SARS-CoV-2 breakthrough infection during pregnancy preferentially elicits IgG4 response and enhanced placental-transfer

Abstract

Background:

Emerging SARS-CoV-2 variants and waning humoral immunity have led to increased occurrences of breakthrough infection. Pregnant individuals and neonates are particularly vulnerable to adverse outcomes associated with COVID-19.

Objective:

We aimed to evaluate the maternal anti-SARS-CoV-2 IgG response and the placental transfer of antibodies to the fetus following breakthrough infection, infection in unvaccinated participants, and vaccination in the absence of infection during pregnancy.

Methods:

Pregnant patients (n=165) were enrolled at Oregon Health & Science University and assigned to three cohorts based on infection and vaccination status. Receptor-binding domain (RBD) and nucleocapsid protein (NP) specific IgG endpoint titers (EPT) and optical density (OD) values were determined via ELISA longitudinally in maternal blood and breastmilk, and at delivery in paired maternal, umbilical cord, and newborn blood.

Results:

Breakthrough infection in vaccinated participants induced a higher maternal RBD-specific IgG response compared to infection or vaccination alone in maternal plasma and breastmilk. As reported with repeated vaccination, breakthrough infection skewed RBD-specific IgG responses toward an IgG4 response. IgG1 and IgG3 were the dominant NP-specific subclasses for both breakthrough infection and unvaccinated/infection groups. The breakthrough infection cohort had the highest newborn-to-maternal ratio of RBD-specific IgG with a strong correlation between maternal and newborn delivery titers.

Conclusion:

Breakthrough infection during pregnancy prolongs pre-existing humoral immunity and facilitates the transplacental transfer of antibodies to the neonate, reflecting an antibody priming effect induced by prior vaccination rather than infection alone. These results warrant continued administration of updated SARS-CoV-2 vaccination during pregnancy to reduce disease severity and provide greater protection to newborns.

INTRODUCTION

In 2020, the World Health Organization declared COVID-19 a pandemic. Today, SARS-CoV-2 is endemic worldwide and new variants are rapidly emerging¹. Current WHO designated variants under monitoring possess mutations within the viral spike protein region permitting evasiveness against the adaptive immunity conferred by previous infection and vaccination². As many current cases are reinfections and breakthrough infections^3,4, a greater understanding of how prior vaccination and infection modulate the anti-SARS-CoV-2 humoral immune response is critical. This is especially important for vulnerable populations such as pregnant individuals and neonates. SARS-CoV-2 infection during pregnancy increases the likelihood of maternal ICU admission, mechanical ventilation, preeclampsia, and preterm birth^5,6. Furthermore, infection risks extend to newborns with infants under six months of age having higher rates of hospitalization than other pediatric patients with COVID^7,8. Maternal vaccination positively impacts pregnancy outcomes⁹ and a large body of evidence supports the safety and benefit of vaccination during pregnancy^10,11. One of the critical benefits is neonatal protection from infection. Newborns exhibit increased susceptibility to infection shortly after birth due to immature humoral and cell mediated immunity¹². During the first 4–6 months of life, newborns are dependent on maternal placental-transferred IgG antibodies and, when available, breastmilk^13,14 for protection against early childhood infections, while their own immune system develops and prior to eligibility for many vaccines.

The American College of Obstetricians and Gynecologists currently recommends SARS-CoV-2 vaccination during pregnancy or breastfeeding to improve both maternal and neonatal outcomes¹⁵. Vaccination has been shown to elicit higher maternal antibody levels than infection¹⁶ and there is evidence of SARS-CoV-2 specific IgG transfer from maternal circulation to the neonate¹⁷. Receiving a booster dose during pregnancy has also been linked to higher antibody levels in both maternal circulation, newborn circulation, and breastmilk^18,19 alongside a lower risk of newborn infection during the first six months of life²⁰.

As of 2024, 76.9% of pregnant individuals in the US have received one or more doses of SARS-CoV-2 vaccine with only 12.9% having received a 2023–2024 vaccine booster dose²¹. As humoral immunity from mRNA vaccination wanes and new variants emerge, the incidence of breakthrough infections increases^22,23. This necessitates an improved understanding of humoral immune response resulting from SARS-CoV-2 breakthrough infections during pregnancy. Here we studied the dynamics of antibody responses following SARS-CoV-2 infection of naive or vaccinated pregnant participants, building on findings from our previous study¹⁹ by characterizing anti-SARS-CoV-2 antibodies longitudinally in maternal blood and breastmilk. Additionally, through comparisons of maternal antibody titers at delivery to matched umbilical cord blood and newborn blood collected within 24 hours of birth, we observed patterns of IgG placental-transfer differentiated by maternal vaccination and infection status. The findings of this study underscore the importance of primary vaccination and booster immunization during pregnancy to achieve improved maternal and newborn health outcomes.

METHODS

Human Subjects

This study was approved by the Institutional Ethics Review Boards at Oregon Health & Science University and University of Kentucky. Written and informed consent for the use of samples in research was obtained from the participants at time of enrollment which occurred from March 2021 to June 2022 at Oregon Health & Science University.

Characteristics of Study Participants

Over a period of 15 months (March 2021–June 2022), we enrolled 165 pregnant participants at a large-tertiary care referral institution (Oregon Health & Science University). Enrolled patients were 18 years-of-age or older, antepartum or lactating, and had received a positive COVID-19 test or COVID-19 vaccination while pregnant or lactating. Patients with an active hepatitis infection, HIV infection, or history of substance use disorder (non-prescription opiate, stimulant, or alcohol) were excluded from enrollment. While most participants were enrolled during the prenatal period, enrollment also extended to patients who had received a vaccination post-partum. This group includes 30 Vaccinated/Uninfected participants the majority of whom were vaccinated shortly after delivery. These enrollments were incorporated to characterize breastmilk and maternal IgG titers post-partum. No samples were collected during pregnancy or at delivery for these participants and infection status was determined solely through COVID-19 tests performed at pre- and post-natal clinic visits routinely during the pandemic.

Participants were divided into three cohorts (Table 1) based on their vaccination and infection status: [1] vaccinated participants who remained uninfected (Vaccinated/Uninfected; n=124); [2] vaccinated participants who developed a SARS-CoV-2 breakthrough infection during pregnancy (Breakthrough Infection; n=25); and [3] participants who were unvaccinated and were infected with SARS-CoV-2 during pregnancy (Unvaccinated/Infected; n=16) (Figure 1).

Table 1. Cohort Demographic Characteristics.

Maternal and gestational ages were presented as mean ± standard deviation. No significant difference between fetal sex was determined with Fisher’s exact test.

		Vaccinated/Uninfected (n=124)	Breakthrough Infection (n = 25)	Unvaccinated/Infected (n = 16)
Age (years)		34.2 ± 4.1	33.2 ± 3.9	31.4 ± 6.4*
Age Range (years)		25 – 45	25 – 40	19 – 40
Body Mass Index (BMI)		25.5 ± 5.6	26.6 ± 6.5	29.2 ± 8.5**
Gestational Age at Delivery (weeks)		38.9 ± 1.6	38.8 ± 1.6	37.7 ± 1.9*
Sex of Fetus (% Female)		48%	56%	59%
Time of First	Pre-Pregnancy	14	23	0
	T1	17	1	0
	T2	47	1	0
	T3	16	0	1
	Postpartum	30	0	9
	No Vaccine	0	0	6
Time of Boost	Pre-Pregnancy	0	4	0
	T1	1	8	0
	T2	1	2	0
	T3	10	3	0
	Postpartum	82	2	2
	No Boost	30	6	14
Time of	Pre-Pregnancy	0	0	0
	T1	0	6	2
	T2	0	5	8
	T3	0	14	6
	Postpartum	39	0	0
	No Infection	85	0	0
Race	American Indian/ .Alaskan Native	2 (1.6%)	1 (4.0%)	1 (6.25%)
	Asian American	9 (7.3%)	2 (8.0%)	0
	Black/African American	5 (4.0%)	0	1 (6.25%)
	Pacific Islander	0	0	1 (6.25%)
	White/Caucasian	105 (84.7%)	20 (80.0%)	11 (68.75%)
	Unknown/Declined	3 (2.4%)	2 (8.0%)	2 (12.5%)
Ethnicity	Hispanic	4 (3.2%)	1 (4%)	6 (37.5%)
	Non-Hispanic	118 (95.2%)	23 (92%)	10 (62.5%)
	Unknown/Declined	2 (1.6%)	1 (4%)	0
Smoking	Never	111 (89.5%)	21 (84.0%)	14 (87.5%)
	Previous	12 (9.7%)	4 (16.0%)	1 (6.25%)
	Current	1 (0.8%)	0	1 (6.25%)

T1: first gestational trimester (< 13 weeks); T2: second gestational trimester (13 to 26.6 weeks); T3: third gestational trimester (≥ 27 weeks)

^*Maternal age and gestational age within the Unvaccinated/Infected cohort was significantly lower than both the Vaccination/Uninfected and Breakthrough Infection cohorts.

^**BMI within the Unvaccinated/Infected cohort was significantly higher than the Vaccination/Uninfected. No significant difference was determined between the Vaccinated/Uninfected and Breakthrough Infection cohorts for maternal age, gestational age, or BMI.

Figure 1. Enrollment and cohort schematic.

Flowchart displaying the number of participants and the division into cohorts based on vaccination and infection status.

Maternal blood was collected longitudinally once every trimester during gestation, at delivery, and postpartum every three months until breastmilk production ceased. Colostrum was collected at delivery and breastmilk was obtained longitudinally postpartum every three months until breastmilk production ceased. Umbilical cord blood (UCB) was collected at delivery. Newborn blood was collected via heel stick within 24 hours of delivery.

Of the vaccinated participants, 85.4% received Pfizer BNT162b2 and 14.6% received Moderna mRNA-1273 for their initial two dose series. Furthermore, 80.9% received Pfizer BNT162b2 and 19.1% received Moderna mRNA-1273 as their 3^rd dose booster vaccine. Of all enrolled vaccinated participants, only 25% (n=37 of 149) received their first COVID vaccine prior to pregnancy. Of the Vaccinated/Uninfected participants, 76% (n=94 of 124) received a primary two-dose series during their pregnancy, and 66% (n=82 of 124) received a booster dose postpartum. To monitor participants for asymptomatic infection throughout the study, COVID-19 nasal swab tests were performed at every clinic visit and sample collection. To identify asymptomatic or unreported infections that may have occurred between visits, all plasma samples from the Vaccinated/Uninfected cohort were screened for anti-SARS-CoV-2 nucleocapsid protein (NP) specific IgG optical density (OD) via ELISA. Individuals in the Vaccinated/Uninfected group who tested positive for NP-specific IgG or reported a positive COVID-19 test were excluded from analysis for timepoints following the positive test (n=39).

Among the Vaccinated/Uninfected participants, 26 participants had detectable NP-specific IgG at one or more timepoints, indicating SARS-CoV-2 infection. Notably, all NP positive results in this group occurred postpartum. Of these 26, only 10 did not report a positive COVID-19 test or symptoms and were therefore considered likely asymptomatic infections. Because these infections occurred exclusively after delivery, samples from these individuals were not included in the Breakthrough Infection cohort, which was defined as participants with confirmed gestational SARS-CoV-2. Within the Breakthrough Infection cohort, 56% (n=14 of 25) were infected after receiving a booster (timing of booster: n=4 pre-pregnancy, n=8 first trimester, n=1 second trimester, n=1 third trimester) while 44% (n=11 of 25) were infected prior to any booster vaccination (timing of booster: n=1 second trimester, n=2 third trimester, n=2 post-partum, n=6 received no booster). Additionally, 84% (n=21 of 25) of Breakthrough Infection participants contracted an infection amidst the Omicron variant surge between January and June 2022²⁴ and 16% (n=4 of 25) contracted a breakthrough infection amidst Beta, Alpha, and Delta variant dominance between December 2020 and October 2021²⁴. Within the Unvaccinated/Infected cohort, 93.75% (n=15 of 16) were infected amidst Beta, Alpha, and Delta variant surges while 6.25% (n=1 of 16) were infected during the Omicron variant surge. Notably, 62.5% (n=10 of 16) of Unvaccinated/Infected participants were subsequently vaccinated post-partum and 37.5% (n=6 of 16) of Unvaccinated/Infected participants remained unvaccinated. The timing of vaccination and infection by trimester is presented for each cohort in Table 1.

Sample Processing

Blood samples were collected in EDTA vacutainers. Maternal and umbilical cord plasma was obtained following centrifugation with Lymphoprep per manufacturer’s recommendation (STEMCELL; Vancouver, Canada). Due to low sample volumes (200ul), newborn heel stick blood was centrifuged for 5 minutes at 664g to retrieve newborn plasma. Breastmilk samples were diluted with an equal volume of HBSS (Corning; Corning, NY) and centrifuged at 810g for 10 minutes. The top fat layer was discarded, and the supernatant was collected. Plasma and breastmilk supernatant samples were stored at −80°C.

Enzyme Linked Immunosorbent Assay

SARS-CoV-2 receptor-binding domain (RBD) and nucleocapsid protein (NP) specific IgG endpoint titers (EPT) were determined by indirect ELISA. High-binding plates (Corning; Corning, NY) were coated with 100μL of either 0.5μg/mL RBD (GenScript, Piscataway, NJ) or 1μg/mL NP (GenScript, Piscataway, NJ) antigen in 1xDPBS (Corning; Corning, NY). Excess antigen was flicked off and plates were blocked with a 5% nonfat dry milk, 0.1% tween, phosphate-buffered saline (PBS) solution for 1 hour. Plates were washed with 0.05% tween-PBS.

To measure total IgG EPT, heat-inactivated maternal and UCB plasma was diluted 1:30 in blocking buffer while newborn plasma was diluted 1:50. To measure IgG subclass EPT, maternal and UCB plasma was diluted 1:30 in blocking buffer for IgG1 and IgG3 and diluted 1:10 for IgG2 and IgG4. Newborn plasma was diluted 1:50 in blocking buffer for IgG1 and IgG3 and diluted 1:10 for IgG2 and IgG4. Following the initial dilution, six 3-fold serial dilutions of were performed for all EPT ELISA. Heat-inactivated breastmilk was diluted 4:1 in blocking buffer to measure total IgG and IgG subclass optical density (OD). All samples were run in duplicate.

Plasma dilutions incubated for 1.5 hours. A secondary horseradish peroxidase (HRP) anti- human IgG (BD Pharmingen, Franklin Lakes, NJ) or IgG subclass specific (IgG1, IgG2, IgG3, or IgG4; SouthernBiotech, Birmingham, Alabama) antibody diluted 1:4000 was added. After 1.5 hours, plates were developed with o-phenylenediamine dihydrochloride and 1M hydrochloric acid stop. Plates were recorded at 490nm absorbance on a SpectraMax iD3 plate reader (Molecular Devices, San Jose, CA). Average absorbance of the blanks was subtracted and EPT were calculated through log-log transformation of the linear portion of the curve with a cutoff of 0.1 absorbance. Batch effect between plates was minimized by normalizing to a positive control ran on each plate.

Statistical Analysis

Statistical analyses were conducted using Prism 10 (GraphPad; Boston, MA). Data sets were tested for normality and subsequently analyzed via parametric and nonparametric test for normal and non-normal distribution respectively. Comparisons between two groups were performed via paired t-test (Wilcoxon test: nonparametric) if samples came from the same participant and unpaired t-test (Mann-Whitney test: nonparametric) if samples were from different participants. Analyses of more than two groups were conducted through one-way ANOVA with Dunn’s multiple comparisons (Kruskal-Wallis test: paired, Friedman test: unpaired). Statistical significance was determined at the p≤0.05 level. P-values between 0.1 and 0.05 were considered trending. Antibody decay rate estimates to half of the initial IgG EPT value for each group (beginning 14 days post antigen exposure) were derived from linear mixed models, which account for repeated measurements from participants while modeling log-transformed antibody levels over time^25,26. A Kenward-Roger adjustment was used, as appropriate, to correct for negative bias in the standard errors and degrees of freedom calculations. Model assumptions were evaluated using a combination of visual plots and formal testing. This portion of the analysis was completed in SAS 9.4 (SAS Institute Inc.; Cary, NC).

Data Availability

The data presented here and supporting the conclusions in this article is available via Mendeley Data Repository (pending).

RESULTS

Vaccine-boosted patients have a more sustained anti-RBD IgG response following an infection.

We previously reported the SARS-CoV-2 maternal antibody responses up to 205 days post booster (DPB) within the Vaccinated/Uninfected cohort¹⁹. Here, we extended this analysis to 496 DPB and expanded our study to include maternal antibody response following gestational SARS-CoV-2 infection in vaccinated versus unvaccinated individuals. Additionally, we evaluated receptor binding domain (RBD) specific IgG transfer to the neonate transplacentally and via breastmilk.

Maternal RBD-specific IgG titers in the Vaccinated/Uninfected cohort showed an over two-fold longer antibody decay rate after the boost immunization relative to days post first vaccination (DP1V) (pre-boost 83.8 DP1V vs post-boost 202.7 DP1V) (Figure 2A). Breakthrough infection after a booster dose further enhanced and prolonged the maternal antibody decay over 6 times to 1307.8 DP1V (Figure 2A). We sought to determine whether the increase in antibody response was due to the breakthrough infection or to variation in the timing between the initial series of vaccination and the booster dose. We compared end-point titers (EPT) between samples from the Vaccinated/Uninfected and Breakthrough Infection cohorts that were matched based on days post booster (DPB). Breakthrough infections resulted in significantly higher RBD-specific total IgG titers post-boost compared to boosted participants who were not infected (Figure 2B).

Figure 2. Breakthrough infection post-boost elicits robust maternal anti-RBD antibody response and IgG4 subclass dominance.

(A) Maternal RBD-specific IgG titers of longitudinal samples relative to days post first vaccination (DP1V) in Vaccinated/Uninfected before booster (light blue, n=398), Vaccinated/Uninfected after booster (dark blue, n=175), and Breakthrough Infection after booster (orange, n=90). The average days between first dose and booster was 269±50 days. The average days between first dose and infection was 332±114 days. (B) Post-booster comparison of maternal RBD-specific total IgG titers (limit of detection (LOD): dotted line) for Vaccinated/Uninfected (n=14) 259±19.7 days post booster (DPB) and Breakthrough Infection (n=15) 262.9±21.5 DPB. (C) Maternal RBD-specific IgG titers relative to days post initial antigen exposure (DPAE) via initial vaccination series or infection. (D) Subclass characterization of maternal RBD-specific IgG at delivery for Breakthrough Infection (left, n=24) and Unvaccinated/Infected (right, n=13). (E) Post-booster comparison of maternal RBD-specific IgG1 and IgG4 titers (LOD: dotted lines) for Vaccinated/Uninfected (n=14) 259±19.7 DPB and Breakthrough Infection (n=13) 262.9±21.5 DPB. (F) Maternal NP-specific IgG titers of longitudinal samples relative to days post infection (DPI) for Breakthrough Infection (orange, n=112) and Unvaccinated/Infected (red, n=32) cohorts. (G) Subclass characterization of maternal NP-specific IgG at delivery for Breakthrough Infection (left, n=24) and Unvaccinated/Infected (right, n=13).

Due to the low number of Unvaccinated/Infected and Vaccinated Post-Infection participants, reliable antibody decay rates were unable to be calculated. Nevertheless, the RBD-specific IgG maternal simple linear regression relative to days post initial antigen exposure (DPAE) for the Vaccinated/Uninfected cohort (longitudinal sample n=398) and Unvaccinated/Infected cohort (longitudinal sample n=27) were comparable (Figure 2C). On the other hand, antibody durability in previously Unvaccinated/Infected participants was increased following vaccination (Figure 2C).

Vaccination skews RBD-specific IgG titers toward an IgG4 response.

We previously demonstrated that Vaccinated/Uninfected participants who received a booster vaccine had a significant increase in all four IgG subclasses, particularly IgG1 and IgG4¹⁹. Here, we observed a similar pattern within our Breakthrough Infection cohort as IgG1 and IgG4 subclass levels were significantly higher than IgG2 and IgG3 in maternal plasma at delivery (Figure 2D). In contrast, the Unvaccinated/Infected cohort generated a response dominated by the IgG3 subclass with IgG2 or IgG4 levels mostly below the limit of detection (Figure 2D). Due to most breakthrough infections occurring post-booster, we compared IgG1 and IgG4 EPT in samples matched by DPB from the Vaccinated/Uninfected and Breakthrough Infection cohorts (same samples used in Figure 2B). As reported for total IgG, breakthrough infections post-boost induced significantly higher IgG1 and trending higher IgG4 RBD-specific EPT compared to the Vaccinated/Uninfected cohort’s booster response (Figure 2E).

SARS-CoV-2 vaccines contain mRNA encoding the viral spike protein (S) to prompt S antigen and RBD-specific antibody production. Natural human infection produces antibodies against RBD as well as other viral antigens including the nucleocapsid protein (NP). We determined NP-specific IgG EPT for the Breakthrough Infection and Unvaccinated/Infected cohorts to evaluate the impact of vaccine status on NP-specific antibody production. NP EPT did not correlate with days post infection (DPI) for either infection cohort (Figure 2F). An NP-specific antibody decay rate of 187.8 DPI was calculated for the Breakthrough Infection cohort (Figure 2F). The low number of Unvaccinated/Infected participants prevented a reliable NP-specific IgG decay rate calculation. IgG subclass analysis of both the Breakthrough Infection and Unvaccinated/Infected groups at delivery showed that NP-specific antibodies were primarily IgG3 (Figure 2G).

Robust transfer of maternal anti-RBD IgG to newborn following breakthrough infection.

To characterize the placental transfer of maternal RBD-specific IgG into newborn circulation, we compared maternal IgG titers at delivery to those detected in UCB and newborn plasma collected within 24 hours of delivery. The transfer of IgG across the placenta begins during the second trimester and increases during the third²⁷. To observe the impact of the trimester of infection or vaccination on the antibody levels at delivery, EPT at delivery were compared against DPI and DP1V within each cohort and each sample type. Both the maternal and newborn delivery samples within the Vaccinated/Uninfected cohort displayed significant inverse correlation with day elapsed since first maternal vaccination (Figure 3A). Conversely, the delivery samples of the Breakthrough Infection cohort did not have decreased levels of RBD-specific IgG as days elapsed from maternal infection, instead, both UCB and newborn titers had a trending positive correlation to DPI (Figure 3A). No correlation was observed among Unvaccinated/Infected delivery samples and DPI, although, low sample numbers prevent a reliable assessment (Figure 3A).

Figure 3. Increased RBD-specific total IgG placental transfer with breakthrough infection and preferential placental transfer of RBD-specific IgG1 regardless of infection status.

RBD-specific IgG EPT in maternal, UCB, and newborn delivery samples relative to days since maternal infection or maternal initial vaccination. (B) RBD-specific IgG titers (LOD: dotted line) in maternal, UCB, and newborn plasma at delivery for Breakthrough Infection (orange, n=25), Unvaccinated/Infected (red, n=13), and Vaccinated/Uninfected (blue, n=63). (C) Correlation of RBD-specific IgG titers in maternal and UCB plasma at delivery for Breakthrough Infection (orange, n=22), Unvaccinated/Infection (red, n=13), and Vaccinated/Uninfected (blue, n=59). (D) Correlation of RBD-specific IgG titers in maternal and newborn plasma at delivery for Breakthrough Infection (orange, n=22), Unvaccinated/Infected (red, n=10), and Vaccinated/Uninfected (blue, n=49). (E) Correlation of RBD-specific IgG titers in newborn and UCB plasma at delivery for Breakthrough Infection (orange, n=19), Unvaccinated/Infected (red, n=10), and Vaccinated/Uninfected (blue, n=48). (F) RBD-specific IgG subclasses (LOD: dotted lines) in maternal, UCB, and newborn plasma at delivery for Breakthrough Infection (orange; IgG1 n=24, IgG2 n=21, IgG3 n=21, IgG4 n=24) and Vaccinated/Uninfected (blue; IgG1 n=17, IgG2 n=16, IgG3 n=17, IgG4 n=17) participants.

In general, RBD-specific EPT were highest among the Breakthrough Infection cohort by at least one order of magnitude (Figure 3B). Within this cohort, the EPT levels observed in newborn plasma were significantly higher than those in maternal circulation (Figure 3B). In contrast, both UCB and newborn RBD-specific IgG titers were significantly lower compared to maternal plasma in the Vaccinated/Uninfected cohort (Figure 3B). Among Unvaccinated/Infected participants, both UCB and newborn RBD IgG titers were modestly lower than maternal levels (Figure 3B). RBD titers in UCB did not significantly correlate with maternal titers at delivery for any of the cohorts (Figure 3C). In contrast, IgG titers in newborn plasma showed a significant positive correlation to maternal titers in both the Breakthrough Infection and Vaccinated/Uninfected cohorts (Figure 3D). Similarly, RBD-specific IgG titers in UCB significantly correlated with those in newborn blood for both the Breakthrough Infection and Vaccinated/Uninfected cohorts (Figure 3E).

Anti-RBD IgG1/4 antibodies robustly transferred to newborn after breakthrough infection.

We next evaluated IgG subclass titers within UCB and newborn samples to determine preferential transfer of RBD-specific IgG subclasses. Due to low volumes of newborn blood collected in the Breakthrough Infection cohort, we could not measure the IgG2 and IgG3 levels within newborn plasma. Instead, only IgG1 and IgG4 levels, which were the highest measured subclasses in maternal plasma, were measured for newborns samples. Therefore, only UCB plasma was used to compare fetal and maternal IgG2 and IgG3 titers. Apart from IgG3, IgG1/2/4 EPT were higher for all maternal and neonatal samples in the Breakthrough Infection cohort compared to the Vaccinated/Uninfected group (Figure 3F). Additionally, IgG4 titers were highest in the Breakthrough Infection group (Figure 3F). For both Breakthrough Infection and Vaccinated/Uninfected cohorts, RBD-specific IgG1 levels were highest in newborn plasma (Figure 3F). No significant difference in IgG2, IgG3, or IgG4 titers was calculated between delivery samples for the Breakthrough Infection cohort (Figure 3F). However, within the Vaccinated/Uninfected group RBD-specific IgG2 titers were significantly lower in UCB compared to maternal plasma (Figure 3F). IgG3 measurements in the Vaccinated/Uninfected cohort displayed a reverse trend with significantly higher UCB IgG3 levels than maternal circulation (Figure 3F). No significant difference in IgG4 was measured between neonatal and maternal samples for either cohort, but newborn IgG4 titers were higher than UCB within the Vaccinated/Uninfected cohort (Figure 3F). UCB is often considered a substitute for newborn blood at delivery and IgG titers between the two sample types would be expected to be similar²⁸; thus, the observed higher titers in newborn plasma compared to UCB plasma were unexpected.

Equivalent maternal and newborn anti-NP IgG titers post-infection.

The NP-specific IgG titers measured in UCB and newborn plasma following both Breakthrough Infection and Unvaccinated/Infection were not significantly different from those observed in maternal plasma at delivery (Figure 4A). Nevertheless, a significant positive correlation was observed between all delivery samples within the Breakthrough Infection group but not the Unvaccinated/Infected cohort (Figure 4B-D). When examining the impact of gestational age at the time of infection on NP-specific IgG transfer to neonate, a significant negative correlation between UCB titers and DPI was observed within the Unvaccinated/Infected cohort but not the Breakthrough Infection (Figure 4E). No significant correlation was observed between titers in newborn plasma and DPI for either cohort (Figure 4F).

Figure 4. Positive correlation of NP-specific IgG titers between newborn and maternal plasma at delivery with Breakthrough Infection.

(A) NP-specific total IgG titers (LOD: dotted line) in maternal, UCB, and newborn plasma at delivery after Breakthrough Infection (orange, n=25) and in Unvaccinated/Infected participants (red, n=12). (B-F) Correlations of NP-specific total IgG titers in Breakthrough Infection (orange, n=22) and Unvaccinated/Infected (red, n=12) cohorts at delivery between (B) maternal and UCB plasma, (C) maternal and newborn plasma, (D) newborn and UCB plasma, (E) UCB plasma and maternal DPI, (F) newborn plasma and maternal DPI.

Anti-RBD IgG1/4 antibodies in breastmilk are higher in breakthrough infection patients.

Since maternal antibodies can also be transferred postpartum via breastmilk, we measured RBD and NP-specific IgG titers in breastmilk samples longitudinally. There was a time-dependent decrease of anti-RBD IgG expression in breastmilk for all the Vaccinated/Uninfected participants (Figure 5A). However, the antibody decay rate was prolonged post-boost immunization (pre-boost 62.0 DP1V vs post-boost 133.0 DP1V) consistent with our observations in plasma (Figure 2A). Breakthrough infection after booster also prolonged the antibody response in breastmilk though patient OD measurements did not correlate with DP1V preventing decay rate calculation (Figure 5A). As observed in maternal circulation, breakthrough infections resulted in significantly higher levels of total IgG in breastmilk compared to the Vaccinated/Uninfected group when matched for DPB (Figure 5B). Vaccination post-infection potentially increased the durability of IgG in the breastmilk according to the calculated simple linear regressions though the low number of samples precludes a decay rate calculation (Figure 5C). As seen in plasma, significantly higher RBD-specific IgG1 and IgG4 levels were observed in breastmilk for the Breakthrough Infection cohort compared to the Vaccinated/Uninfected cohort post booster dose (Figure 5D). Patient NP-specific IgG OD values in breastmilk did not correlate to DPI and a Breakthrough Infection antibody decay rate of 366.7 DPI was calculated (Figure 5E). As described for plasma, the number of data points in the Unvaccinated/Infected group was too low to make any significant determination regarding the NP-specific IgG decay rate.

Figure 5. Postpartum antibody response in breastmilk mirrors longitudinal trends in maternal circulation.

(A) Breastmilk RBD-specific IgG OD measurements of longitudinal samples relative to days post first dose (DP1V) in Vaccinated/Uninfected before booster (light blue, n=398), Vaccinated/Uninfected after booster (dark blue, n=175), and Breakthrough Infection after booster (orange, n=76). The average days between first dose and booster was 269±50 days. The average days between first dose and infection was 332±114 days. (B) Post-booster comparison of breastmilk RBD-specific total IgG OD levels (LOD: dotted line) for Vaccinated/Uninfected (n=14) 259 ± 19.7 DPB and Breakthrough Infection (n=15) 262.9±21.5 days post booster (DPB). (C) Breastmilk RBD-specific IgG optical density relative to days post initial antigen exposure (DPAE) via first vaccination series for Vaccinated/Uninfected cohort (n=197) or infection for Vaccinated Post-Infection group (n=9). (D) Post-booster comparison of breastmilk RBD-specific IgG1 and IgG4 OD measurements (LOD: dotted lines) for Vaccinated/Uninfected 259±19.7 (n=14) days post DPB and Breakthrough Infection (n=15) 262.9±21.5 DPB. (E) NP-specific IgG OD levels in breastmilk of longitudinal samples relative to days post infection (DPI) for Breakthrough Infection (orange, n=49) and Unvaccinated/Infection (red, n=19).

DISCUSSION

Multiple lines of evidence demonstrate waning of antibody response as early as 3 months after mRNA vaccination against SARS-CoV-2 even in the setting of multiple booster doses^29,30. Nevertheless, booster doses have been determined to be highly effective against symptomatic and severe disease³⁰. This longitudinal study evaluated how repeat SARS-CoV-2 antigen exposure through vaccination and infection during pregnancy affected the maternal humoral response and expanded on our previous observations¹⁹. It also examined transplacental antibody transfer to the neonate and levels in breastmilk. Our current findings reemphasize the importance of booster vaccination for extending the antibody response in maternal circulation, neonatal blood, and breastmilk.

We provide evidence of a prolonged RBD-specific IgG antibody response among pregnant women stimulated by repeated antigen exposure through booster vaccination and breakthrough infection. Additionally, breakthrough infection post-boost induced higher titer levels in maternal circulation compared to vaccination alone. The observed immune responses following breakthrough infection reflect a priming effect induced by prior vaccination rather than resulting from infection alone. Interestingly, participants within the Breakthrough Infection cohort were predominantly infected during the Omicron variant surge (late 2021 - early 2022) when there was lack of bivalent vaccine availability³¹. Importantly, none of the participants who contracted a breakthrough infection developed severe symptoms that required medical intervention. Despite lower vaccine effectiveness in preventing Omicron breakthrough infection, primary and especially booster vaccination strongly protects against severe disease, hospitalization, and death³². This highlights the ability of SARS-CoV-2 vaccines to not only prevent severe disease in pregnant women, but to enhance maternal antibody production should infection occur. We also show that primary vaccination after infection yields shorter antibody memory than in previously vaccinated patients with breakthrough infections. Hence, to diminish susceptibility to infection and improve maternal-fetal health outcomes, the current mRNA vaccination strategy should universally include SARS-CoV-2 vaccine boosters during pregnancy.

Antigen-specific IgG subclasses induced by natural infection or vaccination can serve as important determinants of clinical outcome against SARS-CoV-2³³. IgG subclasses play different roles in antigen recognition and neutralization. IgG1 and IgG3 are the most effective at activating complement, promoting pathogen neutralization, and have been associated with anti-viral antibody response³⁴. IgG2 is associated with response to bacterial polysaccharides rather than viral proteins. Finally, IgG4 has anti-inflammatory properties, is less commonly produced after natural infection, and increases after repeated antigen exposure (e.g. allergy immunotherapy)³⁴.

As described in previous studies on various viral infections and SARS-CoV-2^34,35, we show that SARS-CoV-2 infection induced a predominantly IgG1 and IgG3 anti-RBD- and NP-specific antibody response. Additionally, we previously reported that the Vaccinated/Uninfected cohort had an IgG4-predominant response following a third vaccine booster dose¹⁹. In this study, the Breakthrough Infection cohort exhibited robust RBD-specific IgG1 and IgG4 subclass dominance at delivery. In contrast, infection alone led to skewing toward RBD-specific IgG1 and IgG3 subclasses. Furthermore, infection in previously vaccinated patients further enhanced IgG4 production and resulted in higher titers compared to the Vaccinated/Uninfected cohort post-boost. This aligns with recent investigations which report increased IgG4 antibodies in non-pregnant subjects following repeated mRNA vaccination with the highest IgG4 levels being observed after a post-boost breakthrough infection³⁶. It is important to note that these immune changes are specific to mRNA-based vaccines as adenovirus-based SARS-CoV-2 vaccinations do not induce long-term IgG4 response³⁷. Higher proportions of SARS-CoV-2 spike-specific IgG4 antibodies are associated with reduced antibody-dependent cellular phagocytosis (ADCP) and complement deposition (ADCD) despite more efficient neutralization of the Omicron variant³⁶. Overall, it is unclear how IgG4 subclass dominance impacts humoral response to infection. Concerns of growing immune tolerance to mRNA vaccination with increasing IgG4 have been expressed, warranting further investigation to determine the impact of this subclass on vaccine efficacy.

Reportedly, the proportion of IgG4 is higher in pregnant individuals compared to non- pregnant individuals³⁸ with total IgG levels decreasing to below normal IgG range as pregnancy progresses³⁹. Moreover, increased IgG4 production in pregnant women has previously been observed in response to influenza vaccination and was accompanied by decreasing levels of antigen-specific IgG1 and IgG3 titers as gestation progresses⁴⁰. B cell class-switching to IgG4 is Th2 cell mediated and driven by IL-4 which directly correlates with IgG4 production⁴¹. Additionally, IL-10 production by Tregs can also shift antibody response toward IgG4⁴¹. The shift to a Th2/Treg anti-inflammatory milieu with pregnancy⁴² and the accompanying increase in IL-4 and IL-10 cytokine production⁴³ could be a potential explanation for the skewing to IgG4-dominated antibody responses during pregnancy.

Our access to maternal and UCB blood collected at delivery and newborn heel stick collected within 24 hours of birth allowed us to compare IgG levels among these sample types and observe patterns of placental IgG transfer within the cohorts. An important contributing factor to neonatal antibody levels is trimester of maternal antigen exposure as placental transfer of IgG begins during the second trimester and increases into the third⁴⁴. A pertussis vaccination, for example, is recommended during the early third trimester to maximize antibody transfer to the fetus⁴⁵. However, the small sample size of the Breakthrough Infection and Unvaccinated/Infection cohorts did not permit extensive analysis of delivery samples by trimester of infection. Significant correlations within the Vaccinated/Uninfected cohort between maternal and newborn EPT and days post first maternal vaccination suggested heightened placental IgG transfer following third trimester vaccination. Interestingly, UCB titers lacked a strong correlation to days post first maternal vaccination. While UCB is frequently employed as a substitute for newborn blood due to its accessibility⁴⁶, direct comparison studies of IgG levels between these sample types are limited. Our results indicate differential IgG levels within UCB and newborn plasma. Titers in the Breakthrough Infection delivery samples lacked significant correlation to days post maternal infection (the most recent antigen exposure) indicating trimester of infection did not significantly impact placental transfer within this cohort. The prolonged maternal antibody response initiated following breakthrough infection could extend throughout gestational trimesters providing heightened placental transfer regardless of maternal infection timing.

Additionally, the highest neonatal-to-maternal ratio of RBD-specific IgG was observed within the Breakthrough Infection cohort. This highlights the ability of a natural infection to boost pre-existing vaccine-induced humoral immunity and result in a more robust transfer of protective antibodies to the neonate. Indeed, booster doses have previously been reported to induce heightened maternal and neonatal humoral response¹⁸, effectively reducing the infant’s risk for developing symptomatic infection²⁰. We also report that anti-RBD IgG1 levels were higher in UCB and newborn plasma compared to maternal levels indicating more efficient transplacental transfer of this subclass in accord with previous research⁴⁷.

Postpartum passive transfer of antibodies to infants via breastmilk after maternal influenza vaccination has been reported to increase infant influenza-specific IgA titers and decrease occurrence of infection⁴⁸. We were unable to detect RBD-specific anti-SARS-CoV-2 IgA levels in breastmilk potentially due to the fact that breastmilk samples were diluted with equal volume HBSS to completely remove the fat layer. Nevertheless, we were able to detect RBD-specific IgG levels in breastmilk. Although they were much lower than maternal plasma, they were enhanced by breakthrough infection. We propose that an extended transfer of maternal antibodies to infants through breastmilk will provide additional protection against SARS-CoV-2 infection. As with maternal plasma, higher IgG1 and IgG4 levels were observed in the breastmilk of breakthrough infection participants post-boost.

There are limitations to our study. The number of Unvaccinated/Infected participants was significantly lower than the other groups making it challenging to measure reliable antibody decay rates and correlations between delivery samples. While participants in the Breakthrough Infection group were predominantly infected during Omicron variant dominance, due to enrollment occurring during a rapidly evolving pandemic the Unvaccinated/Infected cohort contained mostly participants infected prior to Omicron emergence. These variant differences introduce an additional complexity when comparing the infection cohorts. The Omicron variant contains significant mutations at the RBD, exhibiting heightened immune evasion and reduced neutralizing antibody titers post-infection compared to the earlier variants^49,50,51,52. In addition, we utilized an ancestral SARS-CoV-2 spike protein RBD antigen for antibody detection. Given the rapid mutations which occurred during the pandemic, specifically within the spike protein⁵³, this antigen may not detect all RBD-specific antibodies produced by infection with SARS-CoV-2 variants⁵⁴ . Therefore, our RBD titer measurements likely under-estimate the amount of functional RBD-specific antibodies produced in response to infections by Beta, Delta, and Omicron variants. Additional studies that measure titers against the spike protein RBD of other variants and conduct functional assays to determine antibody-dependent cellular cytotoxicity (ADCC) and neutralization capabilities of participant sera in response to different variants would further enhance the current study. Higher maternal antibody concentrations are associated with increased newborn protection against pertussis and influenza infection⁵⁵. Additionally, higher neonatal SARS-CoV-2 anti-spike antibodies have been associated with a decreased risk of infant COVID-19 up to 6 months of age⁵⁶. To determine the protection acquired by the expanded antibody response presented in this study, future research should explore maternal and fetal neutralization titers against contemporary SARS-CoV-2 variants.

Other limitations of the study include variation in the timing of vaccination and infection relative to the gestational age between participants, but this is unavoidable in human studies. Maternal immune response to vaccination varies across gestational age as does the efficiency of transplacental antibody transfer⁵⁷. Our IgG subclass analysis was limited to delivery time points and did not evaluate changes in subclass prominence during gestation because of limitations imposed at multiple medical centers for human subject research during the pandemic. This calls for retrospective studies evaluating vaccine and infection induced IgG subclass responses with trimester progression. It will also be important to explore a wider range of immune markers beyond IgG subclasses, such as IgA production, T-cell mediated response and antibody neutralizing capacity, to understand the scope of maternal and neonatal response to SARS-CoV-2 infection and vaccination.

In summary, data presented herein emphasize the importance of administering up to date SARS-CoV-2 vaccination during pregnancy. Recurrent antepartum SARS-CoV-2 antigen exposure, either through booster vaccination or post-boost breakthrough infection, extends maternal humoral immunity. In the context of a breakthrough infection, pre-existing maternal vaccination would not only reduce the severity of infection to prevent adverse outcomes but additionally would elicit enhanced placental transfer to the fetus providing greater protection to the newborn postpartum.

Highlights:

• Maternal SARS-CoV-2 infection in boosted individuals extends humoral immunity.

• Repeated RBD antigen exposure skews maternal IgG response toward IgG4 dominance.

• Gestational SARS-CoV-2 breakthrough infection facilitates vaccine-induced IgG transfer to neonates.

Abbreviations:

COVID-19

coronavirus disease 2019

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

IgG

immunoglobulin G

RBD

receptor binding domain

NP

nucleocapsid protein

EPT

endpoint titers

UCB

umbilical cord blood

DP1V

days post first vaccination

DPB

days post booster

DPI

days post infection

DPAE

days post initial antigen exposure

ADCP

antibody dependent cellular phagocytosis

ADCD

antibody dependent complement deposition

ADCC

antibody dependent cellular cytotoxicity