Vaginal infection with Ureaplasma urealyticum accounts for preterm delivery via induction of inflammatory responses

Authors


Correspondence
Kayoko Harada, Department of Obstetrics and Gynecology, Hyogo College of Medicine, 1-1, Mukogawa-cho, Nishinomiya, Hyogo 663-8501, Japan. Tel: +81 798 45 6481; fax: +81 798 46 4163; email: kayoko-h@hyo-med.ac.jp

ABSTRACT

U. urealyticum, a member of the family Mycoplasmataceae, is often detected in the vagina of pregnant women. In this study, the possible association of ureaplasmal infection with preterm delivery was examined, as was the capacity of ureaplasmal LP to stimulate monocytes in vitro to produce pro-inflammatory cytokines relevant to preterm delivery. A hundred cases of normal delivery and 45 cases of preterm delivery were randomly selected. A mAb against U. urealyticum urease, that selectively and positively stained it in vaginal secretions of infected women but not in those of uninfected women, was generated. The preterm delivery group showed a significantly higher incidence of vaginal infection with this bacteria than the normal delivery group. Since the LP of Mycoplasma has potent biological activity, ureaplasmal LP was extracted. THP-1 cells, and human monocytic cells, produced IL-8, a potent pro-inflammatory cytokine associated with preterm delivery, and showed apoptotic cell death in response to the LP in vitro. These results suggest that U. urealyticum infection might play a causative role in preterm delivery via LP-induced IL-8 production and apoptosis.

List of Abbreviations: 
BSA

bovine serum albumin

cDNA

complementary DNA

DAPI

4′,6-diamino-2-phenylindole

DTT

dithiothreitol

ELISA

enzyme-linked immunosorbent assay

FACS

fluorescence activated cell sorter

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

IL

interleukin

LP

lipoprotein

LPS

lipopolysaccharide

mAb

monoclonal antibody

PBS

phosphate –buffered saline

PCR

polymerase chain reaction

PGN

peptidoglycan

PI

propidium iodide

PPLO

pleuropneumonia-like organisms

SD

standard deviation

SDS-PAGE

sulfate-polyacrylamide gel electrophoresis

TLR

Toll-like receptor

TNF

tumor necrosis factor

U. urealyticum

Ureaplasma urealyticum

INTRODUCTION

Both U. urealyticum and the genus Mycoplasma belong to the family Mycoplasmataceae, although they differ in terms of urease production. Both microbes lack a cell wall but have LP in their cell membrane. Mycoplasmal LP has been shown to have the capacity to induce production of various pro-inflammatory cytokines, such as TNF-α, IL-1, IL-6 and IL-8, from innate immune cells including macrophages (1–3). However, it is not known whether LP of U. urealyticum has the same potential.

Several clinical reports suggest that urogenital infection with U. urealyticum may cause abnormal pregnancy outcomes by inducing bacterial vaginosis, cervicitis, chorioamnionitis, intrauterine infection, premature rupture of membranes, preterm delivery and neonatal pneumonia (4–7). During such infection, macrophages and neutrophils in the infection sites may produce various pro-inflammatory cytokines relevant to preterm delivery (8–13). IL-8 has been reported to have the capacity to induce preterm delivery by promoting maturation of the uterine cervix in pregnant rabbits. Indeed, IL-8 can evoke recruitment and degranulation of neutrophils, leading to local release of various proteinases that destroy the fetal membrane (14–16). Furthermore, it has been reported that apoptotic macrophages induced by bacterial infection and/or by stimulation with their LP can prolong inflammatory responses (17, 18).

The TLR family recognizes pathogen-associated molecular patterns produced by various types of microbes (17, 19, 20). TLR4 recognizes LPS of Gram-negative bacteria, while TLR2 senses PGN of Gram-positive bacteria and LP of Mycoplasma. Intriguingly, TLR expression levels are upregulated by signaling via the same or other TLR (21). This might explain severe inflammation after infection with multiple types of microbes.

In this study the possible correlation between vaginal infection with U. urealyticum and preterm delivery in humans was investigated. Mechanisms underlying abnormal pregnancy outcomes in infected women were also studied.

MATERIALS AND METHODS

Study population

The protocols described in this report were approved by the ethical committee of the Hyogo College of Medicine. The study group was selected from pregnant women who delivered at the Department of Obstetrics and Gynecology of the Hospital of Hyogo College of Medicine between January 2006 and July 2007. Women were excluded if they had uterine anomalies, multiple gestations, major fetal anomaly or cervical cerclage. After obtaining informed consent, 45 patients with preterm labor and/or preterm premature rupture of membranes (preterm delivery) and 100 asymptomatic pregnant women (full-term delivery) were selected at random. Vaginal swabs were routinely obtained for general analysis of aerobic and anaerobic bacteria and yeast at the first visit due to symptoms (preterm delivery) or at the onset of labor pains (full-term delivery). At the same time, additional cervical swabs were obtained for analysis of the effects of U. urealyticum. Models were separated into four groups: sole effect of U. urealyticum (U. urealyticum+/other bacteria −); sole effect of other bacteria (U. urealyticum−/other bacteria +); joint effect of U. urealyticum and other bacteria (U. urealyticum+/other bacteria +) and controls; neither U. urealyticum nor other bacteria (U. urealyticum−/other bacteria −).

Culture of U. urealyticum and PCR

Vaginal swabs were placed on ureaplasma culture medium plates at 37oC for three to four days anaerobically. The ureaplasma culture medium (pH 6.4) was PPLO (Becton-Dickinson, Sparks, MD, USA) supplemented with 20% horse serum (Multiser, ThermoTrace, Melbourne, Australia), 10% yeast extract (Becton-Dickinson), 0.1% urea (Wako, Osaka, Japan), phenol red, 15 mM phosphate buffer (K+, Na+), Penicillin G (Meiji Seika Kaisya, Tokyo, Japan), 0.07 mg/ml MnSO4 (Wako) and 1.4% agarose. At the end of the culture, the U. urealyticum-positive plates changed color from orange to red, and the presence of U. urealyticum colonies was morphologically confirmed microscopically. Then DNA was extracted from the colonies and the biovar type was analyzed by PCR-direct sequencing as previously described (22, 23).

Immunofluorescence staining method

U. urealyticum (serotype 3) was obtained from the National Institute of Health (Tokyo, Japan). Ureaplasmal urease-containing bands resolved in denaturing SDS-PAGE were homogenized in TiterMax Gold adjuvant (Funakoshi, Tokyo, Japan) as previously described (24), and subcutaneously injected into Wister rats (Nihon SLC, Shizuoka, Japan). These rats were given booster injections at two week intervals. Three days after the final booster, spleen lymphocytes were isolated and fused with P3U1 myeloma cells using polyethylene glycol 1500. Hybridoma culture supernatants were screened by ELISA using sonicated U. urealyticum as the antigen. Positive hybridomas were cloned. Ascitic fluid was obtained by injecting 2.5 × 106 cells of one of the positive hybridomas intraperitoneally into BALB/c nu/nu mice. The immunoglobulin fraction from the ascitic fluid was purified as described previously (24). This mAb against ureaplasmal urease was identified as an IgG (data not shown).

Vaginal swabs were applied onto slide glass and cells were fixed with CYTOKEEP (Alfresa Pharma, Osaka, Japan). The slides were washed several times with PBS containing 1% BSA, and incubated with anti-U. urealyticum urease mAb (1:1000 dilution) for 60 min at room temperature. They were then incubated with biotinylated anti-rat IgG (Vector Laboratories, Burlingame, CA, USA) (1:500 dilution) for 30 min at room temperature, followed by incubation with Strept Avidin PBXL-1 (Kirkegaard & Perry Laboratories, Maryland, USA) (1:500 dilution) for 5 min. After washing with PBS, the cells were mounted with Vectashield mounting medium containing DAPI (Vector Laboratories) and examined under a Carl Zeiss confocal laser scan microscope LSM 510 (Carl Zeiss, Jena, Germany).

Preparation of Ureaplasmal LP

U. urealyticum was grown anaerobically in the PPLO medium supplemented as described above apart from MnSO4 and agarose. When the color became red, the medium was collected, centrifuged at 30 000 rpm for 60 min and the pellets stored at −80oC until required.

U. urealyticum was suspended in 10 mM Tris-hydrochloride buffer (Sigma, St. Louis, MO, USA) at pH 7.4 and containing 154 mM NaCl and a cocktail of protease inhibitors (Sigma) at a protein concentration of 2 mg/ml (TS buffer) and sonicated lightly to form a uniform suspension. The suspension (0.9 ml) was mixed with 0.1 ml of 20% Triton X-114 (Sigma) on a rotator at 4oC for 2 h and centrifuged at 10 000 g for 10 min at 4oC. The supernatant was incubated at 37oC for 5 min for phase separation, then centrifuged at 10 000 g for 5 min. After discarding the upper aqueous phase, TS buffer (0.9 ml) was added to the Triton X-114 phase and processed again as described above. Nine volumes of methanol were added to the Triton X-114 phase to precipitate lipoproteins, which were collected by centrifugation at 10 000 g for 30 min, redissolved in LPS-free saline and sonicated for mixing (25). The protein concentration of ureaplasmal LP was determined by the method of Dulley and Grieve (26).

Culture of THP-1 cell lines with LP and analysis of cytokine production and apoptosis

A human monocytic cell line, THP-1, was obtained from the Health Science Research Resources Bank (Osaka, Japan). THP-1 cells were grown in RPMI-1640 medium (Invitrogen, Grand Island, NY, USA) supplemented with 10% fetal bovine serum and Antibiotic-Antimycotic (Invitrogen) at 37oC in a humidified atmosphere of 5% CO2. THP-1 cells were cultured at 1 × 106 cells/ml in 24-well plates in a final volume of 1 ml/well. LPS from Escherichia coli Serotype 026: B6 (Sigma-Aldrich, St. Louis, MO, USA) or ureaplasmal LP was added at 100 ng/ml and 10 μg/ml, respectively, and the cells were incubated at 37oC. Conditioned medium was collected after 2, 4, 8, 24 and 48 h of culture and stored at −80oC until required. Assays for IL-1β, IL-6, IL-8 and TNF-α concentrations were performed using commercially-available ELISA reagents (Bio-Plex, Bio-Rad, Tokyo, Japan).

To detect cell death cells were stained with both annexin V and PI. Annexin V shows a high affinity and specificity for phosphatidylserine, which is exposed on the external surfaces of apoptotic and necrotic, but not healthy cells. PI stains DNA of leaky necrotic cells only. Cells that react with both annexin V and PI are necrotic, whereas cells that react with annexin V, but not PI, are apoptotic.

THP-1 cells were incubated with or without LPS (100 ng/ml) or ureaplasmal LP (10 μg/ml) at 37oC for 24 h, washed with PBS and sedimented by centrifugation at 500 g for 5 min at 4oC. The cell pellet was suspended in 500 μl of staining solution containing 5 μl of annexin V-FITC and 5 μl PI (Annexin V-FITC kit, Beckman Coulter, Fullerton, CA, USA). After 10 min of incubation, the cells were analyzed by a flow cytometer (EPICS XL, Beckman Coulter).

Real time PCR

THP-1 cells were cultured with LPS or ureaplasmal LP for 24 h. Total RNA was prepared from THP-1 cells using RNeasy Mini kit (QIAGEN, Tokyo, Japan) according to the manufacturer's instructions. Using random primers, dNTP MIX, first-standard buffer, 0.1 M DTT and Super Script RT (Invitrogen), cDNA was prepared from the total RNA according to the manufacturer's instructions. Each cDNA was prepared from 4 μg of RNA. Quantitative real-time PCR was conducted for TLR2 and GAPDH using TaqMan probes (assay ID Hs00610101_m1, assay ID Hs00266705_g1). The following conditions were used; 50oC for 2 min, 95oC for 10 min, then 40 cycles at 95oC (denature step) for 15 sec, and at 60oC (annealing and extension step) for 1 min. PCR products were assayed according to the predeveloped TaqMan assay reagent protocol (Applied Biosystems, Foster, CA, USA).

Statistical analysis

Rates of vaginal U. urealyticum or other bacterial infections of preterm deliveries and full-term deliveries were analyzed using the χ2 test. Rates of single and joint associations between U. urealyticum, other bacteria and preterm delivery were analyzed using the χ2 test and odds ratio. Real time PCR data, cytokine levels and apoptosis data were analyzed by Student's t-test using Excel software. P < 0.05 was considered as statistically significant.

RESULTS

High incidence of ureaplasmal infection in patients with preterm delivery

The average maternal age of the preterm delivery group (mean ± SD; 30.8 ± 5.0) did not differ significantly from that of the full-term delivery group (32.2 ± 4.8 years). The average gestational age was 31 weeks 2 days ± 4 weeks 3 days for the preterm delivery group and 39 weeks 4 days ± 1 weeks 1 day for the full-term delivery group. The mean birth-weight of infants in the preterm delivery group (1662 ± 688 g) was significantly lower than that of those in the full-term delivery group (3058 ± 411 g).

Vaginal U. urealyticum infection was detected significantly more frequently in the preterm delivery group (51.1%) than in the full-term delivery group (30.0%) (P < 0.02). The rate of mycoplasmal infection was 15.6% in the preterm delivery group and only 6% in the full-term delivery group (P < 0.05). However, the positive rates for yeast-like fungi (13.3% versus 7.0%), group B streptococci (8.9% versus 8.0%), Gardnerella vaginalis (4.4% versus 4.0%), Enterococcus faecalis (8.9% versus 2.0%), Candida (6.7% versus 4.0%) and Escherichia coli (2.2% versus 2.0%) were not significantly different between the two groups (Table 1).

Table 1.  Rates of vaginal infection with U. urealyticum or other bacteria in patients with preterm deliveries and healthy women with full-term deliveries
 Preterm delivery (n = 45)Full-term delivery (n = 100)P value
U. urealyticum51.1% (23)30.0% (30)<0.02
Mycoplasma15.6% (7)6.0% (6)<0.05
Yeast-like fungi13.3% (6)7.0% (7)n.s.
Group B streptococci8.9% (4)8.0% (8)n.s.
Gardnerella vaginalis4.4% (2)4.0% (4)n.s.
Enterococcus faecalis8.9% (4)2.0% (2)n.s.
Candida6.7% (3)4.0% (4)n.s.
Escherichia coli2.2% (1)2.0% (2)n.s.

To examine the correlation between preterm delivery and mixed infection with U. urealyticum and other bacteria, separate models were analyzed as shown in Materials and Methods. There was a sevenfold increase in the risk of preterm delivery in the group infected with U. urealyticum and other bacteria (odds ratio 7.1, 95% confidence interval 2.3–21.7), a threefold increase in the group infected with U. urealyticum alone, and a fourfold increase in the group with other bacterial infections, compared to the control group (Table 2).

Table 2.  Odds ratios and rates of single and joint association among U. urealyticum, other bacteria and preterm delivery
 Preterm deliveryP value
Rate (%)Odds ratio (95% CI)
  1. Figures in parentheses under odds ratio indicate 95% confidence intervals.

U. urealyticum+/other bacteria +(n= 18)55.57.1 (2.3–21.7)<0.001
U. urealyticum+/other bacteria −(n= 35)37.23.3 (1.2–8.8)<0.02
U. urealyticum−/other bacteria +(n= 32)40.63.8 (2.3–6.2)<0.01
U. urealyticum−/other bacteria −(n= 60)15Reference 

All the urease-positive bacterial colonies prepared from the clinical specimens were confirmed as U. urealyticum by PCR analysis. All the U. urealyticum colonies isolated from the preterm delivery group were biovar 1 and those from the full-term delivery group were 90.0% biovar 1 and 10.0% biovar 2, indicating the predominance of biovar 1 in both preterm and full-term deliveries.

Diagnosis of ureaplasmal infection by using anti-ureaplasmal urease monoclonal antibody

mAb against ureaplasmal urease was prepared. Ureaplasmal urease is a homotetramer of subunit of 70 kDa (27). Western blot analysis revealed that our mAb reacted to a protein of about 70 kDa in the U. urealyticum homogenate boiled in SDS containing 2-mercaptoethanol (data not shown). Immunofluorescence staining analysis using this mAb showed positive staining of vaginal secretions obtained from patients infected with U. urealyticum. On the other hand, no staining was observed in vaginal secretions from U. urealyticum-negative cases (Fig. 1).

Figure 1.

Immunofluorescence analysis of vaginal secretions using monoclonal antibody against ureaplasmal urease. Nuclei of the uterine cervix cell and U. urealyticum urease were stained with DAPI (blue) and anti-ureaplasmal urease (red), respectively. Original magnification ×400. (a) Vaginal secretion from U. urealyticum-positive patient.

LP induction of IL-8 production from THP-1 cells and cell death

Since LP is a major bioactive component of the mycoplasmal cell membrane (1, 2, 25), the pathogenic activity of LP as opposed to its parental live bacterium was investigated. THP-1 cells produced undetectable levels of IL-8 under normal culture conditions. Upon stimulation with ureaplasmal LP, the cells started to produce IL-8 at 4 h, and the level was increased thereafter. In response to LPS, the cells promptly produced IL-8 reaching the peak at 8 h (P < 0.05) (Fig. 2).

Figure 2.

Production of IL-8 by THP-1 cells in response to ureaplasmal LP or LPS. THP-1 cells (1 × 106/ml) were incubated in RPMI-1640 in the presence of ureaplasmal LP or LPS for 48 h and IL-8 levels measured by ELISA. Experiments were repeated three times and the average ratios shown as columns. Bars above columns indicate SD. Data were analyzed using Student's t-test. inline image, P < 0.05.

THP-1 cells did not produce IL-1β, IL-6 or TNF-α in response to LP, while they produced significant concentrations of TNF-α in response to LPS (data not shown).

FACS analysis using annexin V-FITC and PI revealed that the proportion of apoptotic cells (annexin V+ PI; lower right) and necrotic cells (annexin V+ PI+; upper right) significantly increased after treatment with LPS or ureaplasmal LP as compared with the untreated cells (P < 0.05) (Fig. 3a, b, c).

Figure 3.

Induction of apoptosis by ureaplasmal LP or LPS. THP-1 cells were incubated in the culture medium only (a), with ureaplasmal LP (b), or with LPS (c) for 24 h and analyzed for apoptosis by FACS following double-staining with annexin V-FITC and PI. Intact cells in the left lower square are both annexin V and PI-negative; apoptotic cells in the right lower square are positive for annexin V only; necrotic cells in the right upper square are positive for both annexin V and PI.

LP upregulates TLR2

THP-1 cells showed unchanged mRNA levels of TLR2 (tlr2) during incubation without LPS or LP for 24 h. On the contrary, in response to ureaplasmal LP, THP-1 cells showed an increase in tlr2 levels and little change was observed in response to LPS (Fig. 4).

Figure 4.

Real-time PCR analysis of TLR2 mRNA in ureaplasmal LP- or LPS- incubated THP-1 cells. THP-1 cells were grown in RPMI-1640 in the presence of ureaplasmal LP (hatched bar) or LPS (closed bars), harvested at the times indicated and analyzed for TLR2 mRNA by real-time PCR.
Experiments were repeated three times and the average ratios shown as columns. Bars above columns indicate SD. Data were analyzed using Student's t-test. inline image, P < 0.05.

DISCUSSION

The relationship between genital colonization with U. urealyticum and pregnancy complications, such as preterm labor and delivery, is controversial (4–6, 28–31). In this study, U. urealyticum was accurately identified by culturing vaginal secretions followed by PCR analysis. U. urealyticum infection rates were found to be significantly higher in the preterm delivery group than in the full-term delivery group (Table 1). Also, biovar 1 infection was found to be much more prevalent than biovar 2 in both groups, indicating that at least biovar 1 of U. urealyticum is associated with the preterm delivery. Thus, vaginal ureaplasmal infection might be an important causative factor for adverse pregnancy outcomes.

Generally, lactobacilli are present in the vagina, keeping the vagina acidic to prevent other bacterial invasion. When U. urealyticum infects the vagina, ureaplasmal urease hydrolyzes urea into carbon dioxide and ammonia. This raises the pH, which facilitates mixed infection with other bacteria. As reported previously, bacterial vaginosis is associated with complications of pregnancy, including premature rupture of membranes, and preterm labor and delivery (30, 32). The results reported here (Table 2) are consistent with these reports. Not only infection solely with U. urealyticum, but also joint infection with other bacteria was significantly related to rate of preterm delivery compared to the control group. Notably, the risk of preterm delivery appeared higher when U. urealyticum was present together with other bacteria in the vagina.

mAb against ureaplasmal urease was raised and made available for immunofluorescence analysis (Fig. 1). To our knowledge, this is the first mAb that is capable of selective detection of U. urealyticum and would therefore be useful for its diagnosis.

Ureaplasmal LP upregulated mRNA expression levels of TLR2 that might recognize this ureaplasmal LP, like mycoplasmal LP (Fig. 4). This suggested a positive circuit between TLR2 concentrations and cytokine production in response to LP, which might plausibly result in the development of serious inflammatory responses. Indeed LP, unlike LPS that lacks capacity to upregulate its receptor TLR4 (data not shown), could induce continuous accumulation of IL-8 (Fig. 2). Ureaplasmal LP induced IL-8 production (Fig. 2), but not TNF-α, IL-1β or IL-6 (data not shown). In parallel it was found that LPS could induce IL-8 (Fig. 2) and TNFα production (data not shown). However, a previous report has shown that both TNF-α and IL-6 are produced by THP-1 cells matured by pholbol 12-myristate 13-acetate after stimulation with live U. urealyticum (33). Thus, LP together with LPS might intensify the inflammatory cascade and the risk of preterm delivery. In fact, the joint effect of U. urealyticum and other bacteria is more intensively related to the rate of preterm delivery than the effect of either of them singly (Table 2).

Intriguingly, the kinetics of IL-8 induction is somewhat different between ureaplasmal LP and LPS. Time-dependent accumulation of IL-8 was observed in the supernatant when THP-1 cells were incubated with LP (Fig. 2). In contrast, IL-8 levels peaked at 8 h and substantially decreased when the cells were stimulated with LPS (Fig. 2). A possible explanation is that LPS-induced IL-8 may be degraded by unidentified mechanisms induced by LPS signaling as well, while LP may not profoundly activate the degradation pathway. Thus, LPS might activate both IL-8 production and degradation pathways, while LP seems to predominantly activate IL-8 production pathway, presumably via its up-regulating TLR2 expressions.

Accumulated lines of evidence indicate that various pathogens, including U. urealyticum and Mycoplasma as described previously (1, 3–6) and in this report, and other bacteria, such as Chlamydia trachomatis (31), trigger preterm delivery by the following scenario: Pathogens induce secretion of diverse proinflammatory cytokines, such as TNF-α, IL-1 and IL-6, and chemokines including IL-8, which in turn evokes production of uterine-contractive prostaglandins by the recruited leukocytes, leading to preterm delivery (8–13).

It has been reported that bacterial lipoproteins and lipopeptides induce apoptosis in macrophages, and that those apoptotic phagocytes might prolong inflammatory responses (17). The present study shows that ureaplasmal LP can also induce apoptosis (Fig. 4). The resultant apoptotic cells might sustain inflammation in the genital tract and promote preterm delivery.

The possibility that contamination with LPS during LP preparation might play a role in its induction of IL-8 production was excluded by incubating the LP preparation with Polymyxin B to inhibit the LPS, and analyzing the ability to induce IL-8 in comparison to untreated LP. No differences between these two preparations were found (data not shown).

In conclusion, U. urealyticum induced TLR2 expression as well as production of IL-8 and cell death. Furthermore, coinfection with other bacteria increased the inflammatory cascade and cytokine production induced by ureaplasmal infection. Collectively, our results suggest that U. urealyticum likely contributes to preterm delivery by these pathogenic mechanisms.

ACKNOWLEDGMENTS

This research work was supported in part by a Grant-in aid for Scientific Research (No.1759765) from the Japan Society for the Promotion of Science, 2005–2007 and a Grant-in-aid of Science Research from Hyogo College of Medicine, 2005–2006.

We thank Dr. T. Tamaoki for help in preparing the manuscript.

Ancillary