Differential vaginal expression of interleukin-1 system cytokines in the presence of Mycoplasma hominis and Ureaplasma urealyticum in pregnant women.

OBJECTIVE: The genital mycoplasmas, Ureaplasma urealyticum and Mycoplasma hominis, are commonly identified in the vagina of healthy pregnant women. However, these microorganisms are the most common isolates from the amniotic fluids of women in preterm labor. The mechanisms responsible for vaginal colonization and ascent to the uterus remain undetermined. We evaluated the association between U. urealyticum and M. hominis vaginal colonization and the presence of pro-inflammatory and anti-inflammatory interleukin-1 system components in asymptomatic pregnant women of different ethnicities. METHODS: Vaginal specimens, obtained from 224 first trimester pregnant women, were assayed for interleukin-1beta (IL-1beta) and IL-1 receptor antagonist (IL-1ra) concentrations by ELISA. U. urealyticum and M. hominis vaginal colonization were identified by polymerase chain reaction (PCR). RESULTS: Vaginal colonization with M. hominis was identified in 37 (16.5%) women, and was more prevalent in black (18.9%) and Hispanic (20.9%) than in white (4.2%) women (p = 0.01). U. urealyticum was present in 84 (37.5%) women and there was no ethnic disparity in its detection. M. hominis colonization was associated with elevated median vaginal IL-1beta concentrations in both black women (p = 0.02) and Hispanic women (p = 0.04), and was unrelated to vaginal IL-1ra concentrations. In marked contrast, U. urealyticum colonization was associated with elevations in vaginal IL-1ra levels, but not with IL-1beta concentrations, in black women (p = 0.02) and Hispanic women (p < 0.0001) and marginally in white women (p = 0.06). CONCLUSION: M. hominis colonization in healthy pregnant women is associated with localized pro-inflammatory immune activation, while U. urealyticum colonization is associated with immune suppression.


INTRODUCTION
Mycoplasma hominis and Ureaplasma urealyticum are frequent colonizers of the vagina during preg-nancy [1][2][3][4] . In most women the presence of these microorganisms is without apparent effect on pregnancy outcome. In some women, however, M. hominis and U. urealyticum ascend to the endometrium, either prior to or during pregnancy, where they may contribute to preterm labor and neonatal morbidity. U. urealyticum and M. hominis are the microorganisms most frequently cultured from amniotic fluids of women in preterm labor 5,6 . A recent study demonstrated an association between U. urealyticum in second trimester amniotic fluids and a subsequent increased rate of premature rupture of fetal membranes (PPROM) and preterm labor and delivery 7 .
The host factors associated with M. hominis and U. urealyticum vaginal colonization and migration to the upper genital tract in some women but not in others have not been extensively examined. Ureaplasma urealyticum vaginal colonization in pregnant women has been related to elevated vaginal concentrations of interleukin-1 receptor antagonist (IL-1ra) 4 . IL-1ra is an anti-inflammatory cytokine and a natural competitive inhibitor of the pro-inflammatory cytokine, interleukin-1b (IL-1b) 8 . Anti-microbial cell-mediated immune defenses are initiated by IL-1b production 9 .
The aim of the present study was to perform a direct comparison, in a group of pregnant women of known ethnic backgrounds, of the relationship between M. hominis and U. urealyticum vaginal colonization and vaginal IL-1b and IL-1ra concentrations.

Specimens
Specimens were obtained from the posterior vaginal walls with a cotton swab and deposited into a test tube containing phosphate-buffered saline. The specimens were kept at 4 o C and transported to the laboratory within 3-4 h. In the laboratory as much liquid as possible was extruded from the swabs. Supernatant and pellet fractions were obtained by microcentrifugation and stored separately at -80 o C until tested.

Detection of M. hominis
An aliquot of the pellet fraction from each subject was tested for M. hominis by a published polymerase chain reaction protocol employing primer-pairs specific for a 324 base pair region of the 16S ribosomal RNA gene 10 , except that digoxigenin-labeled dUTP was added to the reaction mixture. To insure specificity, the PCR amplicons were hybridized to a biotinylated oligonucleotide internal probe (5 1 -biotin-GCC CAC CAA GAC TAT GAT GTT TAG-3 1 ) and the complex detected in duplicate by ELISA utilizing streptavidin-coated wells of a microtiter plate and peroxidase-labeled anti-digoxigenin antibody (Roche Diagnostics, Mannheim, Germany). Purified M. hominis was always processed and analyzed in parallel to the test samples as a positive control. H 2 O blanks served as negative controls.

Detection of U. urealyticum
An aliquot of the pellet fraction was analyzed for U. urealyticum by a published PCR protocol 11 , utilizing digoxigenin-labelled dUTP and primer pairs specific for a 429 base pair region of the urease gene. Analysis was as described above for M. hominis, using 5'-biotin-GAG ATA ATG ATT ATA TGT CAG GAT CA-3' as the biotinylated internal probe. Purified U. urealyticum was always processed and analyzed in parallel to the test samples as a positive control. H 2 O blanks served as negative controls.

Vaginal IL-1ra and IL-1b protein concentrations
The thawed supernatant fractions were tested in duplicate for IL-1ra and IL-1b concentrations by commercial ELISA assays (BioSource International, Camarillo, CA). The mean optical density values were converted to ng/ml (IL-1ra) or pg/ml (IL-1b) by reference to a standard curve utilizing purified cytokines. The lower limit of detection of the assay was 4 pg/ml for IL-1ra and 1 pg/ml for IL-1b.

Statistics
Differences in IL-1b and IL-1ra levels between groups were non-randomly distributed and analyzed by the non-parametric Mann-Whitney test. Comparisons between discrete variables were analyzed by Fisher's exact test. A p value 5 0.05 was considered significant.

M. hominis and U. urealyticum colonization and vaginal cytokine levels
The relation between M. hominis or U. urealyticum vaginal colonization and vaginal concentrations of IL-1ra and IL-1b was analyzed. M. hominis colonization was associated with elevated vaginal IL-1b levels in the subjects as a whole (p = 0.007) and in both blacks subjects (p 5 0.02) and Hispanic subjects (p = 0.04), and a markedly decreased IL-1ra to IL-1b ratio. The number of white women who were M. hominis-positive was too few to analyze. In contrast, there was no relation between the presence of M. hominis and vaginal IL-1ra concentration ( Table 2). U. urealyticum colonization was associated with elevated vaginal IL-1ra levels in the study population as a whole (p 5 0.0001), as well as in black women (p 5 0.02) and Hispanic women (p = 0.002) ( Table 3). IL-1ra was also elevated in whites women that were positive for U. urealyticum, but this did not reach statistical significance (p = 0.06). U. urealyticum colonization was unrelated to vaginal IL-1b levels in all groups; the IL-1ra to IL-1b ratio substantially increased in the presence of U. urealyticum. There were no significant differences in the vaginal concentrations of IL-1b and IL-1ra between the different ethnic groups. Among the 17 women positive for both mycoplasmas the median vaginal concentration of IL-1b was elevated (31.2 pg/ml) as compared to those positive only for U. urealyticum (p = 0.02). Similarly, median vaginal IL-1ra levels were higher (550 ng/ml) as compared to those positive only for M. hominis (p = 0.006).
There was no apparent relationship between U. urealyticum and M. hominis vaginal colonization. Among black subjects, U. urealyticum was detected in 7 of 17 (41.2%) women who were positive for M. hominis. Similarly, among Hispanic subjects, 10 of the 18 women (55.6%) with M. hominis were also positive for U. urealyticum. There was no relationship between detection of M. hominis or U. urealyticum and a prior treated bacterial vaginosis or any other infection.

M. hominis and U. urealyticum colonization and pregnancy outcome
The 14 women in our study who delivered preterm (6.3%), were too few to determine if there was an association between pregnancy outcome and mycoplasmal vaginal colonization. However, the rate of preterm birth was highest when M. hominis was present either alone or in conjunction with U. urealyticum. A preterm birth was documented in 7.3% of women who were positive only for U. urealyticum, 11.1% positive only for M. hominis, 11.8% positive for both microorganisms and 6.7% who were negative for both mycoplasmas.

DISCUSSION
The delineation of immune factors associated with vaginal colonization by individual microorganisms in healthy pregnant women has received scant research attention. Since both M. hominis and U. urealyticum have been implicated in   Differences between M. hominis and U. urealyticum in relation to vaginal colonization and pathogenic effects, and which may reflect their differential effect on induction of IL-1 system components, have been noted by other investigators. Carriage of the specific genotype of the polymorphic IL-1ra gene that is associated with elevated IL-1ra production appears to influence the rate of U. urealyticum vaginal colonization 4 , but not M. hominis colonization 12 . In both asymptomatic pregnant and non-pregnant women, U. urealyticum vaginal colonization is much more prevalent than is M. hominis colonization 3,[13][14][15] . This might reflect the capability of U. urealyticum, but not M. hominis, to inhibit the local pro-inflammatory immune response. Conversely, there is an increase in the rate of vaginal M. hominis colonization, but not U. urealyticum colonization, in association with bacterial vaginosis, or Trichomonas vaginalis or C. trachomatis infections 16,17 . Immune system alterations induced by pathogenic microorganisms may alter local immune defense mechanisms and thereby selectively enhance the likelihood of M. hominis colonization and proliferation. It has also been suggested that M. hominis magnifies the effects of other sexually transmitted infections 17 . Our observation that M. hominis is associated with a pro-inflammatory immune response, i.e. elevated vaginal IL-1b concentration, in the vagina of healthy asymptomatic women suggests that the enhanced proliferation of this mycoplasma in conjunction with other infections results in increased vaginal inflammation. It might also be relevant that the ability of M. hominis but not U. urealyticum, to induce the release of high levels of histamines, a potent inducer of inflammation, from rat mast cells has been noted 18 .
The influence of M. hominis and U. urealyticum on neonatal pathology also appears to differ. Tracheal aspirates of preterm infants are frequently colonized with U. urealyticum, without any apparent effect on the rate of bronchopulmonary dysplasia 19 or chronic lung disease 20 . In one study, the presence of U. urealyticum in tracheal aspirates was actually associated with a reduced rate of cystic periventricular leukomalacia 21 . Colonization of the placenta by U. urealyticum was shown to be unrelated to an increased rate of cerebral damage in preterm infants, while the presence of placental M. hominis was associated with a three-fold increased risk of echolucency 22 . The present data support the suggestion by Dammann et al. that U. urealyticum may be a marker for decreased inflammatory capability 22 .
Our previous study 4 and the present investigation noted no racial differences between white and non-white subjects in U. urealyticum colonization rates. In the present study M. hominis was predominantly detected in non-white women. Racial disparities in microbial vaginal colonization have been reported previously 17,23,24 . A prior study of microbial colonization in black and Hispanic women with a concurrent sexually transmitted infection found M. hominis colonization to be associated with the black race 21 . There are two major differences between this investigation and our study. Beside the absence of sexually transmitted diseases in our population, the Hispanic women analyzed in our study were from Puerto Rico and other Caribbean islands, while in the previous study the Hispanics were of Mexican ancestry.
The rate of adverse pregnancy outcome in this cohort of patients was not outside the expected range for this population. The number of patients was insufficient to be able to comment on whether mycoplasmal vaginal colonization was associated with an increased rate of any adverse pregnancy outcome. In addition, our PCR analysis was only qualitative and not quantitative.
Thus, the present study suggests that M. hominis colonization may be recognized by the immune system in the vagina of pregnant women and results in the induction of pro-inflammatory immunity, while U. urealyticum colonization is associated with induction of the anti-inflammatory cytokine, IL-1ra. Further investigations are needed, however, to definitively differentiate between a direct effect of these microorganisms on vaginal cytokine levels or whether M. hominis and/or U. urealyticum colonization is increased under conditions when immunity in the vaginal milieu is altered due to the presence of other infectious agents or non-infectious factors. Conditions that inhibit pro-inflammatory immunity may lead to increased colonization/proliferation by U. urealyticum, and thereby enhance their capacity to ascend to the endometrium and cause clinical disease.