This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Waites, K. B. Articles by Schelonka, R. L. Search for Related Content PubMed PubMed Citation Articles by Waites, K. B. Articles by Schelonka, R. L.

 Previous Article

Clinical Microbiology Reviews, October 2005, p. 757-789, Vol. 18, No. 4
0893-8512/05/$08.00+0     doi:10.1128/CMR.18.4.757-789.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Mycoplasmas and Ureaplasmas as Neonatal Pathogens

Departments of Pathology,¹ Pediatrics, University of Alabama at Birmingham, Birmingham, Alabama 35249²

SUMMARY

INTRODUCTION

MYCOPLASMAL COLONIZATION AND DISEASE IN THE LOWER UROGENITAL TRACT OF ADULTS

MECHANISMS OF PATHOGENESIS

Localization and Cytadherence

Secretory Products

EFFECT ON INFERTILITY AND PREGNANCY OUTCOME

Infertility

Postpartum Endometritis

Chorioamnionitis, Spontaneous Abortion, and Preterm Labor

Bacterial Vaginosis

VERTICAL TRANSMISSION

RESPIRATORY DISEASES IN INFANTS

Congenital and Neonatal Pneumonia

Pneumonia and Other Respiratory Diseases in Older Infants and Children

Association of Ureaplasma spp. with Development of Chronic Lung Disease in Preterm Neonates

SYSTEMIC INFECTIONS IN THE NEONATE

Bacteremia

Infections of the Central Nervous System

Other Infections

DIFFERENTIAL PATHOGENICITY OF UREAPLASMA UREALYTICUM AND UREAPLASMA PARVUM

HOST DEFENSES IN THE NEONATE

OTHER MYCOPLASMAS FOUND IN THE UROGENITAL TRACT OF ADULTS

Mycoplasma fermentans

Mycoplasma genitalium

Mycoplasma penetrans

Mycoplasma pirum

Mycoplasma pneumoniae

LABORATORY DIAGNOSIS

Culture

Nucleic Acid Amplification

Serology

ANTIMICROBIAL SUSCEPTIBILITY

THERAPEUTIC CONSIDERATIONS

Treatment of Respiratory and Systemic Infections

Antimicrobial Treatment for Association of Ureaplasmas with BPD

CONCLUDING REMARKS

REFERENCES

Top Next References

INTRODUCTION

Top Previous Next References

 
The first report of a mycoplasma to be recovered directly from a human and associated with a pathological condition occurred in 1937, when Dienes and Edsall isolated an organism which was probably the one known now as Mycoplasma hominis from a Bartholin's gland abscess (76). At that time, mycoplasmas were called pleuropneumonia-like organisms because the microbe now known as Mycoplasma mycoides had been shown to cause bovine pleuropneumonia. The term mycoplasma (Greek: mykes = fungus and plasma = formed) was first used to describe the pleuropneumonia-like organisms in the 1950s. This designation was initially intended to describe the growth form of M. mycoides, but the term soon gained widespread usage and was applied to all pleuropneumonia-like organisms of human and animal origin identified at that time. Over subsequent years, several other human mycoplasmal species were described, and in 1954 Shepard provided the first description of T-strain mycoplasmas, later known as ureaplasmas, when he was able to cultivate them in vitro from the urethras of men with nongonococcal urethritis (258). The pleuropneumonia-like organisms were not fully differentiated from bacterial L forms until the 1960s, when it was finally proven that mycoplasmas were unable to produce cell walls under any circumstances, making them unique among the prokaryotes.

The class Mollicutes was established in the 1960s to include the mycoplasmas and related organisms and it now contains four orders, five families, eight genera, and more than 200 known species that have been detected in humans, vertebrate animals, arthropods, and plants. Mollicutes for whom humans are the primary host are listed in Table 1; at least 17 well-documented species are now known to occur, primarily localized in the respiratory or urogenital tracts. Several of these species are considered commensals, but three in the genus Mycoplasma are proven pathogens: M. pneumoniae, M. genitalium, and M. hominis. M. fermentans is an organism which may play a role in human disease in some circumstances. Considerable evidence has accumulated in recent years to suggest it may have an etiologic role as an opportunist in persons with human immunodeficiency virus infection and AIDS (8, 9) and a possible association with chronic arthritic conditions (121, 132). Other organisms such as M. penetrans appear to have the potential for being human pathogens (28), but no conclusive proof demonstrating this has been offered to date. The most recent human mycoplasmal species to be recognized is Mycoplasma amphoriforme, an organism that has been detected in the lower respiratory tract of several immunocompromised persons in association with chronic bronchitis, and investigations are now under way to determine whether a role in human disease can be established with certainty (335). Details of the updated taxonomy of the Mollicutes describing their origin from gram-positive ancestors, their phylogenetic relationships with other bacteria, and their biological properties are available in recently published reviews and reference texts (186, 322, 325).

Within a few years following the first descriptions and characterization of Ureaplasma as a human pathogen implicated in nongonococcal urethritis in 1954, there were reports of a possible association of this organism in adverse pregnancy outcomes and low birth weight in neonates. Since then, additional evidence has accumulated implicating ureaplasmas in infertility, postpartum endometritis, chorioamnionitis, spontaneous abortion, stillbirth, premature birth, perinatal morbidity and mortality, pneumonia, bacteremia, meningitis, and chronic lung disease of prematurity, also known as bronchopulmonary dysplasia (BPD). M. hominis has also been implicated in a number of these conditions affecting pregnant women and their offspring.

Many questions remain unanswered about the role of these organisms as human pathogens for a variety of reasons. These include the high prevalence of mycoplasmas in healthy persons; poor design of many of the earlier research studies that attempted to relate the mere presence of these organisms in the lower urogenital tract to pathology in the upper tract or in offspring; failure to consider other multifactorial aspects of some maternal conditions and potential confounders (e.g., bacterial vaginosis); unfamiliarity of clinicians and microbiologists with the complex and fastidious nutritional requirements necessary for in vitro cultivation; and considering these organisms only as a last resort in conditions thought to be most likely due to other microorganisms.

In recent years, detection of several mycoplasmal species in the urogenital tract such as M. fermentans, M. penetrans, and M. genitalium and improved molecular-based detection methods has mandated a reassessment of the possibilities that mycoplasmas and ureaplasmas may be of clinical significance in a variety of urogenital infections affecting pregnant women and neonates, which are the focus of this review. The availability of the complete genome sequence of Ureaplasma parvum (96) and M. genitalium (87) has greatly improved understanding of their basic biology and pathogenic properties. Unfortunately, the genome of M. hominis has not been completely sequenced and annotated as of late 2005, but this project is currently ongoing.

The topic of perinatal mycoplasmal and ureaplasmal infections (collectively referred to as genital mycoplasmas) was last reviewed in Clinical Microbiology Reviews in 1993 (47) and that publication provided a broad and extensively detailed discussion current as of that time. The present review is not meant to be all inclusive; nor is it intended to repeat information presented in depth in the earlier publication. Instead, most attention will be focused on the following topics related to the genital mycoplasmas: (i) recent work describing the epidemiology and establishment of these organisms as causes of neonatal infections and premature birth; (ii) current evidence linking ureaplasmas and BPD; (iii) recent developments in the taxonomy of the genus Ureaplasma and implications for differential pathogenicity of the 2 biovars, now designated as separate species; (iv) the neonatal host response to infection; (v) advances in laboratory detection of mollicutes; and (vi) therapeutic considerations.

MYCOPLASMAL COLONIZATION AND DISEASE IN THE LOWER UROGENITAL TRACT OF ADULTS

Top Previous Next References

 
In order to understand the potential role of genital mycoplasmas in perinatal and neonatal infections and the reasons why many questions about their significance in these settings remain unanswered, it is necessary to gain an appreciation for the epidemiology of these organisms in adult men and women. Ureaplasma spp. can be found on the mucosal surfaces of the cervix or vagina of 40 to 80% of sexually mature asymptomatic women, whereas M. hominis may occur in 21 to 53%. The incidence of each is somewhat lower in the urethra of males. Colonization is linked to younger age, lower socioeconomic status, sexual activity with multiple partners, African-American ethnicity, and oral contraceptive use (47).

There is now ample evidence from clinical studies involving culture, serology, and more recently from PCR assays in humans, and from experimental infection of laboratory animals that these organisms play etiological roles in a variety of urogenital diseases of men and women as summarized in Table 2. For some conditions, such as nongonococcal urethritis for Ureaplasma spp., Koch's postulates have been fulfilled and a portion of clinical cases of these entities are known to be caused by these respective organisms (293). However, attempts to link inflammatory diseases of the upper urogenital tract with isolation of the organisms in the lower tract are not always successful, complicating our understanding of their significance as pathogens in many conditions. Thus, the importance of genital mycoplasmas in other urogenital conditions such as bacterial vaginosis and prostatitis remains open to debate. The discrepancy sometimes observed between the presence of genital mycoplasmas in the lower tract and disease in the upper tract is apparently due to the fact that upper tract colonization and disease occurs in only a subpopulation of persons who are colonized in the lower tract and that the reasons and risk factors for such upper tract involvement are unknown (293).

MECHANISMS OF PATHOGENESIS

Top Previous Next References

 

Localization and attachment on host cell surfaces is important in the ability of mycoplasmas to colonize and subsequently produce pathological lesions, even if cellular internalization does not occur. The steps in the assembly of the multiple proteins comprising the attachment organelle, the process of cytadherence, and release of inflammatory mediators that cause damage to the respiratory epithelium are complex and they have been studied intensively for M. pneumoniae for more than 20 years (44, 129, 158, 159, 239, 283, 298, 325). Other mycoplasmas such as M. genitalium, M. pirum, and M. penetrans also have a flask-shaped morphology and terminal attachment organelles and knowledge of the cytadherence processes of these mycoplasmas is increasing, due to knowledge gained through study of M. pneumoniae cytadherence (23, 129, 230, 239).

Factors involved with the attachment of M. hominis and Ureaplasma spp. to mucosal surfaces have not been extensively characterized. These mollicutes do not have the prominent attachment tips described in the other species mentioned above, but some investigation has been done in this area. Work by Henrich and colleagues (113) led to the identification and initial characterization of cytadherence proteins in M. hominis. They were able to block adherence of mycoplasmas to HeLa cells using homologous monoclonal antibodies, suggesting that specific proteins may be involved in cytadherence. A major adhesin protein, also known as the variable adherence-associated antigen (Vaa), may undergo antigenic variation and assist M. hominis in evasion of host immune defenses (32, 114, 230). The Vaa antigen is expressed in vivo during chronic active arthritis associated with M. hominis infection and is highly immunogenic in the human host (346).

Variation of mycoplasmal cell surface protein antigens at a high rate may facilitate their persistence in invasive sites (44, 48, 348). Similar to what has been described for the Vaa antigen in M. hominis (32, 114, 230), the MB antigen of Ureaplasma spp. undergoes a high rate of size variation in vitro and is variable in size on invasive ureaplasma isolates. Zheng et al. (347) reported that the MB antigen of Ureaplasma spp. contains serovar-specific and cross-reactive epitopes and is a predominant antigen recognized during human infections.

Ureaplasmas are known to adhere to a variety of human cells including erythrocytes (243), spermatozoa (41), and urethral epithelial cells (260). Ureaplasmas bind spontaneously to neutrophils and directly activate the first component of complement (288, 334). Ureaplasma adhesins are proteins expressed on the surface of the bacterial cell. There may be several of them involved in the cytadherence process, which has not yet been characterized in its entirety (243, 262). Pretreatment of HeLa cell monolayers or human erythrocytes with neuraminidase will reduce ureaplasmal adherence, suggesting that the receptors for ureaplasma adhesins are sialyl residues and/or sulfated compounds, similar to what has been observed with M. pneumoniae and other mycoplasmas (325).

Clinically, Ureaplasma spp. have been cultured directly from renal stones, and these organisms have been isolated from voided urine in 31 of 247 patients (13%) with metabolic stones, compared to 43 of 145 patients (30%) with infection stones (P < 0.001). In the patients for whom stone cultures were performed, ureaplasmas were found in 2 of 125 patients (2%) with metabolic stones, compared to 10 of 64 patients (16%) with struvite stones (P < 0.001). These observations strongly suggest that Ureaplasma spp. are linked to the formation of infection stones in the urinary tract, mediated by urease activity (105). The potential pathogenic effect of ureaplasmal urease and its NH₃ metabolic by-product was demonstrated in a mouse model by Ligon et al. (170) in which they were able to demonstrate toxicity of ureaplasmas injected intravenously that was prevented by injection of fluofamide, a potent urease inhibitor.

Information gained through study of the Ureaplasma genome has produced some interesting, albeit somewhat unexpected, findings regarding other potential virulence factors (96). Immunoglobulin (Ig) A1 protease activity was first described in Ureaplasma spp. more than 20 years ago using radiolabeled IgA and was shown to be present in all 14 serotypes and 34 of 35 wild-type strains (237). This serine protease has been examined and characterized further in additional studies. Five ureaplasma clinical isolates obtained from urine, cervix, vagina, amniotic fluid, and synovial fluid along with 13 serotypes were shown to be positive for IgA1 protease activity by Kilian and coworkers (144). Kapatais-Zoumbos et al. reported IgA1 protease activity in 28 isolates of Ureaplasma spp. and speculated that this enzyme may play a role in host specificity in facilitating mucosal colonization since ureaplasmas from nonhuman hosts could not cleave human IgA and human strains could not cleave murine, porcine, or canine IgA (138). This enzyme has therefore been documented in most ureaplasmal strains tested thus far, but it is absent in M. hominis, M. fermentans, and M. pneumoniae (138, 144, 145, 266).

Since IgA is the predominant immunoglobulin secreted at mucosal surfaces, IgA proteases may facilitate colonization by microorganisms by degrading this important component of the mucosal immune system. However, Robertson et al. (237) emphasized that IgA1 protease may not be a significant virulence factor in ureaplasmas because they were able to detect its presence in men with nongonococcal urethritis and with strains isolated from healthy persons. Glass et al. (96) could not identify the gene for IgA1 protease in the genome of U. parvum serotype 3. They speculated that the ureaplasma enzyme may have diverged so far from orthologues in other bacteria they were unrecognizable, or they may have convergently evolved an enzyme with no recognizable similarity to other enzymes. Even though most of the clinical isolates of Ureaplasma spp. evaluated in the studies cited above demonstrated IgA1 protease activity, the extent to which individual ureaplasma strains may lack functional activity of this enzyme is not known.

The presence of phospholipases A and C in Ureaplasma spp. has been suggested to be the means by which ureaplasmas may initiate preterm labor by liberating arachidonic acid and altering prostaglandin synthesis (69-71). Support for this hypothesis comes from studies that have demonstrated significant elevations of phospholipase A2 in serum and amniotic fluid specimens from women in preterm labor with chorioamnionitis than those undergoing term labor (157). De Silva and Quinn (69-71) identified and characterized phospholipase activities in multiple ureaplasma serotypes and reported that the specific activities of phospholipase A2 differed according to serotype, while the activities of phospholipases A1 and C were similar. They speculated that differences in phospholipase activity might cause differences in pathogenic potential for the various serotypes in terms of adverse pregnancy outcomes. However, Glass and colleagues (96) were unable to identify phospholipase activity in the serotype 3 ureaplasma strain for which the complete genome was sequenced and gave the same possible explanations as those for the apparent lack of IgA1 protease activity. Interestingly, Walther et al. (328) were unable to demonstrate phospholipase activity in M. hominis and they were also unable to detect phospholipase A2 coding sequences in DNA analysis (271). These findings suggest M. hominis is not important in the initiation of premature labor through elaboration of phospholipases and stimulation of prostaglandin activity.

The hemolytic activity of M. pneumoniae is due to production of H₂O₂ and is inhibited by catalase (264). In contrast, the hemolytic activity of ureaplasmas is not inhibited by catalase, suggesting an alternative enzyme system may be responsible (96). Glass et al. (96) reported that Ureaplasma parvum has two hemolysins, encoded by the hlyC and hlyA genes. Since hlyC has an orthologue in M. pneumoniae, in which hemolysis is mediated by H₂O₂, hlyA may function as a virulence factor in Ureaplasma spp. (96). Supporting evidence for this concept lies in the fact that orthologues of the hemolysin hlyA mediate hemolytic and cytotoxic activity in other microbes and some mycobacteria which lack this gene are nonpathogenic (96). Antimicrobial resistance as a virulence factor in genital mycoplasmas is discussed in a subsequent section.

EFFECT ON INFERTILITY AND PREGNANCY OUTCOME

Top Previous Next References

 

The same investigators later provided indirect evidence that ureaplasmas may be involved in post-Cesarean delivery endometritis in a study in which 301 women who received doxycycline plus azithromycin were compared to 297 who received a placebo (16). The interesting finding in that investigation was that prophylaxis with antibiotics having activity against ureaplasmas reduced the length of hospitalization, frequency of endometritis, and wound infections. A recent study from Israel (51) detected no difference in prevalence of Ureaplasma spp. in cervicovaginal swabs of women with and without postpartum endometritis, but they were able to detect a difference quantitatively in that more than twice as many women with endometritis had high numbers (>10⁵ CFU) of organisms detected, suggesting an etiological association. To our knowledge, no studies have been performed that have specifically evaluated the role of M. genitalium in postpartum or postabortion fever and endometritis.

Isolation of Ureaplasma spp. but not M. hominis from the chorioamnion has been consistently associated with histological chorioamnionitis and is inversely related to birth weight, even when adjusting for duration of labor, rupture of fetal membranes, and presence of other bacteria (44, 47). These organisms can invade the amniotic cavity and persist for several weeks when fetal membranes are intact and initiate an intense inflammatory reaction in the absence of labor (43, 85, 103). Moreover, ureaplasmas can then be isolated from the chorioamnion and detected in inflamed areas by immunofluorescence (46). Even though these conditions may be clinically silent, these findings are strongly supportive of a causal role for Ureaplasma spp. in chorioamnionitis.

M. hominis rarely seems to invade the chorioamnion and amniotic fluid in the absence of other microorganisms, and data to support an independent role for this mycoplasma in either histological or clinical amnionitis are modest at best. The extent to which the genital mycoplasmas may produce clinical amnionitis is unclear. As discussed above, both can be detected in endometrial tissue and cause postpartum or postabortion fever, sometimes accompanied by bacteremia. Isolation rates of Ureaplasma spp. and M. hominis in symptomatic and asymptomatic women have been similar, but symptomatic women were more likely to develop a serum antibody response (44).

Intrauterine infection is a major cause of preterm labor and can be detected in approximately half of all preterm births, especially those occurring at less than 30 weeks of gestation. Such infections are often subclinical (148). The earlier the gestational age at delivery, the higher the frequency of intra-amniotic infection (98). This relationship is believed to be related to the concept that uterine contractions may be induced by phospholipases produced by microorganisms, as well as cytokines (190). Cytokines elaborated in the amniotic fluid in response to the presence of microorganisms trigger prostaglandin synthesis in the amnion, chorion, decidua, and myometrium, leading to uterine contractions, cervical dilatation, membrane exposure, and greater entry into the uterine cavity (148).

Vaginal carriage of Ureaplasma spp. is not reliably predictive of preterm labor (42), but there is an association when it is present in the amniotic fluid or placenta (43, 125, 163, 339, 342). Two recent reviews of antibiotic trials involving treatment of pregnant women who were culture positive for ureaplasmas in their vaginas concluded that there is insufficient evidence to recommend administration of antibiotics to women with ureaplasmas in the vagina to prevent preterm birth (148, 229). M. hominis and Ureaplasma spp. can be isolated from endometrial tissue of healthy, nonpregnant women, indicating that they may be present at the time of implantation and might therefore be involved in early pregnancy losses (44). Horowitz et al. (122) reported that women whose cervices were culture positive for Ureaplasma spp. and who had a high level of antibody against ureaplasmas were more likely to develop pregnancy complications than women with a negative culture and absence of antibodies. These investigators also reported that women with amniotic fluids that were culture positive for Ureaplasma spp. and who had elevated antibodies were more likely to experience complications including preterm labor, low birth weight, and fetal death than women without antibody against ureaplasmas (124). In view of the fact that accurate methods for measuring antibody against ureaplasmas are not widely available outside of specialized research laboratories, these findings may not have direct clinical relevance at present, but they are interesting nonetheless.

Studies of women from whom ureaplasmas and M. hominis were isolated from the endometrium or placenta have shown a consistent association with spontaneous abortion, but this has not proven true for studies limited to sampling the lower genital tract (293). Joste et al. (136) reported that ureaplasma cultures were positive in 11 of 42 (26%) early spontaneous abortions versus 0 of 21 elective abortions. Other circumstantial evidence linking ureaplasmas with spontaneous abortion, low birth weight, intrauterine growth retardation, and preterm labor includes reports of successful pregnancies following antimicrobial treatment and serological studies (44). Underlying problems that complicate a complete understanding of any potential role for genital mycoplasmas in low birth weight are that M. hominis and to a lesser extent Ureaplasma spp. can be components of the varied flora that occur with bacterial vaginosis, a condition associated with low birth weight (79, 119, 190), and problems in experimental design of studies including failure to consider potential roles for organisms other than mycoplasmas and ureaplasmas or use of control groups of questionable comparability.

Isolation of Ureaplasma spp. in pure culture from amniotic fluid obtained from women with intact fetal membranes who subsequently experienced fetal loss in the presence of histological chorioamnionitis has been documented by multiple investigators, indicating that in some cases this organism has a causal role in spontaneous abortion (43, 85, 103). Support for use of PCR to detect ureaplasmas was presented in a recent investigation which determined that patients with a positive PCR for Ureaplasma spp. but a negative amniotic fluid culture had a higher rate of significant neonatal morbidity than those with a negative culture and negative PCR (P < 0.05). However, no significant differences in perinatal outcome were observed between patients with a negative culture but positive PCR and those with a positive amniotic fluid culture (342). Another recent study (93) found that preterm labor occurred in 58.6% women with a positive PCR assay for ureaplasmas at 15 to 17 weeks of gestation compared with only 4.4% of women with negative PCR results, suggesting the potential value of PCR testing of second trimester amniotic fluid to identify women at risk for preterm labor and delivery.

Patients with preterm premature rupture of membranes and microbial invasion of the amniotic cavity with Ureaplasma spp. experience a robust host inflammatory response in the fetal, amniotic, and maternal compartments (343). Abele-Horn et al. (4) suggested that the density of ureaplasmal colonization is a factor that correlates with adverse pregnancy outcome, including development of chorioamnionitis and preterm delivery. Logistic regression analyses of demographic and obstetric variables indicate that the presence of U. urealyticum alone or with other bacteria in the chorioamnion is independently associated with birth at <37 weeks of gestation regardless of the duration of labor (44). While the association between ureaplasmal chorioamnion infection and premature birth is strong, this association does not prove a cause-and-effect relationship.

Treatment of pregnant women colonized with Ureaplasma spp. with erythromycin or placebo has shown no significant differences in infant birth weight or gestational age at delivery, frequency of premature rupture of membranes, or neonatal outcome (83). On the basis of current evidence, one might have predicted failure of this trial. First, if Ureaplasma is involved in premature birth, it probably produces an effect via intrauterine infection. If only subgroups of pregnant women are at risk, then it is unlikely that a prospective study based on cervical colonization will show an association. Another major consideration is that no information concerning the efficacy of erythromycin for treating intrauterine infections is available. Erythromycin does not effectively penetrate the amniotic sac nor does it eradicate ureaplasmas from the cervix and vagina, probably because of the normally low vaginal pH. Perhaps a more important reason the treatment trial failed is that the majority of women in this study were treated starting at or beyond week 29 of gestation. It is possible that treatment earlier in pregnancy would have been more effective in preventing invasion of the fetal membranes.

Isolation rates of Ureaplasma spp. from the chorioamnion are higher in infants who weigh <1,500 g at birth and are born before 32 weeks of gestation. Since only 1% of women deliver neonates weighing <1,500 g at birth, a very large number of women would have had to be treated to demonstrate a measurable effect. Two recent reviews of published antibiotic trials involving treatment of pregnant women who were culture positive for ureaplasmas in their vaginas concluded that there is insufficient evidence to recommend administration of antibiotics to women with ureaplasmas in the vagina to prevent preterm birth (148, 229). A detailed review of data relating to the role of genital mycoplasmas in preterm birth through the 1990s has been published elsewhere (44).

It is apparent that no single organism causes BV, but an independent association has been found between BV and four groups of vaginal bacteria: G. vaginalis, Mobiluncus spp., anaerobic gram-negative rods, and M. hominis (117). However, the exact role and significance of M. hominis in BV remain uncertain, as other studies have yielded conflicting results. In one study, eradication of G. vaginalis with metronidazole, a drug inactive against M. hominis, cleared nonspecific vaginitis (now known as BV), whereas eradication of M. hominis alone with doxycycline did not, raising doubts over its role in this condition (217). Conversely, relapse of BV after treatment with metronidazole has been attributed to its lack of activity against M. hominis (66), but if the other organisms are eliminated, M. hominis may also disappear. Arya et al. (18) found no role for M. hominis in the epidemiology of BV in a study of 341 women who harbored the organism in their vaginas, while Keane and colleagues (141) detected no difference in the occurrence of M. genitalium and Ureaplasma spp. in women with or without BV, but found M. hominis significantly more often in women with BV. In contrast, another study isolated M. hominis and Ureaplasma spp. from 17% and 53%, respectively, of women with BV, versus 2% and 13%, respectively, of controls (50). A study using PCR for detection of microorganisms found M. hominis at much lower frequencies than G. vaginalis and found no difference in the frequency of its detection in women with or without BV (345). Although Ureaplasma spp. may not be independently associated with BV, the prevalence of vaginal colonization by ureaplasmas may be increased about twofold, and the intravaginal concentration of these organisms may be increased 100-fold (102).

BV occurs in 15 to 20% of pregnant women (190, 191) and this condition has been associated with premature birth. However, the precise relationship among BV, ureaplasmas, and preterm birth is not known. Some have postulated that the increased intravaginal concentrations of BV organisms may result in increases in the synthesis of phospholipase A2 and the production of prostaglandins, which may lead to preterm labor or premature rupture of membranes (83, 190). Alternatively (83, 190) Bacteroides spp. in the lower genital tract could produce enough proteases to weaken the fetal membrane strength, causing premature rupture of membranes and invasion by other organisms. In addition, it is possible that certain BV-associated microorganisms may be more likely to invade the amniotic sac simply because these organisms are present in larger numbers. However, the latter possibility cannot be the total explanation for the association of ureaplasmas with prematurity, since intravaginal concentrations of Peptococcus spp. are also increased in women with BV but are found infrequently in the chorioamnion and amniotic fluid (83, 118, 190, 191, 265).

The presence of BV is independently and significantly associated with birth at <37 weeks of gestation when cervical organisms and obstetric and demographic factors are taken into consideration (118). However, these studies have not determined whether BV is associated with premature delivery independently of chorioamnion infection (with either organisms associated with BV or those that are not). Hillier et al. (118) performed multiple logistic regression to determine the strength of the relation between the recovery of any organism from the chorioamnion and BV. After adjustment for factors related to both BV and the recovery of organisms from the chorioamnion, BV was significantly associated with the isolation of organisms from the chorioamnion. Due to the small numbers of patients it was not possible to determine the effect of individual organisms, to address the question of whether BV is associated with premature delivery independently of chorioamnion infection, or to determine whether chorioamnion infection by Ureaplasma spp. occurred independently of BV.

VERTICAL TRANSMISSION

Top Previous Next References

 
Interest in the epidemiology of genital mycoplasma infections in infants began more than 30 years ago, when the association was made that colonization of newborn infants was inversely related to birth weight (35, 86, 147). Now it is understood that Ureaplasma spp. and M. hominis can be transmitted from an infected females to the fetus or neonate by at least three different routes (316). First, there can be an ascending intrauterine infection in which the organisms gain access to the amniotic sac, where they multiply and are then passed into the fetal lung. This can occur early in pregnancy, even when fetal membranes are intact, and infection can persist for several weeks. Fetal acquisition of these organisms can also occur through a hematogenous route through placental infection through involvement of the umbilical vessels. Ureaplasma spp. have been isolated directly from maternal and umbilical cord blood at the time of delivery (142). Intrauterine infection with Ureaplasma spp. can result in chorioamnionitis, dissemination to fetal organs, and congenital pneumonia (43). Finally, acquisition of these organisms by the neonate can occur through passage of an infected maternal birth canal with resultant colonization of the skin, mucosal membranes and respiratory tract. Ureaplasmas can be isolated from the endotracheal secretions in up to 40% of newborn infants within 30 min to 24 h after birth (45, 128, 207).

Rates of transmission of ureaplasmas from mother to offspring have been the subject of several studies. The isolation of ureaplasmas from neonates will reflect the frequency of maternal colonization in the lower urogenital tract of women in the population studied. Vertical transmission of Ureaplasma spp. has been reported to range from 18 to 88% and isolation rates vary inversely with gestational age, according to most studies (52, 77, 137, 248, 276). Kafetzis et al. (137) recently demonstrated a vertical transmission rate of 60% for infants with a birth weight of 1,000 g versus only 15.3% for infants with birth weights of 1,500 g. These investigators also found that the overall ureaplasma colonization rate was 10% for full-term infants versus 24% of preterm infants. Records from the Diagnostic Mycoplasma Laboratory at the University of Alabama at Birmingham show that Ureaplasma spp. alone were detected in 56 of 307 (18%) sequential endotracheal aspirates cultured from preterm neonates with respiratory distress. Ureaplasma spp. in combination with M. hominis occurred in 27 specimens (9%) and M. hominis alone was identified in 16 specimens (5%). Some investigators have noted that specimens collected on the day of birth for detection of ureaplasmas may not be positive, but subsequent specimens may eventually demonstrate their presence. Bowman and coworkers cultured endotracheal aspirates twice weekly on preterm neonates undergoing mechanical ventilation and determined the average age for a positive culture was 8 days, but some were not positive until the third week or later, perhaps due to a very low initial inoculum (34).

Despite the likelihood that women who are colonized with genital mycoplasmas will transmit them to their offspring, the mere presence of these organisms in surface cultures of neonates is not evidence of pathogenicity. Although the presence of Ureaplasma spp. has been documented for long periods in the lower respiratory tract of preterm infants (45), surface colonization of full-term infants tends to be transient and declines beyond 3 months of age (86). Recolonization of the lower urogenital tract may occur following puberty and when sexual activity is initiated, or if there is sexual abuse (313). However, genital mycoplasmas may occasionally be isolated from the vagina of healthy prepubescent girls (107).

RESPIRATORY DISEASES IN INFANTS

Top Previous Next References

 

Evidence that Ureaplasma spp. are a cause of congenital pneumonia includes: isolation of the organism in pure culture from amniotic fluid; the affected lungs of neonates less than 14 h after birth, and from the chorioamnion (43, 85); demonstration of a specific IgM response in the neonate (225); presence of histological pneumonia and chorioamnionitis in culture-positive neonates and placentas (43, 46, 103, 318); clinical manifestations of respiratory distress in culture-positive infants (45, 204, 318); radiographic changes indicative of pneumonia in culture-positive infants (60, 209); demonstration of the organisms in lung tissue by immunofluorescence (46); electron microscopy (225); and development of rodent (240, 308) and primate (326, 338) models of pneumonia that resemble disease in humans. Although individual case reports suggest M. hominis may cause pneumonia in newborns, it has not been implicated as a common cause in prospective studies (44).

It is not necessary to reiterate all of the details of case reports and retrospective and prospective studies proving ureaplasmas can cause congenital and neonatal pneumonias that have been reviewed by Cassell et al. (47). However, additional supportive data have been forthcoming during the last few years since this topic was last addressed.

The ability of ureaplasmas to incite an inflammatory response in the bloodstream and lower respiratory tract of neonates has been investigated in an attempt to characterize how these organisms can produce pathological lesions when they gain access to the lung. Ohlsson and coworkers (202) observed an elevation in the peripheral leukocyte count, predominantly in the neutrophil component, in preterm infants from whom Ureaplasma spp. were isolated from the lower respiratory tract. Panero et al. (212) correlated isolation of Ureaplasma spp. in pure culture from endotracheal aspirates and/or blood in preterm neonates with total leukocyte counts and radiographic evidence of pneumonia. They determined that Ureaplasma-positive infants had higher mean total leukocyte counts, absolute neutrophil counts, and band form counts, and greater frequency of pneumonia than infants who were culturally negative.

Additional support for the inflammatory potential for ureaplasmas in preterm neonates was provided by Ollikainen et al. (204), who noted that 11 preterm neonates studied within 12 h of birth who were culturally positive for Ureaplasma spp. in the nasopharynx, trachea, and/or bloodstream had significantly higher peripheral leukocyte counts on the first and second days of postnatal life and more often needed high-frequency oscillatory ventilation than 67 neonates who were culturally negative. Horowitz et al. (123) reported that infants from whom ureaplasmas are isolated from endotracheal aspirates within the first 24 h following delivery were more likely to have neutrophils in their tracheal secretions on day 2 than those who are not colonized. In addition to their contribution to the pathological events in acute pneumonitis, increased numbers of neutrophils in the airways are components of chronic inflammatory lung conditions such as BPD, as discussed in subsequent sections. Figure 1 is a photomicrograph of lung tissue collected from autopsy from a neonate who died with pneumonia and sepsis caused by Ureaplasma spp. (318). The tissue reaction shows an extensive and severe inflammatory response with abundant fibrin deposition.

View larger version (200K): [in this window] [in a new window]   FIG. 1. Photomicrograph of lung (magnification 100x) collected at autopsy of a neonate who died at 6 days of age with pneumonia and sepsis due to Ureaplasma spp. (318). Antemortem cultures of blood, pleural fluid, and tracheal secretions and postmortem cultures of nasopharynx, conjunctiva, and brain were positive for Ureaplasma spp. in pure culture. There is extensive pneumonitis, mixed mononuclear and polymorphonuclear infiltrate with abundant macrophages, and fibrin deposition.

Mere isolation from the upper respiratory tract may not accurately reflect the flora of the lower respiratory tract. Syrogiannopoulos et al. (276) studied 108 full-term infants who were colonized with Ureaplasma spp. at birth. They were unable to demonstrate an increased risk of lower respiratory illness during the first 3 months of postnatal life in ureaplasma-colonized infants compared with infants who did not have pharyngeal ureaplasmal colonization. Matlow and coworkers (192) performed a retrospective microbiological evaluation of respiratory tract specimens including lung tissue, bronchoalveolar lavage, lung and endotracheal aspirates, and sputum, nasopharyngeal, and throat specimens obtained from infants and children with various lower respiratory tract diseases. Among 347 specimens, there were 26 culturally positive for Ureaplasma spp. Among 278 nonneonatal specimens, only 5 (1.8%) were positive for ureaplasmas. Four of these five isolates were detected in cultures from either bronchoalveolar lavages or endotracheal aspirates, and other pulmonary pathogens were present simultaneously. They concluded that ureaplasmas are infrequently encountered as agents of respiratory disease beyond the neonatal period and routine culture for them is not recommended.

Davies et al. (65) tested infants under 6 months of age who were hospitalized with an admitting diagnosis of pneumonia, proven radiologically, and compared the microbiological results for a variety of bacterial and viral pathogens with those for infants hospitalized with bronchiolitis. They evaluated the presence of ureaplasmas by culture of nasopharyngeal secretions and found that 4 of 46 (8.7%) of those with pneumonia versus 4 of 66 (6.1%) with bronchiolitis were culture positive for these organisms. In three cases Ureaplasma spp. occurred simultaneously with Chlamydia trachomatis and/or respiratory viruses. It is difficult to make broad conclusions from this study since the authors were making assumptions regarding infection in the lower respiratory tract based on nasopharyngeal cultures and the culture methods that were described did not include agar media specifically designed and proven to support growth of ureaplasmas. These investigators did not detect M. hominis in any of the nasopharyngeal specimens, which is consistent with findings of other prospective studies, even though a few cases of pneumonia in infants have been reported to be caused by this mycoplasma (44).

Very little information exists to indicate what the long-term consequences may be from neonatal infection by Ureaplasma spp. or M. hominis beyond the period of infancy. This problem is confounded by the fact that most neonates with clinically significant respiratory, bloodstream, and/or cerebrospinal fluid infection with these organisms are born preterm and are therefore at much higher risk for long-term sequelae unrelated to the presence of these microorganisms. Ollikainen et al. (203) determined that 22 infants from whom Ureaplasma spp. were detected in blood experienced significantly more hospital stays and remained hospitalized for more days during the first 12 months of postnatal life than 18 infants without infection (546 days versus 188 days) and noted that the differences observed were related to an increase in respiratory tract disease among the infants who were culturally positive for Ureaplasma spp. These findings most likely represent the greater occurrence of long-term respiratory dysfunction among infants colonized with ureaplasmas. Ureaplasmas are also known to cause lower respiratory infections in immunocompromised children and those receiving therapy for malignancies, but these conditions are not known to be associated with the presence of these organisms in the neonatal period (39, 92, 288).

In recent years, considerable attention has been given to the potential role of M. pneumoniae as a cofactor in development or exacerbation of asthma as summarized by Waites and Talkington (325). There is also some recent evidence that colonization or infection of the lower respiratory tract of infants with ureaplasmas may lead to somewhat similar outcomes (25, 161). A Danish study (25) involving 2,927 women determined that maternal vaginal colonization with Ureaplasma spp. during pregnancy was associated with infant wheezing (odds ratio [OR], 2.0; 95% confidence interval [CI], 1.2 to 3.6), but not with asthma, during the fifth year of life.

Viscardi et al. (308) continued work in this area and adapted one of the same murine strains used by Rudd (240) to a juvenile mouse model of ureaplasmal pneumonia. Through this model they were able to characterize an acute and a chronic phase of infection. Pathological effects attributed to the ureaplasmas included focal loss of ciliated respiratory epithelium and increased interstitial neutrophilic infiltrates, presumably due to ureaplasmal adherence and local release of toxic substances such as NH₃ and H₂O₂. The ability of ureaplasmas to adhere to the alveolar epithelial cells of neonatal mice using in situ DNA hybridization was demonstrated by Benstein et al. (26). It is noteworthy that evidence of ureaplasmal infection can be demonstrated in lung tissue by in situ hybridization or culture even when tracheal cultures are negative (27, 327).

Additional information regarding the pathogenicity of ureaplasmas in the neonatal lung was obtained by Walsh et al. (326) through intratracheal inoculation of ureaplasmas into premature baboons delivered by cesarean section. These animals were maintained on 100% oxygen and mechanically ventilated for 6 days. Two animals inoculated with ureaplasmas developed acute bronchiolitis with epithelial ulceration and neutrophil infiltrates, similar to what has been described in ureaplasmal pneumonia in human neonates (225, 318). These lesions were absent in four control animals who were not inoculated with ureaplasmas, but were treated in the same manner otherwise. Ureaplasma spp. were recovered in culture from multiple sites in both infected animals including blood, tracheal aspirates, nasopharynges, pleural fluid, lung, and/or kidney tissue, indicating the organisms were replicating in this primate host.

Yoder et al. (338) expanded the preterm neonatal primate model of ureaplasmal infection by inoculating 10 pregnant baboons intra-amniotically with U. parvum (serovar 1) and studied the offspring that were delivered electively by cesarean section 48 to 72 h later in comparison to animals that were not exposed to intra-amniotic infection with U. parvum. Infant baboons were treated with artificial surfactant, mechanical ventilation and given supplemental oxygen for 14 days until necropsy. Tests for the presence of ureaplasmas and determination of cytokine levels were performed periodically during that time. Experimental findings showed that preterm baboon infants with 48 to 72 h of intra-amniotic exposure to U. parvum had early elevations of tracheal cytokines and leukocytes. Their clinical and radiographic features were consistent with acute pneumonitis. Animals that failed to clear U. parvum from the lower respiratory tract within the first week had greater risk of lung dysfunction and injury than those who eradicated the organisms, similar to what has been observed in human neonates (49). Histopathological examination of the lungs in the infected animals showed more severe bronchiolitis and interstitial pneumonitis compared with uninfected controls. These findings emphasize the importance of the maternal-fetal immunologic response in the outcome of intrauterine ureaplasmal infections.

Three independent reports associating the presence of Ureaplasma spp. in the lower respiratory tracts with progression to BPD, and even death in very low birth weight infants were published in 1988 (45, 247, 329). These studies have stimulated a great deal of additional work in this area in an attempt to understand the true role of these organisms in this clinically important condition. To appreciate why microbial infection may predispose a preterm infant to develop long-term respiratory dysfunction, it is first important to review what is known about the pathophysiology of BPD and the inflammatory potential of microorganisms such as ureaplasmas in the neonatal lung that may be contributory.

Smaller, more immature neonates survive today due to advances in supportive care and mechanical ventilation (165). The increased survival of these vulnerable newborns results in more infants at risk for morbidity due to conditions such as BPD, an entity that was first described by Northway and associates in 1967 (201). BPD occurs almost exclusively in premature infants who received mechanical ventilation. Its incidence varies considerably from one report to another because of differences in patient susceptibility and management practices in different populations and institutions, as well as the definition employed. BPD has been defined as a requirement for supplemental oxygen at 28 days of age or at 36 weeks postmenstrual age, with characteristic radiographic findings (245). The clinical definition of BPD as a supplemental oxygen requirement at 28 days of age which was once widely used has been criticized, especially for extremely low birth weight infants (birth weights of 500 to 750 g) because oxygen need at 28 days may simply reflect lung immaturity. Therefore, oxygen requirement and the presence of radiographic abnormalities at 36 weeks postmenstrual age may be a better predictor of adverse pulmonary outcome. Bancalari et al. (20) have provided an in-depth discussion of issues related to the definition of BPD and how this can affect incidence figures, as well as complicate interpretation of research studies.

Bancalari et al. (20) reported an incidence of BPD ranging from 67% among infants with birth weights of 500 to 750 g to less than 1% in infants weighing 1,250 to 2,500 g. Thus, BPD is now very uncommon in infants born after 32 weeks of gestation. Widespread use of antenatal steroids has reduced the occurrence of severe respiratory distress syndrome in more mature infants and at the same time has led to enhanced survival of more immature infants who are at higher risk for developing BPD (20). Administration of exogenous surfactant has decreased mortality but has not been shown to affect the incidence of BPD independently of other variables (20). The etiology of BPD is multifactorial and complex. Lung tissues of preterm infants lack sufficient surfactant and have incomplete alveolarization to provide an adequate ventilatory surface. The pulmonary immaturity of preterm neonates leads to diffuse microatelectasis and poor compliance. These factors make the immature lungs more susceptible to oxidant injury from supplemental oxygen delivery and volutrauma to the airways during mechanical ventilation. In recent years, an appreciation for the role of inflammation as a consequence of perinatal infection emerged as important in the pathogenesis of BPD, leading the way for consideration of perinatal pathogens such as Ureaplasma spp. as causal factors (182).

Proinflammatory cytokines are believed to play an important role in mediating pathology in a variety of lung diseases, including BPD, through innate and adaptive immune responses. These include interleukin-1ß (IL-1ß), tumor necrosis factor alpha (TNF-), and IL-6. IL-1ß and TNF- activate the immune system, produce inflammation, and induce the release of IL-6, which affects the proliferation of antibody-producing B cells but also limits pulmonary inflammation associated with pneumonia and hyperoxia (187). In the mature immune system, activation of the inflammatory pathway is opposed by the production of cytokines such as IL-10 which down-regulate inflammation and host defense mechanisms in order to protect from an excessively strong response to stimuli. Small amounts of IL-10 in lung lavages of intubated preterm infants with respiratory distress suggests that the immature immune system has a limited ability to down-regulate the inflammatory response (187).

Higher levels of infection-induced amniotic fluid inflammatory cytokines may initiate lung injury in utero and have been associated with higher rates of BPD in preterm neonates delivered to women within 5 days after having amniocentesis to evaluate for infection (340). Moreover, tracheal aspirate inflammatory cytokine concentrations from infants with BPD are elevated in comparison to infants with self-limited respiratory distress syndrome (187). Elevated tracheal cytokines detected in neonates on the first postnatal day has also been associated with prolonged rupture of fetal membranes and histologic chorioamnionitis (187). Dyke et al. (80) found that the presence of Ureaplasma spp. in gastric aspirates of preterm neonates was associated with a significantly greater risk of developing BPD in those infants delivered by cesarean section but not in those who delivered vaginally, suggesting the possibility that longer exposure to inflammation in utero may be the explanation.

A large study of more than 1,600 very low birth weight infants (306) was designed to determine the contribution made by infection in utero versus infection and inflammation beginning after birth on neonatal outcome. These investigators compared rates of BPD in infants who were mechanically ventilated, in infants with histologic evidence of maternal chorioamnionitis, and in infants with postnatal sepsis. Chorioamnionitis alone reduced the risk of BPD, perhaps by inducing maternal corticosteroid production and hastening fetal lung development. However, in infants exposed to maternal chorioamnionitis and who required more than 7 days of ventilation, the risk for BPD was increased. These data suggest that there is a subset of infants who suffer greater damage from infection in utero, or that there is a subset of pathogens that may cause more severe and lasting damage to fetal lung tissue.

Postnatal sepsis and mechanical ventilation for more than 7 days independently increased the risk of BPD, indicating that continuing inflammatory stimuli from infectious or mechanical causes after birth play a role in the development of BPD. Data from studies such as those described above and summarized by Manimtin and coworkers (187) clearly implicate inflammation from perinatal infection with subsequent development of BPD. They further suggest that an imbalance in the neonatal cytokine milieu in response to inflammation could explain the excessive lung damage seen in infants with BPD, whether induced by mechanical ventilation, or by maternal or fetal infection.

Ureaplasma spp. are the most common microbes isolated from infected amniotic fluid, placentas, and the respiratory tracts of preterm infants and their ability to induce inflammation in these sites is undeniable (2, 44, 45, 47, 150, 339). Knowledge of the biology of ureaplasmas and their behavior in the respiratory tract of preterm neonates suggest that lung disease associated with these organisms is not necessarily due to direct damage from the bacteria themselves, but rather because of their potent stimulation of proinflammatory cytokines (TNF-, IL-1ß, and IL-8) or perhaps blockage of counterregulatory cytokines (IL-6 and 1L-10).

Several recent investigations have examined the relationship between ureaplasmal colonization of the neonatal respiratory tract and release of inflammatory mediators that may be involved in pathogenesis of BPD, including clinical studies (154, 214, 309) evaluation of cell lines from humans or rodents cultivated in vitro and exposed to Ureaplasma antigen (59, 166-169, 187), and animal models (308, 338). Ureaplasma spp. colonization of the respiratory tract in neonates has been consistently associated with increases in proinflammatory cytokines in tracheal secretions, including TNF-, IL-1ß, and IL-8 (68, 106, 214, 309). Blocking expression of IL-6 and/or IL-10 has also been reported in association with ureaplasmal colonization (187), although some reports have noted an increase in IL-6 in association with ureaplasmal colonization (154).

Li et al. (167) demonstrated that human and rodent macrophage cell lines exposed to Ureaplasma antigen will produce TNF- and IL-6. This group subsequently provided additional in vitro evidence that Ureaplasma spp. may be involved in the initiation of pathological changes in BPD by demonstrating that a human macrophage cell line exposed to Ureaplasma antigen releases vascular endothelial growth factor and intercellular adhesion molecule 1 (ICAM-1). Vascular endothelial growth factor is involved in pathological changes in the lung that occur in BPD through modulation of angiogenesis, whereas ICAM-1 mediates neutrophil activation and transendothelial migration of leukocytes to sites of inflammation (166). Moreover, the production as well as the expression of ICAM-1 and vascular endothelial growth factor mRNA were inhibited by steroids.

Manimtin et al. (187) have suggested that the alteration of the host inflammatory cytokine response mediated by Ureaplasma spp. occurs in conjunction with a coinflammatory stimulus such as concurrent bacterial infection or hyperoxia. To test this hypothesis, they measured cytokine release in peripheral blood monocytes that were unstimulated versus those stimulated with Ureaplasma antigen alone, and Ureaplasma antigen in combination with lipopolysaccharide (LPS). The interesting findings of this study were that Ureaplasma alone and in combination with LPS induced changes in cytokine release. In vitro inoculation with a low-inoculum partially blocked the LPS-stimulated IL-6 release by all cells and reduced LPS-stimulated IL-10 release by preterm cells; stimulated TNF- and IL-8 release by preterm cells; and augmented LPS-stimulated TNF- release in all cells. In preterm cells, high inoculum of Ureaplasma stimulated TNF- and IL-8, but not IL-6 or IL-10, release; augmented LPS-stimulated TNF- and IL-8 release; stimulated release of all four cytokines in term cells and IL-8 release in adult cells; and augmented LPS-induced TNF-, IL-10, and IL-8 release in term cells but did not significantly affect LPS-induced cytokine release in adult cells. The authors concluded that the failure to stimulate IL-6 might impair organism specific lymphocyte responses, enhancing persistence of ureaplasmas in the lower respiratory tract and continued expression of the inflammatory cascade.

In addition to stimulating release of cytokines, ureaplasmas have been studied for their abilities to stimulate release of other inflammatory mediators. Nitric oxide is a soluble, short-acting free-radical gas produced by a variety of cells that mediates a number of functions involved in the local inflammatory response. Two studies have demonstrated the ability of Ureaplasma to stimulate rodent macrophage cell lines to release nitric oxide (59, 169). Nitric oxide production induced by Ureaplasma can be down-regulated by administration of corticosteroids (169).

Apoptosis of type II pneumocytes and pulmonary mesenchymal cells has been shown to occur as part of the pathogenesis of BPD in preterm infants (168). When lung epithelial cells undergo apoptosis, pulmonary fibrosis can occur as a consequence. Apoptosis of macrophages may also play a role in development of BPD since this would impact their ability to phagocytose neutrophils. Unchecked, the proliferation of neutrophils at the site of lung infection will lead to prolonged inflammation by means of cytokine production and release of proteases and oxygen free radicals (104). Using human macrophage and lung epithelial cell lines, Li et al. (168) have demonstrated that when these cells are stimulated with Ureaplasma antigen, apoptosis will occur in vitro a