Objective
Genital tract secretions exhibit bactericidal activity against Escherichia coli . We hypothesized that this defense may be modulated during pregnancy.
Study Design
Secretions were collected by vaginal swab from 70 pregnant women (35-37 weeks’ gestation) and 35 nonpregnant controls. We mixed E coli with swab eluants or control buffer and colonies enumerated to measure bactericidal activity. Cytokines, chemokines, and antimicrobial peptides were quantified by multiplex or enzyme-linked immunosorbent assay.
Results
Pregnant women had significantly greater bactericidal activity, higher concentrations of proinflammatory cytokines, and lower levels of beta defensins compared to controls. Seven (10%) pregnant and 8 (23%) nonpregnant women were vaginally colonized with E coli ; colonization was inversely associated with bactericidal activity.
Conclusion
The soluble mucosal immune environment is altered in pregnancy. We speculate that the observed changes may protect against colonization and ascending infection and could provide a biomarker to identify pregnant women at risk for infectious complications including preterm birth.
Prevention of preterm birth and perinatal infection is a global health imperative. Worldwide, 13 million infants (10% of births) are born prematurely each year and 30% of preterm births are associated with infection. Preterm delivery is responsible for almost half of all infant mortality in the United States. The cause of most preterm births is unknown, but evidence strongly implicates infection and changes in the genital tract mucosal immune environment as primary triggers. For example, in the Vaginal Infections and Prematurity Study of >13,000 pregnant women, an increase in Escherichia coli or Klebsiella pneumoniae was associated with an increased risk of preterm birth and a higher adjusted odds ratio (OR) for preterm birth than any other factor identified in the data set (adjusted OR, 2.99; 95% confidence interval, 1.37–6.53). E coli is currently the most frequently isolated pathogen in neonatal early-onset sepsis and meningitis and accounted for 41% of early-onset sepsis in very-low-birthweight preterm infants.
Genital tract secretions from healthy nonpregnant women exhibit variable bactericidal activity against E coli. We speculate that this activity, which likely reflects cumulative interactions among multiple mediators secreted by genital tract epithelial cells, immune cells, and vaginal microbiota, may provide a biomarker of a healthy mucosal immune environment. However, whether this ex vivo bactericidal activity provides clinical protection against infection remains unclear. Moreover, no studies have examined the bactericidal properties of genital tract secretions during pregnancy. To begin to elucidate the role of genital tract mucosal immunity in preventing or promoting preterm birth, the mucosal immune environment in healthy term pregnancy must first be defined.
We therefore designed a cross-sectional study to examine the mucosal immune environment and vaginal E coli colonization in near-term pregnant women compared to nonpregnant controls. We hypothesized that E coli colonization may be inversely associated with bactericidal activity.
Materials and Methods
Participants and sample collection
Following institutional review board approval from the Montefiore Medical Center, informed consent was obtained from all participants. Pregnant women were approached during their routine group B streptococcus (GBS) screening prenatal visit (35-37 weeks of gestation) and nonpregnant women were approached during a routine gynecologic visit. Exclusion criteria included age <18 years, evidence of active genitourinary infection, abnormal Pap smear result within the past year, antibiotic or steroid use within the past month, sexual intercourse within the last 48 hours (as semen may modify the mucosal immune environment), and major chronic medical illnesses such as diabetes mellitus, systemic lupus erythematosus, or human immunodeficiency virus infection. Samples were collected from March 2010 through October 2011.
Demographic and clinical data were collected for each participant. Since self-collected swabs for GBS screening are the standard of care in many obstetrical clinical settings, participants were given the option to have observed self-collection or physician collection of the following study-related vaginal swabs: cotton swab for measurement of vaginal wall pH (Whatman pH paper, pH 3.8-5.5), flocked swab to sample genital tract secretions, and cotton swab for E coli culture.
Sample processing
The swabs for culture were transported to the Montefiore Clinical Microbiology Laboratory and cultured for E coli on MacConkey plates; the plates were evaluated for bacterial colonies after 24 and 48 hours of incubation. The swabs for genital tract secretions were placed in a 1.5-mL sterile Eppendorf tube that had been prefilled with 0.5 mL of sterile saline and transported on ice to the laboratory and processed within 6 hours of collection. The sample was vortexed, the swab was compressed against the side of the tube to maximize elution of the secretions, and clarified by centrifugation at 2000 rpm for 7 minutes at 4°C. The eluant was divided into 50-μL aliquots and stored at –80°C.
Bactericidal activity
E coli (American Type Culture Collection strain 4382627) was grown overnight to stationary phase and then 3 μL of bacteria (∼10 9 colony-forming unit/mL) were mixed with 27 μL of swab eluant or control buffer (20 mmol/L potassium phosphate, 60 mmol/L sodium chloride, 0.2 mg/mL albumin, pH 4.5) and incubated at 37°C for 2 hours. The mixtures were then further diluted in buffer (to yield 800-1000 colonies on control plates) and plated on agar (Acros Organics; Thermo Fisher Scientific Inc., Waltham, MA) enriched with trypticase soy broth (Fluka; Sigma-Aldrich, St. Louis, MO). Colonies were counted using ImageQuant TL v2005 (GE Health Care, Piscataway, NJ) after an overnight incubation at 37°C. All samples were tested in duplicate and the percentage inhibition was determined relative to the colonies formed on the control plates. To evaluate whether viable bacteria were present in the vaginal eluants, random samples were directly plated onto agar and evaluated for bacterial growth. Of note, none of the eluants yielded bacterial growth.
Measurement of protein and immune mediators
Total protein concentration was measured by microBCA Protein Assay kit (Thermo Scientific, Rockford, IL) in swab eluants diluted 1:4. The concentration of interleukin (IL)-1α; IL-1 receptor antagonist (IL-1Ra); IL-1β; IL-6; IL-8; interferon (IFN)-α-2; macrophage inhibitory protein (MIP)-1-α; MIP-1-β; and regulated on activation, normal T-cell expressed and secreted (RANTES) were determined by multiplex assay with beads from Chemicon International, Billerica, MA (measured using multiplex) and analyzed using StarStation (Applied Cytometry Systems, Sacramento, CA) in eluants diluted 1:2. The levels of all other mediators were determined using commercial enzyme-linked immunosorbent assay kits with dilutions as noted: secretory leukocyte protease inhibitor (SLPI) (1:1000 dilution of eluant) (R&D Systems, Minneapolis, MN), human beta defensin (HBD)-1 (1:25), HBD-2 (1:100) and HBD-3 (1:25) (Alpha Diagnostic International, San Antonio, TX), and human neutrophil peptides 1-3 (HNP1-3) (1:200) (Hycult Biotech, Uden, The Netherlands). The lower limits of detection (LOD) for each assay were (pg/mL): HNP1-3, 156; SLPI, 25; HBD-1, 25; HBD-2, 5; HBD-3, 20; IL-1α, 3.5; IL-1β, 0.4; IL-6, 0.3; IL-8, 0.2; IFN-α-2, 24.5; IL-1Ra, 2.9; MIP-1-α, 3.5; MIP-1-β, 4.5; and RANTES, 1.0. The dilutions selected for each analyte were based on previous studies performed in our laboratory. Insufficient material was available to repeat samples that fell below the LOD at a lower dilution. Therefore, concentrations that fell below the LOD were set at the midpoint between zero and the LOD multiplied by the dilution factor. Concentrations that fell above the highest detection concentration were repeated at a higher dilution factor. Mediators with >45% of the samples below the LOD were dichotomized; all other mediators were reported as continuous variables.
Statistical analysis
To reduce skewness in the distribution, log 10 -transformed concentrations of all mediators were used in the analyses. Categorical variables were compared between groups by χ 2 or Fisher exact test. Continuous variables were compared by Student t test or the Mann-Whitney U test, depending on the distribution of the data. Spearman rank order correlation was calculated to assess correlations between mediators and bactericidal activity. A 2-sided P value < .05 was considered significant and a false discovery rate–adjusted P value was reported to account for multiple comparisons. Heat maps were generated to graphically display the correlations among the 2 cohorts of pregnant and nonpregnant women.
A multivariable linear regression was performed to assess for predictors of bactericidal activity against E coli . A backward stepwise regression approach was used and age, race, and smoking were placed in the model a priori; soluble immune mediators that correlated with E coli bactericidal activity at a false discovery rate–adjusted P value of < .05 were also added. Mediators that did not have a significant association with bactericidal activity against E coli and reduced the fit of the model were removed. Regression diagnostics were performed to ensure reasonable estimates and fit. All statistical tests were performed using STATA (version 11.0; StataCorp, Inc., College Station, TX), R version 2.12.2, and GraphPad Prism (version 5; GraphPad Software, Inc, La Jolla, CA).
Results
Description of participants
In all, 105 pregnant women were assessed for eligibility during their third trimester and 70 enrolled. Reasons for exclusion included declined participation (n = 15), abnormal Pap smear finding (n = 10), active infection (n = 5), recent intercourse (n = 3), and persistent asthma requiring steroid therapy (n = 2). In all, 37 nonpregnant women were approached for recruitment in the control arm and 35 enrolled; 2 women were excluded because of recent sexual intercourse. There were no significant differences with respect to age, current smoking status, history of sexually transmitted infections, vaginal wall pH, or method of swab collection ( Table 1 ). The groups did differ with respect to race ( P = .05). Among the nonpregnant women, 76% were using hormonal contraception and the remaining 24% were having normal menstrual cycles; phase of cycle was not determined ( Table 1 ).
| Characterisitc | Pregnant (n = 70) | Nonpregnant (n = 35) | P value |
|---|---|---|---|
| Mean age, y (SD) | 28.3 (6.5) | 30.9 (6.2) | .06 |
| Race, n (%) | .05 | ||
| White | 47 (67) | 30 (86) | |
| Black | 18 (26) | 2 (6) | |
| Other | 5 (7) | 3 (8) | |
| Current smoker, n (%) | 1 (1) | 3 (9) | .1 |
| History, n (%) | |||
| Genital herpes | 4 (6) | 3 (9) | .3 |
| Chlamydia | 6 (9) | 4 (11) | .5 |
| Gonorrhea | 1 (1) | 0 | .5 |
| Genital warts | 4 (6) | 3 (9) | .6 |
| Contraception, n (%) | NA | NA | |
| None | 9 (26) | ||
| Barrier methods | 11 (31) | ||
| Oral contraceptive pills | 5 (14) | ||
| Intravaginal ring | 3 (9) | ||
| Progesterone injectable | 5 (14) | ||
| Progestin-containing IUD | 2 (6) | ||
| Lateral vaginal wall pH, mean (SD) | 4.8 (0.02) | 4.8 (0.37) | .7 |
| Method of swab collection n (%) | .5 | ||
| Physician collection | 44 (63) | 21 (60) | |
| Observed self-collection | 26 (37) | 14 (40) |
Pregnant women have greater E coli bactericidal activity
Genital tract secretions collected from pregnant women had significantly greater bactericidal activity against E coli compared to nonpregnant women (median [range] 65% [17–99] vs 26.2% [–39 to 98], P < .001) ( Figure 1 ). Moreover, the pregnant women were less likely to be vaginally colonized with E coli (10% vs 23%, P = .002) and bactericidal activity was significantly less in women who were colonized compared to women who were not (62.7% [39–99] vs 22% [–30.7 to 91]) ( Figure 2 ), suggesting that activity may protect against vaginal colonization. Pregnancy, E coli colonization, and HBD-3 were significant predictors of bactericidal activity and remained significant after adjusting for age, race, smoking, and immune mediators within the genital tract ( Table 2 ). Pregnancy was associated with a 26% increase in bactericidal activity compared to controls. Independent of pregnancy status, women who were vaginally colonized with E coli had a 20% reduction in their bactericidal activity compared to women whose vaginal cultures did not yield E coli . In addition, in women whose HBD-3 levels were above the LOD, there was an associated 21% increase in bactericidal activity. Of note, vaginal pH was adjusted for in the multivariable logistic regression. However, there was no significant association with vaginal pH and bactericidal activity and it was subsequently removed due to improvement in model fit.
| Linear regression a | β-Coefficient (SE) | P value |
|---|---|---|
| Pregnancy | 25.7 (8.5) | .003 b |
| Escherichia coli colonization | −20.1 (9.3) | .031 b |
| Age | 0.29 (0.5) | .594 |
| Race c | ||
| Black | −6.9 (8.6) | .421 |
| Other | 15.7 (12.8) | .223 |
| Current smoker | −25.1 (17.3) | .151 |
| HBD-2 | .00002 (0.00009) | .838 |
| HBD-3 (detectable) | 20.5 (9.1) | .026 b |
| IL-1α | −0.0003 (0.002) | .899 |
| IL-1β | −0.02 (0.02) | .282 |
| IL-8 | .002 (0.001) | .527 |
a Raw values of mediators were entered into model and HBD-3 was entered as dichotomized variable;
b P < .05 considered significant;
Changes in soluble mucosal immune mediators and correlations with E coli activity
To evaluate whether the increased bactericidal activity observed in pregnant women was associated with differences in mucosal immune mediators, we compared the concentrations of total protein, cytokines, chemokines, and select antimicrobial peptides recovered in vaginal eluants between pregnant and nonpregnant women. There were no significant differences in total protein. However, compared to controls, pregnant women had significantly higher levels of several cytokines including IL-1α, IL-1β, and IL-1Ra, and the chemokines including IL-8 and RANTES ( Figure 3 and Table 3 ). IFN-α-2 was below the LOD in the majority of pregnant and nonpregnant women, and did not differ significantly between the 2 groups. In contrast to the increase in proinflammatory cytokines and chemokines, pregnant women had significantly lower levels of the epithelial beta defensins, HBD-2 ( Figure 3 ) and HBD-3 ( Table 3 ). No differences in HBD-1, HNP1-3, or SLPI were observed.
| Mediator | Percentage of subjects with levels below LOD a (n = 105) | Pregnant (n = 70) | Nonpregnant (n = 35) | FDR-adjusted P value |
|---|---|---|---|---|
| Total protein (μg/mL) | 2 | 3.1 (1.4–3.7) | 3.1 (2.1–3.7) | .743 |
| Alpha and beta defensins | ||||
| HNP1-3 (pg/mL) | 38 | 4.7 (4.1–6) | 4.6 (4.2–5.7) | .697 |
| HBD-1 (pg/mL) | 42 | 3.2 (2.5–4.3) | 3.2 (2.5–4.3) | .543 |
| HBD-2 (pg/mL) | 17 | 3.1 (2.4–5.3) | 4.2 (2.4–5.3) | < .001 c |
| HBD-3, n (%) b | 58 | < .001 c | ||
| Detectable (>250 pg/mL) | 11 (16) | 33 (94) | ||
| Undetectable (≤250 pg/mL) | 59 (84) | 2 (6) | ||
| Proinflammatory mediators | ||||
| IL-1α (pg/mL) | 5 | 2.5 (0.5–4) | 1.9 (0.5–4) | .017 c |
| IL-1β (pg/mL) | 12 | 1.2 (–0.4 to 3.3) | 0.4 (–0.4 to 2.3) | .012 c |
| IL-6 (pg/mL) | 0 | 0.3 (–1.2 to 2.5) | 0.02 (–1.0 to 1.2) | .062 |
| IL-8 (pg/mL) | 0 | 3.2 (1.2-4.0) | 2.5 (–0.4 to 4.0) | .007 c |
| IFN-α-2, n (%) b | 91 | .480 | ||
| Detectable (>24.5 pg/mL) | 6 (9) | 3 (9) | ||
| Undetectable (≤24.5 pg/mL) | 64 (91) | 32 (91) | ||
| MIP-1-α, n (%) b | 73 | .459 | ||
| Detectable (>3.5 pg/mL) | 17 (24) | 11 (31) | ||
| Undetectable (≤3.5 pg/mL) | 53 (76) | 24 (69) | ||
| MIP-1-β, n (%) b | 89 | .747 | ||
| Detectable (>4.5 pg/mL) | 9 (13) | 3 (9) | ||
| Undetectable (≤4.5 pg/mL) | 61 (87) | 32 (91) | ||
| RANTES, n (%) b | 49 | .002 c | ||
| Detectable (>2 pg/mL) | 43 (59) | 11 (31) | ||
| Undetectable (≤2 pg/mL) | 27 (41) | 24 (69) | ||
| Antiinflammatory mediators | ||||
| IL-1Ra (pg/mL) | 1 | 5.1 (2.8–5.7) | 4.0 (0.5–5.6) | .006 c |
| SLPI (pg/mL) | 21 | 5.3 (4.4–6.3) | 5.7 (4.4–6.3) | .447 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
