternal plasma post-boost (~2–6 weeks after the second dose of mRNA vaccine or after a single dose of Ad26.COV2.S vaccine) and at delivery following vaccination in the first (orange) and second (blue) trimester. Lines connect matched samples. Significance was determined by a Wilcoxon signed-rank test (2-sided), *p < 0.05, **p < 0.01. Exact p-value 1st p = 0.016. Exact p-value 2nd p = 0.002. E The dot plot shows the ratio of the IgG Spike titer delivery/post-boost following vaccination in the first (orange) and second (blue) trimesters. Significance was determined by a two-tailed Mann–Whitney test, **p < 0.01. The horizontal line represents the group median. Exact p-value 1st vs 2nd p = 0.0085. Source data are presented in Source Data File 2.