key: cord-0276964-ocfj7hsi
authors: Semmes, E. C.; Miller, I. G.; Jenks, J. A.; Wimberly, C. E.; Berendam, S. J.; Harnois, M. J.; Webster, H.; Hurst, J. H.; Kurtzberg, J.; Fouda, G. G.; Walsh, K. M.; Permar, S. R.
title: Maternal Fc-mediated non-neutralizing antibody responses correlate with protection against congenital human cytomegalovirus infection
date: 2021-12-07
journal: nan
DOI: 10.1101/2021.12.05.21267312
sha: 306b25d171435af7d5ea38133c37a3ef54d12997
doc_id: 276964
cord_uid: ocfj7hsi

Human cytomegalovirus (HCMV) is the most common congenital infection and a leading cause of stillbirth, neurodevelopmental impairment, and pediatric hearing loss worldwide. Development of a maternal vaccine or therapeutic to prevent congenital infection has been hindered by limited knowledge of the immune responses that protect against placental HCMV transmission in maternal primary and nonprimary infection. To identify protective antibody responses, we measured anti-HCMV IgG binding and anti-viral functions in maternal and cord blood sera from HCMV transmitting (n=41) and non-transmitting (n=40) mother-infant dyads identified via a large U.S.-based public cord blood bank. In a predefined immune correlate analysis, maternal monocyte-mediated antibody-dependent cellular phagocytosis (ADCP) and high avidity IgG binding to HCMV envelope glycoproteins were associated with decreased risk of congenital HCMV infection. Moreover, HCMV-specific IgG engagement of Fc{gamma}RI and Fc{gamma}RIIA, which mediate non-neutralizing antibody responses, was enhanced in non-transmitting mother-infant dyads and strongly correlated with ADCP. These findings suggest that Fc effector functions including ADCP protect against placental HCMV transmission. Taken together, our data indicate that future active and passive immunization strategies to prevent congenital HCMV infection should target Fc-mediated non-neutralizing antibody responses.

responses in transmitting versus non-transmitting women. In an exploratory analysis, we applied a 90 systems serology approach to define differences between transmitting and non-transmitting mother-91 infant dyads and quantified HCMV-specific IgG engagement of host FcγRs. Insights from this study of 92 binding and functional antibody responses in HCMV transmitting and non-transmitting mother-infant 93 dyads can inform vaccine and therapeutic development to prevent cCMV and ameliorate a major cause 94 of pediatric morbidity worldwide. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint

Our study included sera from 81 mother-infant dyads recruited as donors to the Carolinas Cord Blood 119 Bank (CCBB), a U.S.-based public cord blood bank with 130,000 collections and 45,000 banked donors 120 to-date (Supplementary Figure 1) . Donor mothers were screened for total anti-HCMV antibodies and if 121 present, cord blood plasma was tested by qPCR for replicating HCMV. Cord blood units that contained 122 replicating HCMV were excluded from banking for cord blood transplantation and were retained for 123 research. Congenital HCMV infection (cCMV) was defined based on the presence of HCMV viremia in 124 the donated cord blood plasma. Forty-one dyads with congenital HCMV infection ("HCMV transmitting") 125 were matched to forty dyads with HCMV IgG seropositive mothers that gave birth to cCMV uninfected 126 infants ("HCMV non-transmitting"). Matching criteria included infant sex, infant race, maternal age (+/-3 127 years), and delivery year (+/-3 years). Baseline demographic and clinical characteristics were 128 comparable between HCMV transmitting and non-transmitting dyads, though transmitting dyads had a 129 non-significantly higher rate of Cesarean section (56% vs. 40%, Fisher's exact p = 0.22, Table 1 (Table 1) . Thus, while the majority of mothers in both groups had high avidity IgG responses and no 140 detectable IgM responses at the delivery timepoint, a higher proportion of mothers in the transmitting 141 group had detectable HCMV-specific IgM, which could indicate a higher rate of primary infection or 142 reinfection during pregnancy in the transmitting group (41, 42). To further assess comparability 143 between groups, we quantified HCMV viral loads in maternal sera and found that a similar proportion of 144 We previously observed that HCMV-specific IgG transfer across the placenta may be compromised in 150 cCMV infection, a phenomenon primarily related to maternal hypergammaglobulinemia in HCMV 151 transmitting women (44). Thus, we first measured placental HCMV-specific IgG transfer in transmitting 152 and non-transmitting mother-infant dyads. We quantified IgG binding against 3 distinct HCMV strains 153 including TB40E (an endotheliotropic strain expressing pentamer), AD169r (a lab-adapted strain with 154 repaired pentamer expression), and Toledo (a low-passage clinical isolate lacking pentamer). Within 155 mother-infant pairs, HCMV-specific IgG binding was lower in cord blood compared to maternal sera in 156 transmitting but not non-transmitting dyads ( Figure 1A) . However, placental IgG transfer efficiency was 157 only slightly lower in transmitting compared to non-transmitting dyads ( Figure 1B) . HCMV transmitting 158 mothers also had higher total IgG levels (i.e., hypergammaglobulinemia), which negatively correlated 159 with HCMV-specific IgG transfer efficiency ( Supplementary Figure 2A-B) . Despite slightly decreased 160 placental IgG transfer ratios, whole virion HCMV-specific IgG titers were comparable in the cord blood 161 of infected and uninfected infants ( Figure 1C ). In contrast, HCMV glycoprotein-specific IgG titers were 162 significantly elevated (2.5-10 fold) in transmitting compared to non-transmitting dyads (Figure 1D-E) . 163

When adjusting for total IgG levels, non-transmitting women demonstrated higher relative binding to 164 HCMV virions and cell-associated native gB, while IgG binding against pentamer, gHgLgO, and gHgL 165 remained significantly higher in transmitting women (Supplementary Figure 2C-D) . 166 167

Transplacental transfer of low-avidity HCMV-specific IgG complexed to HCMV virions has been posited 169 to increase HCMV transmission risk (45) and we previously observed that low avidity HCMV-specific 170 IgG may be preferentially transferred across the placenta in HCMV transmitting pregnancies (44). Thus, 171 we next measured HCMV-specific IgG avidity in paired maternal and cord blood samples. Maternal 172 sera from non-transmitting pairs had higher avidity IgG binding to AD169r and Toledo strains, but not 173 TB40E, compared to transmitting dyads (Figure 2A) . Within mother-infant pairs, whole virion HCMV 174

IgG binding avidity was significantly lower in the cord blood compared to maternal sera in the 175 transmitting but not the non-transmitting groups ( Figure 2B) . HCMV glycoprotein-specific IgG avidity 176 was also lower in the cord blood of infected compared to uninfected infants (Figure 2C) , though no 177 differences in glycoprotein-specific IgG avidity were observed within paired sera from mother-infant 178 dyads. In a sensitivity analysis excluding mothers with HCMV-specific IgM responses as a surrogate 179 biomarker for recent maternal primary infection or reinfection, differences in whole virion HCMV IgG 180 binding avidity persisted but most differences in HCMV glycoprotein-specific IgG binding avidity were 181 no longer statistically significant (Supplementary Figure 3A-C) . In addition to measuring HCMV-specific IgG binding magnitude and avidity, we quantified anti-viral 186 functions of maternal and cord blood sera from transmitting and non-transmitting dyads. Unexpectedly, 187 HCMV neutralizing titers were approximately 1.5-4 fold higher in transmitting versus non-transmitting 188 dyads across HCMV strains and cell types (Figure 3A-D) . Within mother-infant dyads, neutralizing 189 titers were similar in the maternal compared to cord blood sera in both groups. In contrast, ADCP, an 190

Fc-mediated non-neutralizing antibody response, against HCMV was higher in non-transmitting 191 compared to transmitting mothers ( Figure 3E) . This difference was statistically significant for the 192 Toledo HCMV strain (p = 0.0057) and remained significant after correction for multiple comparisons (p 193 = 0.011) with a trend towards statistical significance for TB40E (p = 0.053) and AD169r (p = 0.068). 194

Interestingly, within mother-infant pairs, whole virion HCMV ADCP responses were highly enriched in 195 the cord blood sera compared to maternal sera within both transmitting and non-transmitting groups 196 ( Figure 3F) , suggesting preferential transfer of ADCP-mediating IgG across the placenta. The 197 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. were strongly correlated with each other (Supplementary Figure 4) . Due to this high degree of 219 multicollinearity, LASSO was employed for feature selection. LASSO is an approach to minimize 220 overfitting a regression model that can shrink the coefficients of poorly predictive variables to zero, 221 thereby removing them from the model. First, the cohort was randomly split into a training and test 222 dataset and a k-fold cross-validation approach was employed to train the LASSO model. LASSO-223 selected features included magnitude of IgG binding to pentamer, avidity of gB IgG binding, avidity of 224 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. In an exploratory analysis, we quantified FcR engagement by HCMV-specific IgG in maternal and cord 234 blood sera to examine the role of Fc-mediated antibody responses and placental IgG transfer in 235 congenital HCMV transmission. Since FcRs can transfer immune complexes across the placenta, we 236 explored if differential HCMV-specific IgG binding to FcRs in maternal versus cord blood sera was 237 associated with placental transmission (47). Using a multiplex FcR binding assay, we measured HCMV Since transmitting dyads had higher binding to gB-, pentamer-, gHgLgO-and gHgL-coated multiplex 243 beads at baseline (Figure 1E) , we normalized FcR binding to total antigen-specific IgG binding for each 244 HCMV glycoprotein measured. After normalization, transmitting and non-transmitting dyads had similar 245 magnitude HCMV-specific IgG binding to FcRn, FcγRIIB and FcγRIIIA between and within mother-246 infant pairs (Supplementary Figure 6A , C-E). However, HCMV glycoprotein-specific IgG binding to 247

FcγRI was significantly higher in non-transmitting compared to transmitting dyads ( Figure 4A ). Within 248 mother-infant pairs, HCMV glycoprotein-specific IgG binding to FcγRI was also highly enriched in the 249 cord blood of uninfected infants ( Figure 4B) . HCMV viral loads in the cord blood of infected infants 250 were significantly negatively correlated with FcγRI binding to maternal gHgLgO-(Spearman ρ = -0.3, p 251 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint = 0.019), pentamer-(Spearman ρ = -0.36, p = 0.020), and gB-(Spearman ρ = -0.32, p = 0.037) specific 252

IgG, indicating that HCMV-specific IgG engagement of FcγRI may help control HCMV viremia. 253

Conversely, glycoprotein-specific IgG binding to FcγRIIA was enhanced in the cord blood of infected 254 infants with preferential transfer across the placenta within mother-infant dyads, though this was only 255 observed for the high-affinity variant FcγRIIAH131 (Figure 4C-D, Supplementary Figure 6B) . We next employed a principal components analysis (PCA) across all 50 humoral immune variables 260 tested to define differences in HCMV-specific antibody responses in transmitting versus non-261 transmitting dyads. In the maternal and cord blood sera, PC1 accounted for 57 and 59% of the 262 variance, respectively; however, PC2, which accounted for 16 and 17% of the variance, was superior at 263 delineating between transmitting and non-transmitting dyads (Figure 5A Figure 6A ). Using this previously validated model system with mouse IL-2 as a quantitative read-out 278 for FcγR activation, we quantified HCMV-specific IgG signaling through FcγRI and FcγRIIA. We 279 measured HCMV-specific IgG signaling through FcγRI and FcγRIIA because 1) there were differences 280 in IgG binding to these FcγRs in transmitting versus non-transmitting dyads (Figure 4) and 2) FcγRI 281 and FcγRIIA mediate ADCP, which we identified as associated with protection against congenital 282 HCMV transmission. We first used flow cytometry to confirm that each previously validated cell line was 283 exclusively expressing the FcγR of interest ( Figure 6B ). HCMV-specific IgG signaling through FcγRI 284 was 3-4 fold higher when reporter cells were stimulated with maternal or cord blood sera from non-285 transmitting versus transmitting dyads ( Figure 6C ). HCMV-specific IgG signaling through FcγRIIA was 286 similar in transmitting and non-transmitting mothers, yet there was a trend towards significantly (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint

Despite decades of research, the immune correlates of protection against placental HCMV 307 transmission have remained elusive and there are no licensed therapeutics or vaccines to prevent 308 congenital HCMV infection (12). Using a case-control cohort of cord blood donor HCMV transmitting 309 (n=41) and non-transmitting (n=40) mother-infant dyads, we identified Fc-mediated effector functions 310 and ADCP as novel potential correlates of protection against congenital HCMV transmission. This 311 finding represents a critical step forward towards the goal of preventing cCMV infection worldwide. 312 313 Consistent with our recent study on a subset of this cohort, HCMV-specific IgG transfer across the 314 placenta was modestly decreased in cCMV infection and negatively correlated with maternal 315 hypergammaglobulinemia (44). Despite lower transfer ratios, cord blood from cCMV-infected infants 316 had high levels of HCMV-specific IgG, indicating that reduced transfer of HCMV-specific IgG across the 317 placenta into the fetal circulation is not a risk factor for congenital infection. Since HCMV:IgG immune 318 complex binding to FcRn has been proposed as a mechanism whereby HCMV can bypass the 319 placental barrier (45, 52), we also examined whether HCMV-specific IgG binding to FcRn differed 320 between transmitting and non-transmitting dyads. Normalized pentamer-and gHgLgO-IgG binding to 321

FcRn was slightly higher in transmitting dyads (Supplementary Figure 6A ), yet there were no 322 differences within maternal-cord blood pairs to suggest enhanced placental transfer of anti-HCMV IgG 323 via FcRn in the setting of congenital HCMV transmission. 324 325 Our finding that high-avidity HCMV-specific IgG was enriched in the cord blood of uninfected infants 326 further confirms our previous work and indicates that high-avidity HCMV-specific IgG plays an important 327 role in preventing congenital HCMV infection (44, 53). It is possible that the relative enrichment of low-328 avidity HCMV-specific IgG in the cord blood of infected infants is due to preferential transfer of low-329 avidity IgG across the placenta, which has been proposed as a mechanism of antibody-dependent 330 enhancement (45, 52). However, we speculate that it is more likely an indication of recent maternal 331 primary infection or reinfection leading to an abundance of low avidity antibodies present earlier in 332 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint pregnancy that were then transferred to the fetus in the setting of congenital HCMV transmission. Our 333 regression analyses and PCA of maternal antibody responses provide additional evidence that high-334 avidity HCMV-specific IgG is associated with protection against congenital infection (20, 23). 335 336 HCMV-specific IgG binding and neutralization have been correlated with protection against congenital 337 transmission in some studies (21, 24), yet we unexpectedly found that HCMV glycoprotein-specific IgG 338 binding and neutralizing titers were significantly higher in transmitting versus non-transmitting women. 339

These seemingly discordant results are consistent with a recent study that observed higher gB-and 340 pentamer-specific IgG levels in transmitting versus non-transmitting women with nonprimary infection 341 (25). Dorfman et al. observed that IgG binding against putatively protective epitopes on pentamer were 342 associated with elevated risk of congenital infection, which aligns with our finding that pentamer-specific 343

IgG binding was associated with increased transmission risk. It has been proposed that elevated 344 HCMV-specific IgG levels in transmitters are due to "boosting" from HCMV reactivation or reinfection 345 (25). Nevertheless, we did not observe an association between maternal viral burden and HCMV-346 specific IgG levels or neutralizing titers, as the frequency of detection and levels of HCMV DNAemia 347 were comparable in transmitting and non-transmitting mothers at the delivery timepoint ( Table 1) . 348

Moreover, HCMV-specific IgG binding and neutralizing titers remained significantly elevated in 349 transmitters when excluding HCMV IgM positive mothers as a surrogate biomarker for recent primary 350 infection or reinfection. Though pentamer IgG binding was associated with increased transmission risk, 351 antibody-dependent enhancement of infection is unlikely given the compelling epidemiological and 352 experimental evidence that CMV-specific IgG protects against placental transmission in human and 353 animal studies (2, 54-56). Instead, we infer that boosted IgG levels and in vitro neutralizing titers in 354 transmitting mothers are likely due to active HCMV infection following primary acquisition, reactivation, 355 or reinfection during pregnancy. 356 357 Intriguingly, we found that ADCP, a non-neutralizing antibody effector function, was higher in non-358 transmitting compared to transmitting mother-infant dyads and that maternal ADCP responses were 359 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint HCMV infection in our recent study of adolescent and postpartum gB/MF59 vaccinees, there was a 387 trend towards ADCP being significantly associated with protection against virus acquisition (β = -0.420, 388 p = 0.120) (36). Notably, IgG binding to cell-associated gB, which we identified as an immune correlate 389 of protection for gB/MF59 vaccinees, was significantly correlated with HCMV-specific ADCP and 390

FcγRI/FcγRIIA activation in our study (Supplementary Figure 4) (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. cord blood biospecimens at delivery and mother-infant dyads were identified from over 29,000 CCBB 445 donor records (Supplementary Figure 1) Cases of congenital HCMV infection (cCMV) were defined as mother-infant dyads with cord blood that 453 screened positive for HCMV viremia at birth per PCR testing. "HCMV transmitting" cases with cCMV 454 infection (n=41) were matched to a target of 1 "HCMV non-transmitting" seropositive mother (n=40) with 455 no HCMV viremia detected in the cord blood (Supplementary Figure 1) . Maternal HCMV IgG 456 seropositivity and avidity were confirmed by an in-house whole virion HCMV ELISA and HCMV IgM 457 seropositivity was determined using a clinical diagnostic ELISA (Bio-Rad). Maternal sera was screened 458 for HCMV DNAemia via quantitative PCR (qPCR). For qPCR, 150-200ul of maternal sera was 459 ultracentrifuged then DNA was extracted using the DNA QIAamp Kit (Qiagen) and tested in duplicate 460 using SybrSelect and 300nM of forward and reverse primers designed to amplify the HCMV immediate 461 early-1 (IE1) gene (24). HCMV viral loads were then interpolated from an IE1 plasmid standard curve. 462

Maternal sera was also screened for hypergammaglobulinemia (total IgG concentration >15,000 463 mg/dL) via ELISA as previously described (44). (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Protein Production Facility and biotinylated in-house. Maternal and cord blood sera were diluted 1:500 503 in assay diluent before incubation with HCMV glycoprotein-coated beads, as above. Next, biotinylated 504 human FcRs were complexed with streptavidin-PE (BD Biosciences) then co-incubated with antibody-505 bound beads after washing. For FcRn, the assay buffer pH was adjusted to 6.5 to facilitate FcRn-IgG 506 binding. MFI was acquired on a Bio-Plex 200 and duplicates with CVs >25% were repeated. 507 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. and Toledo were conjugated to fluorochrome AF647 (Invitrogen) to measure ADCP (36). In brief, 545 2.0x10 6 PFU of TB40/E, 1.0x10 7 PFU of AD169r, or 1.0x10 7 PFU of Toledo virions were buffer-546 exchanged with 1X PBS using a 100,000-kDa Amicon filter (Millipore) and conjugated to 10µg of AF647 547 N-hydroxysuccinimide ester reconstituted in DMSO during a 1-hour incubation with constant agitation. 548 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint

The conjugation reaction was quenched with 80ul of 1M tris-HCl (pH = 8.0) and fluorescently-labeled 549 virus was diluted 25X in wash buffer. A serial dilution of HCMV-hyperimmuneglobulin (Cytogam) was 550 included as a positive control while seronegative sera samples and an anti-RSV monoclonal antibody 551 (Synagis) were included as negative controls. In a 96-well plate, fluorescently-labeled virus was co-552 incubated with maternal sera, cord blood sera (1:10) or controls at 37°C for 2 hours to allow immune 553 complex formation before adding 50,000 THP-1 cells per well. Plates were then centrifuged (1200g) at 554 4°C for 1 hour in a spinoculation step before a 1-hour incubation at 37°C to allow for phagocytosis. 555 To quantify FcγR activation, 96-well plates were coated with 20,000 PFU/well of HCMV strain AD169r 574 and incubated at 4°C overnight. After coating, plates were washed with assay buffer (1X PBS + 1% 575 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint FBS) then blocked with buffer at room temperature for 1 hour. After blocking, HCMV-coated plates 576 were co-incubated with maternal or cord blood sera diluted 1:10 in BW cell media at 37°C for 1 hour. 577 HCMV-hyperimmuneglobulin (Cytogam), seronegative and no antibody conditions were included as 578 positive and negative controls. Following immune complex formation, plates were washed with BW 579 media before adding 100,000 FcR1a-expressing or FcR2a-expressing BW cells per well. In separate 580 wells, the parental (non-transfected) BW cells were added as a negative control. Cells were incubated 581 at 37°C with immune complex coated plates for 20 hours before being transferred to V-bottom plates 582 and pelleted (1200 rpm). Cell supernatants were harvested and mouse IL-2 (mIL-2) levels in culture 583 supernatants were measured using ELISA. For the mIL-2 ELISA, 384-well plates were coated with 584 3µg/mL of purified rat anti-mouse IL-2 (BD biosciences) and incubated at 4°C overnight before 585 blocking. A serial dilution of purified mIL2 and BW cell culture supernatants were added in duplicate 586 and incubated for 1 hour at room temperature. After primary incubation, plates were incubated with rat 587 anti-mIL2 conjugated to biotin (BD biosciences; 1:2000) followed by streptavidin-HRP (1:8000) for 1 588 hour and 30 minutes respectively. Plates were developed with TMB/KPL then OD at 450nm was 589 measured via SpectroMax. Mouse IL-2 concentrations were interpolated from a 5-parameter mIL-2 590 standard curve using GraphPad Prism. Duplicates with CVs >30% were repeated. 591 592 Statistics. All primary raw and analyzed data underwent independent data quality control (QC) by a 593 second lab member prior to statistical analysis and inclusion in the study. For all analyses, interpolated 594 antibody or cytokine concentrations, MFI values and neutralization titers were log-transformed to 595 normalize the data distribution. Antibody responses below the limit of detection were set equal to the 596 limit of detection for statistical analyses. To assess differences between HCMV transmitting and non-597 transmitting mother-infant dyads, immune variables were compared between groups using Mann-598 Whitney U/Wilcoxon rank-sum and within dyads using Wilcoxon signed rank tests. Statistical 599 significance was defined a priori as p < 0.05 with a two-tailed test and a Benjamini-Hochberg correction 600 for multiple comparisons. For the primary immune correlate analysis, 13 predefined maternal humoral 601 immune variables were included for univariate logistic regression analysis. Additional sensitivity 602 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint analyses included multivariable models adjusted for 1) maternal hypergammaglobulinemia and 2) 603 maternal HCMV-specific IgM status. Using the 13 predefined immune variables from our primary 604 immune correlate analysis, LASSO regression analysis was performed using the caret R package and 605 a k-fold cross-validation framework. For the LASSO regression, the cohort was randomly split into two 606 independent datasets, which were used for training the LASSO model and testing the predictive 607 performance of the model respectively. All statistical analyses were completed in R v4. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. pentamer-specific antibodies synergize for virus neutralization but do not correlate with 762 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. FDR-corrected P values reported for Mann-Whitney U test or Wilcoxon signed-rank test. * P < 0.05, ** 941 P < 0.01, *** P < 0.001. 942 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. HCMV-specific IgG binding avidity against HCMV strains TB40E, AD169r, and Toledo were measured 946 using whole virion enzyme-linked immunosorbent assay (ELISA) with an additional dissociation step 947 using urea and relative avidity index (RAI) was calculated as (OD with urea/OD without urea)x100%. 948 HCMV glycoprotein-specific IgG binding avidity was measured using a Luminex-based binding antibody 949 multiplex assay with an additional dissociation step with sodium citrate and RAI was calculated as (MFI 950 with sodium citrate/MFI with 1X PBS)x100%. IgG binding avidity in maternal (M) and cord blood (CB) 951 sera were compared between and within HCMV transmitting (red circles, n = 41) and non-transmitting 952 corrected P values reported for Mann-Whitney U test or Wilcoxon signed-rank test. * P < 0.05, ** P < 957 0.01, *** P < 0.001. 958 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. neutralization assays indicate the starting dilution corresponding to the lower limit of detection (ID50 = 972 10). FDR-corrected P values reported for Mann-Whitney U test or Wilcoxon signed-rank test. * P < 973 0.05, ** P < 0.01, *** P < 0.001. 974 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted December 7, 2021. ; https://doi.org/10.1101/2021.12.05.21267312 doi: medRxiv preprint Table 2 . Univariate and LASSO regression analysis of maternal humoral immune correlates of congenital HCMV transmission. Logistic regression analysis on 13 primary variables of maternal sera anti-HCMV antibody responses comparing HCMV transmitting and HCMV IgG seropositive non-transmitting mother-infant dyads. 

Impact of Maternal Immunity on Congenital 685

Advancing Our Understanding of Protective Maternal 688 Immunity as a Guide for Development of Vaccines To Reduce Congenital Cytomegalovirus 689 Infections

The "silent" global burden of 691 congenital cytomegalovirus

In utero cytomegalovirus 693 infection and development of childhood acute lymphoblastic leukemia

Congenital cytomegalovirus infection: clinical outcome

Clinical infectious diseases : an official publication of the Infectious Diseases Society of

Vaccination strategies to enhance immunity in 699 neonates

Initial findings from a 701 novel population-based child mortality surveillance approach: a descriptive study

Antibodies in CMV Infection: Implications for New Therapeutic Approaches

Cytomegalovirus 713 reinfections in healthy seroimmune women. The Journal of infectious diseases

Congenital Human Cytomegalovirus Infection and the Enigma of Maternal Immunity. 716 Journal of virology

Estimation of the 718 worldwide seroprevalence of cytomegalovirus: A systematic review and meta-analysis. Reviews 719 in medical virology

Human cytomegalovirus reinfection is associated with intrauterine transmission in a highly 722 cytomegalovirus-immune maternal population

Intrauterine transmission of 724 cytomegalovirus to infants of women with preconceptional immunity. The New England journal 725 of medicine

Attribution of congenital cytomegalovirus infection 727 to primary versus non-primary maternal infection

Maternal immunity and prevention of congenital 730 cytomegalovirus infection

Maternal Cytomegalovirus Infection: A 15-Year Experience

Congenital Cytomegalovirus Infection Following Primary and Nonprimary Maternal Infection: A 736 congenital transmission

Clinical infectious diseases : an official publication of the Infectious Diseases Society of

Immunoglobulin therapy 768 of fetal cytomegalovirus infection occurring in the first half of pregnancy--a case-control study of 769 the outcome in children

Passive immunization during pregnancy for 771 congenital cytomegalovirus infection. The New England journal of medicine

A randomized 774 trial of hyperimmune globulin to prevent congenital cytomegalovirus. The New England journal 775 of medicine

Hyperimmune Globulin to Prevent Congenital Cytomegalovirus Infection. The New England 778 journal of medicine

Early primary 780 cytomegalovirus infection in pregnancy: maternal hyperimmunoglobulin therapy improves 781 outcomes among infants at 1 year of age. Clinical infectious diseases : an official publication of 782 the Infectious Diseases Society of America

Human cytomegalovirus (HCMV) infection/re-infection: development of a 784 protective HCMV vaccine

HCMV 786 glycoprotein B subunit vaccine efficacy mediated by nonneutralizing antibody effector functions

Vaccine prevention of 790 maternal cytomegalovirus infection

Antibody binding to 793 native cytomegalovirus glycoprotein B predicts efficacy of the gB/MF59 vaccine in humans

Safety and 796 efficacy of a cytomegalovirus glycoprotein B (gB) vaccine in adolescent girls: A randomized 797 clinical trial

Protection from 799 cytomegalovirus viremia following glycoprotein B vaccination is not dependent on neutralizing 800 antibodies

Innate immune defenses at the maternal-fetal interface. Current 803 Opinion in Immunology

Sex differences in pediatric infectious diseases. The Journal of 805 infectious diseases

Congenital Human Cytomegalovirus Infection: A

Narrative Review of Maternal Immune Response and Diagnosis in View of the Development of 808 a Vaccine and Prevention of Primary and Non-Primary Infections in Pregnancy

Immunoglobulin M to cytomegalovirus in primary 811 and reactivation infections in renal transplant recipients

Single-866 cell reconstruction of the early maternal-fetal interface in humans

Phenotypic and 869 functional characterization of first-trimester human placental macrophages, Hofbauer cells. The 870 Journal of experimental medicine

Cytomegalovirus 872 glycoprotein-B vaccine with MF59 adjuvant in transplant recipients: a phase 2 randomised 873 placebo-controlled trial

Polyfunctional Fc-875 effector profiles mediated by IgG subclass selection distinguish RV144 and VAX003 vaccines

Env V1-V2 IgG3 correlates with lower HIV-1 infection risk and declines soon after vaccination

IgG3 881 enhances neutralization potency and Fc effector function of an HIV V2-specific broadly 882 neutralizing antibody

The Immunoregulatory Roles of Antibody Glycosylation. Trends in 886 immunology

Proinflammatory IgG Fc structures in patients with severe COVID-19

CD4+ T cells protect against severe congenital cytomegalovirus disease in a novel nonhuman 892 primate model of placental cytomegalovirus transmission. Proceedings of the National Academy 893 of Sciences of the United States of America

Congenital and maternal cytomegalovirus 895 infections in a London population

Vaccines for the 21st Century: A Tool for Decisionmaking

All rights 899 reserved.; 2000. 900 74. Nelson CS, Herold BC, and Permar SR. A new era in cytomegalovirus vaccinology: 901 considerations for rational design of next-generation vaccines to prevent congenital 902 cytomegalovirus infection

A collaborative study to establish the 1st WHO International 904 Standard for human cytomegalovirus for nucleic acid amplification technology

Humoral Immune Correlates 907 for Prevention of Postnatal Cytomegalovirus Acquisition. The Journal of infectious diseases

Vi-specific serological correlates of 910 protection for typhoid fever

Mediate Selective Placental Transfer of IgG in HIV-Infected Women

HCMV-specific IgG signaling through FcγRI and FcγRIIA is increased in HCMV non-1005 transmitting compared to transmitting mother-infant dyads and strongly correlated with ADCP

CB) sera 1007 from HCMV transmitting (red circles, n = 41) and non-transmitting (blue squares, n = 40) mother-infant 1008 dyads. (A) To quantify HCMV-specific IgG activation of FcγRs, mouse BW cell lines stably expressing 1009 chimeric human FcγRs fused to a mouse CD3ζ signaling domain were co-cultivated with virus:sera 1010 immune complexes for 20 hours. Activation of FcγRs by immune complexes triggered CD3ζ signaling and 1011 mouse IL-2 (mIL-2) secretion

anti-FcγRIII/CD16 (blue) and isotype control (red) PE-conjugated 1014 antibodies to confirm FcγR expression. HCMV-specific IgG activation of (C) FcγRIA and (D) FcγRIIA in 1015 transmitting and non-transmitting dyads. HCMV-specific IgG activation of (E) FcγRIA and (F) FcγRIIA 1016 within paired maternal and cord blood sera

ND = not detected a Maternal HCMV viral copies listed in viral copies/mL, lower limit of detection = 250 copies/mL b Cord blood HCMV viral copies detected listed in IU/mL, lower limit of detection = 137 copies/mL HCMV transmitting and non-transmitting dyads were matched on maternal age

 Tables  1020  1021  Table 1