Microbial Translocation into Amniotic Fluid of Vervet Monkeys is Common and Unrelated to Adverse Infant Outcomes

Abstract

Amniotic fluid was collected from pregnant female African green monkeys (n=20). Analyses indicate microbial translocation into amniotic fluid during pregnancy is typical, and microbial load reduces across gestation. Microbial translocation does not relate to infant outcome or maternal factors. Lastly, we demonstrate that sample contamination is easily introduced and detectable.

Introduction

We have previously reported breeding efficiency in a closed breeding colony of captive African green or vervet monkeys (Chlorocebus aethiops sabeus), where infant mortality rates have consistently been above 30%, with some years exceeding 50% [1]. Maternal factors previously shown to significantly affect infant mortality include glycemic control and social rank, while infant factors such as male sex and macrosomia also contributed to infant mortality [1]. However high rates of neonatal loss were mostly unexplained.

In utero infection is a common cause for abortion and stillbirth in people. There have been many published reports regarding the topic of intra-amniotic sterility; however these reports tend to contradict one another in terms of bacteria and correlation with preterm labor and delivery [2–7]. Our study aimed to analyze the sterility of amniotic fluid from vervet monkeys in a colony with a poor reproductive history. We hypothesized that microbial contamination would be higher in pregnancies that terminate as pre-term or stillbirth.

Materials and Methods

All procedures were approved by the Institutional Animal Care and Use committee of Wake Forest School of Medicine. Pregnant monkeys (n=20) were identified over two birth seasons and amniocentesis was performed during the second and third trimester. Animals were sedated using ketamine or a combination of ketamine/midazolam. After the hair over the abdomen was clipped and aseptic skin preparation using 3 alternating scrubs of chlorhexidine scrub and 70% alcohol was completed, we used ultrasound guidance to aspirate <5ml of amniotic fluid into a sterile plain vacutainer, using a 5ml syringe with a 22-gauge needle. Samples were stored in cryovials at −80°C until near the time of assay, when tubes were moved to −20°C and refrigerator conditions were maintained on assay day. Total protein was measured by bicinchoninic acid protein assay (Bio-Rad, Hercules, CA) in order to normalize for outcomes and assess for potential blood contamination. Prior to DNA isolation, amniotic fluid samples were centrifuged at 15000g for 15 minutes to concentrate the sample. DNA was isolated from amniotic fluid using the QIAamp DNA Mini Kit (Qiagen, Germantown, MD) and following the manufacturer’s instructions. The concentration of nucleic acids was quantitated spectrophotometrically (Nanodrop, ThermoFisher Scientific, Waltham, MA). PCR was performed using forward and reverse primers that identify universal bacterial 16s DNA using methods previously described [8, 9]. A positive sample was defined as having an amplification value (Ct) as more than 3 standard deviations below the mean Ct calculated for triplicate blank samples. PCR was performed once after 20 samples had been collected, as this number was considered the minimum for normality assumptions to be met for statistics assessments.

Differences between birth year was considered a factor as a subset of samples (N=10 from one year of collection) had undergone a single additional thaw for subsampling, which requires the vial to be opened briefly to the environment and an additional 12 months in storage. All samples were consistently handled in a comparable, clinically clean manner suitable for biohazardous specimens. Maternal factors assessed included age, waist circumference and pre-pregnancy body weight. Infant factors assessed included birth outcome, body weight, head circumference and sex. Analysis of variance was used to compare outcomes and Pearson’s correlations indicated strengths of association, and non-parametric tests were used for categorical outcomes (Statistica V11, StatSoft, Tulsa OK). Significance was set as p<0.05.

Results

Our results indicate that microbial translocation into the amniotic fluid during pregnancy was typical, as 85% of samples tested positive for bacteria (p=0.02, Table 1). Data analysis showed that microbial load within the amniotic fluid decreased over time during gestation (Figure 1A), and microbial load demonstrated no relation to infant viability (p=0.33, Table 1). Lastly, we noted that sample contamination is detectable after a prolonged storage and additional air exposure and thus easily introduced (p=0.002, N=10/group; Figure 1B). As such, careful handling of biospecimens needs to be conducted to interpret data. There were no associations of microbial load in amniotic fluid and infant or maternal factors.

Table 1.

Maternal and infant factors relative to amniotic fluid categorized as positive or negative for microbial detection. Microbial translocation into the amniotic fluid is common (85% of samples) and unrelated to maternal or infant factors. Means with standard error of the means in parentheses are reported.

Amniotic fluid category	Maternal age (yrs)	Maternal weight (kg)	Maternal waist (cm)	Infant weight (kg)	Infant sex (%female)	Infant head circumference (cm)	Infant survival (%)	Amniotic total protein (g/dL)
16s positive (n=17)	10.4 (3.88)	5.59 (0.85)	34.7 (1.19)	0.316 (0.02)	20	11.5 (0.22)	60.2 (0.09)	2.96 (0.40)
16s negative (n=3)	7.97 (3.31)	5.65 (0.88)	36.2 (4.07)	0.332 (0.02)	80	11.2 (0.67)	83.3 (0.10)	2.40 (0.04)
ANOVA p-value	0.40	0.92	0.65	0.77	0.14*	0.34	0.33*	0.58

^*p-value calculated from Chi square statistics for differences between proportions.

Figure 1.

A Microbial load, measured as normalized 16s gene abundance, in amniotic fluid decreases across gestation.

B Raw 16s gene abundance estimates demonstrate that a single air exposure results in significant microbial contamination (N=10/condition).

Discussion

The sterility of the intra-amniotic environment during pregnancy has come into question regarding pregnancy outcome and induction of preterm labor in humans [10]. Amniotic fluid from low-risk individuals at mid-trimester had been tested specifically for bacteria associated with genital tract infections, as well as for bacteria in general using broad-range PCR primers. These samples were found to be sterile based on PCR assays [11]. Other studies have detected bacteria utilizing PCR assays and culture, but differ on their findings of correlation with preterm labor and delivery [2–7]. Therefore, this study has great implications beyond correlating to infant mortality within a closed colony due to the differing and opposing previously reported results.

Current thought is that there are no sterile spaces in the body and our brief report substantiates microbial translocation is a normal process [12]. The intra-amniotic environment significantly impacts the development of the gut microbiome of infants, supporting the “in utero colonization hypothesis”. Distinct microbial populations have been identified in amniotic fluid, the most prevalent being Proteobacteria, which corresponds to bacteria found in meconium in infants [13]. Accordingly, gut colonization may begin in utero through the swallowing of amniotic fluid. Early induction of the gut microbiome has been shown to protect against numerous diseases later in life, and enhances development of the immune system [14]. In humans, the in utero environment is also affected by the nutritional and metabolic status of the mother and can have detrimental influence on the metabolism of the infant later in life [15, 16]. Because the related maternal factors, glycemic control and social status/increased stress, were found to impact the infant mortality rate in a closed colony of vervet monkeys [1], characterizing the intra-amniotic environment in these monkeys was of great interest in the context of healthy, comparable nutritional environments. We found maternal factors did not substantially drive microbial translocation. In conclusion, our data has shown that microbial translocation into amniotic fluid during pregnancy is typical. We also found that microbial load decreases with time during gestation and does not relate to infant outcome. Lastly, we noted that sample contamination is detectable and very easily introduced which is an issue receiving increasing attention as biospecimens from previously considered ‘sterile’ spaces are being evaluated [17]. As such, handling of biospecimens needs to be carefully conducted and documented to accurately interpret data.