Maternal Inflammation Likely Drives Impaired Immune Responses to Respiratory Syncytial Virus in HIV-Exposed Uninfected Infants

Abstract

Background

Respiratory syncytial virus (RSV) is the leading cause of lower respiratory tract infection and a major contributor to morbidity and mortality in HIV-exposed, uninfected (HEU) infants. The mechanisms underlying HEU infants’ increased susceptibility to RSV are unclear.

Methods

We recruited pregnant women with and without HIV, plus HEU and HIV-unexposed (HUU) children at 12–18 months of age and collected peripheral and cord blood. We measured innate immune responses of infants to RSV using an in vitro model of human respiratory infection. These responses were correlated with markers of inflammation in maternal and cord blood. We also incubated HUU cord blood cells in plasma from women living with HIV (WLHIV) to recapitulate the RSV-specific responses in HEU cord blood cells.

Results

We enrolled 30 WLHIV, 61 HIV-negative women, 19 HEU and 20 HUU children. At birth, HEU infants demonstrated lower expression of IL-12 by dendritic cells (P < .0001) and IFNγ by natural killer (NK) cells (P = .007); the difference in IL-12 expression persisted to 12–18 months of age (P = .015). WLHIV had high concentrations of multiple inflammatory molecules in peripheral blood; these correlated inversely with infant RSV-specific immune responses. Incubation of cord blood cells from HUU infants in maternal plasma from WLHIV significantly lowered RSV-specific NK cytotoxicity and antigen-presenting cell activation compared to incubation in HIV-negative maternal plasma.

Conclusions

Maternal inflammation is a likely driver of innate immune dysregulation in HEU infants and predisposes to an increased susceptibility to RSV infection.

Among HIV-exposed, uninfected infants, immune responses to RSV are impaired in an in vitro model. The RSV responses appear to be driven by maternal inflammation and can be partially reproduced by incubation of cord blood cells in plasma from women with HIV.

(See the Editorial Commentary by Cohen and Moore on pages 417–9.)

Each year, >3500 infants in the United States and >1 million worldwide are born to women living with HIV-1 (WLHIV) [1–3]. As a result of successful perinatal HIV prevention programs, the vast majority of these infants are born HIV-exposed, but uninfected (HEU) [2, 4]. Despite not acquiring HIV, HEU infants experience high rates of morbidity and mortality, primarily from infections with common childhood pathogens [5–13]. Respiratory syncytial virus (RSV) is the leading cause of lower respiratory tract infection and a major contributor to morbidity and mortality in HEU infants, especially in the first 6 months of life [5–10, 14]. Among infants with RSV, HIV exposure is associated with an increased incidence of hospitalization, longer duration of hospitalization, and increased incidence of complications and deaths [6–10, 14].

Infants rely heavily on innate immune defenses for protection from RSV, as their adaptive immune system is antigen-inexperienced and biased toward tolerance [15, 16]. The early immune response to RSV is highly dependent upon a pathway that requires signaling from antigen-presenting cells (APCs) such as monocytes and dendritic cells (DCs), which subsequently triggers recruitment, activation, and cytotoxic responses from natural killer (NK) cells [16–18]. We and others have demonstrated altered innate immunity in HEU infants, including increased activation and exhaustion of APCs and increased basal activation of NK cells, associated with decreased cytotoxic responses to stimulation [7, 19–25]. However, these changes have not previously been studied in the setting of RSV infection.

The mechanisms driving immune dysfunction in HEU infants are uncertain. HIV infection is characterized by excess immune activation and inflammation, even in the setting of viral suppression with antiretrovirals, and this phenomenon has also been demonstrated during pregnancy [21, 26–30]. Multiple other inflammatory maternal conditions, including obesity, type 1 diabetes, and infections such as hepatitis B, SARS-CoV-2, and malaria, are known to impact fetal immune development, although the mechanism is unclear [31–37]. We hypothesized that in utero exposure to the inflammatory environment associated with maternal HIV infection imprints HEU fetal immune cells during a critical developmental stage, subsequently attenuating their immune responses to RSV infection.

METHODS

Study Population

We enrolled pregnant women and children from 3 health systems in Denver, Colorado: Children's Hospital Colorado, Denver Health and Hospital Authority, and University of Colorado Health. All participants were enrolled between August 2020 and June 2023, prior to the availability of vaccines or the monoclonal antibody nirsevimab to prevent RSV infection. Pregnant women with known HIV status were enrolled into the Birth Cohort in the 2nd trimester of pregnancy. See eligibility criteria, Supplementary Table 1. Samples from participants who subsequently developed exclusion criteria were excluded. Maternal blood was collected at entry and cord blood was collected at delivery.

Children with known HIV exposure status were enrolled into the 12–18 Month Cohort. Infants enrolled in the Birth Cohort were not subsequently eligible for the 12–18 Month Cohort (Supplementary Table 1). Peripheral blood was collected from children on the day of enrollment. Anti-RSV IgG ELISA (IBL America, IB79887) was performed on plasma in real time and children seropositive for RSV were excluded.

All HEU infants and children enrolled in this study were determined to be HIV-uninfected according to standard definitions [38]. WLHIV did not breastfeed their infants. Infant feeding method affects the gut microbiome, which plays a role in infant immune development [39]. However, the microbiome of breastfed and formula-fed infants converges after breastfeeding cessation. Therefore, we excluded HUU children from the 12 to 18 Month Cohort who breastfed beyond 6 months. Exclusion criteria based on feeding method were irrelevant to the Birth Cohort, because cord blood samples were collected prior to the infants’ first feeding.

This study was approved by the Colorado Multiple Institutional Review Board (protocol #19-1911) and all participants or caregivers provided written informed consent. Pregnant participants signed consent for themselves and their unborn child. Research was conducted in accordance with the ethical standards of the Helsinki Declaration.

Cytokine/Chemokine Measurement in Plasma

Cord or peripheral blood was processed within 8 hours of collection and plasma was aliquoted and cryopreserved at −80°C until analysis. Cytokine and chemokine concentrations were measured using custom multiplex electrochemiluminescence microarrays (Meso Scale Diagnostics).

RSV Infection Coculture Model

To test immune responses to RSV, we used an in vitro model of human respiratory infection that incorporated A549 human alveolar epithelial cell cultures (ATCC) infected with RSV originally isolated from a clinical sample, as previously described [40]. We isolated peripheral or cord blood mononuclear cells (PBMCs/CBMCs) from whole blood using ficoll-histopaque (Sigma) gradient centrifugation and cryopreserved the cells in FBS containing 10% Dimethyl Sulfoxide (Sigma). Cells were stored at −150°C until analysis. After thawing, PBMCs/CBMCs were washed and resuspended in Xvivo^TM media containing gentamicin (Lonza). Cell counts and viability were obtained using a Guava easyCyte^TM 5HT System (Cytek Biosciences) and samples with a minimum of 80% viability and 3E6 viable cells were tested. PBMCs/CBMCs at 1.5 × 10⁶ cells in 500 µL were added to paired wells in 24-well plates containing A549 cultures infected with RSV or mock-infected. Cocultures were incubated at 37°C, 5% CO₂ for 20 hours.

Treatment of CBMCs With Maternal Plasma

Thawed CBMCs from each HUU infant were divided into 8 aliquots of 4E6 cells and each aliquot resuspended in RPMI 1640 with L-glutamine (Corning) supplemented with 10% plasma from either a WLHIV or an HIV-negative (HIVneg) mother and 1% penicillin–streptomycin (GeminiBio). CBMCs were incubated at 37°C, 5% CO₂ for 48 hours after which cells were resuspended in Xvivo^TM and used in the RSV coculture model.

Flow Cytometry

For cell cultures intended for flow cytometry, A549s were resuspended in phosphate-buffered saline (PBS) at 10⁶ cells/mL and stained using a 1:10 dilution of Cell Trace Violet^TM (Invitrogen) for 20 minutes at room temperature. After incubation, cells were washed, resuspended in MEM with 10% FBS, seeded into 24-well plates, RSV-infected or mock-infected, and cocultured with PBMCs/CBMCs as described above.

For the final 4 hours of incubation, 5 µL of CD107a-BV711 (Biolegend) and 5 µg/mL of Brefeldin A (Sigma) and Monensin (BD Biosciences) were added to each well.

At the end of incubation, PBMCs/CBMCs and supernatants were harvested via pipette aspiration. Accutase (Millipore) was used per manufacturer's instructions to collect adherent cells. Cells were washed, resuspended in PBS and transferred to 96-well round-bottom polypropylene plates for staining.

Cells were stained with the live/dead discriminator Zombie Aqua (Biolegend) prior to being incubated with an Fc Receptor block (Biolegend). Cells were stained extracellularly with monoclonal antibodies (mAbs) against the following: CD3-, CD19-, CD20-BV605; CD56-BV570; CD14-Alexa 700; CD16-PE-Cy5; HLADR-PE; CD83-PE-Cy7; (Biolegend); and PD-L1-BV650 (BD Horizon). Cells were fixed and permeabilized with FACS lysing and permeabilizing solutions (BD) and stained for intracellular markers with mAbs against the following: RSV-FITC (Invitrogen); IL-12-APC; (BD Pharmingen); Perforin-APC-Cy7; IFNγ-BV786; Granzyme B-PE-CF594. Samples were acquired using a Novocyte Quanteon 4025^TM (Agilent) flow cytometer and analyzed using FlowJo version 10.10 software (BD). See gating strategy, Supplementary Figure 1. RSV knockdown, defined as the proportion of RSV-infected A549s when cocultured with PBMCs/CBMCs, subtracted from the proportion of RSV-infected A549s when cultured alone, was measured in each well as a marker of the ability of PBMCs/CBMCs to control RSV replication. Markers of activation, maturation, and degranulation in PBMCs/CBMCs were reported as fold-change in the proportion of parent population expressing the marker or in median fluorescence intensity (MFI) between RSV-infected and mock-infected cultures.

Antibody-dependent Cellular Cytotoxicity (ADCC)

ADCC assays were prepared using supernatant from A549 cultures that were either RSV-infected or mock-infected. Supernatant was UV-inactivated on ice for 30 minutes, aliquoted and stored at −80°C. UV-inactivated supernatant was added to flat bottom 96 well polystyrene culture plates (Corning) at a 1:10 dilution in PBS (Corning) and incubated for 20 hours at 4°C on a plate shaker at 500 RPM. The plate was washed with wash buffer (PBS with 0.05% Tween-20) and blocked with wash buffer containing 0.05% w/v powdered milk (Carnation), 20% FBS and 5% bovine serum albumin (MilliporeSigma) for 2 hours at 37°C on a plate shaker at 500 RPM. Plasma from a laboratory control with high anti-RSV antibodies was added at 100 μL/well at the preoptimized dilution of 1:10 in blocking solution. The plate was incubated at 37°C on a plate shaker for 2 hours at 500 RPM. We thawed CBMCs and isolated NK cells using the EasySep™ Human NK Cell Isolation Kit (STEMCELL) according to manufacturer instructions. NK cells from samples containing ≥200 000 NK cells were added to the ADCC plate at a preoptimized concentration of 100 000 cells/well in 200 uL Xvivo™ and incubated for 20 hours at 37°C. For the final 4 hours of incubation, Brefeldin A, Monensin, and CD107a-BV711 were added as above. At the end of incubation, cells were washed and stained for flow cytometry.

Statistical Analysis

Statistical analyses were performed in GraphPad Prism, version 10.4.2, or R version 4.3.1. P < .05 defined significance. A priori power analyses established that 20 infants per group provided a power of ≥0.85 to detect significant differences. Demographic and clinical characteristics were compared between HEU and HUU infants using χ², Fisher's exact, or Mann–Whitney tests, as appropriate. In experiments that measured the expression of markers of activation, maturation, or degranulation in CBMCs/PBMCs after exposure to RSV-infected versus mock-infected respiratory epithelial cells, we used Mann–Whitney tests to compare expression between HEU versus HUU, or between HUU CBMCs incubated with maternal plasma from WLHIV versus HIVneg mothers. We used Mann–Whitney tests to compare concentrations of plasma cytokines and chemokines between WLHIV versus HIVneg mothers, and between HEU versus HUU infants. We corrected for multiple comparisons using the Bonferroni method. Among participants with a paired maternal and cord blood sample, we identified correlations between the cytokine/chemokine concentrations in maternal or cord plasma and infant cellular RSV responses at birth using Spearman's correlation coefficients. Subsequently, we used linear regression to model the relationship between the cytokine/chemokine concentrations in maternal or cord plasma and infant cellular RSV responses at birth, adjusting for HIV exposure status of the infant.

RESULTS

The birth cohort enrolled 30 WLHIV and 61 HIVneg women, among whom a cord blood sample was collected from 23 HEU and 61 HUU infants (Table 1 and Figure 1). Mothers with and without HIV were well-matched for age, but there were more White mothers in the HIVneg group, and more Hispanic mothers in the WLHIV. HIVneg mothers were more likely to have a hypertensive disorder, but WLHIV were more likely to use alcohol or tobacco during pregnancy. WLHIV had well-controlled HIV, with the median CD4 count >500 cells/mm³ at entry and delivery, and >90% of viral loads at delivery <20 copies/mL.

Table 1.

Maternal and Infant Demographic and Clinical Characteristics, Birth Cohort

Maternal Characteristics Median (IQR) or count (%)	WLHIV (n = 30)	HIV-negative (n = 61)	P-value
Maternal age (y)	29.5 (26–32)	30 (26–34)	.70
Maternal race/ethnicity			.005
White non-Hispanic	4 (13.3%)	24 (39.3%)
Black non-Hispanic	12 (40.0%)	28 (45.9%)
Hispanic	13 (43.3%)	8 (13.1%)
Unknown/Other	1 (3.3%)	1 (1.6%)
Gestational age on date of maternal blood draw (wks)a	25.1 (20.0–31.9)	28.3 (27.0–29.3)	.07
Obesity	11 (36.7%)	12 (19.7%)	.12
Hypertensive disorders (gestational or pre-gestational)	0 (0%)	8 (13.1%)	.049
Diabetes (gestational or pre-gestational)	5 (16.7%)	5 (8.2%)	.29
Substance use during pregnancy
Tobacco	5 (16.7%)	2 (3.3%)	.04
Alcohol	3 (10.0%)	0 (0%)	.03
Marijuana	2 (6.7%)	1 (1.6%)	.25
Other	1 (3.3%)	0 (0%)	.33
CD4 count at entry (cells/mm3)	570 (353–830)	–	–
CD4 count at delivery (cells/mm3)	580 (425–743)	–	–
Viral load at entry (copies/mL)	38 (0–2698)	–	–
Viral load <20 copies/mL at entry	13 (43.3%)	–	–
Viral load at delivery (copies/mL)	0 (0–10)	–	–
Viral load <20 copies/mL at delivery	28 (93.3%)	–	–

Infant characteristics Median (IQR) or count (%)	HEU (n = 23)	HUU (n = 61)	P-value
Mechanism of delivery
Vaginal	18 (78.3%)	48 (78.7%)	>.99
Cesarean	5 (21.7%)	13 (21.3%)
Gestational age at delivery (wks)	39.1 (38.8–39.4)	39.1 (38.1–39.6)	.80
Infant sex, male	11 (47.8%)	31 (50.8%)	.13
Infant birth weight, grams	3290 (3050–3528)	3235 (2955–3610)	.93

Abbreviations: HEU, infant who is HIV-exposed but uninfected; HUU, infant who is HIV-unexposed; IQR, interquartile range; WLHIV, women living with HIV.

^a N = 27 WLHIV and N = 60 HIV-negative mothers. Mothers whose plasma was not included in analyses were excluded from this calculation.

Figure 1.

Flowchart of the study population. The number of study participants included in each experiment is shown. The number of samples included in each experiment was determined by a priori power calculations and sample availability. Abbreviations: CBMC, cord blood mononuclear cells; HEU, HIV-exposed uninfected; HIVneg, HIV negative; HUU, HIV-unexposed uninfected; RSV, respiratory syncytial virus; WLHIV, women living with HIV.

The 12–18-month cohort enrolled 19 HEU and 20 HUU children (Table 2 and Figure 1). The groups were well-matched for age and sex, but there were more Black children in the HEU group, and more Hispanic children in the HUU group. None of the HEU children and 65% of the HUU children were breastfed for <6 months. Children enrolled were healthy; the most common comorbid condition reported in both groups was atopic disease (asthma, allergies, or eczema), but there was no difference in the proportions of children with these diagnoses between groups.

Table 2.

Child Demographic and Clinical Characteristics, 12–18 Month Cohort

Child Characteristics Median (IQR) or Count (%)	HEU (n = 19)	HUU (n = 20)	P-value
Child age (months) at enrollment	14 (13.5–16)	15 (15–15)	.23
Child race/ethnicity
White non-Hispanic	3 (15.8%)	1 (5.0%)	.02
Black non-Hispanic	10 (52.6%)	3 (15.0%)
Hispanic	4 (21.1%)	12 (60.0%)
Unknown/other	2 (10.5%)	4 (20.0)
Infant sex, male	10 (52.6%)	12 (60%)	.75
Preterm delivery (<37 wks’ gestation)	4 (21.1%)	1 (5.0%)	.18
Breastfeeding
Never	19 (100%)	7 (35%)	<.0001
1–6 m	0 (0%)	13 (65%)
Atopic disease (asthma/allergies/eczema)	5 (26.3%)	8 (40%)	.50

Abbreviations: HEU, infant who is HIV-exposed but uninfected; HUU, infant who is HIV-unexposed; IQR, interquartile range.

Diminished Immune Responses to RSV in HEU Infants Exist at Birth and Persist to 12–18 Months of Age

We compared innate immune responses to RSV between 23 HEU versus 30 HUU infants at birth and between 19 HEU versus 20 HUU RSV-seronegative children at 12–18 months of age, using our respiratory infection model [40]. At birth, CBMCs from HEU infants demonstrated lower frequencies of IL-12-expressing DCs than HUU infants [median 1.7-fold increase (IQR 1.6–1.9) versus 2.0-fold increase (IQR 1.9–2.4) between RSV- and RSV + cultures, P < .0001] and of IFNγ-expressing NK cells [median 14-fold increase (IQR 8–54) versus 37-fold increase (IQR 24–67), P = .007] (Figure 2A). Among PBMCs from children 12–18 months of age, the difference in IL-12-expressing DCs persisted, but was attenuated [median 1.4-fold increase (IQR 1.4–1.6) versus 1.6-fold increase (IQR 1.5–1.8) P = .015], and the difference in IFNγ-expressing NK cells was no longer apparent (Figure 2A). We did not identify any other differences between HEU and HUU infants in the expression of activation or degranulation markers in NK cells, activation or maturation markers in DCs or monocytes, nor RSV knockdown (Figure 2A and 2B).

Figure 2.

Anti-RSV responses among HIV-exposed uninfected (HEU) versus -unexposed (HUU) infants. A, Fold-change in expression of markers of activation, maturation, and degranulation among cord or peripheral blood mononuclear cells from HEU or HUU infants after exposure to RSV-infected versus mock-infected respiratory epithelial cells. B, RSV knockdown, defined as the proportion of RSV-infected respiratory epithelial cells when cocultured with PBMCs/CBMCs, subtracted from the proportion of RSV-infected A549s when cultured alone, in HEU versus HUU infants. Experiments at birth were performed in N = 23 HEU and N = 30 HUU infants; experiments at 12–18 m were performed in N = 19 HEU and N = 20 HUU children. C, Results of antibody-dependent cellular cytotoxicity (ADCC) experiments: fold-change in expression of CD107a among purified natural killer (NK) cells at birth after exposure to immobilized RSV antigen/antibody complexes versus mock antigen/antibody complexes. Data are from N = 8 HEU and N = 8 HUU infants. Significant differences are indicated by asterisks, where * represents P = .01 to .05, ** P = .001 to .01, *** P = .0001 to .001, and **** P < .0001. Abbreviations: CBMCs, cord blood mononuclear cells; DCs, dendritic cells; MFI, median fluorescence intensity; NK, natural killer; PBMCs, peripheral blood mononuclear cells; RSV, respiratory syncytial virus.

To determine whether HIV exposure had an independent effect on NK cell function (in the absence of signaling from APCs such as DCs), we tested antibody-dependent cellular cytotoxicity (ADCC) using purified NK cells from 8 HEU versus 8 HUU infants at birth and anti-RSV antibodies bound to immobilized RSV antigen. We found that NK cells from HEU infants expressed less CD107a than HUU infants in the ADCC model [median 1.6-fold increase (IQR 1.1–1.8) versus 1.0-fold increase (IQR 0.8–1.4) between RSV- and RSV + conditions, P = .03] (Figure 2C).

HIV in Pregnancy is Associated With Peripheral Inflammation

To explore our hypothesis that maternal inflammation drives the dysregulated immune responses to RSV among HEU infants, we measured the concentrations of inflammatory and regulatory cytokines and chemokines in the plasma of WLHIV (N = 27) and HIVneg women (N = 60) in the 2nd trimester of pregnancy, and in the cord blood of HEU (N = 21) and HUU (N = 61) infants at delivery. We identified multiple differences between WLHIV versus HIVneg women (Figure 3). In contrast, there were few differences in the concentrations of plasma cytokines and chemokines between HEU versus HUU infants. After adjustment for multiple comparisons, none of the infant differences persisted, but the following maternal cytokines or chemokines remained significantly higher in WLHIV versus HIVneg women: IL-1β, CXCL10 (IP-10), CCL22 (MDC), sCD163, TGFβ, and TNFα.

Figure 3.

Concentrations of cytokines and chemokines in plasma from pregnant women in the 2nd trimester and infants at delivery. Values are represented as median concentration (pg/mL) and interquartile range. Data are from N = 60 HIV-negative (HIVneg) women, N = 27 women living with HIV (WLHIV), N = 61 HIV-unexposed (HUU) infants, and N = 21 HIV-exposed (HEU) infants. Significant differences are indicated by asterisks, where * represents P = .01 to .05, ** P = .001 to .01, *** P = .0001 to .001, and **** P < .0001. Abbreviations: FABP, fatty acid binding protein; IFN, interferon; IL, interleukin; IP, interferon gamma induced protein; MCP, monocyte chemoattractant protein; MDC, macrophage-derived chemokine; MIP, macrophage inflammatory protein; TGF, transforming growth factor; TNF, tumor necrosis factor.

Markers of Inflammation in Maternal and Cord Plasma Correlate Inversely With Infant Immune Responses to RSV

Among participants with a paired maternal peripheral blood and cord blood sample (N = 30 HIVneg/HUU dyads and N = 23 WLHIV/HEU dyads), we examined correlations between the concentrations of cytokines and chemokines in maternal or cord plasma with infant cellular responses to RSV or RSV knockdown, defined as the proportion of RSV-infected respiratory epithelial cells that are cleared by CBMCs. We identified several inverse correlations between the concentration of inflammatory cytokines or chemokines in maternal or cord plasma and infant RSV responses (Figure 4). Among all correlations identified, only 2 remained significant after adjustment for multiple comparisons: the inverse associations between the inflammatory molecules CXCL10 (IP-10) and TNFα in cord plasma with infant IFNγ-expressing NK cells.

Figure 4.

Heatmap depicting correlations between concentrations of cytokines and chemokines in cord blood plasma or maternal plasma, and infant innate immune responses to RSV or RSV knockdown. A linear regression model was fit to identify correlations, adjusting for HIV exposure status of the infant. Red indicates a positive correlation and blue indicates a negative correlation. Significant correlations are indicated by asterisks, where * represents P = .01 to .05, ** P = .001 to .01, *** P = .0001 to .001, and **** P < .0001. Data are from N = 30 HIV-negative mothers/HIV-unexposed infants and N = 23 women living with HIV/HIV-exposed infants. Abbreviations: DC, dendritic cell; FABP, fatty acid binding protein; IFN, interferon; IL, interleukin; IP, interferon gamma induced protein; MCP, monocyte chemoattractant protein; MDC, macrophage-derived chemokine; MIP, macrophage inflammatory protein; NK, natural killer cell; RSV, respiratory syncytial virus; TGF, transforming growth factor; TNF, tumor necrosis factor.

Incubation of HUU CBMCs in Maternal Plasma From WLHIV Impairs Immune Responses to RSV

To determine whether exposure to soluble factors in maternal plasma may precipitate the diminished immune responses to RSV that we identified in HEU infants, we performed paired experiments in which we incubated aliquots of CBMCs from 5 HUU infants in maternal plasma from 4 WLHIV in parallel with maternal plasma from 4 HIVneg women for 48 hours. Subsequently, we compared innate immune responses to RSV using the respiratory infection model. Strikingly, we identified that incubation in plasma from WLHIV lowered several RSV-specific responses compared with incubation in HIVneg plasma: CD107a in NK cells [median 1.3-fold increase (IQR 0.8–3.2) versus 2.3-fold increase (IQR 1.3–3.9) between RSV- and RSV + cultures, P = .04], IL-12 in monocytes [median 1.0-fold increase (IQR 0.4–1.3) versus 1.8-fold increase (IQR 1.2–2.7), P = .0004], PD-L1 in DCs [median 3.2-fold increase (IQR 1.1–4.8) versus 6.6-fold increase (IQR 4.9–13.0), P = .002], and CD83 in DCs [median 1.7-fold increase (IQR 1.3–3.0) versus 5.1-fold increase (IQR 1.9–7.8), P = .01] (Figure 5).

Figure 5.

Fold-change in expression of markers of activation, maturation, and degranulation among cord blood mononuclear cells from HIV-unexposed (HUU) infants at birth, after 48 h incubation with plasma from pregnant women with or without HIV and subsequent exposure to RSV-infected versus mock-infected respiratory epithelial cells. Data are from N = 5 HUU infants, each incubated with plasma from N = 4 WLHIV and N = 4 HIVneg women. Significant differences are indicated by asterisks, where * represents P = .01 to .05, ** P = .001 to .01, *** P = .0001 to .001, and **** P < .0001. Abbreviations: DCs, dendritic cells; MFI, median fluorescence intensity; NK, natural killer; RSV, respiratory syncytial virus.

DISCUSSION

Using an in vitro model of human respiratory infection, we demonstrated specific innate immune defects in the cellular response to RSV infection among HEU infants, which likely contribute to the increased morbidity and mortality associated with RSV in this population. The most striking difference that we identified was decreased production of IL-12 by HEU versus HUU DCs, with the greatest difference present at birth, and differences persisting but to a lesser degree at 12–18 months of age in RSV-seronegative children. IL-12 production by antigen-presenting cells (including DCs) is an integral component of the innate immune response to RSV [16–18]. The secreted IL-12 acts on nearby NK cells, increasing their recruitment, proliferation, cytotoxicity, and production of antiviral cytokines such as IFNγ. Accordingly, we also identified decreased expression of IFNγ by HEU versus HUU NK cells in response to RSV infection, but this difference was only apparent at birth and not 12–18 months. The improved immune response to RSV around 12–18 months of age mirrors the epidemiology of infections in HEU infants, in which the highest rates of morbidity and mortality are observed in the first 6–12 months of life and then gradually subside over time [41 , 42]. This phenomenon suggests that the immune dysfunction observed in HEU infants is a result of in utero programming that is gradually reversible over the first year of life.

To distinguish whether the NK cell abnormalities that we identified at birth were a result of decreased signaling from APCs versus defects intrinsic to the NK cells, we performed experiments measuring ADCC among isolated NK cells from HEU and HUU infants at birth. We identified decreased degranulation in NK cells from HEU versus HUU infants (represented by decreased CD107a expression) upon exposure to immobilized RSV antigen/antibody complexes. This finding suggests that both neonatal APCs and NK cells are directly affected by in utero exposure to an environment affected by maternal HIV.

We hypothesized that the maternal factor driving this in utero programming is HIV-associated inflammation. This study confirmed previous publications demonstrating that pregnant WLHIV experience higher rates of inflammation and immune activation than HIVneg women [21, 26–30]. We identified the elevation of multiple inflammatory cytokines and chemokines in WLHIV, despite receiving antiretroviral therapy and having well-controlled HIV. In contrast, we found few differences in the concentrations of inflammatory markers in cord blood. The relatively modest degree of inflammation detected in HEU infants in this study may be a result of improved HIV control in their mothers or other demographic or clinical differences between cohorts. Alternatively, it is possible that the differences between the cord blood of HEU and HUU infants were masked by the release of inflammatory molecules that is known to occur during labor and delivery [43]. Maternal plasma was intentionally obtained during the 2nd trimester to avoid confounding of the maternal inflammatory markers by this phenomenon.

We identified multiple inverse correlations between maternal and cord blood inflammatory markers and infant cellular RSV-specific responses. Of these correlations, the 2 that remained significant after adjustment for multiple comparisons were between TNFα and CXCL10 (IP-10) concentrations in cord blood and infant IFNγ production by NK cells in response to RSV. This finding is especially notable given that NK cells from HEU versus HUU infants expressed less IFNγ in the respiratory infection model and suggests that this difference may be a result of exposure of HEU immune cells to inflammatory molecules. The mechanism whereby maternal inflammatory signals affect the fetus is not clearly understood. Nevertheless, it is important to note that maternal plasma inflammatory molecules were measured during the 2nd trimester of pregnancy, suggesting that the fetus was exposed to these signals while their immune system was developing.

We performed a mechanistic experiment to determine whether incubation of CBMCs from HUU infants in maternal plasma from WLHIV would induce similar innate immune defects as we observed in CBMCs from HEU infants. Remarkably, we observed that incubation in plasma from WLHIV compared to plasma from HIVneg women resulted in significantly lower RSV-specific responses among DCs, NK cells, and monocytes. While the immune defects induced by the experimental conditions were not identical to those observed in CBMCs from HEU infants, this observation was proof of concept that exposure to soluble factors in maternal plasma, which may include inflammatory molecules or other substances, may drive alterations in neonatal innate immune cells. Taken together, our findings suggest that incubation in an inflammatory environment in utero primes the fetal immune cells for a postnatal fate of suboptimal cellular responses to RSV.

Strengths of this study include that the HEU and HUU cohorts were relatively well-matched. We used a novel in vitro model of human respiratory infection to evaluate innate immune responses to RSV infection between HEU and HUU infants. This model closely recapitulates RSV infection in vivo because it allows direct cell-to-cell communication between innate immune cells and RSV-infected respiratory epithelial cells. We used CBMCs to study the early immune response to RSV, which is an appropriate sample in which to study immune defects that occur in utero, before the cells have an opportunity to be confounded by environmental exposures. Limitations include that we only collected maternal blood once during pregnancy, which limited our ability to directly compare concentrations of inflammatory biomarkers between mothers and infants. Additionally, the type of maternal antiretroviral drugs administered during pregnancy may have differed between participants, which may have impacted infant immune responses; however, prior literature suggests that receipt of ART is associated with reduced inflammation in pregnant women and their infants, regardless of the specific drugs included in the ART regimen [7, 44–46].

In conclusion, maternal inflammation associated with HIV infection is a likely driver of innate immune dysregulation in HEU infants and predisposes to their increased susceptibility to morbidity and mortality upon exposure to RSV. These findings were identified despite the fact that the WLHIV in our cohort were receiving antiretroviral therapy and had excellent HIV control. More research is needed to dissect the mechanistic process through which inflammatory signals are transferred across the placenta and imprint the fetal immune cells.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. C.S. and A.W. contributed to conceptualization. C.S., K.C., A.B., S.B., K.N., J.F., N.L., and J.S. involved in data collection. C.S., K.C., and A.W. performed analysis. C.S. involved in writing—original draft. C.S., K.C., A.B., S.B., K.N., J.F., N.L., J.S., L.L., and A.W. involved in writing: review and editing.

Acknowledgments. We would like to thank the infants and families who enrolled in these studies and generously provided research samples. We would also like to thank Jocelyn Phipers, Trisha Adkins, and Amy Lamprecht for their assistance in enrolling participants and collecting research samples for this study.

Financial support. This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development at the National Institutes of Health Grant Number 1K08HD100205 to C.S.; the National Center for Advancing Translational Sciences at the National Institutes of Health Grant Numbers UL1 TR002535, UM1 TR004399; and grant 2020141 from the Doris Duke Charitable Foundation and University of Colorado School of Medicine to C.S. L.L.L. received support from the National Institute of Allergy and Infectious Diseases at the National Institutes of Health Grant Number R01 AI178925. The funding sources had no role in the study design, analysis or writing of the manuscript. Contents are the authors’ sole responsibility and do not necessarily represent official NIH views.

Prior presentations at scientific meetings.

1. Smith, C., Curtis, K., Bonham, A., Boyer, S., and Weinberg, A. Immune Responses to Respiratory Syncytial Virus among HIV-Exposed Uninfected Infants from the United States Correlate with Maternal Inflammation during Pregnancy. International AIDS Society 12th Conference on HIV Science, Brisbane, AU, July 2023 (Poster Presentation)

2. Smith, C., Curtis, K., Bonham, A., Boyer, S., and Weinberg, A. Immune Responses to Respiratory Syncytial Virus among HIV-Exposed Uninfected Infants from the United States Correlate with Maternal Inflammation during Pregnancy. International Workshop on HIV & Pediatrics 2023, Brisbane, AU, July 2023 (Oral Presentation)