Isolation of Separate Ureaplasma Species From Endotracheal Secretions of Twin Patients
Isolation of Ureaplasma spp. from preterm neonates and the association with development of bronchopulmonary dysplasia has been previously investigated. However, few studies have contrasted the nature of infection in twins. In this article, we report that dizygotic twins (1 girl, 1 boy) born at 24 weeks gestation both yielded culturable Ureaplasma from endotracheal secretions. The samples were part of a serial blind collection cohort of ventilated premature neonates, and analysis of repeat cultures showed stable, separate infections over a period of 17 and 21 days, respectively. Immunoblot and probe-specific quantitative polymerase chain reaction analysis determined that Twin 1 was solely infected with Ureaplasma parvum (specifically, serovar 6 by gene sequencing), whereas Twin 2 was solely infected with Ureaplasma urealyticum (specifically, genotype A- serovars 2, 5, and 8 by gene sequencing). Immunoblot analysis found that the major surface antigen (multiple-banded antigen) altered relative mass for both strains during the course of infection. Quantitative polymerase chain reaction analysis of extracted endotracheal aspirates confirmed no evidence of mixed infection for either twin. Failure of sentinel ventilated preterm infants on the same ward to acquire Ureaplasma infection after the first week of birth suggests no cot-to-cot transfer of Ureaplasma infection occurred. This study demonstrated not only a contrasting clinical outcome for a set of twins infected with 2 separate species of Ureaplasma, but also the first real-time demonstration of multiple-banded antigen alteration and evolution of Ureaplasma over the course of a clinical infection.
- CCU —
- color changing unit
- ePTB —
- early preterm birth
- MBA —
- multiple-banded antigen
- NEC —
- necrotising enterocolitis
- PCR —
- polymerase chain reaction
- qPCR —
- quantitative polymerase chain reaction
The Office for National Statistics has reported a steady rate of preterm birth (<37 weeks gestation) of 7.1% to 7.3% since 2009 in the United Kingdom. Those born <28 weeks gestation (early preterm birth [ePTB]) are of greatest clinical significance; 66% of these cases are associated with chorioamnionitis, and Ureaplasma spp. are the most common microorganisms isolated from cases of ePTB.1,2 Furthermore, respiratory Ureaplasma spp. infection is found to be concentrated (up to 65%) in ventilated ePTB postnatally.3 However, the mechanism by which Ureaplasma gain entry to the amniotic sac as well as the differential pathogenicity of the 2 species is still under debate.
Only 2 previous twin studies on intrauterine infection or chorioamnionitis have examined Ureaplasma spp, but preferential infection of the presenting twin has been reported.4,5 In this article, we present a unique case study examining repeated endotracheal samples where each twin was solely infected with a separate species of Ureaplasma, with different clinical outcomes, as well as evidence of bacterial evolution during infection.
Patient Details and Ethical Approval
The mother was fit and well, it was her first pregnancy, and she was a nonsmoker during pregnancy with no abnormal antenatal serology. She presented with spontaneous onset of labor at 24 weeks plus 4 days gestation followed by vaginal delivery of twins (Twin 1, cephalic, girl; Twin 2, breech, male).
Twin 1 was transferred from the birth hospital to the nearest tertiary NICU on day 1. She received amoxicillin and gentamicin, which were stopped after 48 hours when blood cultures taken at birth were negative and serial C-reactive protein measurements were not elevated. She was extubated on day 20 to nasal continuous positive airways pressure after commencing diuretics, but as a result of increasing apneas, bradycardias, and desaturations, she required reintubation on day 23 accompanied by administration of teicoplanin and gentamicin. Twin 1 was extubated on day 31.
Because of high ventilation requirements and recurrent pulmonary hemorrhages, Twin 2 was too unstable to transfer on day 1 and arrived at the NICU on day 5. Similarly to Twin 1, he was started on amoxicillin and gentamicin, but because of poor clinical condition, he was changed to cefotaxime for 3 days until transfer. He then received a 4-day course of ceftazidime and teicoplanin for presumed sepsis on day 8 and an additional 7-day course from day 22 for presumed sepsis, followed by a course of flucloxacillin from day 28. This infant remained on ventilation until his death on day 32. Primary cause of death was determined as necrotizing enterocolitis (NEC) with secondary causes listed as prematurity, chronic lung disease of prematurity, and intracranial hemorrhage. No data on chorioamnioitis was available.
Ethical consent was received and approved by a local ethics board.
Sample Transport and Culture Methods
Endotracheal aspirate samples were taken as part of the UREAtrack study and were transported to University Hospital of Wales (Cardiff, United Kingdom) for analysis. Infectious Ureaplasma spp. titer was determined by culture in Ureaplasma-selective media (Mycoplasma Experience, Surrey, United Kingdom) in triplicate on arrival, while aliquots were frozen (–80°C) for later DNA extraction and quantitative polymerase chain reaction (qPCR) determination of U. parvum, U. urealyticum, and Mycoplasma hominis genomes at Public Health England by published methods.6,7
Analysis of Ureaplasma Isolates
Positive cultures from each sample were subjected to Ureaplasma spp. PCR genotyping analysis, and results were confirmed by sequencing the purified DNA amplicons (Eurofins MWG, Ebersberg, Germany)8 Cultured Ureaplasma isolates from each positive patient sample were also analyzed by previously published immunoblot analysis, using a panel of characterized monoclonal antibodies (Virostat, Portland, ME) that recognize the serovar-determining multiple-banded antigen (MBA)9 that can differentiate between U. parvum and U. urealyticum as well as some serovars.10 MBA protein mass was determined relative to PageRuler protein mass standards (ThermoFisher, Hemel Hempstead, United Kingdom).
Culture and qPCR Data Graphs
Bacterial loads were quantified by both culture and multiplex real-time qPCR. Values for bacterial load over the course of infection followed a similar pattern by each method. Endotracheal aspirate samples from Twin 1 were negative for U. parvum on day 1 (Fig 1), but progressively increased in titer with a maximum load of 104 color changing units (CCU) on day 10. Titers then dropped by day 24 with a final negative result. The first sample for Twin 2 on day 12 was strongly positive for Ureaplasma with 104 CCU, but became increasingly positive with high titers of 107 CCU (2.8 × 105 copies/µl U. urealyticum) until the patient died.
Species Determination of P130 and P131
Analysis methods (primer-probe pairs from qPCR, amplification by species-specific PCR primers, and species-specific monoclonal antibodies by immunoblot analysis) demonstrated that all culture-positive samples from Twin 1 contained only U. parvum (designated strain P130), whereas all culture-positive samples from Twin 2 contained only U. urealyticum (designated strain P131). There was no contamination of any sample by the alternate Ureaplasma species or M. hominis as confirmed by qPCR.
Analysis of the MBA Surface Antigen Over the Course of Infection
Immunoblot analysis using monoclonal antibodies that detect both U. urealyticum and U. parvum (clone 6522; Fig 2A) or specifically U. parvum (clones 6523; Fig 2B) found that the surface MBA protein changed considerably over the course of infection. A single discrete band of ∼80 kDa was initially noted for strain P130 on day 4 postbirth, which was accompanied by a second larger band of 100 kDa by day 6. At day 10, an additional 2 bands were apparent (36 kDa and 50 kDa) and were the predominant MBA forms until day 23. For isolate P131, a single MBA form of 60 kDa was seen on days 12 and 18 postbirth, but this was replaced by 2 bands of 50 kDa and 130 kDa. Amplification of the mba gene by PCR gave results where amplicon length was consistent with protein mass for each isolate (data not shown). The larger 130 kDa band was then present until the final sample before the patient’s death.
We present a case study of dizygotic twins both with Ureaplasma-positive endotracheal aspirate samples, but interestingly harboring 2 separate species. Of particular interest is the observed alteration to the bacterial major surface antigen mass over the course of infection.
For Twin 1, the titers of U. parvum (P130) obtained from samples rapidly increased from the first sample, but a sample taken on day 24 failed to grow Ureaplasma by culture, despite detection of residual bacterial genomic DNA by the sensitive PCR methods. Although this culture-negative result coincided with the administration of teicoplanin and gentamicin, this will not have an impact on the Ureaplasma colonization status because of the pathogen’s intrinsic resistance to these antibiotics.11
The first sample for Twin 2 was not taken until 10 days after birth because poor patient clinical status precluded transport to the research hospital. Although U. urealyticum was not detected in the initial sample, high levels of U. urealyticum were present in a sample taken 3 days later and in all subsequent samples tested until the death of the patient 21 days later as a result of NEC. Ureaplasma spp. infection have been reported to contribute to the development of NEC. Ureaplasma has been detected in gastric aspirate samples from preterm human neonates, and experimental intrauterine Ureaplasma infection in pregnant sheep was found to impair development of the fetal ovine gut in an IL-1–dependent manner.12,13 In addition, Okogbule-Wonodi et al,14 have reported a 2-fold increased NEC prevalence in Ureaplasma-infected neonates born <33 weeks and a 3.3-fold increase in those born <28 weeks with Ureaplasma infection.
The greater pathogenic potential of U. urealyticum relative to U. parvum has been reported among urethritis patients as well as individuals presenting with miscarriage, which is consistent with the worse clinical outcome for Twin 2 compared with Twin 1.15,16
The repeated sampling nature of this study gave a unique opportunity to track the evolution of antigen variability of each Ureaplasma strain throughout the course of the infection. Altered mass of the MBA, caused by insertion or deletion of repeat units in the external portion of the protein, is likely in response to host immune response, especially because immunoglobulin and complement have been shown to be important to Ureaplasma-cidal activity of human sera.13,17 This is the first real-time demonstration of MBA alteration in a clinical infection: the U. parvum MBA from the strain infecting Twin 1 decreased in mass, whereas the U. urealyticum from Twin 2 increased. An increase in molecular weight of the major surface antigen has been shown to be important to evading complement-mediated killing by serum in the closely related species Mycoplasma pulmonis.18 Future studies will collect serum samples from mothers and neonates, in parallel to endotracheal aspirate samples, to determine if MBA alteration coincides with altered serum killing of matched Ureaplasma isolates.
Previous reports have shown discordance in microbial infection between twins. In 1990, Romero et al4 performed amniocentesis on women carrying twins with intact membranes, presenting with preterm labor. Of the 5 pairs with evidence of microbial invasion of the amniotic cavity that subsequently delivered preterm, Ureaplasma spp. was cultured from 3 of these pairs and invariably only from the presenting (never from the nonpresenting) twin. Unfortunately, this study predated the differentiation of Ureaplasma species. In a second study by Mazor et al19 examining microbial invasion of preterm twins, only the presenting twin was infected in 5 out of 9 cases, and the microbial titer was always higher for the presenting twin in the remaining cases where both twins were infected. Ureaplasma spp. were the most commonly observed pathogen in this study. These data suggest that microbial invasion occurs via ascending infection, and a retrospective study on chorioamnionitis and funisitis examining 1156 twins found that dichorionic placentas conferred significant protection against the spread of chorioamnionitis from the presenting to the nonpresenting gestational sac.5
We report the first known case of twins independently infected with different Ureaplasma spp., where different clinical outcomes were observed and both bacterial species showed antigenic evolution across serial samples.
We thank the family of these twins for consenting to the use of their children’s samples in this study, as well as the neonatal unit staff for the help in collection of samples.
- Accepted May 6, 2016.
- Address correspondence to Dr Michael L. Beeton, PhD, Department of Biomedical Sciences, Cardiff School of Health Sciences, Cardiff Metropolitan University, Cardiff, CF5 2YB, United Kingdom. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relative to this article to disclose.
FUNDING: Supported by UREAtrack and the Microbiology and Infection Translational Research Group (MITREG).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Phung DT,
- Blickstein I,
- Goldman RD,
- Machin GA,
- LoSasso RD,
- Keith LG
- Beeton ML,
- Chalker VJ,
- Maxwell NC,
- Kotecha S,
- Spiller OB
- Payne MS,
- Goss KC,
- Connett GJ,
- Legg JP,
- Bruce KD,
- Chalker V
- Teng LJ,
- Zheng X,
- Glass JI,
- Watson HL,
- Tsai J,
- Cassell GH
- Beeton ML,
- Maxwell NC,
- Davies PL, et al
- Abele-Horn M,
- Wolff C,
- Dressel P,
- Pfaff F,
- Zimmermann A
- Zheng X,
- Teng LJ,
- Watson HL,
- Glass JI,
- Blanchard A,
- Cassell GH
- Simmons WL,
- Dybvig K
- Copyright © 2016 by the American Academy of Pediatrics