Soluble CD14: An Independent Biomarker for the Risk of Mother-to-Child Transmission of HIV in a Setting of Preexposure and Postexposure Antiretroviral Prophylaxis

Abstract

Elevated soluble CD14 (sCD14) concentrations, a marker of monocyte activation, predicts adverse outcomes in human immunodeficiency virus (HIV)–infected adults. To examine the association of sCD14 concentrations with the risk of mother-to-child transmission (MTCT) of HIV, we nested a case-control study (49 pairs of infants and their HIV-infected mothers) within the Six-Week Extended-Dose Nevirapine trial. Median peripartum maternal log₂ sCD14 concentration was higher among transmitters (defined as pairs in which maternally transmitted HIV infection occurred by 12 months of age) than nontransmitters (20.29 pg/mL vs 19.41 pg/mL; P = .005). There was an increased odds of MTCT for every log₂ increase in maternal sCD14 concentration, after adjustment for maternal HIV load, CD4 count and cART exposure (adjusted odds ratio, 3.51; 95% confidence interval, 1.21–10.21). Maternal monocyte activation may adversely influence the risk of MTCT of HIV.

Systemic immune activation and inflammation have emerged as key features of chronic human immunodeficiency virus (HIV) infection and have been linked to HIV disease progression and to HIV-related morbidity and mortality among adults [1–3]. However, there is a lack of studies examining the association between maternal immune activation during the peripartum period and infant outcomes.

Serum soluble CD14 (sCD14), released from monocytes in response to lipopolysaccharide (LPS) activation, is one of the clinically relevant biomarkers for the process of microbial translocation and HIV-associated immune activation [1–5]. During acute HIV infection, early disruption and irreversible damage to intestinal mucosa enable ongoing translocation of microbial products, particularly LPS, a component of commensal gram-negative bacteria that triggers inflammation and immune activation with monocyte activation [4, 6].

Previous studies have shown that children affected by HIV have poor intestinal integrity [7], increased microbial translocation and monocyte activation [5], and, in later childhood, is associated with a low CD4⁺ T-cell count [8]. However, studies are needed to determine whether peripartum maternal monocyte activation also influence infant outcomes such as MTCT of HIV.

To begin to address this research gap, our study aimed to determine whether the peripartum maternal sCD14 concentration is associated with an increased risk of MTCT of HIV. We performed this analysis on HIV-infected mother-infant pairs enrolled in the India Six-Week Extended-Dose Nevirapine (SWEN) study, an early National Institutes of Health–funded randomized clinical trial for prevention of HIV type 1 transmission via breast milk.

METHODS

Study Population and Design

A nested case-control study of 49 HIV-infected mother-infant pairs was conducted within the India SWEN study (clinical trials registration NCT00061321), a randomized controlled trial of single-dose versus extended-dose nevirapine to prevent HIV transmission via breast milk, conducted from 2002 to 2007, before Option B+ (universal antiretroviral therapy [ART] for HIV-positive pregnant women) roll out. The SWEN trial methods have been previously described [9]. In brief, HIV-infected pregnant women who intended to breastfeed were enrolled prior to delivery, infants were randomly assigned to one of 2 treatment arms, and mother-infant pairs were followed through postpartum month 12. Infant plasma samples collected at the following times were evaluated for HIV infection, using HIV DNA–specific polymerase chain reaction analysis: within 48 hours of birth; 1, 2, 4, 6, 10, and 14 weeks after birth; and 6, 9, and 12 months after birth. HIV-positive status was confirmed by detection of plasma HIV RNA. Written informed consent was obtained from all women and, in the case of infants, from parents or legal guardians.

Cases were defined as mother-infant pairs in which HIV was identified by 12 months of age (hereafter, “transmitters”); controls were defined as mother-infant pairs with infant HIV uninfected status during the study period (hereafter, “nontransmitters”). Cases were those with available samples, while controls were a random subsample of 10% of the nontransmitters. Maternal sCD14 concentration was evaluated by commercially available enzyme-linked immunosorbent assay–based techniques (R&D Systems, Minneapolis, Minnesota), using maternal plasma samples collected from cases and controls between 32 weeks of gestation to within 24 hours of delivery [9, 10]. Most (79%) of the samples were collected 0–24 hours after delivery, with a median of 0 days from birth. Samples were tested in duplicate, and the optical density of each well was determined using a Softmax VersaMaxPLUS ROM v1.21 reader; the sCD14 concentration was determined by interpolation to a nonlinear standard curve of sCD14 dilutions, using GraphPad Prism software (version 5).

Statistical Analyses

Baseline demographic, clinical, and laboratory data were summarized on the basis of infant HIV infection status by 12 months of age. Maternal sCD14 concentrations were log transformed to base 2 values, and all categorical and continuous variables were compared between cases and controls by using the Fisher exact test and the Wilcoxon rank sum test, respectively. Multivariable logistic regression analyses were used to assess the odds of MTCT of HIV for every log₂ increase in maternal sCD14 concentration; the analysis was adjusted for known maternal risk factors for MTCT, including maternal plasma HIV RNA load and CD4⁺ T-cell count (measured at the same time during which the sCD14 concentration was assessed) and ART exposure. This study was powered at 80% to detect a minimum odds ratio (OR) of 2.0 at a 5% significance level. Pearson rank correlations were used to assess the relationship between maternal log₂ sCD14 concentration with both maternal HIV RNA and maternal CD4⁺ T-cell count.

RESULTS

Study Population Characteristics

Among 49 mother-infant pairs included in this analysis, there were 29 cases and 20 controls. Of the 29 cases, the timing of MTCT was within 48 hours of birth for 7 infants (24%), between 48 hours and 6 weeks after birth for 17 infants (52%), and >6 weeks after birth for 7 infants (24%). Transmitters had a significantly higher median maternal plasma HIV RNA load than nontransmitters (4.53 vs 3.86 log₁₀ copies/mL; P = .01). Median maternal CD4⁺ T-cell count was lower in transmitters, compared with nontransmitters (298 vs 425 cells/mm³; P = .06).

Study groups were otherwise similar with respect to maternal and infant characteristics (Table 1). Overall, median maternal age was 23 years (interquartile range, 21–27 years), few mothers (14%) had received combination ART (cART) by the time of delivery, and most mothers (82%) received intrapartum single-dose nevirapine for prevention of peripartum HIV transmission.

Table 1.

Maternal-Infant Pair Characteristics Overall and by Presence (Cases) or Absence (Controls) of Human Immunodeficiency Virus Type 1 in Infants by Age 12 Months

Characteristic	Overall (n = 49)	Cases (n = 29)	Controls (n = 20)	P Valuea
Maternal
Age, y	23 (21–27)	22 (22–26)	24 (21–28)	.90
Primary education level or less	31 (63)	17 (59)	14 (70)	.55
Primigravida	19 (39)	11 (38)	8 (40)	.84
Received ZDV during pregnancy	16 (33)	8 (28)	8 (40)	.54
Received cART during pregnancy	7 (14)	3 (10)	4 (20)	.42
Received intrapartum NVR	40 (82)	23 (79)	17 (85)	.72
CD4+ T-cell count,b cells/mL
≥350	24 (49)	13 (45)	11 (55)
200–350	14 (29)	9 (31)	5 (25)	.80
<200	11 (22)	7 (24)	4 (20)
Viral load,b log10 copies/mL	4.30 (3.59–4.73)	4.53 (3.99–5.13)	3.86 (2.95–4.42)	.007
sCD14 concentration,b log2 pg/mL	20.08 (19.30–20.53)	20.29 (19.97–20.56)	19.41 (18.77–20.29)	.005
Infant
Male sex	22 (46)	13 (45)	9 (47)	.99
Birth weight <2500 g	20 (42)	12 (41)	8 (42)	.99
Received 6-wk extended-dose NVR	21 (43)	10 (34)	11 (55)	.24
Breastfed >4 mo	24 (49)	14 (48)	10 (50)	.99

Data are no. (%) of subjects or median value (interquartile range).

Abbreviations: cART, combination antiretroviral therapy; NVR, nevirapine; ZDV, zidovudine.

^a Calculated to assess differences between cases and controls among categorical and continuous variables, using the Fisher exact test and Wilcoxon rank sum test, respectively.

^b Measured at or near delivery.

Maternal Immune Activation and Outcomes

Median peripartum maternal log₂ sCD14 concentration was 20.08 pg/mL in the overall cohort and was significantly higher among transmitters than nontransmitters (20.29 vs 19.41 pg/mL; P = .005; Table 1 and Supplementary Figure). Maternal sCD14 concentration was not correlated with maternal HIV RNA (R = 0.24; P = .09) or maternal CD4⁺ T-cell count (R =−0.02; P = .88), using Pearson rank correlations. In univariate and multivariate analyses, maternal sCD14 concentration was associated with an increased risk of MTCT (Table 2). For every log₂ increase in maternal sCD14 concentration, there was an increased odds of MTCT (adjusted OR, 3.51; 95% confidence interval [CI], 1.21–10.21; P = .02), after adjustment for maternal factors, namely CD4⁺ T-cell count, HIV RNA load, and maternal treatment. Limiting the analyses to transmitters with MTCT occurring only after 48 hours following delivery did not change this association (data not shown). Similar results were also observed when the analyses was limited to samples (for sCD14 assessment) collected 0–48 hours from delivery (adjusted OR, 3.09; 95% CI, 1.04–9.10). Maternal CD4⁺ T-cell count or cART exposure were not significantly associated with MTCT in multivariable models, while there was a strong trend for viral load (P = .05; Table 2).

Table 2.

Association Between Maternal Soluble CD14 (sCD14) Levels and the Odds of Maternal-Infant Transmission of Human Immunodeficiency Virus

Maternal Characteristic	Univariable Model, OR (95% CI)	Multivariable Model, Adjusted ORa (95% CI)
CD4+ T-cell count, cells/mL
≥350	Reference	Reference
200–350	1.52 (.39–5.91)	0.82 (.15–4.53)
<200	1.48 (.34–6.42)	2.83 (.16–49.45)
cART use
No	Reference	Reference
Yes	0.46 (.09–2.34)	0.22 (.008–6.44)
Log10 viral load	2.55 (1.22–5.34)	2.65 (.99–7.10)
Log2 sCD14 level	3.95 (1.56–10.0)	3.51 (1.21–10.21)

Abbreviations: cART, combination antiretroviral therapy; CI, confidence interval; OR, odds ratio.

^a Adjusted for log₂ maternal sCD14 level, maternal viral load, maternal CD4⁺ T-cell count, and maternal cART use.

DISCUSSION

We observed that an elevated maternal sCD14 concentration in the peripartum period was an independent predictor of the risk of MTCT of HIV among pairs of infants and HIV-infected mothers in India and that the association was independent of maternal HIV load, CD4⁺ T-cell count, and cART exposure. Our results extend previous findings of inflammation-associated adverse adult HIV outcomes to mother-infant pairs and highlight the potentially important role of monocyte activation in infant outcomes and HIV transmission among untreated women, which warrant further study with more-contemporaneous approaches to MTCT.

Importantly, to our knowledge, this is the first study to show increased risk of HIV transmission in association with increased peripartum concentrations of markers of monocyte activation. Understanding the mechanisms of enhanced risk of HIV transmission in the context of microbial translocation and immune activation is critical for designing interventions to avert this process. Similarly, understanding the increased risk of HIV acquisition associated with monocyte activation is also important. A potential mechanistic explanation, as suggested by studies in pediatric populations [7] but not yet in pregnant women is that the increase in HIV transmission is due to disrupted mucosal integrity, with the increased levels of circulating LPS and sCD14 representing the subsequent microbial translocation [7]. Future studies should also explore the interplay of breast milk inflammation and systemic immune activation in influencing MTCT.

Consistent with other MTCT studies, our analysis suggest that maternal HIV RNA load is associated with the risk of HIV transmission [11]. However, whereas previous studies in nonpregnant, HIV-infected individuals have shown the sCD14 concentration to be directly correlated with HIV load and indirectly correlated with CD4⁺ T-cell count [2], we found no correlation between sCD14 concentration and either HIV RNA load or CD4⁺ T-cell count among the pregnant HIV-infected women included in this analysis. Whether this is because the sCD14 concentration is influenced by pregnancy, which is associated with several dynamic immune and inflammation changes, remains unknown [12].

While among the first studies to explore the impact of HIV-associated immune activation during pregnancy, this analysis has limitations. First, the cohort is small and is representative of only Indian women enrolled in the SWEN trial, where HIV subtype C predominates. In addition, sample volume limited our ability to assess infant sCD14 concentration and other markers of microbial translocation, including proinflammatory cytokines, among mother-infant pairs. We may have not accounted for all potential confounders but did adjust for plasma HIV load, CD4⁺ T-cell count, and maternal cART exposure, which are the key correlates of MTCT of HIV. Finally, most of the women in our study had not received cART, as this was not standard of care more than a decade ago, at the time of the SWEN study, but did receive peripartum prophylaxis for prevention of MTCT of HIV as part of the parent study. It is likely that immune activation during pregnancy is reduced with cART and in turn associated with reduced risk of MTCT independent of HIV load, but this requires further study. Despite these limitations, our study provides evidence that maternal monocyte activation may be relevant to the pathogenesis of infant HIV infection. We also acknowledge that further studies are needed to focus on how maternal immune activation during pregnancy affects maternal health in addition to the infant's health.

In summary, our results suggest that high maternal sCD14 concentration during the peripartum period comes with a higher odds of MTCT of HIV, independent of maternal HIV load, CD4⁺ T-cell count, and cART exposure. If future studies in HIV-infected pregnant women receiving cART under Option B+ confirm these findings, approaches to reduce microbial translocation and immune activation may have applications for further reducing risk of HIV transmission to HIV-exposed infants in conjunction with cART. Our findings may also have implications for assessing the risk of HIV transmission in adult populations.

Supplementary Data

Supplementary materials are available at http://jid.oxfordjournals.org. Consisting of data provided by the author to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the author, so questions or comments should be addressed to the author.