Prenatal Infection by Respiratory Viruses Is Associated with Immunoinflammatory Responses in the Fetus

Abstract

Rationale

Respiratory viral infections can be transmitted from pregnant women to their offspring, but frequency, mechanisms, and postnatal outcomes remain unclear.

Objectives

The aims of this prospective cohort study were to compare the frequencies of transplacental transmission of respiratory syncytial virus (RSV) and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), analyze the concentrations of inflammatory mediators in maternal and fetal blood, and assess clinical consequences.

Methods

We recruited pregnant women who developed upper respiratory infections or tested positive for SARS-CoV-2. Maternal and cord blood samples were collected at delivery. Study questionnaires and electronic medical records were used to document demographic and medical information.

Measurements and Main Results

From October 2020 to June 2022, droplet digital PCR was used to test blood mononuclear cells from 103 mother–baby dyads. Twice more newborns in our sample were vertically infected with RSV compared with SARS-CoV-2 (25.2% [26 of 103] vs. 11.9% [12 of 101]; P = 0.019). Multiplex ELISA measured significantly increased concentrations of several inflammatory cytokines and chemokines in maternal and cord blood from newborns, with evidence of viral exposure in utero compared with control dyads. Prenatal infection was associated with significantly lower birth weight and postnatal weight growth.

Conclusions

Data suggest a higher frequency of vertical transmission for RSV than SARS-CoV-2. Intrauterine exposure is associated with fetal inflammation driven by soluble inflammatory mediators, with expression profiles dependent on the virus type and affecting the rate of viral transmission. Virus-induced inflammation may have pathological consequences already in the first days of life, as shown by its effects on birth weight and postnatal weight growth.

At a Glance Commentary

Scientific Knowledge on the Subject

Common respiratory viruses like influenza and respiratory syncytial virus (RSV) have been shown to cross the placental barrier and spread hematogenously from the respiratory tract of pregnant women to the developing fetus. Since late 2019, the Coronavirus Disease 2019 (COVID-19) pandemic has added the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) to the number of respiratory pathogens that increase the risk of pregnancy complications and pose a potential threat to the offspring. In addition to the physical transmission of replicating viral particles to the fetus, a maternal infection - especially within critical developmental windows - may damage the fetus even without direct invasion, but rather via activation of maternal inflammation and immunity.

What This Study Adds to the Field

The findings reported in this article represent the first systematic assessment of transplacental transmission of the two most common respiratory viruses, RSV and SARS-CoV-2, of the inflammatory response they cause in the mother and the fetus, and the pathophysiologic consequences, especially on the offspring’s weight growth. The evidence reported here has important implications for our understanding of the influence of prenatal exposure to respiratory viruses on immune and lung development in early life.

Common respiratory viruses such as influenza and respiratory syncytial virus (RSV) have been shown to cross the placental barrier and spread hematogenously from the respiratory tract of a pregnant woman to the developing fetus (1). Since late 2019, the coronavirus disease (COVID-19) pandemic has added the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) to the number of respiratory pathogens that increase the risk of pregnancy complications and pose a potential threat to offspring (2). In addition to the physical transmission of replicating viral particles to the fetus, a maternal infection, especially within critical developmental windows, may damage the fetus even without direct invasion but rather via activation of maternal inflammation and immunity, transfer of soluble inflammatory mediators, and mobilization of cellular effectors or by impairing placental function and limiting the passage of nutrients and other factors essential for fetal growth (3).

Despite the obvious public health importance of vertical infections by respiratory viruses, the precise dynamics and mechanisms are still poorly understood. In particular, it is unknown how frequent is the transplacental transmission of specific pathogens, whether coinfections are possible or common, and the immune responses and inflammatory pathways activated or suppressed in the mother, fetus, and newborn. Even more important, little is known about the direct and indirect pathophysiologic consequences of maternal viral respiratory infections for the fetus and the newborn, especially regarding clinical outcomes, postnatal development, and weight growth.

To address these questions, we began a multisite prospective cohort study of 105 dyads, each comprising a pregnant woman delivering a full-term newborn between 2020 and 2022 in greater New Orleans, Louisiana, including clinical outcomes recorded in the perinatal period (e.g., weight, length of hospital stay). Some of the results of these studies have been previously reported in abstract form (4, 5).

Methods

Additional details on methods are provided in the online supplement.

Study Design

We conducted a multisite study of pregnant mothers and their offspring, TTRSV (Transplacental Transmission of RSV) (ClinicalTrials.gov identifier NCT 05443607), aimed at studying the transplacental transmission of RSV and resulting immunoinflammatory and clinical outcomes. To this end, we recruited a prospective cohort of full-term newborns delivered between October 2020 and June 2022 at Tulane Lakeside Hospital and Ochsner Baptist Hospital. Both medical facilities are located in greater New Orleans and at the time of the study served a population including 59% Black, 31% White non-Latino/a/x, 4% White Latino/a/x, and 3% Asian residents. Although the study was initially planned to focus on the vertical transmission of RSV, the onset of the COVID-19 pandemic shortly before starting the recruitment of subjects prompted the inclusion of pregnant women who tested positive for SARS-CoV-2 on rapid or PCR tests done in prenatal clinics.

Blood Collection and Analysis

A total of 287 deidentified blood samples were collected by trained nursing staff members. Of these, 127 peripheral blood samples were drawn by phlebotomy from mothers admitted to labor and delivery (L&D). A subgroup of 39 mothers had additional blood samples drawn during prenatal clinic visits. In addition, 121 cord blood (CB) samples (i.e., the fetal blood left in the umbilical cord after birth) were drawn in L&D by venipuncture of the umbilical cord and placental vessels immediately after delivery and strictly following a standardized sterile protocol.

Droplet digital PCR (ddPCR) was used to measure viral RNA copy numbers in maternal and cord peripheral blood mononuclear cells using specific primer/probe sets for RSV-A and RSV-B (Genesig; Primerdesign) and for SARS-CoV-2 (Integrated DNA Technologies).

The expression of inflammatory cytokines, chemokines, and soluble growth factors was measured using the Quantibody Human Cytokine Antibody Array 440 (RayBiotech), an array-based multiplex ELISA system that can simultaneously and quantitatively detect the expression amounts of 440 proteins (6). Because of sample volume limitations, two maternal blood (MB) samples collected in the clinic and five CB samples could not be tested with this array.

Data Analysis

All data were uploaded into REDCap software, a secure web-based application designed to support data capture for clinical research (7), and were analyzed using SAS version 9.4 (SAS Institute), JMP Pro version 16.2.0 (SAS Institute), or Prism version 9.5.1 for Windows (GraphPad Software). Proportions were compared using Fisher’s exact test. Bivariate differences between groups were compared using one-way ANOVA for continuous data and Fisher’s exact test for categorical data. In cases in which the assumptions of ANOVA were not met, the nonparametric Kruskal-Wallis test was performed. Individual concentrations of cytokines, chemokines, and growth factors were compared using the Mann-Whitney U test. Regression analysis was performed to test the association between infection status and infant weight growth. Statistical modeling with the least absolute shrinkage and selection operator or probabilistic latent semantic analysis was used to test whether viral load or inflammation could predict neonatal birth weight or postnatal weight change (i.e., birth weight minus body weight measured at hospital discharge). All P values were two sided, and significance was set at an α error <0.05.

Results

We included women ⩾18 years of age who were pregnant for ⩾12 weeks with a singleton gestation, reported two or more respiratory symptoms and/or tested positive for COVID-19 while pregnant, were fluent in English or Spanish, and delivered at ⩾34 weeks’ gestation in the L&D facilities at Tulane Lakeside Hospital or Ochsner Baptist Hospital. We excluded women positive for HIV, using chronic immunosuppressive medications (e.g., corticosteroids), with multiple pregnancies, or with prenatal diagnoses of any fetal anomaly. We also deliberately excluded preterm-born infants because of the concern that prematurity and secondary comorbidities could modify inflammatory signatures and clinical outcomes, introducing unacceptable noise-to-data ratio and making the interpretation of the data impossible.

Of the 287 blood samples collected, 38 were excluded because of incomplete source documents or insufficient collection volume, 12 because the MB or CB sample of the dyad was missing, 2 because they were duplicate samples, and 7 because of incomplete data (Figure 1). Thus, 209 blood samples collected in L&D from 105 mother–baby dyads were processed for ddPCR viral assays. However, two dyads were excluded from the ddPCR analysis because the samples did not yield enough cells or RNA, and for an additional two dyads, the sample volume was sufficient only for RSV but not SARS-CoV-2 amplification. Ultimately, the final analysis included complete RSV ddPCR data for 103 dyads and complete SARS-CoV-2 ddPCR data for 101 dyads. In addition, 19 MB samples collected during prenatal clinic visits were tested to assess differences with the corresponding L&D samples.

Figure 1.

Overview of blood sample collection and analysis for maternal–baby dyads included in the TTRSV (Transplacental Transmission of RSV) study. Of the 108 dyads with complete blood collections, 5 dyads were excluded from the ddPCR analysis because the samples did not yield enough cells or RNA, and for an additional 2 dyads, the sample volume was sufficient only for RSV but not SARS-CoV-2 amplification. ddPCR = droplet digital PCR; L&D = labor and delivery; RSV = respiratory syncytial virus; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

We compared mother–baby dyads with RSV (see Table E1 in the online supplement) and/or SARS-CoV-2 (see Table E2) infection detected by ddPCR in MB and/or CB (i.e., fetal) versus mother–baby dyads without any evidence of infection. We also ran comparisons among four groups on the basis of the RSV (see Table E3) or SARS-CoV-2 (see Table E4) infection status of MB and/or CB: MB⁻/CB⁻ (no virus detected: negative control), MB⁻/CB⁺ (virus detected only in CB: vertical transmission to the fetus of a maternal infection resolved before birth), MB⁺/CB⁻ (virus detected only in MB: maternal infection without vertical transmission to the fetus), and MB⁺/CB⁺ (virus detected in both MB and CB: vertical transmission of maternal infection to the fetus).

No significant differences were noted among the groups on the basis of maternal demographic characteristics (age, race, ethnicity, education, employment status, source of insurance) or clinical history (gestation, mode of delivery, neonatal sex, Apgar score), except a slight prevalence of White versus non-White subjects in the SARS-CoV-2 analysis (P < 0.05). Also, no statistically significant differences were noted in neonatal clinical outcomes from birth to discharge from the hospital (respiratory distress or symptoms, supplemental oxygen requirement, neonatal ICU admission, and length of stay). Postnatal hospital stay was significantly longer for newborns with prenatal SARS-CoV-2 exposure compared with noninfected control subjects (P < 0.01), whereas no difference was found for newborns prenatally exposed to RSV.

Blood samples were drawn from a subgroup of women (n = 19) attending prenatal care clinics in their first, second, or third trimester of gestation (see Figure E1A). All other MB samples included in the final analysis were drawn after admission to L&D and paired with the respective offspring’s CB sample for virology and immunoinflammatory analysis. Also, ddPCR results from individual MB or CB samples were compared with the epidemic activity of RSV and SARS-CoV-2 (by specific variants) in the greater New Orleans area on the basis of weekly reports compiled by the Microbiology Laboratory at the Children’s Hospital of New Orleans and the chronology of dominant COVID-19 variants tracked by the Louisiana Department of Health for Orleans and Jefferson parishes (see Figure E1B). As expected from other epidemiologic studies (see Discussion), there was an inverse relationship between RSV and SARS-CoV-2 prevalence in the community, and an atypical peak of RSV cases was observed in the summer of 2021, just before the COVID-19 spike caused by the delta variant of SARS-CoV-2. Yet most RSV-positive and dual-positive CB samples were collected outside the epidemic surge. Heat maps reporting virology results for each blood sample (see Figure E1C) show that the distribution of COVID-19–positive tests generally reflected the epidemic activity in the community, and there was a similar distribution of positive tests across all trimesters of pregnancy. Also, the trimester of infection and timing of blood sampling for RSV and SARS-CoV-2 did not affect the viral copy numbers measured in MB and CB samples.

Viral RNA for RSV, SARS-CoV-2, or both was found in MB and/or CB in 61 (59.2%) dyads, whereas 42 (40.8%) dyads had no evidence of infection and were used as negative controls (see Table E5). Both RSV and SARS-CoV-2 were susceptible to being vertically transmitted in utero, as determined by ddPCR detection of viral RNA in CB samples, albeit in significantly different proportions. When vertical transmission occurred, the viral copy numbers measured for RSV (see Figure E2A) and SARS-CoV-2 (see Figure E2B) in MB and CB samples were not statistically different. Of the infected dyads, 11 (10.7%) had both RSV and SARS-CoV-2 detected by ddPCR in the same MB or CB sample (see Figure E2C).

We found positive viral ddPCR in both MB and CB samples for RSV in 11 (10.7%) dyads (Figure 2A) and for SARS-CoV-2 in 8 (7.9%) dyads (Figure 2B). Also, 15 (14.6%) RSV-infected newborns and 4 (3.9%) SARS-CoV-2–infected newborns were born to mothers without evidence of viremia at delivery, suggesting that maternal infection had occurred earlier during gestation but cleared before the blood sample was drawn in L&D. In total, more than twice as many newborns (26 [25.2%]) in our sample were vertically infected with RSV in utero compared with those with SARS-CoV-2 (12 [11.9%]) (P = 0.019). Twenty-six (25.2%) RSV-infected mothers and 15 (14.9%) SARS-CoV-2–infected mothers delivered babies without direct evidence of the infection, suggesting a lack of physical transfer of the virus. Coinfection with both RSV and SARS-CoV-2 was detected in seven (6.8%) mothers, six (5.8%) newborns, and two (1.9%) dyads (Figure 2C).

Figure 2.

Maternal–fetal transfer of respiratory syncytial virus (RSV) and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) copies. (A) RSV copy numbers measured using droplet digital PCR (ddPCR) in peripheral blood mononuclear cell (PBMCs) separated from maternal blood drawn in prenatal clinic or labor and delivery (L/D) versus cord blood. Data from the 103 dyads with complete RSV ddPCR data are shown. Dotted lines represent 15 cord blood samples with virus-positive ddPCR born to virus-negative mothers. (B) SARS-CoV-2 copy numbers measured using ddPCR in PBMCs separated from maternal blood drawn in prenatal clinic or L/D versus cord blood. Data from the 101 dyads with complete SARS-CoV-2 ddPCR data are shown. Dotted lines represent four cord blood samples with virus-positive ddPCR born to virus-negative mothers. (C) RSV or SARS-CoV-2 copy numbers measured using ddPCR in maternal and cord blood from the 11 dyads with at least one sample with dual-positive viral detection. Blue symbols represent samples testing positive for both RSV and SARS-CoV-2.

A significant increase in the plasma concentration of several inflammatory mediators was measured in CB samples obtained from newborns with evidence of viral exposure in utero (i.e., viremia detected in MB and/or CB samples), compared with control dyads with no detectable viremia in either MB or CB (see Figure E3A). The most significant increase was measured for a set of potent markers of monocyte/macrophage activation, such as soluble CD163 (cluster of differentiation 163), as well as immune cells chemoattractants such as PF4 (platelet factor 4; CXCL4 [C-X-C motif chemokine ligand 4]), ALCAM (activated leukocyte cell adhesion molecule), I-TAC (IFN-inducible T-cell α chemoattractant; CXCL11 [C-X-C motif chemokine ligand 11]), and DR6 (death receptor 6). Furthermore, significant differences in the expression patterns of inflammatory mediators were detected among CB samples (see Figure E3B) and MB samples (see Figure E3C) depending on the presence of RSV, SARS-CoV-2, or both viruses.

Interestingly, in virus-positive dyads, newborns with evidence of physical transmission of either virus in utero (i.e., CB ddPCR positive) tended to have higher MB concentrations of proinflammatory cytokines such as IL-11, IL-17, and eotaxin (Figures 3A, 3B, and E5), combined with lower concentrations of DKK-3 (Dickkopf-related protein 3) (a negative modulator of T-cell responses), GNLY (granulysin), and Siglec-5 (sialic acid–binding immunoglobulin-type lectin 5) (an inhibitory T-cell immune checkpoint). Similar patterns were noted when maternal samples from RSV-infected (Figure 3C) or SARS-CoV-2–infected (Figure 3D) dyads were analyzed separately. RSV-positive CB exhibited higher maternal expression of IL-6 in addition to IL-11 and IL-17, as well as lower expression of the homeostatic hormone adiponectin and SCF-R (stem cell factor receptor; c-Kit, CD117). SARS-CoV-2–positive CB exhibited higher maternal expression of T-helper cell type 1–type cytokines such as IFN-γ and TSLP (thymic stromal lymphopoietin) that promotes T-helper cell type 2–type immune responses, as well as lower expression of the other homeostatic hormone leptin. Several of these soluble factors were also found differentially expressed in CB by vertical transmission status (see Figures E4 and E5), including granulysin, DKK-3, adiponectin, and IL-6R (IL-6 receptor).

Figure 3.

Comparison of soluble mediators’ concentrations in maternal blood (MB) from dyads with vertical transmission of virus versus dyads without vertical transmission. (A) Volcano plot showing a comparison of soluble mediators’ concentrations between MB samples from dyads with vertical transmission of virus (i.e., cord blood [CB] positive for any virus [n = 36], respiratory syncytial virus [RSV] [n = 27], or severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2] [n = 13]) versus MB samples from dyads without vertical transmission of virus (i.e., only maternal sample positive for any virus [n = 35], RSV [n = 31], and SARS-CoV-2 [n = 18]). For this analysis, we excluded any dyad that had no detectable viremia. Statistical analysis was performed using Mann-Whitney comparison of individual cytokines, with the significance threshold of P < 0.05 indicated by dotted lines. Adjusted comparisons were not significant. (B) Expression of soluble mediators with the highest expression in maternal samples from dyads with (MB^+/−CB⁺) versus without (MB⁺CB⁻) vertical transmission for any virus, by box-and-whisker plots (minimum to maximum) showing all subjects (each comparison was significant at P < 0.05 using Mann-Whitney tests). *P < 0.05, **P < 0.01. (C) Expression of soluble mediators with the highest expression in maternal samples from dyads with (MB^+/−CB⁺) versus without (MB⁺CB⁻) vertical transmission for RSV, by box-and-whisker plots (minimum to maximum) showing all subjects (each comparison was significant at P < 0.05 using Mann-Whitney tests). *P < 0.05, **P < 0.01. (D) Expression of soluble mediators with the highest expression in maternal samples from dyads with (MB^+/−CB⁺) versus without (MB⁺CB⁻) vertical transmission for SARS-CoV-2, by box-and-whisker plots (minimum to maximum) showing all subjects (each comparison was significant at P < 0.05 using Mann-Whitney tests). *P < 0.05, **P < 0.01. CD = cluster of differentiation; diff. = difference; DKK-3 = Dickkopf-related protein 3; SCF-R = stem cell factor receptor; SARS = severe acute respiratory syndrome coronavirus 2; Siglec-5 = sialic acid–binding immunoglobulin-type lectin 5; TNFα = tumor necrosis factor-α; TSLP = thymic stromal lymphopoietin; VEGF = vascular endothelial growth factor.

Prenatal viral infection and the consequent inflammatory response significantly affected newborns’ weight gain. Specifically, statistical modeling revealed that CB concentrations of inflammatory mediators such as CXCL5 (P < 0.01) and MIP-1β (macrophage inflammatory protein; P < 0.05), as well as metabolic hormones such as leptin (P < 0.05) and Pref-1 (preadipocyte factor 1; P < 0.01) independently predicted birth weight (Figure 4). Also, CB concentrations of SCF-R (P < 0.01), VEGF (vascular endothelial growth factor; P < 0.01), IL-1RII (IL-1 receptor type 2; P < 0.05), the acute-phase reactant CRP (C-reactive protein; P < 0.05), and RSV copy number (P < 0.01) predicted neonatal body weight loss before discharge from the hospital (Figure 5). Furthermore, regression analysis (see Table E6) showed a significant association between postnatal weight change and RSV copy numbers in MB (P < 0.05), as well as SARS-CoV-2 titers in both MB (P < 0.01) and CB (P < 0.05). To visualize these complex interactions, Figure E6 shows a multivariate correlation analysis among weight growth, viral copy number, and the soluble inflammatory markers tested in this study.

Figure 4.

Relationship between cord blood inflammatory mediators and birth weight. (A) Overview of feature selection model used to identify viral or cytokine changes that predict neonatal birth weight or change in weight at hospital discharge using the least absolute shrinkage and selection operator and probabilistic latent semantic analysis (PLSA). (B) PLSA for birth weight, including variable importance (VIP) score, model coefficient (for VIP score > 0.8), and predicted versus actual model fit. (C) Feature-selected soluble mediators (CXCL5, MIP-1β, leptin, preadipocyte factor 1) that were individually significant by the line of fit for weight at birth, with R² and P values indicated. CXCL5 = C-X-C motif chemokine ligand 5; MB = maternal blood; MIP-1β = macrophage inflammatory protein; RSV = respiratory syncytial virus; SARS = severe acute respiratory syndrome coronavirus 2.

Figure 5.

Relationship between cord blood inflammatory mediators and viral copy numbers versus weight change from birth to hospital discharge. (A) Results of probabilistic latent semantic analysis for change in weight, including variable importance (VIP) score, model coefficient (for VIP score > 0.8), and predicted versus actual model fit. (B) Feature-selected soluble mediators (VEGF, SCF-R, IL-1 RII, and CRP) that were individually significant by the line of fit for neonatal weight change from birth to hospital discharge, with R² and P values indicated. (C) Correlation between RSV and SARS-CoV-2 cord and maternal blood viral copy numbers versus neonatal weight change from birth to hospital discharge, with the line of fit R² and P values indicated. CB = cord blood; CRP = C-reactive protein; IL-1 RII = IL-1 receptor II; RSV = respiratory syncytial virus; SARS = severe acute respiratory syndrome coronavirus 2; SCF-R = stem cell factor receptor; VEGF = vascular endothelial growth factor.

Discussion

This study confirms that maternal viral respiratory infections during pregnancy can be vertically transmitted to the fetus and detected in fetal blood at birth. This result is consistent with previous findings in animal models (8, 9) and small human samples reported by our laboratory and others (10, 11) and highlights the potential importance of even subtle events occurring in pregnant women (e.g., apparently inconsequential head colds) for the future development and well-being of the offspring. In addition, important new information emerging from the data presented here is that different respiratory viruses cross the placenta with different frequencies. Specifically, maternal–fetal transfer of RSV infection in our birth cohort (1 in 4 newborns) was much more frequent than that of SARS-CoV-2 (1 in 10 newborns), which is consistent with the limited incidence and severity of perinatal COVID-19 (12).

It should be noted that the rate of RSV vertical transmission in this cohort was lower than in previous studies using a similar methodology (11), possibly because of the unusually low prevalence and atypical seasonality of RSV in the community during the COVID-19 pandemic (13). Indeed, according to data collected by the CDC, no typical winter RSV epidemic occurred in the United States during 2020 and 2021, possibly because of pandemic mitigation measures, and the atypical 2021–2022 epidemic began in May and lasted until January. Furthermore, the 2022 season started in June (i.e., after the end of recruitment for our cohort) and peaked in November. Notably, 2022–2023 data show a global resurgence of RSV infections with record-breaking hospitalization rates, suggesting that children and pregnant mothers without repeated exposure to RSV experience waning immunity, which predisposes them to more severe disease (14). Collectively, these observations imply that the frequency of RSV maternal–fetal transmission may vary in time and space depending on the current prevalence of the virus in the local community.

When RSV and/or SARS-CoV-2 crossed the placenta, similar copy numbers of each virus were detected in CB and MB. Also, the timing of maternal upper respiratory infection symptoms or blood sampling across the three trimesters of gestation did not translate into significant differences in viral load, frequency of vertical transmission, or clinical outcomes. Interestingly, most of the CB samples with positive viral PCR were collected outside of the epidemic peaks reported by the local microbiology laboratory, suggesting that although the clinical cases of viral infection were low in the local community, the causative viruses were still present and being passed from infected mothers to their offspring. Finally, simultaneous vertical transmission of both RSV and SARS-CoV-2 is possible, albeit relatively uncommon, as it was detected in approximately 1 in 20 newborns in our cohort.

Prenatal exposure to respiratory viruses resulted in robust immune activation and inflammation in the fetus and newborn, as shown by significant upregulation of proinflammatory cytokines and chemokines with potent chemoattracting activity on monocytes/macrophages, neutrophils, and T cells. However, a limitation of our approach was that we did not create families of soluble factors to adjust for multiple comparisons within the 440 analytes. A similar inflammatory signature was noted not only after the physical transplacental passage of viral particles but also in the context of maternal immune activation without measurable virus transfer. Of particular relevance, the high IL-6 concentrations measured in CB have been widely used to diagnose fetal inflammatory responses, chorioamnionitis, and early-onset neonatal infections (15, 16). Also, both PF4 (CXCL4) and DR6 (a member of the TNF receptor superfamily) have independently been observed in cellular responses to RSV, and the former has been proposed as a marker of clinical severity (17, 18).

The antimicrobial and cytolytic peptide granulysin seemed to exert a protective effect against vertical virus transmission, whereas the proinflammatory and neutrophil-recruiting cytokine IL-17 seemed to facilitate it. Importantly, different inflammatory and immune activation patterns were observed in the mother and the newborn depending on whether the causative viral pathogen was RSV or SARS-CoV-2. On the basis of these data, we speculate that the interplay between the infecting virus and the mother’s immunoinflammatory response to the invading pathogen is a major determinant of the frequency and severity of vertical transmission to the offspring. Hence, active maternal immunization with specific vaccines holds promise not only as a protective measure against postnatal infections but also as a preventive strategy against prenatal exposures.

Our SARS-CoV-2 data add to previous studies reporting CB cytokine analysis in much smaller maternal–infant samples infected by the same virus (19, 20). The cytokine signatures found in our study show several similarities but also some differences that may reflect the timing of infection during pregnancy, the acuity of COVID-19 at the time of maternal testing, the dominant viral variant, the multiplex panels adopted, and several other variables. Overall, however, these studies collectively support the hypothesis that a respiratory viral infection during pregnancy is associated with robust expression of inflammatory cytokines measurable in fetal blood. This fetal inflammation may result in part from a placental immune reaction to the virus or passive transfer of maternal cytokines to the fetal circulation, but the different expression of many cytokines observed in MB compared with CB suggests an independent response of the developing fetal immune system to the infection transmitted from the mother.

Nevertheless, additional studies are warranted to identify the exact source of intrauterine inflammation in the context of maternal respiratory infections, as this distinction will be important to understand the potential impact on the long-term health of the offspring and guide future prophylactic and therapeutic strategies. In particular, special attention should be directed toward the placenta, as we have previously shown that RSV infects migratory dendritic cells (Hofbauer cells) resident in the human placenta, which are then able to transfer the virus through a contact-dependent mechanism to fetal airway epithelial cells (21). Moreover, Hofbauer cells support long-term infection with this virus and produce and secrete proinflammatory T-helper cell type 1 cytokines such as TNF-α (tumor necrosis factor-α), IFN-γ, and IL-12, as well as IL-6.

Intrauterine exposure to respiratory viruses and the ensuing immunoinflammatory response had clinical consequences already in the immediate neonatal period, as shown by the significant correlation of viral loads and multiple inflammatory mediators with birth weight and postnatal weight change. This finding suggests that virus-induced perinatal inflammation does alter postnatal development and metabolism and may constitute an early origin of pathological outcomes presenting later in childhood and possibly adulthood. However, the present study was deliberately designed to analyze the pregnancy and immediate postpartum up to discharge of the infant from the nursery, and we did not detect immediate adverse effects on the newborn’s clinical outcomes or respiratory status. Prospective follow-up of the cohort established by this study has been planned to continue until school entry and will provide critical information to understand the long-term consequences of respiratory viral infections transmitted from mother to offspring during intrauterine life. In particular, it will be important to determine whether perinatal virus-induced inflammation can predispose to chronic respiratory pathology such as recurrent wheezing and asthma or more subtle functional abnormalities in lung flows and volumes during childhood.

Among the potential limitations of our study is the sample size. However, the TTRSV cohort exceeded the initial recruitment goal according to the power analysis of preliminary data. Furthermore, a careful search of all major literature databases found no evidence of any other prospective cohort looking at the vertical transmission of RSV, except for a study done in Bangladesh on 149 mother–infant dyads but focusing only on antibody transfer measured in samples collected during the last trimester of pregnancy (22). Similarly, the few prospective cohorts looking at vertical transmission of SARS-CoV-2 were all significantly smaller than ours. Thus, we believe the sample presented in this paper is currently the largest birth cohort looking prospectively at the vertical transmission of respiratory viruses in mother–infant dyads.

Another potential limitation is the unlikely possibility of false positives due to contamination of CB samples during collection in the L&D suite. Although this risk cannot be ruled out completely, it is deemed minimal because the venipunctures of the umbilical cord and placental vessels were performed immediately after delivery by highly experienced and trained nurses following a strictly standardized sterile protocol. Also, the robust inflammatory response shown by our data would not have been present in the case of limited microbial exposure occurring immediately before sample collection. Likewise, false negatives cannot be ruled out completely because, although the diagnosis of infection was based on the detection of active viremia, tissue infection could have been present with subthreshold viral copy count in circulating blood. However, given the extremely high sensitivity of ddPCR, this possibility would imply negligible viral loads unlikely to have any significant biological effect, and in any case, it would have been technically and ethically impossible to detect the infection in peripheral tissues of live newborns after hematogenous transmission.

Concerning the clinical outcomes, some of the blood samples collected during the study had to be excluded from the final analysis because of incomplete collection or information, which might have introduced bias, and we excluded premature-born infants, which could have influenced weight at discharge and length of hospital stay. Finally, as the study was conducted during the COVID-19 pandemic, longer postnatal stays for offspring of mothers positive for SARS-CoV-2 could have been due to isolation practices and exposure precautions, rather than actual illnesses of the infants.

Overall, our data provide new insight into the growing evidence that prenatal exposure to respiratory viruses can profoundly affect immune development in early life and predispose to increased morbidity later in life. These concepts are a logical application to the respiratory system of the paradigm known as developmental origins of health and disease (23) and highlight the critical role that infection prevention by active or passive prophylaxis can play during pregnancy and infancy, potentially protecting respiratory health later in life. This topic is especially timely because the first ever vaccine for maternal immunization and a new long-acting monoclonal antibody against RSV were recently approved by the U.S. Food and Drug Administration to protect infants from birth throughout the first years of life (24, 25). The imminent availability of these long-sought interventions to prevent or decrease the severity of the initial RSV infections, together with the relentless evolution of COVID-19 vaccines, promises to usher in a new era when the growing prevalence of pediatric respiratory disease can be significantly reduced at the population level by primary prophylaxis.

Conclusions

This study is the first to compare the vertical transmission of RSV and SARS-CoV-2 from pregnant women with acute respiratory infections to their unborn babies, showing that RSV has a much higher transmission rate. It is also the first to show that viral copy numbers in fetal blood are similar to those measured in the respective mothers, that transmission can occur in any trimester of gestation, and that cotransmission of both viruses is possible, albeit relatively uncommon. Perhaps more important, a robust and virus-specific cytokine- and chemokine-mediated inflammatory response was measured in the CB of newborns exposed to respiratory pathogens in utero, even in the absence of physical transfer of viral particles, and the pattern of immunoinflammatory activation seemed to affect the rate of vertical transmission. Finally, our data suggest that virus-induced inflammation may have had pathological consequences already in the first days of life, as shown by the reported effects on birth weight and postnatal weight loss.