2012 · 4 abstract the human ureaplasma species are the most frequently isolated bacteria from the...
TRANSCRIPT
UREAPLASMA PARVUM: UNDERSTANDING THE COMPLEXITIES
OF INTRA-AMNIOTIC INFECTION IN AN OVINE MODEL
Samantha Joan Dando
Bachelor of Applied Science (Honours IA)
A thesis submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
2012
Institute of Health and Biomedical Innovation
School of Biomedical Sciences
Faculty of Health
Queensland University of Technology
2
3
LIST OF KEY WORDS:
Ureaplasma parvum; Ureaplasma urealyticum; intra-amniotic infection; amniotic
fluid; chorioamnion; chorioamnionitis; fetus; pregnancy; sheep; preterm birth;
adverse pregnancy outcomes; erythromycin; macrolide; minimum inhibitory
concentration; 23S ribosomal RNA; genetic variation; in vivo selection; virulent;
avirulent; multiple banded antigen; inflammation; pathogenesis.
4
ABSTRACT
The human Ureaplasma species are the most frequently isolated bacteria from the
upper genital tract of pregnant women and can cause clinically asymptomatic, intra-
uterine infections, which are difficult to treat with antimicrobials. Ureaplasma
infection of the upper genital tract during pregnancy has been associated with
numerous adverse outcomes including preterm birth, chorioamnionitis and neonatal
respiratory diseases. The mechanisms by which ureaplasmas are able to
chronically colonise the amniotic fluid and avoid eradication by (i) the host immune
response and (ii) maternally-administered antimicrobials, remain virtually
unexplored. To address this gap within the literature, this study investigated
potential mechanisms by which ureaplasmas are able to cause chronic, intra-
amniotic infections in an established ovine model.
In this PhD program of research the effectiveness of standard, maternal
erythromycin for the treatment of chronic, intra-amniotic ureaplasma infections was
evaluated. At 55 days of gestation pregnant ewes received an intra-amniotic
injection of either: a clinical Ureaplasma parvum serovar 3 isolate that was sensitive
to macrolide antibiotics (n = 16); or 10B medium (n = 16). At 100 days of gestation,
ewes were then randomised to receive either maternal erythromycin treatment (30
mg/kg/day for four days) or no treatment. Ureaplasmas were isolated from amniotic
fluid, chorioamnion, umbilical cord and fetal lung specimens, which were collected
at the time of preterm delivery of the fetus (125 days of gestation). Surprisingly, the
numbers of ureaplasmas colonising the amniotic fluid and fetal tissues were not
different between experimentally-infected animals that received erythromycin
treatment or infected animals that did not receive treatment (p > 0.05), nor were
there any differences in fetal inflammation and histological chorioamnionitis between
these groups (p > 0.05). These data demonstrate the inability of maternal
erythromycin to eradicate intra-uterine ureaplasma infections. Erythromycin was
5
detected in the amniotic fluid of animals that received antimicrobial treatment (but
not in those that did not receive treatment) by liquid chromatography-mass
spectrometry; however, the concentrations were below therapeutic levels (<10 – 76
ng/mL). These findings indicate that the ineffectiveness of standard, maternal
erythromycin treatment of intra-amniotic ureaplasma infections may be due to the
poor placental transfer of this drug.
Subsequently, the phenotypic and genotypic characteristics of ureaplasmas isolated
from the amniotic fluid and chorioamnion of pregnant sheep after chronic, intra-
amniotic infection and low-level exposure to erythromycin were investigated. At 55
days of gestation twelve pregnant ewes received an intra-amniotic injection of a
clinical U. parvum serovar 3 isolate, which was sensitive to macrolide antibiotics. At
100 days of gestation, ewes received standard maternal erythromycin treatment (30
mg/kg/day for four days, n = 6) or saline (n = 6). Preterm fetuses were surgically
delivered at 125 days of gestation and ureaplasmas were cultured from the amniotic
fluid and the chorioamnion. The minimum inhibitory concentrations (MICs) of
erythromycin, azithromycin and roxithromycin were determined for cultured
ureaplasma isolates, and antimicrobial susceptibilities were different between
ureaplasmas isolated from the amniotic fluid (MIC range = 0.08 – 1.0 mg/L) and
chorioamnion (MIC range = 0.06 – 5.33 mg/L). However, the increased resistance
to macrolide antibiotics observed in chorioamnion ureaplasma isolates occurred
independently of exposure to erythromycin in vivo. Remarkably, domain V of the
23S ribosomal RNA gene (which is the target site of macrolide antimicrobials) of
chorioamnion ureaplasmas demonstrated significant variability (125 polymorphisms
out of 422 sequenced nucleotides, 29.6%) when compared to the amniotic fluid
ureaplasma isolates and the inoculum strain. This sequence variability did not occur
as a consequence of exposure to erythromycin, as the nucleotide substitutions were
identical between chorioamnion ureaplasmas isolated from different animals,
6
including those that did not receive erythromycin treatment. We propose that these
mosaic-like 23S ribosomal RNA gene sequences may represent gene fragments
transferred via horizontal gene transfer. The significant differences observed in (i)
susceptibility to macrolide antimicrobials and (ii) 23S ribosomal RNA sequences of
ureaplasmas isolated from the amniotic fluid and chorioamnion suggests that the
anatomical site from which they were isolated may exert selective pressures that
alter the socio-microbiological structure of the bacterial population, by selecting for
genetic changes and altered antimicrobial susceptibility profiles.
The final experiment for this PhD examined antigenic size variation of the multiple
banded antigen (MBA, a surface-exposed lipoprotein and predicted ureaplasmal
virulence factor) in chronic, intra-amniotic ureaplasma infections. Previously defined
‘virulent-derived’ and ‘avirulent-derived’ clonal U. parvum serovar 6 isolates (each
expressing a single MBA protein) were injected into the amniotic fluid of pregnant
ewes (n = 20) at 55 days of gestation, and amniotic fluid was collected by
amniocentesis every two weeks until the time of near-term delivery of the fetus (at
140 days of gestation). Both the avirulent and virulent clonal ureaplasma strains
generated MBA size variants (ranging in size from 32 – 170 kDa) within the amniotic
fluid of pregnant ewes. The mean number of MBA size variants produced within the
amniotic fluid was not different between the virulent (mean = 4.2 MBA variants) and
avirulent (mean = 4.6 MBA variants) ureaplasma strains (p = 0.87). Intra-amniotic
infection with the virulent strain was significantly associated with the presence of
meconium-stained amniotic fluid (p = 0.01), which is an indicator of fetal distress in
utero. However, the severity of histological chorioamnionitis was not different
between the avirulent and virulent groups. We demonstrated that ureaplasmas were
able to persist within the amniotic fluid of pregnant sheep for 85 days, despite the
host mounting an innate and adaptive immune response. Pro-inflammatory
cytokines (interleukin (IL)-1β, IL-6 and IL-8) were elevated within the chorioamnion
7
tissue of pregnant sheep from both the avirulent and virulent treatment groups, and
this was significantly associated with the production of anti-ureaplasma IgG
antibodies within maternal sera (p < 0.05). These findings suggested that the
inability of the host immune response to eradicate ureaplasmas from the amniotic
cavity may be due to continual size variation of MBA surface-exposed epitopes.
Taken together, these data confirm that ureaplasmas are able to cause long-term in
utero infections in a sheep model, despite standard antimicrobial treatment and the
development of a host immune response. The overall findings of this PhD project
suggest that ureaplasmas are able to cause chronic, intra-amniotic infections due to
(i) the limited placental transfer of erythromycin, which prevents the accumulation of
therapeutic concentrations within the amniotic fluid; (ii) the ability of ureaplasmas to
undergo rapid selection and genetic variation in vivo, resulting in ureaplasma
isolates with variable MICs to macrolide antimicrobials colonising the amniotic fluid
and chorioamnion; and (iii) antigenic size variation of the MBA, which may prevent
eradication of ureaplasmas by the host immune response and account for
differences in neonatal outcomes. The outcomes of this program of study have
improved our understanding of the biology and pathogenesis of this highly adapted
microorganism.
8
LIST OF PUBLICATIONS AND MANUSCRIPTS
The following is a list of manuscripts that have been prepared in conjunction with
this thesis.
Dando SJ, Nitsos I, Newnham JP, Jobe AH, Moss TJM, Knox CL (2010) Maternal
administration of erythromycin fails to eradicate intrauterine ureaplasma infection in
an ovine model. Biol Reprod 83: 616-622.
Dando SJ, Nitsos I, Polglase GR, Newnham JP, Jobe AH, Knox CL (2012) Genetic
variability and antimicrobial resistance of Ureaplasma parvum in response to
maternal erythromycin treatment: a study in pregnant sheep. Manuscript in
preparation.
Dando SJ, Nitsos I, Kallapur SG, Newnham JP, Polglase GR, Pillow JJ, Jobe AH,
Timms P, Knox CL (2012) The role of the multiple banded antigen of Ureaplasma
parvum in intra-amniotic infection: major virulence factor or decoy? PLoS Pathogens
7: e29856.
9
TABLE OF CONTENTS
Title page 1
Keywords 3
Abstract 4
List of publications 8
Table of contents 9
List of abbreviations 14
Statement of original authorship 16
CHAPTER 1 INTRODUCTION 17
1.1 Description of the scientific problem investigated 18
1.2 Specific aims of this study 19
1.3 Progress of research linking the scientific papers 20
1.4 Literature cited 22
CHAPTER 2 LITERATURE REVIEW 23
2.1 Introduction 24
2.2 Historical perspectives and taxonomy 25
2.3 Ureaplasma colonisation of the lower genital tract 26
2.3.1 Female lower genital tract colonisation 26
2.3.2 Male lower genital tract colonisation 28
2.4 In utero ureaplasma infections 29
2.4.1 Routes of in utero infection 31
2.4.2 In utero ureaplasma infection and adverse pregnancy outcomes 33
2.4.3 In utero ureaplasma infection is associated with neonatal sequelae 35
2.4.4 Long term sequelae of in utero ureaplasma infection 36
2.4.5 Ureaplasmas: controversial pathogens? 37
2.5 Virulence factors of Ureaplasma spp. 39
2.5.1 The multiple banded antigen (MBA) 40
2.5.2 Urease 45
2.5.3 IgA protease 46
2.5.4 Phospholipase A and C 47
2.6 The host response to in utero ureaplasma infection 48
10
2.6.1 Innate immunity 49
2.6.2 Adaptive immunity 51
2.7 Antimicrobial treatment of in utero infections 52
2.7.1 Tetracycline treatment of in utero infections 53
2.7.2 Fluoroquinolone treatment of in utero infections 55
2.7.3 Macrolide treatment of in utero infections 56
2.8 Animal models for the study of in utero infections 62
2.9 Concluding remarks 66
2.10 Literature cited 67
Figures
2.1 Phylogenetic tree of selected members of the Mollicutes based on 25 16S rRNA sequence comparison
2.2 Ascending route of infection 32
2.3 Size variability of the multiple banded antigen gene (mba) 41
2.4 Predicted mechanism of MBA phase variation in ureaplasmas 44
Tables
2.1 Comparison of 3’ mba repeat sequences in U. parvum and U. urealyticum 42
2.2 Comparison of outcomes associated with maternal erythromycin treatment 61
of pregnant women
2.3 Comparison of animal models of intra-uterine infection 63
CHAPTER 3 Maternal administration of erythromycin fails to eradicate intrauterine
ureaplasma infection in an ovine model 87
Statement of joint authorship 89
Abstract 90
Introduction 91
Materials and methods 93
Results 98
Discussion 106
Acknowledgements 111
11
References 112
Figures
3.1 Ureaplasma colonisation of amniotic fluid and fetal tissues before 99 and after maternal erythromycin treatment
3.2 Histological chorioamnionitis and fetal inflammation induced by U. parvum 101
3.3 Quantitation of erythromycin within amniotic fluid after maternal 105 erythromycin treatment
Tables
3.1 Fetal measurements at 125 days of gestation 104
CHAPTER 4 Genetic variability and antimicrobial resistance of Ureaplasma parvum in
response to maternal erythromycin treatment: a study in pregnant sheep 117
Statement of joint authorship 119
Abstract 120
Author summary 121
Introduction 122
Materials and methods 125
Results 132
Discussion 142
Acknowledgements 151
References 152
Figures
4.1 Minimum inhibitory concentrations of ureaplasmas isolated from the 135 amniotic fluid and chorioamnion of pregnant sheep
4.2 23S ribosomal RNA gene variation between amniotic fluid and 137 chorioamnion ureaplasmas
4.3 Detection of macrolide resistance genes in clinical U. parvum isolates 141
12
Tables
4.1 PCR primers used for the amplification of 23S ribosomal RNA gene 129 sequences and macrolide resistance genes
4.2 Comparison of minimum inhibitory concentrations and minimum 133 biofilm inhibitory concentrations between ureaplasmas isolated from the amniotic fluid and chorioamnion of pregnant sheep
CHAPTER 5 The role of the multiple banded antigen of Ureaplasma parvum in intra-
amniotic infection: major virulence factor or decoy? 158
Statement of joint authorship 160
Abstract 161
Introduction 163
Materials and methods 166
Results 174
Discussion 192
Acknowledgements 201
References 202
Figures
5.1 Colonisation of amniotic fluid and fetal tissues with virulent and 177 avirulent clonal ureaplasma strains
5.2 Histological chorioamnionitis and fetal inflammation as a result 178 of intra-amniotic ureaplasma infection
5.3 Size variation of the MBA throughout the gestation of pregnancy 181
5.4 Demonstration of a maternal and fetal serum anti-ureaplasma IgG 184 humoral response
5.5 An elevated pro-inflammatory cytokine response was significantly 189 associated with the production of anti-ureaplasma IgG antibodies in pregnant sheep
5.6 Phase variation of the MBA in vitro 191
Tables
5.1 PCR primers used for quantitative reverse transcriptase 170 PCR of selected Toll-like receptors and cytokines
13
5.2 Pregnancy outcomes of pregnant sheep that were intra-amniotically 175 infected with virulent or avirulent ureaplasma clonal isolates
5.3 The molecular weights of ureaplasmal proteins detected by 185 anti-ureaplasma IgG antibodies in maternal serum
5.4 Relative expression of Toll-like receptors and cytokines in the 188 chorioamnion and fetal lung tissue
CHAPTER 6 GENERAL DISCUSSION 210
6.1 Discussion 211
6.2 Conclusions 227
6.3 Future directions 230
6.4 Literature cited 232
Figures
6.1 A proposed model of chronic, intra-amniotic ureaplasma infection 229
Tables
6.1 Comparison of the placental transfer and anti-ureaplasmal activity 215
of Category A antibiotics
14
LIST OF ABBREVIATIONS
442S Clinical U. parvum serovar 3 isolate originally obtained from the semen of an infertile man
AF Amniotic fluid
ANOVA Analysis of Variance
AZM Azithromycin
BLAST Basic Local Alignment Search Tool
bp Base pairs
BPD Bronchopulmonary dysplasia
CAM Chorioamnion
CBC Complete blood count
CFU Colony forming unit
Cmax Maximum concentration
CT Cycle threshold
CSF Cerebrospinal fluid
d Days of gestation
DAB 3’, 3’-diaminobenzidine tetrahydrochloride
E22 5.8.1 Clonal avirulent U. parvum serovar 6 isolate
E24 3.2.1 Clonal virulent U. parvum serovar 6 isolate
erm(B) Erythromycin ribosome methylase-B gene
ERY Erythromycin
GBS Group B Streptococcus
H&E Haematoxylin and eosin
HGT Horizontal gene transfer
HRP Horse radish peroxidase
IgA Immunoglobulin A
IgG Immunoglobulin G
IgM Immunoglobulin M
IL Interleukin
IM Intra muscular
IV Intra venous
15
IVF In vitro fertilisation
KDa Kilo Daltons
LC-MS Liquid chromatography-mass spectrometry
LPS Lipopolysaccharide
M 10B medium group
mba Multiple banded antigen gene
MBA Multiple banded antigen
MBIC Minimum biofilm inhibitory concentration
M/E 10B medium + erythromycin group
MIC Minimum inhibitory concentration
msr(A, B, C, D) Macrolide streptogramin resistance gene
PCR Polymerase chain reaction
PPROM Preterm prelabour rupture of membranes
ROX Roxithromycin
rRNA Ribosomal RNA
SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
SEM Standard error of the mean
SP-A Surfactant protein A
tet(M) Tetracycline resistance gene
TNF Tumour necrosis factor
TLR Toll-like receptor
Up Ureaplasma only group
Up/E Ureaplasma + erythromycin group
UU376 Surface exposed lipoprotein adjacent to the MBA
V-1 Variable antigen 1 of Mycoplasma pulmonis
16
STATEMENT OF ORIGINAL AUTHORSHIP
The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institute. To the best
of my knowledge and belief, the thesis contains no material previously published or
written by another person, except where due reference is made
_________________________
Samantha Dando
17
Chapter 1
INTRODUCTION
18
1.1 DESCRIPTION OF THE SCIENTIFIC PROBLEM INVESTIGATED
According to the latest perinatal statistics, 7.4% of babies born in Australia (Laws et
al. 2010) and 12.3% of babies born in the US (Hamilton et al. 2010) are delivered
preterm. Infection of the amniotic fluid and fetal membranes during pregnancy is a
major risk factor for preterm birth and neonatal morbidity and mortality (Bibby and
Stewart 2004). Of preterm births, 30% are predicted to occur due to infection of the
upper genital tract during pregnancy; however, this may be a conservative figure as
a number of microorganisms, which cause intra-amniotic infections, have fastidious
nutritional requirements and are difficult to detect by conventional microbiological
culture (Goldenberg et al. 2008).
The human Ureaplasma spp. (U. parvum and U. urealyticum) are the most
frequently isolated microorganisms from infected amniotic fluids and placentas and
have been associated with numerous adverse pregnancy outcomes and neonatal
sequelae (Cassell et al. 1993). Unlike other microorganisms, which can cause
rapidly fatal intra-amniotic infections, ureaplasmas are capable of causing chronic,
asymptomatic infections of the amniotic fluid. Although clinically silent, intra-amniotic
ureaplasma infections have been associated with histological chorioamnionitis,
funisitis, preterm birth and fetal death (Cassell et al. 1983). Due to the sub-clinical
nature of these infections, ureaplasmas are not routinely screened for during
pregnancy, nor are they suspected as aetiological agents of upper genital tract
infections. However, there is overwhelming evidence that ureaplasmas are the
microorganisms most associated with preterm delivery and chorioamnionitis
(Viscardi 2010).
Ureaplasmas are wall-less prokaryotes with minimal genomes (Glass et al. 2000)
and appear to be highly adapted to the urogenital tract. Remarkably, these
microorganisms are able to chronically colonise the amniotic fluid despite the
19
development of a maternal/fetal immune response and targeted antimicrobial
therapies. This suggests that ureaplasmas may have highly evolved mechanisms,
which enable them to persist long-term in utero; however, these mechanisms have
not been characterised. Therefore, the overall objective of this PhD project was to
characterise potential mechanisms by which ureaplasmas are able to cause
chronic, intra-amniotic infections and evade eradication by the host immune
response and antimicrobial treatment. The hypotheses of this study were that (i)
current treatment options are ineffective due to the poor placental transfer of
antibiotics, which may promote the emergence of antimicrobial resistant strains; and
(ii) the host immune system is unable to eliminate ureaplasmas from the amniotic
cavity due to antigenic variation of the multiple banded antigen (MBA, a
ureaplasma-specific, surface-exposed lipoprotein). These hypotheses were
investigated using an ovine model of chronic, intra-amniotic ureaplasma infection.
By investigating these aspects of chronic, intra-amniotic ureaplasma infections, this
study may improve our understanding of ureaplasmal pathogenesis and inform
improved therapeutic options.
1.2 SPECIFIC AIMS OF THE STUDY
1. To investigate the efficacy of maternally-administered erythromycin in
eradicating chronic, intra-amniotic U. parvum infection in a sheep model
(Chapter 3).
2. To determine if standard erythromycin treatment of chronic, intra-amniotic
ureaplasma infections can induce genetic markers of macrolide resistance in
amniotic fluid and chorioamnion ureaplasma clinical isolates, resulting in
changes to antimicrobial susceptibility profiles (Chapter 4).
3. To determine if MBA size variation is associated with the virulence of clonal
ureaplasma strains and if variable expression of this surface-exposed
20
antigen enables ureaplasmas to avoid eradication by the host immune
response (Chapter 5).
1.3 PROGRESS OF RESEARCH LINKING THE SCIENTIFIC PAPERS
The three papers presented in this thesis are directly linked to the topic of chronic,
intra-amniotic ureaplasma infection. Combined, these papers advance our
understanding of the mechanisms by which ureaplasmas are able to cause long-
term in utero infections in an established ovine model.
Erythromycin is the standard antimicrobial used for the treatment of intra-amniotic
infections and preterm prelabour rupture of membranes. However, researchers and
clinicians are not in agreement as to whether maternal erythromycin treatment is
able to effectively eradicate microorganisms from the amniotic cavity. In Chapter 3,
the ability of maternally-administered erythromycin to eradicate intra-amniotic
ureaplasma infections was investigated in pregnant sheep. Erythromycin treatment
failed to eradicate intra-uterine ureaplasma infection or reduce ureaplasma
colonisation and fetal inflammation. Quantitative liquid chromatography-mass
spectrometry analysis of amniotic fluid samples demonstrated that erythromycin
was present in low concentrations within the amniotic fluid of treated animals,
suggesting that this antimicrobial was not effectively transported across the
placental barrier.
Due to the limited placental transfer of erythromycin, microorganisms present within
the amniotic fluid may be exposed to sub-inhibitory concentrations of antimicrobials,
which may promote the emergence of antibiotic resistance (Zhanel 2005).
Therefore, in Chapter 4, the effects of standard erythromycin treatment (resulting in
sub-inhibitory concentrations within the amniotic fluid) on genotypic and phenotypic
markers of macrolide resistance in ureaplasmas were investigated. Chronic, intra-
21
amniotic infection with a single U. parvum clinical isolate resulted in amniotic fluid
and chorioamnion isolates with variable minimum inhibitory concentrations to
macrolide antimicrobials. The genetic mechanisms of macrolide resistance were
investigated by polymerase chain reaction and sequencing. Significant genetic
variability was found in the 23S rRNA gene of chorioamnion ureaplasma isolates,
but surprisingly, not in the same gene of amniotic fluid ureaplasma isolates. The
23S rRNA sequence variability within chorioamnion ureaplasma isolates occurred
independently of exposure to erythromycin in vivo. Therefore it was suggested that
the anatomical site of infection and the associated microenvironment exert selective
pressures that result in the selection of ureaplasma sub-populations in utero.
To address the second component of the overall hypothesis of this study, the role of
the MBA was investigated in chronically infected pregnant sheep (Chapter 5). Serial
amniocenteses were performed to collect amniotic fluid from 55 days of gestation
until the time of surgical delivery of the fetus (140 days of gestation, term = 150
days). Ureaplasmal MBA size variation occurred in all experimentally-infected
animals, as demonstrated by western blot, regardless of the intensity of the innate
and adaptive immune responses. This suggests that ureaplasmal MBA size
variability does not prevent recognition by host pattern recognition receptors. Size
variation of the MBA of clonal U. parvum strains did not correlate with different
severities of histological chorioamnionitis, although subtle differences in fetal
outcomes were observed between animals infected with clonal ureaplasma strains.
The generation of numerous MBA variants throughout gestation provided evidence
that size variability of this surface-exposed antigen may prevent the host immune
response from eradicating ureaplasmas from the amniotic cavity.
Taken together, the results presented in Chapters 3, 4 and 5 demonstrate that
ureaplasmas have evolved sophisticated mechanisms to establish and maintain
clinically asymptomatic in utero infections.
22
1.4 LITERATURE CITED
Bibby E, Stewart A (2004) The epidemiology of preterm birth. Neuro Endocrinol Lett 25: 43-47.
Cassell GH, Davis RO, Waites KB, Brown MB, Marriott PA, Stagno S, Davisk JK (1983) Isolation of Mycoplasma hominis and Ureaplasma urealyticum from amniotic fluid at 16-20 weeks of gestation: potential effect on outcome of pregnancy. Sex Transm Dis 10: 294-302.
Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R (1993) Ureaplasma urealyticum intrauterine infection: role in prematurity and disease in newborns. Clin Microbiol Rev 6: 69-87.
Glass JI, Lefkowitz EJ, Glass JS, Heiner CR, Chen EY, Cassell GH (2000) The complete sequence of the mucosal pathogen Ureaplasma urealyticum. Nature 407: 757-762.
Goldenberg RL, Culhane JF, Iams JD, Romero R (2008) Epidemiology and causes of preterm birth. Lancet 5: 75-84.
Hamilton BE, Martin JA, Ventura SJ (2010) Births: Preliminary Data for 2008. National Vital Statistics Reports 58:1-17. Laws PJ, Li Z, Sullivan EA (2010) Australia’s mothers and babies. Perinatal statistics series. Australian Institute of Health and Welfare. Available at: http://www.aihw.gov.au/publication-detail/?id=6442472399. Viscardi RM (2010) Ureaplasma species: role in diseases of prematurity. Clin Perinatol 37: 393-409.
Zhanel GG (2005) Antibacterial drivers of resistance. Treat Respir Med 4: 13-18.
23
Chapter 2
LITERATURE REVIEW
24
2.1 INTRODUCTION
The genus Ureaplasma contains seven host-specific species (U. parvum, U.
urealyticum, U. canigenitalium, U. cati, U. diversum, U. felinum and U. gallorale).
The Ureaplasma spp. are members of the class Mollicutes, and are classified within
the order Mycoplasmatales (originally proposed by Edward and Freundt 1956) and
the family Mycoplasmataceae (Brown 2010). The Ureaplasma spp., which infect
human hosts (U. parvum and U. urealyticum, commonly referred to as ‘the
ureaplasmas’) are closely related to Mycoplasma spp. and are phylogenetically
clustered in the M. pneumoniae group (Figure 2.1). Ureaplasmas and mycoplasmas
are unique bacteria as they lack a cell wall and are bounded only by a plasma
membrane. These free-living microorganisms are also characterised by small
genomes and high A+T content (Glass et al. 2000). The minimal genomes of
ureaplasmas and mycoplasmas are thought to have arisen by degenerative
evolution from low G+C Gram positive bacteria (Maniloff 1983). Phlyogenetic
analysis of rRNA sequences suggested that Clostridium innocuum and Clostridium
ramosum are the closest relatives of the ureaplasmas and mycoplasmas (Woese et
al. 1980; Rogers et al. 1985; Olsen et al. 1994). Woese et al. (1980) also
determined that ureaplasmas and mycoplasmas are peripherally related to Bacillus
spp., Lactobacillus spp., and Streptococcus spp. However, more recently Wolf et al.
(2004) demonstrated that Streptococcus spp. and Lactobacillus spp. may be the
closest relatives of the ureaplasmas and mycoplasmas based on comparative
phosphoglycerate kinase sequencing.
Ureaplasmas are phenotypically distinguished from Mycoplasma spp. by their ability
to hydrolyse urea to produce 95% of their ATP requirements (Smith et al. 1993;
Glass et al. 2000). Urea hydrolysis by the urease enzyme causes the production of
ammonia, which results in an increase in proton electrochemical potential and de
novo ATP synthesis (Smith et al. 1993). In primary culture, ureaplasma colonies are
25
FIGURE 2.1: Neighbour joining phylogenetic tree of selected members of the Mollicutes based on 16S
rRNA sequence comparison (Sirand-Pugnet et al. 2007). A bootstrap of 500 replicates was performed; the number indicated on each node represents the percentage with which each branch topology was supported. Candidatus phytoplasma asteris (Onion Yellows strain) and Aster Yellows phytoplasma were used as the outgroup species. H = hominis cluster, P = pneumoniae cluster, S = spiroplasma cluster, M = mycoides cluster. In this figure, Ureaplasma urealyticum is representative of the 14
ureaplasma serovars.
significantly smaller than mycoplasma colonies, and usually range from 5 µm to
20µm in diameter (Shepard 1956). Ureaplasmas are pleomorphic, due to the lack of
structural integrity provided by a cell wall, and individual bacterial cells typically
range in size from 100 nm to 1 µM (Shepard and Masover 1979). Interestingly, the
ureaplasmas are the only free-living bacteria that lack the cell division FtsZ protein,
which forms a constricting ‘Z’ ring between dividing cells (Glass et al. 2000).
Therefore, the genetic mechanism of cell division in ureaplasmas is currently
unknown. Similar to most eubacteria, Mycoplasma spp. reproduce by binary fission,
but cytoplasmic division frequently lags behind genome replication resulting in the
formation of multinuclear filaments (Razin 1996). In contrast, cell division of
ureaplasmas is predicted to involve budding of daughter cells to produce single
cells, pairs, small aggregates or filamentous elements (Shepard et al. 1974).
2.2 HISTORICAL PERSPECTIVES AND TAXONOMY
26
Ureaplasmas were first discovered in 1954 in agar cultures of urethral exudates
from male patients with non-gonococcal urethritis (Shepard 1954). They were
initially identified as tiny-form pleuropneumonia-like organisms and subsequently
referred to as T-mycoplasmas. Based on the unique urease activity of T-
mycoplasmas, Shepard et al. (1974) proposed that a separate genus within the
Mycoplasmataceae family should be established for the classification of these
microorganisms. Ureaplasma urealyticum was proposed as a single human species
containing eight antigenically distinct serovars. The number of recognised U.
urealyticum serovars was increased to 14 after Robertson and Stemke (1982)
demonstrated further unique specificities of antisera generated against human
isolates by metabolic inhibition tests and colony indirect epifluorescence assays.
The 14 serovars of U. urealyticum were divided into two distinct biovars, the parvo
biovar and the T960 biovar, based on DNA-DNA hybridisation homology
(Christiansen et al. 1981), restriction endonuclease cleavage patterns (Razin et al.
1983), polyacrylamide gel electrophoresis of cellular proteins (Swenson et al. 1983),
sequences of 16S rRNA, urease and multiple banded antigen genes (Kong et al.
1999a; Teng et al. 1994; Knox et al. 1998) and genome size (Kakulphimp et al.
1991). Based on the accumulation of phenotypic and genotypic evidence
suggesting significant differences between the parvo and T960 biovars, a
reclassification of U. urealyticum into two separate species was proposed
(Robertson et al. 2002). Serovars 1, 3, 6 and 14 were regrouped into a new species:
U. parvum; and U. urealyticum was emended to include serovars 2, 4, 5 and 7-13.
This remains the current accepted classification system for the human ureaplasmas,
although this nomenclature has not been consistently adopted within the literature.
2.3 UREAPLASMA COLONISATION OF THE LOWER GENITAL TRACT
2.3.1 FEMALE LOWER GENITAL TRACT COLONISATION
27
Ureaplasmas can be isolated from the mucosal surfaces of vagina or cervix from
40-80% of sexually active females (Cassell et al. 1993). Ureaplasma colonisation of
the lower genital tract has been associated with numerous factors including African-
American ethnicity (McCormack et al. 1986a), the number of recent sexual partners
(McCormack et al. 1986b; Nelson et al. 2007) and the use of non-barrier
contraceptives (Knox et al. 1997). Ureaplasmas are considered to be commensal
microorganisms of the lower genital tract of women. In a study of 162 women, the
rates of ureaplasma colonisation from urethral and cervical swabs were not different
between symptomatic women attending a venereal disease clinic (64 out of 85,
74%), or women with no urogenital symptoms and normal findings at pelvic
examination (55 out of 77, 71%, Møller et al. 1985). Similarly, Casari et al. (2010)
reported that there were no differences in the rates of endocervical ureaplasma
colonisation between women with symptoms of genital tract infection (27 out of 556,
4.86%) and asymptomatic women (15 out of 396, 3.79%). U. parvum is consistently
isolated more frequently from the lower genital tract of females (81-96%, Abele-
Horn et al. 1997; Knox and Timms 1998; Kong et al. 1999; Kong et al. 2000) than U.
urealyticum, and serovar 3 is the most common serovar isolated from both males
and females in Australia and the United States (Knox and Timms 1998; Cassell et
al. 1993).
Although considered to be commensals of the female lower genital tract, the
ureaplasmas have been associated with symptomatic vaginitis (Zdrodowska-
Stefanow et al. 2006a; De Francesco et al. 2009), urinary tract infections
with/without pyuria (Ganzàez-Pedraza et al. 2003; Latthe et al. 2008; Reyes et al.
2009) and bacterial vaginosis (Hillier et al. 1993; Haggerty et al. 2009). Ureaplasma
colonisation of the female lower genital tract has also been identified as a risk factor
for preterm delivery in pregnant women (Abele-Horn et al. 2000; Vogel et al. 2006;
Harada et al. 2008). In a study of 877 women, U. parvum and U. urealyticum were
28
detected in 52% and 8.7% of vaginal swabs respectively. U. parvum colonisation
was detected in 16 out of 21 women (76.2%) who delivered preterm, and was
identified as a risk factor for late abortion or early preterm birth (Kataoka et al.
2006). However, in this same study U. parvum was detected in vaginal swabs
collected from 440 out of 856 women (51.4%) who delivered term babies. Similarly,
Breugelmans et al. (2010) isolated Ureaplasma spp. from cervical swabs of 52 out
of 97 women (53.6%) who delivered preterm, and from 783 out of 1891 women
(41.1%) who delivered at term. Although these two studies suggested that lower
genital tract ureaplasma colonisation was associated with preterm birth, the data are
confounded due to: (i) large inequalities in the size of preterm delivery and term
delivery groups; and (ii) the high levels of detection of ureaplasmas from women
who delivered at term. Other studies have reported that positive vaginal or cervical
ureaplasma cultures were not associated with spontaneous preterm birth or low
birth weight (Lee et al. 2009; Donders et al. 2009). Therefore, it is still generally
accepted that lower genital tract ureaplasma colonisation is not a significant
predictor of preterm birth.
2.3.2 MALE LOWER GENITAL TRACT COLONISATION
Ureaplasmas can be present as asymptomatic colonisers of the urethra in up to
50% of males (Volgmann et al. 2005). Ureaplasma colonisation of the lower genital
tract has also been associated with non-gonococcal urethritis in the absence of
other microorganisms (Yoshida et al. 2005; Zdrodowska-Stefanow et al. 2006b;
Couldwell et al. 2010) and chronic prostatitis (Skerk et al. 2002; Badalyan et al.
2003). David Taylor-Robinson confirmed that ureaplasmas were aetiological agents
of non-gonococcal urethritis after he inoculated his own urethra with U. urealyticum
serovar 5 and subsequently experienced dysuria, frequency of urination and pyuria
(Taylor-Robinson et al. 1977). Numerous studies have also detected ureaplasmas
in the seminal fluid of both fertile (de Jong et al. 1990; Wang et al. 2005) and
29
infertile men (Zeighami et al. 2009; Golshani et al. 2007; Gdoura et al. 2007;
Gdoura et al. 2008) and it has been suggested that ureaplasmas may be associated
with infertility.
The presence of ureaplasmas within semen has been associated with andrology
outcomes that may adversely affect fertility. These include: increased or decreased
sperm motility (Naessens et al. 1986; Rose and Scott 1994; Nύñez-Calonge et al.
1998; Reichart et al. 2001; Knox et al. 2003; Golshani et al. 2007); reduced sperm
concentration (Upadhyaya et al. 1984; Wang et al. 2006; Golshani et al. 2007);
reduced concentrations of seminal plasma immunosuppressive factors and semen
pH (Wang et al. 2005); a decrease in the inducibility of the acrosome reaction (Köhn
et al. 1998) and sperm chromatin decondensation and DNA damage (Reichart et al.
2000). Additionally, ureaplasmas have been shown to remain attached to the
surface of spermatozoa after standard assisted reproductive technology semen
washing procedures (Knox et al. 2003). It is predicted that ureaplasmas are able to
attach to spermatozoa by binding to sulfogalactoglycerolipid, which is a component
of the germ cell membrane (Lingwood et al. 1990). Despite these findings, there are
studies which suggest that ureaplasmas are not associated with sperm impairment
and infertility (de Jong et al. 1990; Martens et al. 1993; Andrade-Rocha 2003).
Therefore, the role of these microorganisms in infertility is controversial (Waites et
al. 2005) and requires further investigation in both fertile and infertile couples.
2.4 IN UTERO UREAPLASMA INFECTIONS
Ureaplasmas are the most frequently isolated microorganisms from the amniotic
fluid (Yoon et al. 1998; Yoon et al. 1999; Gerber et al. 2003; Perni et al. 2004) and
placentas of pregnant women (Kundsin et al. 1984; Hillier et al. 1988; Gray et al.
1992; Cassell et al. 1993). Ureaplasmas have been detected in the amniotic fluid of
pregnant women as early as the 16th week of pregnancy, in the presence of intact
30
fetal membranes and in the absence of other microorganisms (Cassell et al. 1983).
Furthermore, it has been demonstrated that ureaplasmas can cause clinically silent
intra-amniotic infections, associated with histological chorioamnionitis and funisitis,
which can persist for as long as two months in humans (Cassell et al. 1983). Due to
the clinically asymptomatic nature of intra-amniotic ureaplasma infections and the
fastidious growth requirements of these microorganisms, pregnant women are not
routinely screened for ureaplasmas and they are often not suspected as aetiological
agents of upper genital tract infection.
The amniotic fluid is a proteinaceous biological fluid (Tsangaris et al. 2006), which
undergoes dynamic change throughout pregnancy. Early in gestation, the protein
composition of amniotic fluid resembles that of maternal serum (albeit at lower
concentrations, Gao et al. 2009); however, fetal urine is a major source of amniotic
fluid in the second half of pregnancy (Modena and Fieni 2004). A comprehensive
proteomic analysis (Michaels et al. 2007) demonstrated that the human amniotic
fluid proteome contained proteins that function in immune defence (25%), cell
communication/transport (24%), metabolism (18%), enzyme activity (9%), signal
transduction (7%), development/cell differentiation (7%), cell proliferation (4%), cell
organisation (1%) and others of unknown function (5%). Of particular relevance,
amniotic fluid contains an abundant source of urea, which has been shown to
increase in concentration in a linear fashion over gestational age (Sozanskii 1961;
Gulbis et al. 1998). As urea is the sole source of energy for ureaplasmas, amniotic
fluid is able to support the persistent growth of these microorganisms and
represents a niche environment.
Ureaplasmas are also frequently isolated from the chorioamnion of pregnant women
(Quinn et al. 1987; Hillier et al. 1991; Joste et al. 1994). The chorioamnion is
anatomically part of the placenta; however it is composed entirely of fetal tissue
(often referred to as the ‘fetal membranes’, Bourne 1962). The chorion layer is
31
composed of collagen, trophoblasts and mesenchymal cells such as fibroblasts;
whereas the inner amnion layer is largely acellular and consists mainly of
connective tissue bordered by epithelial cells (Calvin and Oyen 2007). Both Toll-like
receptor (TLR) 2 and TLR4 are expressed by epithelial cells lining the amnion
(Abrahams 2005) and numerous natural antimicrobial peptides and defensins are
present in both the chorion and amnion (Horne et al. 2008). Activation of these host
innate immune factors is associated with pro-inflammatory cytokine production,
neutrophil infiltration and the development of histological chorioamnionitis (Yoon et
al. 1999).
2.4.1 ROUTES OF IN UTERO INFECTION
The female upper genital tract is traditionally considered to be a sterile anatomical
site (Romero et al. 2007). Microorganisms causing infections of the upper genital
tract during pregnancy are predicted to gain access to the chorioamnion, amniotic
fluid and fetus by numerous mechanisms. Goldenberg et al. (2000) suggested that
bacteria are able to invade the female upper genital tract during pregnancy by
migration from the abdominal cavity through the Fallopian tubes, iatrogenic needle
contamination at the time of amniocentesis or chronic villus sampling,
haematogenous spread through the placenta, or by an invasive ascending infection.
Of these routes, an ascending infection from the vagina is predicted to be the most
common mechanism resulting in intra-amniotic infection. Kundsin et al. (1996)
demonstrated that the recovery of ureaplasmas from the chorioamnion increased
with the duration of rupture of fetal membranes, which suggested that ascension
from the lower genital tract may be a primary source of infection. Zervomanolakis et
al. (2007) provided evidence of rapid ascension from the lower genital tract, after
radioactively-labelled particles deposited into the vagina of women were detected in
the uterus within 2 minutes. Figure 2.2 demonstrates the predicted mechanism by
which microorganisms are able to pass through the cervix, infect the maternal
32
(decidua) and fetal (chorioamnion) layers of the placenta and access the amniotic
fluid.
Figure 2.2: Ascension from the vagina is predicted to be a common mechanism by which
microorganisms are able to infect the fetal membranes and amniotic fluid, resulting in chorioamnionitis and fetal infection. Source: Goldenberg et al. (2000).
Microorganisms may also gain access to the female upper genital tract via
attachment to spermatozoa. Quinn et al. (1993) reported the case history of
fraternal twins (developed from separately fertilised ova), in which the placenta and
respiratory tract of one infant (who died shortly after birth) were colonised with U.
urealyticum serovar 5. There was no evidence of ureaplasma infection in the other
surviving twin, nor were ureaplasmas isolated from the mother, suggesting that the
source of infection may have been from an infected spermatozoan (proposed by
33
Knox 1998). This also suggests that ureaplasmas may infect the embryo from the
time of conception.
It has also been demonstrated that microorganisms can colonise the endometrium
of non-pregnant women and therefore may infect the embryo at the time of
implantation. Ureaplasmas have been isolated from the endometrium of non-
pregnant women undergoing diagnostic laparoscopy for infertility, tubal ligation or
tubal reanastomosis (Cassell et al. 1993). In these women, ureaplasma colonisation
of the endometrium was not associated with inflammation or clinical signs of
endometritis, indicating that ureaplasmas were present as asymptomatic colonisers.
More recently, Onderdonk et al. (2008) demonstrated high levels of bacterial
colonisation in the second-trimester placental parenchyma. This study
demonstrated that up to 79% of placentas were colonised with bacteria at 23 weeks
of gestation. Combined, these data challenge the view that the female upper genital
tract is a sterile anatomical site and suggest another potential source of intra-uterine
infection.
2.4.2 IN UTERO UREAPLASMA INFECTION IS ASSOCIATED WITH ADVERSE
PREGNANCY OUTCOMES
Ureaplasma infection of the amniotic fluid and chorioamnion has been associated
with adverse pregnancy outcomes including chorioamnionitis (Kundsin et al. 1984;
Cassell et al. 1993; Namba et al. 2010), funisitis (Egawa et al. 2007), preterm
prelabour rupture of membranes (Witt et al. 2005), postpartum endometritis (Chaim
et al. 2003), spontaneous abortion (Joste et al. 1994), stillbirth (McClure and
Goldenberg 2009) and low fetal birth weight (Bayraktar et al. 2010). Intra-amniotic
ureaplasma infection is also associated with preterm birth (Cassell et al. 1993; Perni
et al. 2004; Taylor-Robinson and Lamont 2011), which is the leading cause of
neonatal death in the developed world (Klein and Gibbs 2004), and accounts for
34
70% of perinatal mortality (Goldenberg et al. 2000) and more than half of the long
term infant and childhood morbidity (McCormick 1985). Approximately 30% of all
preterm births are caused by an infectious aetiology (Goldenberg et al. 2008).
Microbial pathogens such as Streptococcus agalactiae, Escherichia coli,
Gardenerella vaginalis, Fusobacterium spp., Staphylococcus spp.,
Propionibacterium spp., Peptostreptococcus spp., Pseudomonas spp., Proteus
spp., and Klebsiella spp. are commonly isolated from the amniotic fluid of women
who deliver preterm (Faye-Peterson 2008). However, ureaplasmas are the
microorganisms most frequently associated with preterm birth (Viscardi 2010) and
are considered to be important predictors of adverse pregnancy outcomes. The
extremely low gestational age newborn study (ELGAS) demonstrated that
ureaplasmas can be isolated from the placental parenchyma from 52 out of 866
(6%) singleton pregnancies that end before 28 weeks of gestation. This large
gestational-age-defined prospective study demonstrated that ureaplasma
colonisation of the placental parenchyma was associated with preterm labour,
preterm prelabour rupture of membranes, as well as umbilical cord, fetal vessel,
membrane and parenchymal inflammation (Olomu et al. 2009).
Inflammation-mediated preterm birth (associated with intra-amniotic infection) is
predicted to occur due to microbial invasion of the choriodecidual space, which
stimulates the production of cytokines such as tumour necrosis factor-alpha (TNF-
α), interleukin (IL)-1α, IL-1β, IL-6, IL-8 and granulocyte-macrophage colony-
stimulating factor. These cytokines, in combination with microbial virulence factors
and phospholipases, stimulate prostaglandin synthesis, neutrophil infiltration and
the release of metalloproteases. The upregulation of prostaglandin causes uterine
contractions, whereas the metalloproteases weaken the chorioamnion, leading to
membrane rupture and ripening of the cervix (Goldenberg et al. 2000). A causal
relationship was recently demonstrated between intra-amniotic U. parvum serovar 1
35
infection and preterm birth in a rhesus macaque model (Novy et al. 2009). In this
animal model, ureaplasma infection was associated with increased amniotic fluid
concentrations of TNF-α, IL-1β, IL-6, IL-8, prostaglandin E2, prostaglandin F2α,
matrix metalloproteinase 9 and leukocytes. The mean time from inoculation-to-
labour onset in animals intra-amnioticially inoculated with ureaplasmas was 6.4 ±
2.5 days, compared to 24.8 ± 1.6 days in animals that were exposed to media or
saline. These data confirmed that ureaplasmas, as sole pathogens, cause
inflammation within the amniotic cavity and preterm birth in a non-human primate
model of intra-uterine infection.
2.4.3 IN UTERO UREAPLASMA INFECTION IS ASSOCIATED WITH NEONATAL
SEQUELAE
The respiratory tract, blood stream and cerebrospinal fluid (CSF) of the fetus can
become colonised with ureaplasmas in utero due to continuous swallowing and
inspiration of infected amniotic fluid. It should also be noted that ureaplasmas may
be vertically transferred from the lower genital tract to the neonate during passage
through the birth canal (Schelonak and Waites 2007). Ureaplasmas are the
microorganisms most frequently isolated from the CSF of neonates (Waites et al.
1988) and can cause meningitis (Garland and Murton 1987), echolucent brain
lesions (Olomu et al. 2009) and intraventricular haemorrhage (Ollikainen et al.
1993). In a study of 313 very low birth weight infants (<1501 g), ureaplasmas were
isolated from the CSF of 74 infants (23.6%) and this was associated with an
increased risk of severe intraventricular haemorrhage (Viscardi et al. 2008).
Ureaplasmas can be detected in 23% of umbilical cord blood cultures from preterm
infants (Goldenberg et al. 2008b) and have been associated with sepsis and
neonatal death (Pinna et al. 2006). Ureaplasma colonisation of the neonatal
respiratory tract is associated with pulmonary diseases such as pneumonia (Quinn
et al. 1985; Viscardi et al. 2002; Morioka et al. 2010) and bronchopulmonary
36
dysplasia (BPD), which can be defined as the requirement for oxygen
supplementation at 36 weeks postmenstrual age and the presence of radiographic
abnormalities (Schelonka and Waites 2007). The link between ureaplasmas and
BPD was first established in 1988, after three independent studies demonstrated
that ureaplasma lower respiratory tract colonisation was associated with BPD in
very low birth weight infants (Cassell et al. 1988; Sanchez and Regan 1988; Wang
et al. 1988). Since these initial reports, there have been numerous studies, which
have provided further evidence that ureaplasma colonisation of the neonatal
respiratory tract may be a risk factor for BPD (Abele-Horn et al. 1997; van Waarde
et al. 1997; Colaizy et al. 2007; Beeton et al. 2011; Kasper et al. 2011; Sung et al.
2011).
Recent research has been aimed at characterising the mechanisms of lung
inflammation and injury, which lead to BPD. Viscardi and Hasday (2009) proposed
that in utero ureaplasma infection stimulates a number of fetal and maternal derived
cytokines, which recruit inflammatory cells and alter transforming growth factor- β1
developmental signalling in the fetal lung. This results in arrested alveolar septation,
capillary development, apoptosis of type II pneumocytes, disordered myofibroblast
proliferation and excessive collagen and elastin deposition. Clinically, infants that
develop BPD are born with relatively mature lungs (and thus a decreased risk of
respiratory distress syndrome) in comparison to those infants without BPD due to
the increased expression of surfactant proteins in response to intra-uterine
inflammation (Jobe and Ikegami 2001). However, BPD is associated with significant
neonatal morbidity and mortality (Gien and Kinsella 2011) and an increased risk of
obstructive lung diseases later in life (Kwinta and Peitzyk 2010).
2.4.4 LONG TERM SEQUELAE OF IN UTERO UREAPLASMA INFECTION
37
The long term effects associated with intra-amniotic ureaplasma infection have not
been determined. This is primarily due to the lack of follow-up of study populations
beyond the neonatal period, but also due to the fact that outcomes are often multi-
factorial, which presents numerous confounding variables (Waites et al. 2005).
During the period from 23 to 32 weeks of gestation, both the fetal lung and brain are
vulnerable to injury mediated by inflammation, which may alter developmental
signalling and result in long-term sequelae (Jobe and Ikegami 2001). Intra-amniotic
inflammation (characterised by elevated levels of IL-6 and IL-8 in amniotic fluid) has
been identified as a potential risk factor for the development of cerebral palsy at
three years of age (Yoon et al. 2000). Berger et al. (2009) demonstrated that
neonates exposed to ureaplasmas in utero had a significantly higher risk of adverse
neuromotor outcome at two years of age, when compared to those who had not
been exposed to ureaplasmas. Furthermore, a murine model of intra-uterine
ureaplasma infection demonstrated that the brains of newborn mice showed
evidence of microglial activation, delayed myelination and disturbed neuronal
development (Normann et al. 2009). These findings suggest that there may be long-
term neurological effects associated with intra-amniotic ureaplasma infection and
highlights the need for further research.
2.4.5 UREAPLASMAS: CONTROVERSIAL PATHOGENS?
Although intra-amniotic ureaplasma infection has been associated with adverse
pregnancy outcomes such as preterm birth, the pathogenic role of ureaplasmas in
the female upper genital tract is complicated by the fact that not all women
colonised with ureaplasmas experience adverse pregnancy outcomes. Rather, the
literature suggests that only sub-populations of women with intra-amniotic
ureaplasma infections deliver preterm babies. Gerber et al. (2003) conducted a
study of 254 pregnant women and detected ureaplasmas within the amniotic fluid of
29 women (11.4%). Of these 29 women, only seven (24.1%) delivered preterm
38
babies. Similarly, Horowitz et al. (1995a) detected intra-amniotic ureaplasma
infection in six pregnant women (2.8%), but only three women (50%) experienced
preterm birth. Whilst both of these studies concluded that ureaplasma infection of
the amniotic fluid is a significant risk factor for preterm birth and adverse pregnancy
outcomes, they failed to acknowledge that a large number of ureaplasma-
infected/colonised women did not experience any clinical signs of adverse
pregnancy outcome.
To potentially explain the inconsistent relationship between intra-amniotic
ureaplasma infection and adverse pregnancy outcomes, it was suggested that some
ureaplasma serovars may be more virulent than others. This has been
demonstrated for other bacterial pathogens, such as Haemophilus influenzae. There
are six antigenically distinct capsular types of H. influenzae, labelled serovars a-f,
however, H. influenzae serotype b is the serotype responsible for 95% of invasive
diseases in children (Chandran et al. 2005). In contrast, there has been very little
evidence to support the hypothesis that some ureaplasma serovars are more
virulent than others, and research findings have not been reproducible. Two
separate investigations have proposed that U. urealyticum serovar 4 is highly
virulent as it was the most frequently isolated serovar from women with recurring
abortion (Quinn et al. 1983; Naessens et al. 1988). Others have suggested that U.
urealyticum serovar 8 is more associated with preterm birth and may be more
invasive due to increased phospholipase production (DeSilva and Quinn 1986;
DeSilva and Quinn 1991). In contrast, Knox and Timms (1998) demonstrated that U.
parvum serovar 6 was significantly associated with preterm birth, and was also the
serovar most adherent to spermatozoa after standard assisted reproductive
technology semen washing procedures (Knox et al. 2003). Furthermore, Zheng et
al. (1992) serotyped ureaplasmas isolated from the CSF of neonates and found that
serovars 1, 3, 6, 8 and 10 were capable of systemic infection. Therefore, there are
39
no conclusive data to suggest that that virulence is limited to specific ureaplasma
serovars. The different rates of serovar detection between these studies were most
likely influenced by the serotyping methods used and geographical differences in
the distribution and prevalence of ureaplasma serovars.
Based on these findings, Zheng et al. (1992) predicted that the property of
invasiveness was not likely to be limited to particular serovars. This group also
demonstrated that clinical ureaplasma isolates of the same serovar were capable of
expressing antigenic size variants. They therefore suggested that antigenic variation
and host factors may be the most important determinants of ureaplasma
pathogenicity. Antigenic variation of surface exposed lipoproteins occurs in several
Mycoplasma spp. and is predicted to contribute to pathogenesis by modulating
interactions between the bacterium and host cells (Citti et al. 2010). Surface-
exposed antigens often contain pathogen-associated molecular patterns, which are
recognised by pattern recognition receptors, such as Toll-like receptors. Therefore,
variation in the expression of these antigens can interfere with recognition of
microbial antigens and the subsequent immune response (Hornef et al. 2002). In a
sheep model of intra-amniotic ureaplasma infection, our research group
demonstrated an inverse relationship between the number of antigenic size variants
produced by a clinical strain of U. parvum serovar 6 and the severity of histological
chorioamnionitis (Knox et al. 2010). These data support the original hypothesis of
Zheng et al. (1992) and provide evidence that antigenic variation may be a predictor
of ureaplasmal virulence.
2.5 VIRULENCE FACTORS OF UREAPLASMA SPP.
Ureaplasmas and mycoplasmas are considered to be microorganisms of low
virulence due to their commensal role in the lower genital tract of females. Five
ureaplasmal proteins have been proposed as virulence factors, which may
40
contribute towards the pathogenesis of ureaplasma infections of the upper genital
tract during pregnancy. These include the multiple banded antigen (MBA), urease,
immunoglobulin A (IgA) protease, phospholipase A and phospholipase C proteins
(Glass et al. 2000). Momynaliev et al. (2007) also predicted that U. parvum contains
a hypervariable plasticity zone, which encodes a putative pathogenicity island.
However, there has been limited investigation into the role of these predicted
virulence factors and the specific mechanisms of ureaplasma pathogenesis remain
unclear.
2.5.1 THE MULTIPLE BANDED ANTIGEN
The MBA was first described by Watson et al. (1990), who demonstrated that
human sera collected from patients infected with ureaplasmas predominantly
recognised a 71 kDa ureaplasmal protein. Further analysis of this antigen using
monoclonal antibodies demonstrated a unique electrophoretic profile, associated
with less intensely stained bands of lower molecular weight, which formed a
symmetrical laddering pattern. These investigators also demonstrated that the MBA
was capable of structural size variation (Figure 2.3), contained both serovar-specific
and cross-reactive epitopes and was expressed by invasive ureaplasma isolates
(Watson et al. 1990; Zheng et al. 1992, Zheng et al. 1994).
Cloning and sequencing of the MBA gene (mba) from the U. parvum serovar 3
reference strain demonstrated that the mba consisted of one large open reading
frame of 1230 bp, which encoded 409 amino acid residues (Zheng et al. 1995). The
N terminus of the MBA consists of a signal peptide followed by a membrane
lipoprotein lipid attachment site. Shimizu et al. (2008) confirmed that the MBA is a
lipoprotein, which could be extracted in the detergent phase by Triton X-114 phase
partitioning. Whilst the 5’ region of the mba is conserved in all 14 ureaplasma
serovars, it contains species-specific nucleotide polymorphisms, which have been
41
Figure 2.3: Immunoblot of the MBA demonstrating size variability of this antigen from five U. parvum serovar 3 isolates. The characteristic laddering pattern is also shown in lanes 1-4. Lane 1: U. parvum serovar 3 reference strain; lanes 2-4: U. parvum serovar 3 clones generated from a clinical isolate; lane 5: U. parvum serovar 3 amniotic fluid isolate. Source: Zheng et al. 1994.
exploited in polymerase chain reaction (PCR) based methods of detection and
speciation (Teng et al. 1994). The 3’ region of the mba encodes multiple tandem
repeat units, which vary in length and copy number between ureaplasma serovars
(Table 2.1). Direct sequencing of mba size variants indicated that increases or
decreases in the number of tandem repeat units was responsible for size variation
of the MBA (Zheng et al. 1995). Hydrophobicity plots of the amino acid sequence of
the MBA predicted the carboxy repeat region to be hydrophilic, surface-exposed
and antigenic (Zheng et al. 1995). Peptide scanning analysis demonstrated that the
dominant MBA epitope recognised by antibodies in human sera is defined by the
amino acid sequence PAGK (Zheng et al. 1996). Whilst initial studies demonstrated
that the MBA is a size variable protein, more recent investigations have shown that
the MBA can undergo phase variation (alternating on/off expression) in vitro. Two
separate studies have demonstrated that selective antibody pressure against the
MBA can result in the generation of MBA-negative escape variants in serial transfer
experiments (Monecke et al. 2003; Zimmerman et al. 2009). MBA-negative
ureaplasmas were detected after two passages in culture medium containing anti-
1 2 3 4 5
68 kDa
53 kDa
47 kDa
39 kDa
36 kDa
42
UREAPLASMA SEROVAR
ORIGIN GENBANK
ACCESSION NUMBER
mba TANDEM REPEAT SEQUENCE (5’ – 3’)
LENGTH OF
TANDEM REPEAT UNIT (nt)
NUMBER OF
TANDEM REPEATS IN GENE
Serovar 1
ATCC 27813
AFO56983
GTAAAGAACAACAACCAG 18 18
Serovar 3
ATCC
700970
L20329
GGTAAAGAACAACCAGCA 18 41
Serovar 6
ATCC 27818
AF056984 GGTAAAGAACCA 12 30
Serovar 14
ATCC 33697
AF056982 GGTAAAGAACAACAACCAGCA 21 31
Serovar 2
ATCC 27814
AF055362
GGTGAAACTACAAAACCAGGAAGT 24 16
Serovar 4
ATCC 27816
AF055363 GGTACAACAAGCCCAGAAAAACCAGGCAAT 30 13
Serovar 5
ATCC 27817
AF055364 GGTGAAACTACAAAACCAGGAAGT 24 18
Serovar 7
ATCC 27819
AF055365 No repeat unit - -
Serovar 8
ATCC 27618
AF055366 GGTGAAACTACAAAACCAGGAAGT 24 18
Serovar 9
ATCC 33175
AF055367 No repeat unit - -
Serovar 10 ATCC 33699
AF055358
GGTTCAACTACACAACCAGGAAGT 24 16
Serovar 11
ATCC 33695
AF055359
No repeat unit - -
Serovar 12
ATCC 33696
AF055360
GGTACAACAAGCCCAGAAAAACCAGGCAAT 30 13
Serovar 13
ATCC 33698
AF055361
GGTACAACAAGCCCAGAAAAACCAGGCAAT 30 13
Table 2.1: Comparison of 3’ mba tandem repeat sequences in U. parvum and U. urealyticum serovars.
Differences in the length and number of tandem repeat units were determined by performing gene alignments of the mba from all 14 ATCC ureaplasma strains (sequences available for download from
Genbank). Alignments were performed using Clustal W. nt = nucleotides.
43
MBA antibodies (Monecke et al. 2003), indicating that this antigen is capable of
rapid phase variation. Zimmerman et al. (2009) hypothesised that expression of the
MBA (locus UU375) is alternated with expression of an adjacent locus (UU376),
which encodes a ureaplasma-specific conserved hypothetical protein. Using
hyperimmune rabbit polyclonal antisera generated against the conserved N
terminus of the MBA (non-repetitive region), the repeat region of the MBA and
UU376, these authors demonstrated that antibody treatment led to the emergence
of escape variants, which expressed the protein that had not been the target of
selective pressure. Specifically, selective antibody pressure targeted against UU376
yielded ureaplasmas predominantly expressing the MBA, whereas selective
pressure against the MBA yielded ureaplasmas predominantly expressing UU376.
Based on Southern blot analysis of ureaplasma clones before and after antibody
treatment, three possible mba locus configurations were proposed (Figure 2.4).
These configurations were predicted to occur due to DNA inversion events, in which
the non-repetitive region of the mba and it’s putative promoter region are opposed
to either the repeat region of the mba or UU376, resulting in alternate expression of
these proteins.
Although the MBA is predicted to be a major virulence factor of ureaplasmas, there
has been minimal investigation into the role of this surface-exposed antigen in
ureaplasmal pathogenesis. Monecke et al. (2003) suggested that the MBA may
function in adhesion to erythrocytes and HeLa cells, as selective pressure against
cytadherence to these cell types resulted in the emergence of MBA-negative
ureaplasma isolates. Surface-exposed antigens of Mycoplasma spp. (such as the V-
1 antigen of M. pulmonis) have been shown to function in bacterial adhesion. In M.
pulmonis, changes in the number of repeat units resulted in altered hydrophobicity
of the V-1 protein, which affected cellular adhesion (Watson et al. 1993). Therefore,
it is possible that the MBA may function in cytadherence; however, further
44
Figure 2.4: Possible DNA inversion events within the mba locus resulting in alternate expression of the MBA and UU376. Locus configurations of (a) U. parvum serovar 3 ATCC 700970 strain (expressing the MBA), (b) U. parvum serovar 3 ATCC 27815 strain (expressing the MBA) and (c) U. parvum serovar 3 MBA-negative escape variant (expressing UU376). Abbreviations: ir = intergenic region; nr = non-repetitive region of the mba. Black triangles represent inverted repeat sequences that are putative recombination sites. Source: Zimmerman et al. 2009.
experimental evidence is required to support this hypothesis. Size variation of the
MBA has also been associated with different severities of histological
chorioamnionitis in a sheep model of intra-amniotic ureaplasma infection (Knox et
al. 2010). In this study, a non-clonal, clinical U. parvum serovar 6 strain produced
MBA size variants within the amniotic fluid of pregnant sheep. The production of ≤ 5
MBA size variants within the amniotic fluid was associated with severe
chorioamnionitis characterised by tissue fibrosis, whereas the production of a higher
number of MBA size variants was associated with minimal/no evidence of
histological chorioamnionitis. Therefore, it was suggested that the number of MBA
size variants produced within the amniotic fluid may contribute to the pathogenesis
of intra-uterine ureaplasma infection. It has also been suggested that size/phase
variation of the MBA may be a mechanism by which ureaplasmas are able to avoid
recognition by the host immune response (Zheng et al. 1996). However, specific
interactions between the MBA and elements of the host immune response have not
45
been well studied. As the MBA is predicted to be the major virulence factor of
ureaplasmas, further investigation is required to determine the role of this surface-
exposed antigen.
2.5.2 UREASE
The ureaplasmal urease enzyme is 30 - 180 fold more efficient than that reported
for other bacterial ureases (Mobley et al. 1995) and was demonstrated to be highly
lethal after intravenous injection in mice (Ligon and Kenny 1991). Urease is a key
virulence factor of urinary tract pathogens, such as Proteus mirabilis, and its
virulence is associated with ammonia production (Mobley et al. 1995). Takebe et al.
(1984) demonstrated that the urease enzyme of U. urealyticum serovar 8 caused
urolithiasis (stone formation) in human urine, and this was preventable by the
addition of urease inhibitors. Interestingly, ureaplasmas (and Blochmannia vafer)
are the only sequenced bacteria which encode the urease enzyme but lack the
ability to assimilate ammonia into glutamine or glutamate (Williams and Wernegreen
2010). This could potentially explain why the intra-cellular ammonia concentration of
ureaplasmas is very high (measured at 21 times the extracellular concentration,
Smith et al. 1993).
The ureaplasma urease gene cluster was found to have a similar genetic
organisation to other ureolytic bacteria. Similar to E. coli, P. mirabilis, Klebsiella
pneumoniae and Klebsiella aerogenes, the ureA, ureB and ureC genes encode the
structural subunit of the ureaplasmal urease complex (Neyrolles et al. 1996). The
ureA, ureB and ureC genes were respectively demonstrated to share 95%, 85% and
92% homology between U. parvum and U. urealyticum. Further downstream, the
ureE, ureF, ureG and ureD genes encode accessory proteins, which are involved in
the synthesis of the nickel metallo-centre (Neyrolles et al. 1996). The urease
complex constitutes a major component of the ureaplasmal cytoplasm (Blanchard et
46
al. 1988) and the urease α-subunit contains species-specific epitopes, which can be
identified by monoclonal antibodies under denaturing conditions (MacKenzie et al.
1996).
Whilst the urease enzyme has been identified as a key virulence factor in the
pathogenesis of urinary tract infections, there are no studies investigating the role of
urease in the amniotic cavity of pregnant women. Due to the cytoplasmic
localisation of the urease enzyme, it is unlikely that this complex would stimulate an
immune response or mediate inflammation within the chorioamnion and fetal
tissues. Recent experiments by our research group have demonstrated that chronic
intra-amniotic ureaplasma infection resulted in increases in the pH of amniotic fluid
and fetal lung fluid in sheep (Robinson, personal communication 2011). This
observed increase in pH was most likely attributed to increased levels of ammonia,
due to the enzymatic activity of the ureaplasmal urease. The associated effects of
increased pH and ammonia concentration (as a result of intra-amniotic ureaplasma
infection) on fetal development are yet to be determined.
2.5.3 IgA PROTEASE
Robertson et al. (1984) first published evidence that U. urealyticum produced an IgA
protease capable of cleaving IgA1. These findings were confirmed by Kilian and
Freundt (1984) who demonstrated that the IgA protease of ureaplasmas caused
specific cleavage of human IgA1, resulting in intact Fab and Fc fragments. More
specifically, the ureaplasma IgA protease (a serine protease) was shown to cleave
human IgA between the proline and threonine residues (235 and 236) in the hinge
region of the heavy chain (Spooner et al. 1992). All 14 ureaplasma serovars
possess IgA protease activity, but do not have proteolytic activity against IgA2, IgG
or IgM antibodies (Kilian et al. 1984). Clinical ureaplasma strains isolated from the
cervix, urine, vagina, synovial fluid and amniotic fluid also demonstrated IgA
47
protease activity, but related Mycoplasma spp. are not capable of cleaving IgA
(Kelian et al. 1984, Kapatais-Zoumbos et al. 1985).
IgA is a primary component of the mucosal immune system of the genital tract,
therefore cleavage of IgA may enable ureaplasmas to colonise and invade the
cervix and upper genital tract of pregnant women. Curiously, complete genome
sequencing of U. parvum serovar 3 failed to identify genes encoding for an IgA
protease (Glass et al. 2000). It was suggested that these genes may have diverged
so far from orthologues in other bacteria that they are unrecognisable, or that
ureaplasmas may have convergently evolved an IgA protease with no recognisable
sequence similarity to known enzymes.
2.5.4 PHOSPHOLIPASE A AND C
Phospholipases are a diverse subgroup of lipolytic enzymes, which hydrolyse ester
linkages in phospholipids and have phosphodiesterase and acyl hydrolase activity
(Istivan and Coloe 2006). Phospholipase A1 and A2 catalyse the hydrolysis of sn-1
and sn-2 acyl ester bonds in 1,2-diacyl-sn-glycero-3-phospholipids, and
phospholipase A2 also plays a role in the production of prostaglandin precursors.
Phospholipase C is a phosphorylhydrolase that catalyses the hydrolysis of the
phosphodiester bond in phospholipids, which results in the production of 1,2-
diglyceride and phosphorylester (Van den Bosch 1980). Phospholipases have also
been identified as virulence factors for pathogenic microorganisms such as
Clostridium perfringens, Listeria monocytogenes, Legionella pneumophila,
Pseudomonas aeruginosa, Staphylococcus aureus and Yersinia enterocolitica
(Schmiel and Miller 1999). Pathogenesis occurs due to the cytolytic activity of
phospholipases, which results from the accumulation of membrane-destabilising
products or by extensive destruction of host cell membrane phospholipids (Istivan
and Coloe 2006). Endogenous phospholipases A1, A2 and C have been detected in
48
U. parvum serovar 3 and U. urealyticum serovars 4 and 8 (DeSilva and Quinn 1986;
Desilva and Quinn 1991; DeSilva and Quinn 1999). These phospholipases
demonstrated higher activity in exponentially growing ureaplasmas, when compared
to stationary phase cells, and initial findings suggested that ureaplasma
phospholipases were membrane bound (and not secreted, DeSilva and Quinn
1991). These authors demonstrated that phospholipase A2 activity was three-fold
higher in U. urealyticum serovar 8, when compared to U. urealyticum serovar 4 and
U. parvum serovar 3 (DeSilva and Quinn 1986). As the activity of phospholipase A1
results in the production of prostaglandins, which play a key role in inflammation
and initiation of labour in pregnant women, it was proposed that differences in
phospholipase A1 activity may account for differences in virulence among
ureaplasma serovars (DeSilva and Quinn 1991). Similar to the ureaplasmal IgA
protease, genome sequencing of U. parvum serovar 3 did not identify genes
encoding phospholipases (Glass et al. 2000), suggesting that these genes may also
have undergone significant divergent or convergent evolution.
2.6 THE HOST IMMUNE RESPONSE TO IN UTERO UREAPLASMA INFECTION
Compared to other microorganisms that cause disease in humans, very little is
known about the host immune response generated during in utero ureaplasma
infection. Throughout pregnancy the immune system integrates the maternal
immune response and the fetal-placental immune response, and is associated with
both pro-inflammatory and anti-inflammatory stages (Mor et al. 2011). The
developing fetus is generally considered to be immune naïve; however, the
expression of fetal innate immune factors increases over gestational age and
maternal IgG antibodies are transported across the placenta and reach 10% of
circulating maternal levels by 17-22 weeks of gestation and 50% by 28-32 weeks of
gestation (van den Berg et al. 2011). IgM is also detectable within umbilical cord
49
blood at 22 weeks of gestation (Daffos et al. 1984); therefore, the fetus is capable of
mounting an immune response in utero.
2.6.1 INNATE IMMUNITY
During pregnancy high numbers of leukocytes (approximately 70% natural killer
cells, 20-25% macrophages and 1.7% dendritic cells) are present in the placenta
and uterus (Mor and Cardenas 2010). Additionally, numerous antimicrobial peptides
and defensins (such as human neutrophil peptides 1-4, human defensin 5, human β
defensins 1-3, elafin and secretory leukocyte protease inhibitor) have been detected
in the amniotic fluid, chorioamnion and placenta of pregnant women (King et al.
2007). Animal models have demonstrated that intra-uterine ureaplasma infection is
associated with increased inflammatory cell influx within the chorioamnion, umbilical
cord and fetal lung tissue, characterised predominantly by neutrophils and
monocytes/macrophages (Moss et al. 2008; Novy et al. 2009; Collins et al. 2010;
Knox et al. 2010). In vitro studies have demonstrated that stimulation of human
monocytes with ureaplasmas causes increased expression of pro-inflammatory
cytokines/chemokines, including IL-1β, IL-8 and TNF-α; and vascular endothelial
growth factor and intracellular adhesion molecule-1 (Li et al. 2000; Li et al. 2002;
Peltier et al. 2008). The macrophage-stimulating activity of U. urealyticum occurs in
response to ureaplasmal lipoproteins (including the MBA, Peltier et al. 2007), which
activate nuclear factor-kappa beta though TLR1, TLR2 and TLR6 (Shimizu et al.
2008). Studies in humans and animal models have demonstrated that intra-amniotic
ureaplasma infection is associated with increased levels of IL-1β, IL-6, IL-8 and
TNF-α within the amniotic fluid and fetal membranes (Yoon et al. 1998; Yoon et al.
2003; Jacobsson et al. 2009; Novy et al. 2009; Thomakos et al. 2010; Kasper et al.
2010; Marconi et al. 2011), which are predicted to contribute towards the onset of
preterm birth.
50
Ureaplasma colonisation also stimulates the production of pro-inflammatory
cytokines within the fetal lung, which contribute to lung inflammation and injury
(Viscardi et al. 2006). Previously, chronic intra-amniotic ureaplasma infection in a
sheep model was shown to increase pulmonary surfactant production and relative
expression of surfactant protein (SP)-A, SP-B and SP-C in fetal lung tissue (Moss et
al. 2005). Whilst SP-B and SP-C play important roles in the physiological functions
of surfactant (Kuroki et al. 2007), SP-A (a member of the collectin family of proteins)
enhances bacterial opsonisation and phagocytosis, and modulates pulmonary
inflammation in a ligand-specific manner (Kingma and Whitsett 2006). Famuyide et
al. (2009) demonstrated that SP-A-deficient mice with experimental U. parvum
pneumonia exhibited delayed clearance and an exaggerated inflammatory
response. Moreover, SP-A was recently demonstrated to increase the
ureaplasmacidal activity of murine RAW 264.7 macrophages in vitro (Okoguble-
Wonodi et al. 2011). Taken together, these data suggest that SP-A may be
important in modulating the pulmonary immune response generated against in utero
ureaplasma infection.
The inflammatory response generated against intra-amniotic ureaplasma infection
can be highly variable. As previously mentioned, Knox et al. (2010) demonstrated
that intra-amniotic infection with a clinical U. parvum serovar 6 isolate was
associated with marked differences in the severity of chorioamnionitis. Although
experimentally-infected sheep were inoculated with the same ureaplasma strain,
histopathology of the chorioamnion demonstrated: (i) fibrosis and tissue scarring; (ii)
mild inflammation; or (iii) no inflammation. Furthermore, pro-inflammatory cytokines
within the amniotic fluid can be highly up-regulated (Holst et al. 2005; Witt et al.
2005), moderately up-regulated (Menon et al. 2009), or not elevated (Perni et al.
2004) during intra-amniotic ureaplasma infection. Kasper et al. (2010) demonstrated
a positive correlation between the bacterial load of U. parvum within the amniotic
51
fluid and levels of IL-8, and suggested that variability of the host immune response
may be associated with the concentration of ureaplasmas within the amniotic fluid.
However, studies in sheep and rats have demonstrated that inflammation occurs
independently of the ureaplasma inoculum dose in experimental models of intra-
amniotic infection (Knox et al. 2010) and urinary tract infection (Reyes et al. 2009).
Therefore, it is more likely that antigenic variability of the MBA (and possibly other
surface-exposed proteins) account for differences in the intensity of the innate
immune response.
2.6.2 ADAPTIVE IMMUNITY
Early studies have demonstrated that both pregnant women and newborn infants
can produce anti-ureaplasma IgG (Liepmann et al. 1988; Dinsmoor et al. 1989), IgM
(Liepmann et al. 1988) and IgA (Cunningham et al. 1996) antibodies. Analysis of
maternal serum collected at study enrolment (≤30 weeks of gestation) and at the
time of delivery, demonstrated that the immunoreactivity of serum antibodies (based
on the number of bands produced in immunoblots) can change over the duration of
pregnancy (Cunningham et al. 1996). As the MBA is the predominant antigen
recognised by human serum (Watson et al. 1990), it is possible that antibodies may
be produced against numerous MBA size variants.
Anti-ureaplasma antibodies were suggested to be a specific marker of infection
during pregnancy; however, these antibodies have also been detected in women
who were culture-negative (Liepmann et al. 1988; Dinsmoor et al. 1989). The
presence of anti-ureaplasma antibodies in maternal serum has been significantly
associated with pregnancy loss (Quinn et al. 1983), preterm delivery, low birth
weight, fetal death (Horowitz et al. 1995b) and postpartum fever (Lee and Kenny
1987). In a study of 424 neonates, anti-ureaplasma antibodies were present in the
sera of: 77% of stillborn babies; 58% of neonates with respiratory disease; and 69%
52
of neonates who died, when compared to 6.5% of healthy term neonates (Quinn
1986). Therefore, the presence of a humoral response to ureaplasmas during
pregnancy may identify women at high risk of adverse pregnancy outcomes.
Despite the reported associations between the presence of anti-ureaplasma
antibodies and adverse pregnancy outcomes, these antibodies have also been
detected in healthy neonates and in populations of women who deliver term babies
(Horowitz et al. 1995b). Similar to the innate immune response generated against
intra-amniotic ureaplasma infection, there is variability in the humoral immune
response, which may also be attributed to antigenic variation. Further research is
required to better characterise humoral immunity in response to intra-amniotic
ureaplasma infections, and to determine the role of cell-mediated immunity (which
at yet has not been investigated).
2.7 ANTIMICROBIAL TREATMENT OF IN UTERO INFECTIONS
Preterm birth and adverse pregnancy outcomes caused by microbial invasion of the
amniotic cavity may be prevented by antimicrobial treatment of pregnant women.
Antimicrobials are routinely administered to women with preterm prelabour rupture
of membranes, as large randomised-controlled trials and meta-analyses have
demonstrated that antibiotic administration can reduce the risk of respiratory
distress syndrome, chorioamnionitis, early onset postnatal infection and also delay
preterm birth (Kenyon et al. 2001a; Kenyon et al. 2010; Yudin et al. 2009; Cousens
et al. 2010). However, there are numerous challenges associated with the
administration of antimicrobial agents to pregnant women, which have hindered the
success of this approach. The ORACLE II randomised trial demonstrated that
erythromycin, co-amoxiclav or combined erythromycin/co-amoxiclav treatment did
not reduce the rates of neonatal death, chronic lung disease or major cerebral
abnormality in neonates delivered by women in spontaneous preterm labour with
53
intact fetal membranes, when compared to women who received placebo (Kenyon
et al. 2001b). One possible explanation of these results is that the timing of
administration of antibiotics was too late for there to be any beneficial effects. In a
review of drug therapies for the prevention of preterm birth, Lamont and Jaggat
(2007) wrote: “if the right antibiotics with activity against those organisms known to
constitute an increased risk of preterm birth are used in the right women at the right
time in pregnancy (before 22 weeks of gestation), before inflammation and feto-
maternal tissue damage occurs, the incidence of preterm birth can be reduced by
40-60% even in low risk women”. However, this scenario is complicated by the fact
that (i) intra-amniotic infections can be clinically asymptomatic and can cause fetal
inflammation early in gestation, (ii) a large range of microorganisms (with different
antimicrobial susceptibilities) are capable of invading the amniotic fluid and (iii) the
amniotic fluid of pregnant women is not routinely cultured or screened for infection
during pregnancy. Therefore, antimicrobial treatment of pregnant women has been
associated with variable success rates and as a result, some authors do not
recommend the routine administration of antibiotics to pregnant women (Gibbs
2001; van den Broek et al. 2009).
Treatment options for pregnant women are limited due to the potential tetratogenic
and harmful effects associated with the use of some antimicrobials during
pregnancy. Even fewer options are available for the treatment of intra-amniotic
ureaplasma infections, as ureaplasmas are inherently resistant to the beta-lactam
and glycopeptide antimicrobials (due to the lack of cell wall), and also demonstrate
resistance to trimethoprim, sulfonamides and rifampicin (Waites et al. 2009).
Antimicrobials that are potentially active against ureaplasmas include the
tetracyclines, fluoroquinolones and macrolides (Waites et al. 2009); however, not all
of these antibiotics are appropriate to administer during pregnancy.
2.7.1 TETRACYCLINE TREATMENT OF IN UTERO INFECTIONS
54
Tetracyclines are antimicrobials that bind to the bacterial 30S ribosomal subunit and
inhibit binding of aminoacyl-tRNA to the mRNA-ribosome complex. Tetracyclines
demonstrate high levels of ureaplasmacidal activity, with minimum inhibitory
concentrations (MICs) typically ranging between 0.25 µg/mL and 2.0 µg/mL (Beeton
et al. 2009a). However, resistance to the tetracyclines is becoming more common
due to ureaplasma isolates harbouring the tet(M) resistance gene (Roberts and
Kenny 1986), which confers high-level resistance (MICs > 64 µg/mL) by binding to
the ribosome and inducing dissociation of the tetracycline-ribosome complex in a
GTP-dependent manner (Zakeri and Wright 2008). Tetracycline resistance (due to
the tet(M) gene) has also been associated with concurrent erythromycin resistance
(Robertson et al. 1988; Xiao et al. 2011). The ureaplasmal tet(M) gene
demonstrates significant homology to the streptococcal tet(M) determinant (Roberts
and Kenny 1986), suggesting that ureaplasmas may have acquired tetracycline
resistance via conjugative transposons.
The tetracyclines are contraindicated in the second and third trimesters of
pregnancy, as they have caused or are suspected to have caused, an increased
incidence of human fetal malformations or irreversible damage (pregnancy category
D, Antibiotic Expert Group 2010). After the 18th week of pregnancy, tetracyclines
can interfere with teeth mineralisation processes until the child reaches eight years
of age and can induce hypoplasia of the enamel, discolouration of the teeth and
accumulate in the growing skeleton (Antibiotic Expert Group 2010). Milch et al.
(1957) were the first to report that tetracyclines were incorporated into growing and
newly mineralised bones. These authors demonstrated that the skeletons of rats,
which received tetracycline, exhibited fluorescence (due to absorption of fluorescent
tetracycline) under ultra-violet light for several months after drug exposure.
Therefore, the tetracyclines are not often administered to pregnant or breastfeeding
women; however, have been used to successfully eradicate ureaplasmas from the
55
genitourinary tract of patients with non-gonococcal urethritis (Oriel and Ridgway
1983) and vulvovaginitis (Ventolini and Lee 2011).
2.7.2 FLUOROQUINOLONE TREATMENT OF IN UTERO INFECTIONS
Fluoroquinolones are antimicrobials that inhibit bacterial DNA gyrase (encoded by
the gyrA and gyrB genes) or the topoisomerase IV enzyme (encoded by the parC
and parE genes), thereby inhibiting bacterial DNA replication and transcription
(Madurga et al. 2008). Examples of antimicrobials belonging to this class include
moxifloxacin, ofloxacin, levofloxacin, trovafloxacin and grepafloxacin, all of which
exhibit high activity against ureaplasmas, with MICs ≤ 2.0 µg/mL (Bébéar et al.
2000; Bébéar et al. 2008; Krause and Schubert 2010; Samra et al. 2011).
Ciprofloxacin, another common fluoroquinolone, is less active against ureaplasmas,
as 65% of ureaplasma isolates collected over a 20 year period were resistant to this
antimicrobial (Krause and Schubert 2010). Bébéar et al (2003) demonstrated that
ureaplasmal fluoroquinolone resistance was associated with amino acid substitution
mutations in gyrA and parC. However, the precise mechanisms of fluoroquinolone
resistance in ureaplasmas are unknown, as numerous non-resistance associated
amino acid substitutions have also been reported in gyrA, gyrB, parC and parE
(Beeton et al. 2009b).
Fluoroquinolones (as second, third and fourth generation quinolones) have only
been discovered and approved for use in the last 10 to 20 years (Andriole 2005).
Therefore, there are limited data regarding the use of these antimicrobials during
pregnancy. Ciprofloxacin, moxifloxacin and norfloxacin are classified as Pregnancy
Category B3 drugs, indicating that only a limited number of pregnant women have
been administered these drugs without an increase in the frequency of malformation
or other direct or indirect harmful effects on the fetus (Antibiotic Expert Group 2010).
A meta-analysis, which evaluated the risks of fluoroquinolone treatment of pregnant
56
women, demonstrated that the use of fluoroquinolones during the first trimester of
pregnancy did not represent an increased risk for major malformations recognised
after birth, stillbirth, preterm birth or low birth weight (Bar-Oz et al. 2009). However,
a cohort study of 87 Danish women who received fluoroquinolone treatment at any
time during pregnancy demonstrated that prenatal fluoroquinolone exposure was
associated with an increased risk of bone malformation (Wogelius et al. 2005).
These inconsistent findings highlight the need for large randomised-controlled trials
to assess the effects of fluoroquinolone exposure on the developing fetus.
2.7.3 MACROLIDE TREATMENT OF IN UTERO INFECTIONS
Macrolides are antimicrobials that inhibit bacterial protein biosynthesis by binding to
the 50S subunit of ribosomes and inhibiting translocation of peptidyl transfer RNA.
Macrolides block the tunnel, which guides the chain of amino acids through the
ribosome, and thus prevents the synthesis of full length proteins (Mankin 2008). The
most common macrolides include erythromycin, azithromycin, clarithromycin and
roxithromycin, which demonstrate broad-spectrum bacteriostatic activity. Macrolide
antibiotics differ in the number of carbon atoms within the lactone ring and newer
macrolides (such as azithromycin and roxithromycin) accumulate in high
concentrations in tissues, macrophages and polymorphonuclear leukocytes (Jain
and Danziger 2004). In vitro, macrolide antimicrobials are highly active against
ureaplasmas; however, MIC testing is complicated due to the optimal acidic pH
required for ureaplasmal growth. Kenny and Cartwright (1993) demonstrated that
erythromycin MICs were 4 – 16 fold higher at pH 6.0, when compared to pH 7.0,
which may lead to inaccurate reporting of resistance in these microorganisms.
The rates of ureaplasmal macrolide resistance appear to vary between geographical
locations. The percentage of erythromycin-resistant ureaplasmas isolated from
patients attending an outpatient clinic in China decreased from 63.9% to 20.0%
57
between 1999 and 2000 (Xie and Zhang 2006). In Spain, 100% of ureaplasma
strains isolated from the urethra of symptomatic and healthy men were susceptible
to erythromycin (García-Castillo et al. 2008), whereas erythromycin and
azithromycin were inactive against all ureaplasmas isolated from vaginal and
endocervical swabs from women in Greece (Kechagia et al. 2008). Despite the
emergence of resistance, macrolides are usually first-line antimicrobials used for the
treatment of ureaplasma infections in neonates and adults (Waites et al. 2005).
The mechanisms of macrolide resistance in ureaplasmas are similar to those
reported for other bacteria. Specifically, macrolide resistance can occur by: i) point
mutations in 23S rRNA genes and ribosomal protein L4 and L22 genes; ii)
methylation of 23S rRNA by erythromycin-ribosome methylase (erm) resistance
genes and (iii) transport of macrolides out of the bacterial cell by efflux pumps
(Goldman and Scaglione 2004).
The lactone ring of macrolides interacts hydrophobically with the crevice formed by
23SrRNA bases 2057, 2058 and 2059 (E. coli numbering); therefore a mutation in
any of these nucleotides can inhibit macrolide binding. Single mutations in
ribosomal protein L4 and L22 genes can also allosterically affect macrolide binding
(Gaynor and Mankin 2005). Point mutations in domain V of the 23S rRNA gene that
have been associated with macrolide resistance in ureaplasmas include G2056U,
G2057U and A2058G (E. coli numbering, Pereyre et al. 2007) and C2443 (T or C,
U. urealyticum numbering, Dongya et al. 2008). A 6 bp deletion in the ribosomal
protein L4 gene of U. parvum was also associated with high level resistance to
erythromycin (MIC > 64 mg/L, Beeton et al. 2009a). Macrolide resistance can also
occur by the activity of erm genes, which post-translationally methylate 23S rRNA
genes. The erm gene family catalyse the transfer of one or two methyl groups to the
exocyclic nitrogen of nucleotide 2058 in domain V of the 23S rRNA gene. The
methyl group sterically hinders binding of macrolide and lincosamide antimicrobials
58
to nucleotide 2058, thus rendering bacteria resistant to these drugs (Maravić 2004).
Expression of erm genes can be induced by exposure of microorganisms to
macrolides. In the presence of erythromycin, the ribosome stalls at the 8th or 9th
codon of the leader peptide of the erm cassette, which triggers a conformational
rearrangement in mRNA, resulting in opening of the erm ribosome binding site and
subsequent erm translation (Gaynor and Mankin 2005). In a study of 72
ureaplasmas strains isolated from the urethra or cervix of patients with non-
gonococcal urethritis or mucopurulent cervicitis, Lu et al. (2010) detected erm(B) in
21 isolates (29%). Previous studies have demonstrated that the erm(B) was strongly
associated with detection of conjugative transposons, such as Tn1545 and Tn916
(Seral et al. 2001). A significant association was found between the detection of
erm(B) and int-Tn (a genetic marker of a transposon) in ureaplasma isolates, which
suggests that the ureaplasmal erm(B) may be located on a transposon (Lu et al.
2010).
The macrolide-streptogramin resistance (msr) genes encode efflux pumps, which
(unlike the erm genes) do not alter macrolide target sites, but instead pump
macrolide and streptogramin antimicrobials out of the cell. The msr genes are
putative members of the class 2 ATP-binding cassette transporter superfamily
(Roberts et al. 1999) and have been detected in a large number of both Gram
positive and Gram negative bacteria (Roberts 2008). Four subtypes of the msr gene
family have been described: msr(A), msr(B), msr(C) and msr(D). Of these, msr(A),
msr(B) and msr(D) have been detected in Ureaplasma spp. (Lu et al. 2010). In this
study, 37 out of 72 ureaplasma isolates (51%) harboured at least one msr gene
subtype and macrolide MICs were generally higher against ureaplasmas isolates
that carried more than one of these genes. The high rates of detection of msr
subtypes by Lu et al. (2010) suggest that macrolide resistance by the activity of drug
efflux pumps may be common in ureaplasmas; however, further studies are
59
required to confirm these findings and to determine the mechanisms by which
ureaplasmas are able to acquire these genes.
Despite the potential for induced macrolide resistance in ureaplasmas, macrolides
(in particular, erythromycin) are the most common antimicrobials administered to
pregnant women with intra-amniotic infection or preterm prelabour rupture of
membranes. Erythromycin (Pregnancy Category A) has been administered to a
large number of pregnant women and women of childbearing age without any
proven increase in the frequency of malformations or other direct or indirect harmful
effects on the fetus being observed (Antibiotic Expert Group 2010). Erythromycin
became standard treatment for preterm prelabour rupture of membranes after the
ORACLE I trial demonstrated that erythromycin treatment of pregnant women was
associated with numerous health benefits for the neonate. In this randomised trial,
4826 women with preterm prelabour rupture of membranes were randomly assigned
to receive erythromycin (n = 1197), co-amoxiclav (n = 1212), erythromycin and co-
amoxiclav (n = 1192) or placebo (n = 1225) four times daily for 10 days, or until
delivery. The use of erythromycin was associated with prolongation of pregnancy,
reduction in neonatal surfactant treatment, decreased rates of BPD, fewer cerebral
abnormalities and fewer positive blood cultures (Kenyon et al. 2001a). Macrolides
are also considered to be potent immunomodulators, which decrease the production
of mucus by pulmonary epithelial cells and prevent the production of pro-
inflammatory cytokines by inhibiting nuclear factor-kappa beta (Giamarellos-
Bourboulis 2008). Therefore, in addition to exhibiting broad-spectrum bacteriostatic
activity, macrolides may also down-regulate fetal inflammation in utero.
Whilst erythromycin appears to be an ideal antimicrobial for the treatment of
pregnant women, a number of studies have demonstrated that maternal
erythromycin treatment (either alone or in combination with other antimicrobials) is
ineffective at eliminating intra-amniotic infections and/or improving pregnancy
60
outcomes (Table 2.2). Furthermore, a follow-up of childhood outcomes after the
ORACLE II trial demonstrated that prescription of erythromycin to women in
spontaneous preterm labour with intact membranes was associated with an
increased risk of cerebral palsy among their children at 7 years of age (Kenyon et
al. 2008). These findings suggest that in utero erythromycin exposure may have
harmful effects on the developing fetus. Conversely, some studies have
demonstrated that maternal erythromycin treatment is capable of eradicating intra-
amniotic infection and improving neonatal outcomes (Table 2.2); however, there is a
lack of controlled studies investigating the efficacy of erythromycin in intra-amniotic
infections. Recent studies have investigated the use of azithromycin in pregnancy.
The APPLe study, a randomised placebo-controlled trial of azithromycin for the
prevention of preterm birth, demonstrated that azithromycin treatment of pregnant
women (n = 1096) did not affect gestational age at delivery, fetal birth weight or the
rates of perinatal death, when compared to women who received placebo (n =
1087, van den Broek et al. 2009).
Whilst differences in the timing and duration of antimicrobial treatment may have
influenced the outcomes associated with macrolide treatment, it has been
suggested that the failure of macrolides to eradicate intra-amniotic infection and
improve pregnancy outcomes in some women is due the poor placental transfer of
these antimicrobials. Initial studies demonstrated that the concentration of
erythromycin (after multiple, maternal oral doses) in the serum of aborted fetuses
was only 2% of the concentration measured in the maternal serum at the time of
abortion (Kiefer et al. 1955). Heikkinen et al. (2000) confirmed these findings after
ex vivo perfusion of a human placenta and demonstrated that the transplacental
transfer of erythromycin, azithromycin and roxithromycin was 3.0%, 2.6% and 4.3%
respectively. Therefore, the placenta may be an effective barrier against the transfer
of macrolide antimicrobials from the maternal circulation to the amniotic fluid and
61
Table 2.2: Comparison of outcomes associated with maternal erythromycin treatment of pregnant women. pPROM = preterm prelabour rupture of membranes.
REFERENCE ANTIMICROBIAL TREATMENT POPULATIONS STUDIED OUTCOMES E
RY
TH
RO
MY
CIN
DO
ES
NO
T I
MP
RO
VE
OU
TC
OM
ES
Gomez et al. 2007
Erythromycin and ampicillin (7 days) or ceftriaxone, clindamycin and erythromycin (10–14 days)
Pregnant women with pPROM and intra-amniotic inflammation (n = 28) or no intra-amniotic inflammation (n = 18)
Maternal antibiotic treatment rarely eradicated intra-amniotic inflammation (3/18, 16.7%) and did not prevent subsequent intra-amniotic inflammation
Eschenbach et al. 1991 Erythromycin (333 mg) or placebo, three times daily starting between 26 and 30 weeks of gestation and continuing until 35 weeks
Pregnant women with lower genital tract ureaplasma colonisation
There were no differences between erythromycin and placebo treated women in terms of: birth weight, gestational age at delivery, pPROM and neonatal health
Kenyon et al. 2001b (ORACLE II Trial)
Erythromycin (250 mg), co-amoxiclav (250 mg amoxicillin and 125 mg clavulanic acid), combined erythromycin and co-amoxiclav, or placebo
Pregnant women in spontaneous preterm labour with intact membranes (erythromycin n = 1611, co-amoxiclav n = 1550, combined n = 1565, placebo n = 1569)
Neonatal death, chronic lung disease and cerebral abnormalities were not different between women treated with antibiotics or placebo
Ogasawara and Goodwin 1997
Erythromycin (oral) ± ampicillin (intra-venous) for 7 days
51 pregnant women (between 22 and 35 weeks of gestation) with pPROM or preterm labour
Erythromycin did not prevent vertical transmission of U. urealyticum to the neonate or reduce the rates of histological chorioamnionitis
ER
YT
HR
OM
YC
IN D
OE
S
IMP
RO
VE
OU
TC
OM
ES
Mazor et al. 1993
Erythromycin (10 days)
A pregnant woman with premature contractions at 32 weeks of gestation
Erythromycin successfully eradicated ureaplasmas from the amniotic fluid and delivery was prolonged until 39 weeks of gestation
Antsaklis et al. 1997
Erythromycin (500 mg) every 8 hours for 10 days (n = 18) or no antibiotics (n = 17)
Women in preterm labour (between 26 and 34 weeks of gestation)
Erythromycin treatment prolonged delivery and was associated with higher birth weight, lower neonatal morbidity and shorter hospitalisation stays (although none of these associations were statistically significant)
Berg et al. 1999
Oral erythromycin (10 days, n = 44) or no antibiotics (n = 9)
Retrospective study of ureaplasmas infection in genetic amniocentesis specimens
Mid-trimester pregnancy loss was reduced in erythromycin group (11.4%), when compared to untreated women (44.4%); however, the rates of preterm delivery were not different between groups (19.4% and 20%)
62
fetal circulation, which suggests that the fetus may be exposed to low-level
concentrations of these drugs in utero. Moreover, microorganisms residing in the
amniotic fluid may also be exposed to low levels of macrolides, which may be a
significant driver of antimicrobial resistance. Despite these potential limitations,
erythromycin is still the drug of choice in clinical obstetrics. There is little agreement
amongst authors regarding the effectiveness of erythromycin (and newer macrolides
such as azithromycin) for the treatment of intra-amniotic infections, which suggests
that further controlled studies are required.
2.8 ANIMAL MODELS FOR THE STUDY OF IN UTERO INFECTIONS
Animal models have contributed towards our understanding of inflammation-
mediated preterm birth and adverse neonatal outcomes. Mice, rats, guinea pigs,
sheep and non-human primates have been used to investigate intra-uterine
infection. These animal models are compared in Table 2.3. Rats and mice are
attractive animal models as they are relatively economical, widely available and
easily housed. However, research findings in rodents cannot be directly translated
to humans, due to numerous differences in reproduction and parturition. Whilst it
has been demonstrated that lipopolysaccharide (Fidel et al. 1994) and heat-
inactivated E. coli (Hirsch et al. 2002) can induce preterm birth and increased
maternal IL-1 serum concentrations in mice, the underlying mechanism of parturition
in mice differs to that in humans. In mice and rats, parturition occurs due to
progesterone withdrawal and luteolysis, which results from the actions of
prostaglandin F2α (Mitchell and Taggart 2009). In contrast, maternal serum
progesterone continues to increase over the course of gestation in humans
(Boroditsky et al. 1978) and parturition occurs due to increased production of
prostaglandins and oxytocin (a hormone that is released in large quantities after
distension of the cervix and uterus during labour) and their receptors (Mendelson
2009). Therefore, mice and rats have not been extensively used as experimental
63
Table 2.3: Comparison of reproductive characteristics and parturition in animal models and humans. Adapted from Mitchell and Taggart 2009 and Adams Waldorf et al. 2011
MOUSE RAT GUINEA PIG SHEEP
NON-HUMAN PRIMATES
HUMAN
Length of gestation
20 ± 1 days 22 ± 1 days 67 ± 3 days 147 ± 4 days 160 – 180 days 266 ± 14 days
Litter size
10 ± 5 10 ± 6 3 ± 2 1 - 2 1 1
Placental anatomy
Haemotrichorioal Haemotrichorial Haemomonochorial Synepitheliochorial Haemomonochorial Haemomonochorial
Mechanism of induction of preterm birth
Antiprogestin, ovariectomy, LPS
Antiprogestin, ovariectomy
Antiprogestin plus oxytocin
Progesterone withdrawal due to activation of the fetal hypothalamic-pituitary-adrenal axis and glucocorticoids
Intra-amniotic inoculation of bacteria, LPS or cytokines
Postaglandin, antiprogestin, oxytocin
Advantages
Low cost, availability of genetically modified breeds, availability of commercial reagents and microarray platforms
Low cost, easy to house
Similar placental anatomy to humans
Chronic infection models are possible and both ewe and fetus are amenable to surgical interventions and long-term catheterisation
Similar placental anatomy, gestational length and litter size to humans
Directly translational
Disadvantages
Significant differences to humans, limited ability for fetal instrumentation and longitudinal analyses
Significant differences to humans, limited ability for fetal instrumentation and longitudinal analyses
Spontaneous preterm birth can occur
High cost, require specialised holding facilities, different placental anatomy and different mechanism of parturition compared to humans
Very high cost, require specialised holding facilities and care, ethical considerations, low availability
Often limited to small, cross-sectional studies
64
models of intra-amniotic ureaplasma infection, although have been used to
investigate lung injury and inflammation (Viscardi et al. 2002; Famuyide et al. 2009),
urinary tract infections (Reyes et al. 2009; Allam et al. 2011) and infertility (Engel et
al. 1988; Audring et al. 1989; Xu et al. 1997; Wang et al. 2010) associated with
ureaplasma infection.
In contrast to rodents, larger animal models, such as sheep, provide opportunities to
study the effects of chronic intra-uterine infection, due to increased gestational
length (147 ± 4 days). Additionally, pregnant sheep tolerate invasive surgical
procedures and sophisticated techniques have been developed for functional
studies in utero after full recovery from anaesthesia and surgery (Carter 2007). As a
result, fetal physiology in sheep has been well characterised, but ovine immunology
is not well understood (Kemp et al. 2010). A disadvantage of the sheep model is
that intra-uterine inflammation does not cause preterm birth. This is because
parturition in sheep is dependent on progesterone withdrawal, which occurs due to
maturation of the fetal hypothalamic-pituitary axis, secretion of adrenocorticotropic
hormone and increased cortisol production. Cortisol stimulates maturation of the
fetal lungs and signals placental enzymes to convert progesterone to estrogen,
which initiates labour (Mitchell and Taggart 2009). This pathway occurs
independently of intra-uterine inflammation in sheep, therefore increased levels of
cytokines and prostaglandins within the amniotic cavity do not precipitate preterm
birth. However, this feature does enable researchers to study the long term effects
of intra-amniotic infection on the developing fetus. Additionally, the in utero
development of ovine fetuses is similar to that of human fetuses, and newborn
lambs and human babies weigh approximately the same at the time of birth (Mitchell
and Taggart 2009). Our research group has pioneered the use of a sheep model of
chronic, intra-amniotic ureaplasma infection. Studies in pregnant ewes have
provided evidence that in utero ureaplasma exposure is associated with fetal lung
65
maturation and increased surfactant production (Moss et al. 2005; Moss et al. 2008;
Polglase et al. 2010), histological chorioamnionitis and systemic spread to the fetal
cerebrospinal fluid (Knox et al. 2010).
Non-human primates exhibit the closest resemblance to humans with respect to
gestational length, uterine anatomy and electromyographic activity. Furthermore,
parturition in non-human primates is not associated with progesterone withdrawal
and is similar to parturition in humans (Mitchell and Taggart 2009). Unlike sheep,
preterm birth can occur due to intra-uterine infection in non-human primates.
Gravett et al. (1994) initially characterised inflammation-mediated preterm birth in
rhesus macaques experimentally infected with group B Streptococcus (GBS). This
study demonstrated that preterm labour occurred (on average) 28 hours after intra-
amniotic infusion of 106 colony forming units of GBS, whereas control animals
delivered 30 days later. Intra-amniotic GBS infection in rhesus macaques was
associated with increased levels of IL-1β, IL-6, TNF-α and prostaglandins E2 and
F2α, which stimulated uterine contractility. Similar data published by Novy et al.
(2009) demonstrated that intra-amniotic inoculation of U. parvum serovar 1 in
rhesus macaques resulted in preterm delivery, chorioamnionitis and fetal
pneumonia. A study in pregnant baboons also demonstrated that antenatal U.
parvum respiratory tract infection stimulates proinflammatory, profibrotic responses
in the preterm lung (Viscardi et al. 2006). In this study, ureaplasma-infected baboon
lungs demonstrated extensive fibrosis, increased alpha-smooth muscle actin and
transforming growth factor- β1 staining and increased pro-inflammatory cytokine
expression. These findings suggested that a prolonged inflammatory response
initiated by intra-uterine ureaplasma infection may contribute to pulmonary fibrosis
and altered developmental signalling in the immature lung. Despite the advantages
of these models, the use of non-human primates is associated with ethical, financial
and infrastructural constraints, which limit their use.
66
2.9 CONCLUDING REMARKS
Ureaplasma infection of the amniotic fluid and fetal membranes is associated with
numerous problems in human reproduction. These microorganisms can cause
chronic, intra-amniotic infections, which can persist in utero for lengthy periods of
time and cause neonatal morbidity and mortality. Ureaplasmas possess a small
suite of proposed virulence factors, which may facilitate colonisation of the upper
genital tract during pregnancy and prevent eradication by the immune response.
The variable rates of success of macrolide treatment for intra-amniotic ureaplasma
infections may be due to a combination of host factors (such as the placental
barrier) and microbial factors, including the emergence of induced antimicrobial
resistance.
Although there are a number of animal models used to study intra-uterine infections,
the sheep model is well established and is associated with numerous advantages.
As pregnant sheep are protected against inflammation-induced preterm birth,
researchers are able to perform longitudinal studies in which gestational age can be
stringently controlled. This feature is unique to the sheep model, and provides an
excellent opportunity for the study of long term, in utero ureaplasma infections and
the associated fetal outcomes, which closely mimic those observed in human
pregnancies.
This literature review has highlighted the need for further research to characterise
chronic, intra-amniotic ureaplasma infections and potential treatment options. The
following results chapters of this thesis report the findings of investigations
performed in a sheep model of intra-amniotic ureaplasma infection, to address
these knowledge gaps.
67
2.10 LITERATURE CITED
Abele-Horne M, Genzel-Boroviczény O, Uhlig T, Zimmermann A, Peters J, Scholz M (1998) Ureaplasma urealyticum colonization and bronchopulmonary dysplasia: a comparative prospective multicentre study. Eur J Pediatr 157: 1004-1011.
Abele-Horn M, Scholz M, Wolff C, Kolben M (2000) High-density vaginal Ureaplasma urealyticum colonization as a risk factor for chorioamnionitis and preterm delivery. Acta Obstet Gynecol Scand 79: 973-978.
Abele-Horn M, Wolff C, Dressel P, Pfaff F, Zimmermann A (1997) Association of Ureaplasma urealyticum biovars with clinical outcome for neonates, obstetric patients, and gynaecological patients with pelvic inflammatory disease. J Clin Microbiol 35: 1199-1202.
Abrahams VM (2005) Toll-like receptors in the cycling female reproductive tract and during pregnancy. Curr Womens Health Rev 1: 35-42.
Adams Waldorf KM, Rubens CE, Gravett MG (2011) Use of nonhuman primate models to investigate mechanisms of infection-associated preterm birth. BJOG 118: 136-144.
Allam AB, Alvarez S, Brown MB, Reyes L (2011) Ureaplasma parvum infection alters filamin A dynamics in host cells. BMC Infect Dis 11: 101.
Andrade-Rocha FT (2003) Ureaplasma urealyticum and Mycoplasma hominis in men attending for routine semen analysis. Prevalence, incidence by age and clinical settings, influence on sperm characteristics, relationship with the leukocyte count and clinical value. Urol Int 71: 377-81. Andriole VT (2005) The quinolones: past, present and future. Clin Infect Dis 41: S113-S119.
Antibiotic Expert Group (2010) Therapeutic Guidelines: Antibiotic Version 14. Melbourne: Therapeutic Guidelines Limited pp. 374-380.
Antsaklis A, Daskalakis G, Michalas S, Aravantinos D (1997) Erythromycin treatment for subclinical Ureaplasma urealyticum infection in preterm labor. Fetal Diagn Ther 12: 89-92.
Audring H, Klug H, Bollmann R, Sokolowska-Kohler W, Engel S (1989) Ureaplasma urealyticum and male infertility: an animal model: II. Morphologic changes of testicular tissue at light microscopic level and electron microscopic findings. Andrologia 21: 66-75. Badalyan RR, Fanarjyan SV, Aghajanyan IG (2003) Chlamydial and ureaplasmal infections in patients with nonbacterial chronic prostatitis. Andrologia 35: 263-265. Bar-Oz B, Moretti ME, Boskovic R, O’Brien L, Koren G (2009) The safety of quinolones- a meta analysis. Eur J Obstet Gynecol 143: 75-78. Bayraktar MR, Ozerol IH, Gucluer N, Celik O (2010) Prevalence and antibiotic susceptibility of Mycoplasma hominis and Ureaplasma urealyticum in pregnant women. Int J Infect Dis 14: e90-e95.
68
Bébéar CM, de Barbeyrac B, Pereyre S, Renaudin H, Clerc M, Bébéar C (2008) Activity of moxifloxacin against the urogenital mycoplasmas Ureaplasma spp., Mycoplasma hominis and Mycoplasma genitalium and Chlamydia trachomatis. Clin Microbiol Infect 14: 797-812. Bébéar CM, Renaudin H, Charron A, Clerc M, Pereyre S, Bébéar C (2003) DNA gyrase and topoisomerase IV mutations in clinical isolates of Ureaplasma spp. and Mycoplasma hominis resistant to fluoroquinolones. Antimicrob Agents Chemother 47: 3323-3325. Bébéar CM, Renaudin H, Charron A, Gruson D, Lefrancois M, Bébéar C (2000) In vitro activity of trovafloxacin compared to those of five antimicrobials against mycoplasmas including Mycoplasma hominis and Ureaplasma urealyticum fluoroquinolone-resistant isolates that have been genetically characterized. Antimicrob Agents Chemother 44: 2557-2560.
Beeton ML, Chalker VJ, Kotecha S, Spiller OB (2009b) Comparison of full gyrA, gyrB, parC and parE gene sequences between all Ureaplasma parvum and Ureaplasma urealyticum serovars to separate true fluoroquinolone antibiotic resistance mutations from non-resistance mutations. J Antimicrob Chemother 64: 529-538.
Beeton ML, Chalker VJ, Maxwell NC, Kotecha S, Spiller OB (2009a) Concurrent titration and determination of antibiotic resistance in Ureaplasma species with identification of novel point mutations in genes associated with resistance. Antimicrob Agents Chemother 53: 2020-2027.
Beeton ML, Maxwell NC, Davies PL, Nuttall D, McGreal E, Chakraborty M, Brad Spiller O, Kotecha S (2011) Role of pulmonary infection in the development of chronic lung disease of prematurity. Eur Respir J 37: 1424-1430. Berg TG, Philpot KL, Welsh MS, Sanger WG, Smith CV (1999) Ureaplasma/mycoplasma-infected amniotic fluid: pregnancy outcome in treated and nontreated patients. J Perinatol 19: 275-277. Berger A, Witt A, Haiden N, Kaider A, Klebemasz K, Fuiko R, Langgartner M, Pollak A (2009) Intrauterine infection with Ureaplasma species is associated with adverse neuromotor outcome at 1 and 2 years adjusted age in preterm infants. J Perinat Med 37: 72-78. Blanchard A, Razin S, Kenny GE, Barile MF (1988) Characteristics of Ureaplasma urealyticum urease. J Bacteriol 170: 2692-2697. Boroditsky RS, Reyes FI, Winter JS, Fairman C (1978) Maternal serum estrogen and progesterone concentrations preceeding normal labor. Obstet Gynecol 51: 686-691. Bourne G (1962) The foetal membranes: A review of the anatomy of normal amnion and chorion and some aspects of their function. Postgrad Med J 38: 193-201. Brown DR (2010) Phylum. Tenericutes. In: Whitman WB, editor. Bergeys’ Manual of Systemic Bacteriology. Volume Four: The Bacteroidetes, Spirochaetes, Tenericutes (Mollicultes), Acidobacteria, Fibrobacteres, Fusobacteria, Dictyoglomi,
69
Gemmatimonadetes, Lentisphaerae, Verrucomicrobia, Chlamydiae, and Plantomycetes. New York: Springer. pp. 567-723.
Breugelmans M, Vancutsem E, Naessens A, Lauback M, Foulon W (2010) Association of abnormal vaginal flora and Ureaplasma species as risk factors for preterm birth: a cohort study. Acta Obstet Gynecol Scand 89: 256-260.
Calvin SE, Oyen ML (2007) Microstructure and mechanics of the chorioamnion membrane with an emphasis on fracture properties. Ann N Y Acad Sci 1101: 166-185.
Carter AM (2007) Animal models of human placentation- a review. Placenta 28: S41-S47.
Casari E, Ferrario A, Morenghi E, Montanelli A (2010) Gardnerella, Trichomonas vaginalis, Candida, Chlamydia trachomatis, Mycoplasma hominis and Ureaplasma urealyticum in the genital discharge of symptomatic fertile and asymptomatic infertile women. New Microbiol 33: 69-76.
Cassell GH, Davis RO, Waites KB, Brown MB, Marriott PA, Stagno S, Davis JK (1983) Isolation of Mycoplasma hominis and Ureaplasma urealyticum from amniotic fluid at 16-20 weeks of gestation: potential effect on outcome of pregnancy. Sex Transm Dis 10: 294-302.
Cassell GH, Waites KB, Crouse DT, Rudd PT, Canupp KC, Stagno S, Cutter GR (1988) Association of Ureaplasma urealyticum infection of the lower respiratory tract with chronic lung disease and death in very-low-birth-weight infants. Lancet 30: 240-245.
Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R (1993) Ureaplasma urealyticum intrauterine infection: role in prematurity and disease in newborns. Clin Microbiol Rev 6: 69-87.
Chaim W, Horowitz S, David JB, Ingel F, Evinson B, Mazor M (2003) Ureaplasma urealyticum in the development of postpartum endometritis. Eur J Obstet Gynecol Reprod Biol 109: 145-148.
Chandran A, Watt JP, Santosham M (2005) Prevention of Haemophilus influenzae type b disease: past success and future challenges. Expert Rev Vaccines 4: 819-827.
Christiansen C, Black FT, Freundt EA (1981) Hybridization experiments with deoxyribonucleic acid from Ureaplasma urealyticum serovars I to VIII. Int J Syst Bacteriol 31: 259-262.
Citti C, Nouvel L-X, Baranowski E (2010) Phase and antigenic variation in mycoplasmas. Future Microbiol 5: 1073-1085.
Colaizy TT, Morris CD, Lapidus J, Sklar RS, Pillers DA (2007) Detection of ureaplasma DNA in endotracheal samples is associated with bronchopulmonary dysplasia after adjustment for multiple risk factors. Pediatr Res 61: 578-583.
Collins JJP, Kallapur SG, Knox CL, Nitsos I, Polglase GR, Pillow JJ, Kuypers E, Newnham JP, Jobe AH, Kramer BW (2010) Inflammation in fetal sheep from intra-amniotic injection of Ureaplasma parvum. Am J Physiol Lung Cell Mol Physiol 299: L852-L860,
70
Couldwell DL, Gidding HF, Freedman EV, McKechnie ML, Biggs K, Sintchenko V, Gilbert GL (2010) Ureaplasma urealyticum is significantly associated with non-gonococcal urethritis in heterosexual Sydney men. Int J STD AIDS 21: 337-341.
Cousens S, Blencowe H, Gravett M, Lawn JE (2010) Antibiotics for pre-term pre-labour rupture of membranes: prevention of neonatal deaths due to complications of pre-term birth and infection. Int J Epidemiol 39: 134-143.
Cunningham CK, Bonville CA, Hagen JH, Belkowitz JL, Kawatu RM, Higgins AM, Weiner LB (1996) Immunoblot analysis of anti-Ureaplasma urealyticum antibody in pregnant women and newborn infants. Clin Diagn Lan Immunol 3: 487-492.
Daffos R, Forestier F, Grangeot-Keros L, Pavlovsky MC (1984) Gradual appearance of total IgM in fetal serum during the second trimester of pregnancy. Application to the prenatal diagnosis of fetal infections. Ann Biol Clin 42: 135-139.
De Francesco MA, Negrini R, Pinsi G, Peroni L, Manca N (2009) Detection of ureaplasma biovars and polymerase chain reaction-based subtyping of Ureaplasma parvum in women with or without symptoms of genital infections. Eur J Microbiol Infect Dis 28: 641-646.
De Jong Z, Pontonnier F, Plante P, Perie N, Talazac N, Mansat A, Chabanon G (1990) Comparison of the incidence of Ureaplasma urealyticum in infertile men and in donors of semen. Eur Urol 18: 127-131. DeSilva NS, Quinn PA (1986) Endogenous activity of phospholipases A and C in Ureaplasma urealyticum. J Clin Microbiol 23: 354-359. DeSilva NS, Quinn PA (1991) Localization of endogenous activity of phospholipases A and C in Ureaplasma urealyticum. J Clin Microbiol 29: 1498-1503. DeSelvia NS, Quinn PA (1999) Characterization of phospholipase A1, A2, C activity in Ureaplasma urealyticum membranes. Mol Cell Biochem 201: 159-167. Dinsmoor MJ, Ramamurthy RS, Cassell GH, Gibbs RS (1989) Neonatal serologic response at term to the genital mycoplasmas. Pediatr Infect Dis J 8: 487-491. Donders GG, Van Calsteren K, Bellen G, Reybrouck R, Van den Bosch T, Riphagen I, Van Lierde S (2009) Predictive value for preterm birth of abnormal vaginal flora, bacterial vaginosis and aerobic vaginitis during the first trimester of pregnancy. BJOG 116: 1315-1324.
Dongya M, Wencheng X, Xiaobo M, Lu W (2008) Transition mutations in 23S rRNA account for acquired resistance to macrolides in Ureaplasma urealyticum. Microb Drug Resist 14: 183-186.
Edward DG, Freundt EA (1956) The classification and nomenclature of organisms of the pleuropneumonia group. J Gen Microbiol 14: 197-297.
Engel S, Bollmann R, Sokolowska-Kohler W, Audring H, Klug H (1988) Ureaplasma urealyticum and male infertility: an animal model: I. Artificial infection, breeding experiments and histological preparation of organs. Andrologia 20: 467-471.
Egawa T, Morioka I, Morisawa T, Yokoyama N, Nakao H, Ohashi M, Matsuo M (2007) Ureaplasma urealyticum and Mycoplasma hominis presence in umbilical cord is associated with pathogenesis of funisitis. Kobe J Med Sci 53: 241-249.
71
Eschenbach EA, Nugent RP, Rao AV et al. (1991) A randomized placebo-controlled trial of erythromycin for the treatment of Ureaplasma urealyticum to prevent premature delivery. The Vaginal Infections and Prematurity Study Group. Am J Obstet Gynecol 164: 734-742.
Famuyide ME, Hasday JD, Carter HC, Chesko KL, He JR, Viscardi RM (2009) Surfactant protein-A limits ureaplasma-mediated lung inflammation in a murine pneumonia model. Pediatr Res 66: 162-167.
Faye-Peterson OM (2008) The placenta in preterm birth. J Clin Pathol 61: 1261-1275.
Fidel PL Jr, Romero R, Wolf N, Cutright J, Ramirez M, Araneda H, Cotton DB (1994) Systemic and local cytokine profiles in endotoxin-induced preterm parturition in mice. Am J Obstet Gynecol 170: 1467-1475.
Gao T, Zablith N, Burns DH, Koski KG, Skinner CD (2009) Identification and quantities of human amniotic fluid components using capillary zone electrophoresis. Anal Biochem 388: 155-157.
García-Castillo M, Morosini M, Gálvez M, Baquero F, del Campo R, Meseguer M (2008) Differences in biofilm development and antibiotic susceptibility among clinical Ureaplasma urealyticum and Ureaplasma parvum isolates. J Antimicrob Chemother 62: 1027-1030.
Garland SM, Murton LJ (1987) Neonatal meningitis caused by Ureaplasma urealyticum. Pediatr Infect Dis 6: 868-870.
Gaynor M, Mankin AS (2005) Macrolide antibiotics: Binding site, mechanism of action, resistance. Curr Top Med Chem 3: 949-961.
Gdoura R, Kchaou W, Ammar-Keskes L, Chakroun N, Sellemi A, Znazen A, Rebai T, Hammami A (2008) Assessment of Chlamydia trachomatis, Ureaplasma urealyticum, Ureaplasma parvum, Mycoplasma hominis, and Mycoplasma genitalium in semen and first void urine specimens of asymptomatic male partners of infertile couples. J Androl 29: 198-206.
Gdoura R, Kchaou W, Chaari C, Znazen A, Keskes L, Rebai T, Hammami A (2007) Ureaplasma urealyticum, Ureaplasma parvum, Mycoplasma hominis and Mycoplasma genitalium infections and semen quality of infertile men. BMC Infect Dis 7: 129-137.
Gerber S, Vial Y, Hohlfeld P, Witkin SS (2003) Detection of Ureaplasma urealyticum in second-trimester amniotic fluid by polymerase chain reaction correlates with subsequent preterm labor and delivery. J Infect Dis 187: 518-521.
Giamarellos-Bourboulis (2008) Macrolides beyond the conventional antimicrobials: a class of potent immunomodulators. Int J Antimicrob Agents 31: 12-20.
Gibbs RS (2001) The relationship between infections and adverse pregnancy outcomes: an overview. Ann Periodontol 6: 153-163.
Gien J, Kinsella JP (2011) Pathogenesis and treatment of bronchopulmonary dysplasia. Curr Opin Pediatr 23: 305-313.
72
Glass JI, Lefkowitz EJ, Glass JS, Heiner CR, Chen EY, Cassell GH (2000) The complete sequence of the mucosal pathogen Ureaplasma urealyticum. Nature 407: 757-762.
Goldenberg RL, Andrews WW, Goepfert AR, Faye-Petersen O, Cliver SP, Waldemar AC, Hauth JC (2008b) The Alabama preterm birth study: umbilical cord blood Ureaplasma urealyticum and Mycoplasma hominis cultures in very preterm newborn infants. Am J Obstet Gynecol 198: 43.e1-43.e5.
Goldenberg RL, Culhane JF, Iams JD, Romero R (2008a) Epidemiology and causes of preterm birth. Lancet 371: 75-84.
Goldenberg RL, Hauth JC, Andrews WW (2000) Intrauterine infection and preterm delivery. New Engl J Med 342: 1500-1507.
Goldman RC, Scaglione F (2004) The macrolide-bacterium interaction and its biological basis. Curr Drug Targets Infect Disord 4: 241-260.
Golshani M, Eslami G, Ghobadloo SM, Fallah F, Goudarzi H, Rahbar AAS, Fayaz F (2007). Detection of Chlamydia trachomatis, Mycoplasma hominis and Ureaplasma urealyticum by multiplex PCR in semen sample of infertile men. Iranian J of Publ Health 36: 50-57.
Gomez R, Romero R, Nien JK, Medina L, Carstens M, Kim YM, Espinoza J, Chaiworapongsa T, Gonzalez R, Iams JD, Rojas I (2007) Antibiotic administration to patients with preterm premature rupture of membranes does not eradicate intra-amniotic infection. Matern Fetal Neonatal Med 20: 167-173.
Gonzàlex-Pedraza A, Ortiz C, Mota R, Dàvilla R, Dickinson E (2003) Role of bacteria associated with sexually transmitted infections in the etiology of lower urinary tract infection in primary care. Enfem Infecc MIcrobiol Clin 21: 89-92.
Gravett MG, Witkin SS, Haluska GJ, Edwards JL, Cook MJ, Novy MJ (1994) An experimental model for intraamniotic infection and preterm labor in rhesus monkeys. Am J Obstet Gynecol 171: 1660-1667.
Gray DJ, Robinson HB, Malone J, Thomson RB Jr. (1992) Adverse outcome in pregnancy following amniotic fluid isolation of Ureaplasma urealyticum. Prenat Diagn 12: 111-117. Gulbis B, Gervy C, Jauniaux E (1998) Amniotic fluid biochemistry in second-trimester trisomic pregnancies: relationships to fetal organ maturation and dysfunction. Early Hum Dev 52: 211-219. Haggerty CL, Totten PA, Ferris M, Martin DH, Hoferka S, Astete SG, Ondondo R, Norori J, Ness RB (2009) Clinical characteristics of bacterial vaginosis among women testing positive for fastidious bacteria. Sex Transm Infect 85: 242-248.
Harada K, Tanaka H, Komori S, Tsuji Y, Nagata K, Tsutsui H, Koyama K (2008) Vaginal infection with Ureaplasma urealyticum accounts for preterm delivery via induction in inflammatory responses. Microbiol Immunol 52: 297-304.
Heikkinen T, Laine K, Neuvonen PJ, Ekblad U (2000) The transplacental transfer of the macrolide antibiotics erythromycin, roxithromycin and azithromycin. BJOG 107: 770-755.
73
Hillier SL, Krohn MA, Kiviat NB, Watts DH, Eschenback DA (1991) Microbiologic causes and neonatal outcomes associated with chorioamnion infection. Am J Obstet Gynecol 165: 955-961.
Hillier SL, Krohn MA, Rabe LK, Klebanoff SJ, Eschenbach DA (1993) The normal vaginal flora, H2O2-producing lactobacilli, and bacterial vaginosis in pregnant women. Clin Infect Dis 16: S273-S281.
Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, Eschenback DA (1988) A case-control study of chorioamnion infection and histological chorioamnionitis in prematurity. N Engl J Med 319: 972-978.
Hirsch E, Muhle RA, Mussalli GM, Blanchard R (2002) Bacterially-induced preterm labor in the mouse does not require maternal interleukin-1 signalling. Am J Obstet Gynecol 186: 523-530.
Holst RM, Mattsby-Baltzer I, Wennerholm UB, Hagberg H, Jacobsson B (2005) Interleukin-6 and interleukin-8 in cervical fluid in a population of Swedish women in preterm labor: relationship to microbial invasion of the amniotic fluid, intra-amniotic inflammation, and preterm delivery. Acta Obstet Gynecol Scand 84: 551-557.
Horne AW, Stock SJ, King AE (2008) Innate immunity and disorders of the female reproductive tract. Reproduction 135: 739-749.
Hornef MW, Wick MJ, Rhen M, Normark S (2002) Bacterial strategies for overcoming host innate and adaptive immune responses. Nat Immunol 3: 1033-1040. Horowitz S, Mazor M, Horowitz J, Porath A, Glezerman M (1995b) Antibodies to Ureaplasma urealyticum in women with intraamniotic infection and adverse pregnancy outcome. Acta Obstet Gynecol Scand 74: 132-136. Horowitz S, Mazor M, Romero, R, Horowitz J, Glezerman M (1995a) Infection of the amniotic cavity with Ureaplasma urealyticum in the midtrimester of pregnancy. J Reprod Med 40: 375-379. Istivan TS, Coloe PJ (2006) Phospholipase A in Gram-negative bacteria and its role in pathogenesis. Microbiology 152: 1263-1274. Jacobsson B, Mattsby-Baltzer I, Hagberg H (2005) Interleukin-6 and interleukin-8 in cervical and amniotic fluid: relationship to microbial invasion of the chorioamniotic membranes. BJOG 112: 719-724 Jain R, Danziger LH (2004) The macrolide antibiotics: a pharmacokinetic and pharmacodynamics overview. Curr Pharm Des 10: 3045-3053. Jobe AH, Ikegami M (2001) Antenatal infection/inflammation and postnatal lung maturation and injury. Respir Res 2: 27-32.
Joste NE, Kundsin RB, Genest DR (1994) Histology and Ureaplasma urealyticum culture in 63 cases of first trimester abortion. Am J Clin Pathol 102: 729-732.
Kakulphimp J, Finch LR, Robertson JA (1991) Genome sizes of mammalian and avian ureaplasmas. Int J Syst Bacteriol 41: 326-327.
74
Kapatais-Zoumbos K, Chandler DKF, Barile MF (1985) Survey of immunoglobulin A protease activity among selected species of Ureaplasma and Mycoplasma: specificity for host immunoglobulin A. Infect Immun 47: 704-709.
Kasper DC, Mechtier TP, Böhm J, Petricevic L, Gleiss A, Spergser J, Witt A, Herkner KR, Berger A (2011) In utero exposure to Ureaplasma spp. is associated with increased rate of bronchopulmonary dysplasia and intraventricular haemorrhage in preterm infants. J Perinat Med 39: 331-336.
Kasper DC, Mechtler TP, Reischer GH, Witt A, Langgartner M, Pollak A, Herkner KR, Berger A (2010) The bacterial load of Ureaplasma parvum in amniotic fluid is correlated with an increased intrauterine inflammatory response. Diagn Microbiol Infect Dis 67: 117-121.
Kataoka S, Yamada T, Chou K, Nishida R, Morikawa M, Minami M, Yamada H, Sakuragi N, Minakami H (2006) Association between preterm birth and vaginal colonization by mycoplasmas in early pregnancy. J Clin Microbiol 44: 51-55.
Kechagia N, Bersimis S, Chatzipanagiotou S (2008) Incidence and antimicrobial susceptibilities of genital mycoplasmas in outpatient women with clinical vaginitis in Athens, Greece. J Antimicrob Chemother 62: 122-125.
Kemp MW, Saito M, Newnham JP, Nitsos I, Okamura K, Kallapur SG (2010) Preterm birth, infection, and inflammation advances from the study of animal models. Reprod Sci 17: 619-628.
Kenny GE, Cartwright FD (1993) Effect of pH, inoculum size, and incubation time on the susceptibility of Ureaplasma urealyticum to erythromycin. Clin Infect Dis 17: S215-S218.
Kenyon SL, Boulvain M, Neilson JP (2010) Antibiotics for preterm rupture of membranes. Cochrane Database Syst Rev: CD001058.
Kenyon SL, Pike K, Jones DR, Brocklehurst P, Marlow N, Salt A, Taylor DJ (2008) Childhood outcomes after prescription of antibiotic to pregnant women with spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet 372: 1319-1327.
Kenyon SL, Taylor DJ, Tarnow-Mordi W, the ORACLE Collaborative Group (2001a) Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLE I randomised trial. Lancet 2001 357: 979-985.
Kenyon SL, Taylor DJ, Tarnow-Mordi W, the ORACLE Collaborative Group (2001b) Broad-spectrum antibiotics from spontaneous preterm labour: the ORACLE II randomised trial. Lancet 357: 981-990.
Kiefer L, Rubin A, McCoy JB, Foltz EL (1955) The placental transfer of erythromycin. Am J Obstet Gynecol 69: 147-177.
Killan M, Brown MB, Brown TA, Freundt EA, Cassell GH (1984) Immunoglobulin A1 protease activity in strains of Ureaplasma urealyticum. Acta Pathol Microbiol Immunol Scand B 92: 61-64.
75
Killan M, Freundt EA (1984) Exclusive occurrence of an extracellular protease capable of cleaving the hinge region of human immunoglobulin A1 strains of Ureaplasma urealyticum. Isr J Med Sci 20: 938-941.
King AE, Kelly RW, Sallenave JM, Bocking AD, Challis JRG (2007) Innate immune defences in the human uterus during pregnancy. Placenta 28: 1099-1106.
Kingma PS, Whitsett JA (2006) In defense of the lung: surfactant protein A and surfactant protein D. Curr Opin Pharmacol 6: 277-283.
Klein LL, Gibbs RS (2004) Use of microbial cultures and antibiotics in the prevention of infection-associated preterm birth. Am J Obstet Gynecol 190: 1493-1502.
Knox CL (1998) Molecular subtyping, phylogeny and clinical relevance of Ureaplasma urealyticum isolates from pregnant women. PhD Thesis, Queensland University of Technology.
Knox CL, Allan JA, Allan JM, Edirisinghe WR, Stenzel D, Lawrence FA, Purdie DM, Timms P (2003) Ureaplasma parvum and Ureaplasma urealyticum are detected in semen after washing before assisted reproductive technology procedures. Fertil Steril 80: 921-929. Knox CL, Cave DG, Farrell DJ, Eastment HT, Timms P (1997) The role of Ureaplasma urealyticum in adverse pregnancy outcome. Aust NZ J Obstet Gynaecol 37: 45-51.
Knox CL, Dando SJ, Nitsos I, Kallapur SG, Jobe AH, Payton D, Moss TJ, Newnham JP (2010) The severity of chorioamnionitis in pregnant sheep is associated with in vivo variation of the surface-exposed multiple-banded antigen/gene of Ureaplasma parvum. Biol Reprod 83: 415-426.
Knox CL, Giffard P, Timms P (1998) The phylogeny of Ureaplasma urealyticum based on the mba gene fragment. Int J Syst Bacteriol 48: 1323-1331.
Kohn FM, Erdmann I, Oeda T, el Mulla KF, Schiefer HG, Schill WB (1998) Influence of urogenital infections of sperm functions. Andrologia 30: 72-80. Kong F, James G, Zhenfang M, Gordon S, Wang B, Gilbert GL (1999a) Phylogenetic analysis of Ureaplasma urealyticum- support for the establishment of a new species, Ureaplasma parvum. Int J Syst Bactiol 49: 1879-1889.
Kong F, Ma Z, James G, Gordon S, Gilbert G (2000) Species identification and subtyping of Ureaplasma parvum and Ureaplasma urealyticum using PCR-based assays. J Clin Microbiol 38: 1175-1179.
Kong F, Zhu X, Wang W, Zhou X, Gordon S, Gilbert GL (1999b) Comparative analysis and serovar-specific identification of multiple-banded antigen genes of Ureaplasma urealyticum biovar 1. J Clin Microbiol 37: 538-543.
Krausse R, Schubert S (2010) In vitro activities of tetracyclines, macrolides, fluoroquinolones and clindamycin against Mycoplasma hominis and Ureaplasma spp. isolated in Germany over 20 years. Clin Microbiol Infect 16: 1649-1655. Kundsin RB, Driscoll SG, Monson RR, Yeh C, Biano SA, Cochran WD (1984) Association of Ureaplasma urealyticum in the placenta with perinatal morbidity and mortality. N Engl J Med 310: 941-945.
76
Kundsin RB, Leviton A, Allred EN, Poulin SA (1996) Ureaplasma urealyticum infection of the placenta in pregnancies that ended prematurely. Obstet Gynecol 87: 122-127.
Kuroki Y, Takahashi M, Nishitani C (2007) Pulmonary collectins in innate immunity of the lung. Cell Microbiol 9: 1871-1879.
Kwinta P, Pietrzyk JJ (2010) Preterm birth and respiratory disease in later life. Expert Rev Respir Med 4: 593-604.
Lamont RF, Jaggat AN (2007) Emerging drug therapies for preventing spontaneous preterm labor and preterm birth. Expert Opin Investig Drugs 16: 337-345.
Latthe PM, Toozs-Hobson P, Gray J (2008) Mycoplasma and ureaplasma colonisation in women with lower urinary symptoms. J Obstet Gynaecol 28: 519-521.
Lee GY, Kenny GE (1987) Humoral immune response to polypeptides of Ureaplasma urealyticum in women with postpartum fever. J Clin Microbiol 25: 1841-1844.
Lee SE, Romero R, Kim EC, Yoon BH (2009) A high Nugent score but not a positive culture for genital mycoplasmas is a risk factor for spontaneous preterm birth. J Matern Fetal Neonatal Med 22: 212-217.
Li YH, Brauner A, Jensen JS, Tullus K (2002) Induction of human macrophage vascular endothelial growth factor and intracellular adhesion molecule-1 by Ureaplasma urealyticum and downregulation by steroids. Biol Neonate 82: 22-28.
Li YH, Brauner A, Jonsson B, van der Ploeg I, Söder O, Hoist M, Jensen JS, Lagercrantz H, Tullus K (2000) Ureaplasma urealyticum-induced production of proinflammatory cytokines by macrophages. Pediatr Res 48: 114-119.
Liepmann MF, Wattre P, Dewilde A, Papierok G, Delecour M (1988) Detection of antibodies to Ureaplasma urealyticum in pregnant women by enzyme-linked immunosorbent assay using membrane antigen and investigation of the significance of the antibodies. J Clin Microbiol 26: 2157-2160.
Ligon JV, Kenny GE (1991) Virulence of ureaplasmal urease for mice. Infect Immun 59: 1170-1171.
Lingwood VA, Quinn PA, Wilansky S, Nutikka A, Ruhnke HL, Miller RB (1990) Common sulfoglycolipid receptor for mycoplasmas involved in animal and human infertility. Biol Reprod 43: 694-697.
Lu C, Ye TI, Zhu Gx, Feng Py, Ma H, Lu Rb, Lai W (2010) Phenotypic and genetic characteristics of macrolide and lincosamide resistant Ureaplasma urealyticum isolated in Guangzhou, China. Curr Microbio 61: 44-49.
MacKenzie CR, Henrich B, Hadding U (1996) Biovar-specific epitopes of the urease enzyme of Ureaplasma urealyticum. J Med Microbiol 45: 366-371.
Madurga S, Sànchez-Céspedes J, Belda I, Vila J, Giralt E (2008) Mechanism of binding of fluoroquinolones to the quinolone resistance-determining region of DNA gyrase: towards an understanding of the molecular basis of quinolone resistance. Chem Bio Chem 9: 2081-2086.
77
Maniloff (1983) Evolution of wall-less prokaryotes. Ann Rev Microbiol 37: 477-99.
Mankin AS (2008) Macrolide myths. Curr Opin Microbiol 11: 414-421.
Marconi C, Ramos BRA, Peraçoli JC, Donders GGG, Silva MG (2011) Amniotic fluid interleukin-1 beta and interleukin-6, but not interleukin-8 correlate with microbial invasion of the amniotic cavity in preterm labor. Am J Reprod Immunol. 65: 549-556.
Martens MG, Young RL, Uribe M, Buttram VC, Faro S (1993) Presence of chlamyida, mycoplasma, ureaplasma, and other bacteria in the upper and lower genital tracts of fertile and infertile populations. Infect Dis Obstet Gynecol 1: 85-90.
Marvić G (2004) Macrolide resistance based on the Erm-mediated rRNA methylation. Curr Drug Targets Infect Disord 4: 193-202.
Mazor M, Chaim W, Horowitz S, Leiberman JR, Glezerman M (1993) Successful treatment of preterm labour by eradication of Ureaplasma urealyticum with erythromycin. Arch Gynecol Obstet 253: 215-218.
McClure EM, Goldenberg RL (2009) Infection and stillbirth. Semin Fetal Neonatal Med 14: 182-189.
McCormack WM (1986b) Ureaplasma urealyticum: ecological niche and epidemiological considerations. Pediatr Infect Dis 5: S232-S233.
McCormack WM, Rosner B, Alpert S, Evrard JR, Crockett VA, Zinner SH (1986a) Vaginal colonization with Mycoplasma hominis and Ureaplasma urealyticum. Sex Transm Dis 13: 67-70.
McCormick MC (1985) The contribution of low birth weight to infant mortality and childhood morbidity. N Engl J Med 312: 82-90.
Mendelson CR (2009) Minireview: fetal-maternal hormone signalling in pregnancy and labor. Mol Endocrinol 23: 947-954.
Menon R, Peltier MR, Eckardt J, Fortunato SJ (2009) Diversity in cytokine response to bacteria associated with preterm birth by fetal membranes. Am J Obstet Gynecol 201: 306.e1-6.
Michaels JA, Dasari S, Pereira L, Reddy AP, Lapidus JA, Lu X, Jacob T, Thomas A, Rodland M, Roberts CT Jr, Gravett MG, Nagalla SR (2007) Comprehensive proteome analysis of the human amniotic fluid proteome: gestational age-dependent changes. J Proteome Res 6: 1277-1285.
Milch RA, Rall DP, Tobie JE (1957) Bone localization of the tetracyclines. J Natl cancer Inst 19: 87-93.
Mitchell BF, Taggart MJ (2009) Are animal models relevant to key aspects of human parturition? Am J Physiol Regul Integr Comp Physiol 297: R525-545.
Mobley HLT, Island MD, Hausinger RP (1995) Molecular biology of microbial ureases. Microbiol Rev 59: 451-480.
78
Modena AB, Fieni S (2004) Amniotic fluid dynamics. Acta Bio Medica Ateneo Parmense 75: 11-13.
Møller BR, Jørgenses AS, From E, Stenderup A (1985) Chlamydia, mycoplasmas, ureaplasmas and yeasts in the lower genital tract of females. Acta Obstet Gynecol Scand 64: 145-149.
Momynaliev K, Klubin A, Chelysheva V, Selezneva O, Akopian T, Gavorun V (2007) Comparative genome analysis of Ureaplasma parvum clinical isolates. Res Microbiol 158: 371-378.
Monecke S, Helbig JH, Jacobs E (2003) Phase variation of the multiple banded protein in Ureaplasma urealyticum and Ureaplasma parvum. Int J Med Microbiol 293: 203-211.
Mor G, Cardenas I (2010) The immune system in pregnancy: a unique complexity. Am J Reprod Immunol 63: 425-433.
Mor G, Cardenas I, Abrahams V, Guller S (2011) Inflammation and pregnancy: the role of the immune system at the implantation site. Ann N Y Acad Sci 1221: 80-87.
Morioka I, Fujibayashi H, Enoki E, Yokoyama N, Yokozaki H, Matsuo M (2010) Congenital pneumonia with sepsis caused by intrauterine infection of Ureaplasma parvum in a term newborn: a first case report. J Perinatol 30: 359-362.
Moss TJ, Knox CL, Kallapur SG, Nitsos I, Theodoropoulos C, Newnham JP, Ikegami M, Jobe AH (2008) Experimental amniotic fluid infection in sheep: effects of Ureaplasma parvum serovars 3 and 6 on preterm or term fetal sheep. Am J Obstet Gynecol 198: 122.e1-122.e8.
Moss TJ, Nitsos I, Ikegami M, Jobe AH, and Newnham JP (2005) Experimental intrauterine ureaplasma infection in sheep. Am J Obstet Gynecol 192: 1179-1186.
Naessens A, Foulon W, Breynaert J, Lauwers S (1988) Serotypes of Ureaplasma urealyticum isolated from normal pregnant women and patients with pregnancy complications. J Clin Microbiol 26: 319-322. Naessens A, Foulon W, Debrucker P, Devroey P, Lauwers S (1986) Recovery of microorganisms in semen and relationship to semen evaluation. Fertil Steril 45: 101-105.
Namba F, Hasegawa T, Nakayama M, Hamanaka T, Yamashita T, Nakahira K, Kimoto A, Nozaki M, Nishihara M, Mimura K, Yamada M, Kitajima H, Suehara N, Yanagihara I (2010) Placental features of chorioamnionitis colonized with Ureaplasma species in preterm delivery. Pediatr Res 67: 166-172.
Nelson Am Press N, Bautista CT, Arevalo J, Quiroz C, Calderon M, Campos K, Bryant A, Shantz-Dunn J, Dahodwala N, Vera M, Vivar A, Saito M, Gilman RH (2007) Prevalence of sexually transmitted infections and high-risk sexual behaviours in heterosexual couples attending sexually transmitted disease clinics in Peru. Sex Trans Dis 34: 344-361.
Neyrolles O, Ferris S, Behbahani N, Montagnier L, Blanchard A (1996) Organization of Ureaplasma urealyticum urease gene cluster and expression in a suppressor strain of Escherichia coli. J Bacteriol 178: 647-655.
79
Normann E, Lacaze-Masmonteil T, Eaton F, Schwendimann L, Gressens P, Thébaud B (2009) A novel mouse model of ureaplasma-induced perinatal inflammation: effects on lung and brain injury. Pediatr Res 65: 430-436.
Novy MJ, Duffy L, Axthelm MK, Sadowsky DW, Witkin SS, Gravett MG, Cassell GH, Waites KB (2009) Ureaplasma parvum or Mycoplasma hominis as sole pathogens cause chorioamnionitis, preterm delivery, and fetal pneumonia in rhesus macaques. Reprod Sci 16: 56-70.
Nύñez-Calonge R, Caballero P, Redondo C, Baquero F, Martinez-Ferrer M, Meseguer MA (1998) Ureaplasma urealyticum reduces motility and induces membrane alterations in human spermatozoa. Hum Reprod 13: 2756-2761.
Ogasawara KK, Goodwin TM (1997) The efficacy of prophylactic erythromycin in preventing vertical transmission of Ureaplasma urealyticum. Am J Perinatol 14: 233-237.
Okogbule-Wonodi AC, Chesko KL, Famuyide ME, Viscardi RM (2011) Surfactant protein-A enhances ureaplasmacidal activity in vitro. Innate Immun 17: 145-151.
Ollikainen J, Hiekkaniemi H, Korppi M, Katila ML, Heinonen K (1993) Ureaplasma urealyticum cultured from brain tissue of preterm twins who died on intraventricular hemorrhage. Scand J Infect Dis 25: 529-531.
Olomu IN, Hecht JL, Onderdonk AO, Allred EN, Leviton A; Extremely Low Gestational Age Newborn Study Investigators (2009) Perinatal correlates of Ureaplasma urealyticum in placenta parenchyma of singleton pregnancies that end before 28 weeks of gestation. Pediatrics 123: 1329-1336.
Olsen GJ, Woese CR, Overbeek R (1994) The winds of (evolutionary) change: breathing new life into microbiology. J Bacteriol 176: 1-6.
Onderdonk AB, Hecht JL, McElrath TF, Delaney ML, Allred EN, Leviton A, ELGAN Study Investigators (2008) Colonization of second-trimester placental parenchyma. Am J Obstet Gynecol 199: 52.e1-52.e10.
Oriel JD, Ridgway GL (1983) Comparison of tetracycline and minocycline in the treatment of non-gonococcal urethritis. Br J Vener Dis 59: 245-248.
Peltier MR, Freeman AJ, Mu HH, Cole BC (2007) Characterization of the macrophage-stimulating activity from Ureaplasma urealyticum. Am J Reprod Immunol 57: 186-192.
Peltier MR, Tee SC, Smulian JC (2008) Effect of progesterone on proinflammatory cytokine production by monocytes stimulated with pathogens associated with preterm birth. Am J Reprod Immunol 60: 346-353.
Pereyre S, Métifiot M, Cazanave C, Renaudin H, Charron A, Bébéar C, Bébéar CM (2007) Characterisation of in vitro-selected mutants of Ureaplasma parvum resistant to macrolides and related antibiotics. Int J Antimicrob Agents 29: 207-211.
Perni SC, Vardhana S, Korneeva I, Tuttle SL, Paraskevas L, Chasen ST, Kalish RB, Witkin SS (2004) Mycoplasma hominis and Ureaplasma urealyticum in midtrimester amniotic fluid: association with amniotic fluid cytokine levels and pregnancy outcome. Am J Obstet Gynecol 191: 1382-1386.
80
Pinna GS, Skevaki CL, Kafetzis DA (2006) The significance of Ureaplasma urealyticum as a pathogenic agent in the paediatric population. Curr Opin Infect Dis 19: 283-289.
Polglase GR, Dalton RGB, Nitsos I, Knox CL, Pillow JJ, Jobe AH, Moss TJM, Newnham JP, Kallapur SG (2010) Pulmonary vascular and alveolar development in preterm lambs chronically colonized with Ureaplasma parvum. Am J Physiol Lung Cell Mol Physiol 299: L232-L241.
Quinn PA (1986) Evidence of an immune response to Ureaplasma urealyticum in perinatal morbidity and mortality. Pediatr Infect Dis 5: S282-S287.
Quinn PA, Butany J, Taylor J, Hannah W (1987) Chorioamnionitis: its association with pregnancy outcome and microbial infection. Am J Obstet Gynecol 156: 379-387.
Quinn PA, Gillan JE, Markestad T, St John MA, Daneman A, Lie K, Li HC, Czegled-Magy E, Klein A (1985) Intrauterine infection with Ureaplasma urealyticum as a cause of fatal neonatal pneumonia. Pediatr Infect Dis 4: 538-543.
Quinn PA, Li HC, Th’ng C, Dunn M, Butany J (1993) Serological response to Ureaplasma urealyticum in the neonate. Clin Infect Dis 17: S136-S143.
Quinn PA, Shewchuk AB, Shuber J, Lie KI, Ryan E, Sheu M, Chipman ML (1983) Serologic evidence of Ureaplasma urealyticum infection in women with spontaneous pregnancy loss. Am J Obstet Gynecol 145: 245-250. Razin S (1996) Mycoplasmas. In: Baron S, editor. Medical Microbiology, volume 4. Galveston: University of Texas Medical Branch.
Razin S, Harasawa R, Barile MF (1983) Cleavage patterns of the mycoplasma chromosome, obtained by using restriction endonucleases, as indicators of genetic relatedness among strains. Int J Syst Bacteriol 33: 201-206.
Reichart M, Kahane I, Bartoov B (2000) In vivo and in vitro impairment of human and ram sperm nuclear chromatin integrity by sexually transmitted Ureaplasma urealyticum infection. Biol Reprod 63: 1041-1048.
Reichart M, Levi H, Kahane I, Bartoov B (2001) Dual energy metabolism-dependent effect of Ureaplasma urealyticum infection on sperm activity. J Androl 22: 404-412.
Reyes L, Reinhard M, Brown MB (2009) Different inflammatory responses are associated with Ureaplasma parvum-induced UTI and urolith formation. BMC Infect Dis 9: 1-12.
Roberts MC (2008) Update on macrolide-lincosamide-streptogramin, ketolide, and oxazolidinone resistance genes. FEMS Microbiol Lett 282: 147-159.
Roberts MC, Kenny GE (1986) Dissemination of the tetM tetracycline resistance determinant to Ureaplasma urealyticum. Antimicrob Agents Chemother 29: 350-352.
Roberts MC, Sutcliffe J, Courvalin P, Jensen LB, Rood J ,Seppala H (1986) Nomenclature for macrolide and macrolide-lincosamide-streptogramin B resistance determinants. Antimicrob Agents Chemother 43: 2823-2830.
81
Robertson JA, Stemke GW (1982) Expanded serotyping scheme for Ureaplasma urealyticum strains isolated from humans. J Clin Microbiol 15: 873-878.
Robertson JA, Stemke GW, Davis JW, Jr, Harasawa R, Thirkell D, Kong F, Shepard MC, Ford D (2002) Proposal of Ureaplasma parvum sp. nov. and emended description of Ureaplasma urealyticum (Shepard et al. 1974) Robertson et al. 2001. Int J Syst Evol Microbiol 52: 587-597.
Robertson JA, Stemke GW, Maclellan SG, Taylor DE (1988) Characterization of tetracycline-resistant strains of Ureaplasma urealyticum. J Antimicrob Chemother 21: 319-332.
Robertson JA, Stemler ME, Stemke GW (1984) Immunoglobulin A protease activity of Ureaplasma urealyticum. J Clin Microbiol 19: 255-258.
Rogers MJ, Simmons J, Walker RT, Weisburg WG, Woese CR, Tanner RS, Robinson IM, Stahl DA, Olsen G, Leach RH, Maniloff J (1985) Construction of the mycoplasma evolutionary tree from 5S rRNA sequence data. Proc Natl Acad Sci 82: 1160-1164.
Romero R, Espinoza J, Goncalves LF, Kusanovic JP, Friel LA, Nien JK (2007) Inflammation in preterm and term labour and delivery. Semin Fetal Neonatal Med 11: 317-326.
Rose BI, Scott B (1994) Sperm motility, morphology, hyperactivation, and ionophore-induced acrosome reactions after overnight incubation with mycoplasmas. Fertil Steril 61: 341-348.
Samra Z, Rosenbert S, Dan M (2011) Susceptibility of Ureaplasma urealyticum to tetracycline, doxycycline, erythromycin, roxithromycin, clarithromycin, azithromycin, levofloxacin and moxifloxacin. J Chemother 23: 77-79.
Sanchez PJ, Regan JA (1988) Ureaplasma urealyticum colonization and chronic lung disease in low birth weight infants. Pediatr Infect Dis 7: 542-546.
Schelonka RL, Waites KB (2007) Ureaplasma infection and neonatal lung disease. Semin Perinatol 31: 2-9.
Schmiel DH, Miller VL (1999) Bacterial phospholipases and pathogenesis. Microbes Infect 1: 1103-1112.
Seral C, Castillo FJ, Rubio-Calvo MC, Fenoll A, García C, Gómez-Lus R (2001) Distribution of resistance genes tet(M), aph3’-III, catpC194 and the integrase gene of Tn1545 in clinical Streptococcus pneumoniae harbouring erm(B) and mef(A) genes in Spain. J Antimicrob Chemother 47: 863-866.
Shepard MC (1954) The recovery of pleuropneumonia-like organisms from Negro men with and without non-gonococcal urethritis. Amer J Syph Gonor Vener Dis 38: 113-124.
Shepard MC (1956) T-form colonies of pleuropneumonia-like organisms. J Bacteriol 71: 362-369.
Shepard MC, Lunceford CD, Ford DK, Purcell RH, Taylor-Robinson D, Razin S, Black FT (1974) Ureaplasma urealyticum gen. nov., sp. nov.: proposed nomenclature for the human T (T-strain) mycoplasmas. Int J Syst Bacteriol 24: 160-171.
82
Shepard MC, Masover GK (1979) Special features of ureaplasmas. In: Barile MF and Razin S, editors. The Mycoplasmas, vol. 1. New York: Academic Press. pp. 452-494.
Shimizu T, Kida Y, Kuwano K (2008) Ureaplasma parvum lipoproteins, including MB antigen, activate NF-κB through TLR1, TLR2 and TLR6. Microbiology 154: 1318-1325.
Sirand-Pugnet P, Lartigue C, Marenda M, Jacob D, Barŕe A, Barbe V ,Schenowitz C, Mangenot S, Coulox A, Segurens B, de Daruvar A, Blanchard A, Citti C (2007) Being pathogenic, plastic, and sexual while living with a nearly minimal bacterial genome. PLoS Genetics 3: 0744-0758.
Skerk V, Schönwald S, Krhen I, Markovinović L, Beus A, Kuzmanović NS, Kružić V, Vince A (2002) Aetiology of chronic prostatitis. Int J Antimicrob Agents 19: 471-474.
Smith DGE, Russell WC, Ingledew WJ, Thirkell D (1993) Hydrolysis of urea by Ureaplasma urealyticum generates a transmembrane potential with resultant ATP synthesis. J Bacteriol 175: 3253-3258.
Sozanskii AM (1961) The biochemical composition of amniotic fluid and of maternal and fetal blood at various periods of pregnancy. Biull Eksp Biol Med 51: 323-326.
Spooner RK, Russell WC, Thirkell D (1992) Characterization of the immunoglobulin A protease of Ureaplasma urealyticum. Infect Immun 60: 2544-2546.
Sung TJ, Xiao L, Duffy L, Waites KB, Chesko KL, Viscardi RM (2011) Frequency of ureaplasma serovars in respiratory secretions of preterm infants at risk for bronchopulmonary dysplasia. Pediatr Infect Dis J 30: 379-383.
Swenson CE, Van Hamont J, Dunbar BS (1983) Specific protein differences among strains of Ureaplasma urealyticum as determined by two-dimensional gel electrophoresis and a sensitive silver stain. Int J Syst Bacteriol 33: 417-421.
Takebe S, Numata A, Kobashi K (1984) Stone formation by Ureaplasma urealyticum in human urine and its prevention by urease inhibitors. J Clin Microbiol 20: 869-873.
Taylor-Robinson D, Csonka GW, Prentice MJ (1977) Human intra-urethral inoculation of ureaplasmas. Q J Med 46: 309-326.
Taylor-Robinson D, Lamont RF (2010) Mycoplasmas in pregnancy. BJOG 118: 164-174.
Teng LJ, Zheng X, Glass JI, Watson HL, Cassell GH (1994) Ureaplasma urealyticum biovar specifity and diversity are encoded in multiple-banded antigen gene. J Clin Microbiol 32: 1464-1469.
Thomakos N, Daskalakis G, Papapanagiotou A, Papantoniou N, Mesogitis S, Antsaklis A (2010) Amniotic fluid interleukin-6 and tumor necrosis factor-α at mid-trimester genetic amniocentesis: relationship to intra-amniotic microbial invasion and preterm delivery. Eur J Obstet Gynecol Reprod Biol 148: 147-151.
Tsangaris GT, Kolialexi A, Karamessinis PM, Anagnostopoulos AK, Antsaklis A, Fountoulakis M, Mavrou A (2006) The normal human amniotic fluid supernatant proteome. In vivo 20: 479-490.
83
Upadhyaya M, Hibbard BM, Walker SM (1984) The effect of Ureaplasma urealyticum on semen characteristics. Fertil Steril 41: 304-308.
Van Den Berg JP, Westerbeek EA, van der Klis FR, Berbers GA, van Elburg RM (2011) Transplacental transport of IgG antibodies to preterm infants: a review of the literature. Early Hum Dev 87: 67-72.
Van Den Bosch H (2003) Intracellular phospholipases A Biochim Biophys Acta 30: 191-246.
Van Den Broek NR, White SA, Goodall M, Ntonya C, Kayira E, Kafulafula G, Neilson JP (2009) The APPLe study: a randomized, community-based, placebo-controlled trial of azithromycin for the prevention of preterm birth, with meta-analysis. PLoS Med 6: e1000191.
Van Waarde WM, Brus F, Okken A, Kimpen JLL (1997) Ureaplasma urealyticum colonization, prematurity and bronchopulmonary dysplasia. Eur Respir J 10: 886-890.
Ventolini G, Lee M (2011) Clinical antibiotic response of mycoplasma and ureaplasma in patients with symptomatic recurrent vulvovaginal colonization. J Reprod Med 56: 138-141.
Viscardi RM (2010) Ureaplasma species: role in diseases of prematurity. Clin Perinatol 37: 393-409.
Viscardi RM, Atamas SP, Luzina IG, Hasday JD, He JR, Sime PJ, Coalson JJ, Yoder BA (2006) Antenatal Ureaplasma urealyticum respiratory tract infection stimulates proinflammatory, profibrotic responses in the preterm baboon lung. Pediatr Res 60: 141-146.
Viscardi RM, Hasday JD (2009) Role of Ureaplasma species in neonatal chronic lung disease: epidemiologic and experimental evidence. Pediatr Res 65: 84-90.
Viscardi RM, Hashmi N, Gross GW, Sun CC, Rodriguez A, Fairchild KD (2008) Incidence of invasive ureaplasma in VLBW infants: relationship to severe intraventricular haemorrhage. J Perinatol 28: 759-766.
Viscardi RM, Kaplan J, Lovchik JC, He JR, Hester L, Rao S, Hasday JD (2002) Characterization of a murine model of Ureaplasma urealyticum pneumonia. Infect Immun 70: 5721-5729.
Viscardi RM, Manimtim WM, Sun CC, Duffy L, Cassell GH (2002) Lung pathology in premature infants with Ureaplasma urealyticum infection. Pediatr Dev Pathol 5: 141-150.
Vogel I, Thorsen P, Hogan VK, Schieve LA, Jacobsson B, Ferre CD (2006) The joint effect of vaginal Ureaplasma urealyticum and bacterial vaginosis on adverse pregnancy outcomes. Acta Obstet Gynecol Scand 85: 778-785.
Volgmann T, Ohlinger R, Panzig B (2005) Ureaplasma urealyticum-harmless human reproduction? A review. Arch Gynecol Obstet 273: 133-139.
Waites KB, Katz B, Schelonka RL (2005) Mycoplasmas and ureaplasmas as neonatal pathogens. Clin Microbiol Rev 18: 757-789.
84
Waites BK, Rudd PT, Crouse DT, Canupp KC, Nelson KG, Ramsey C, Cassell GH (1988) Chronic Ureaplasma urealyticum and Mycoplasma hominis infections of central nervous systems in preterm infants. Lancet 2: 17-21. Waites KB, Schelonka RL, Xiao L, Grigsby PL, Novy MJ (2009) Congenital and opportunistic infections: Ureaplasma species and Mycoplasma hominis. Semin Fetal Neonatal Med 14: 190-199.
Wang EE, Frayha H, Watts J, Hammerberg O, Chernesky MA, Mahony JB, Cassell GH (1988) Role of Ureaplasma urealyticum and other pathogens in the development of chronic lung disease of prematurity. Pediatr Infect Dis 7: 547-551. Wang Y, Han XD, Hou YY, Chen JX (2005) Ureaplasma urealyticum infection related to seminal plasma immunosuppressive factors, semen pH and liquefaction duration. Arch Androl 51: 267-270. Wang Y, Liang C, Wu J, Xu C, Qin S, Goa E (2006) Do Ureaplasma urealyticum infections in the genital tract affect semen quality? Asian J Androl 8: 562-568. Wang Y, Wu ZW, Zhang LF, Wu XK, Yi L, Han X (2010) Effects of Ureaplasma urealyticum infection on the male reproductive system in experimental rats. Andrologia 42: 297-301. Watson HL, Blalock DK, Cassell GH (1990) Variable antigens of Ureaplasma urealyticum containing both serovar-specific and serovar-cross-reactive epitopes. Infect Immun 58: 3679-3688. Watson HL, Zheng X, Cassell GH (1993) Structural variations and phenotypic switching of mycoplasmal antigens. Clin Infect Dis 17: S183-S186. Williams LE, Wernegreen JJ (2010) Unprecedented loss of ammonia assimilation capability in a urease-encoding bacterial mutualist. BMC Genomics 11: 687-700. Witt A, Berger A, Gruber CJ, Petricevic L, Apfalter P, Husslein P (2005) IL-8 concentrations in maternal serum, amniotic fluid and cord blood in relation to different pathogens within the amniotic cavity. J Perinat Med 33: 22-26. Witt A, Berger A, Gruber CJ, Petricevic L, Apfalter P, Worda C, Husslein P (2005) Increased intrauterine frequency of Ureaplasma urealyticum in women with preterm labor and preterm premature rupture of the membranes and subsequent caesarean delivery. Am J Obstet Gynecol 193: 1663-1669. Woese CR, Maniloff J, Zablen LB (1980) Phylogenetic analysis of the mycoplasmas. Proc Natl Acad Sci 77: 494-498.
Wogelius P, Nøgaard M, Gislum M, Pedersen L, Schønheyder HC, Sørensen HT (2005) Further analysis of the risk of adverse birth outcome after maternal use of fluoroquinolones. Int J Antimicrob Agents 26: 323-326.
Wolf M, Mϋller T, Dandekar T, Pollack JD (2004) Phylogeny of Firmicutes with special reference to Mycoplasma (Mollictes) as inferred from phosphoglycerate kinase amino acid sequence data. Int J Syst Evol Microbiol 54: 871-875.
Xiao L, Crabb DM, Duffy LB, Paralanov V, Glass JI, Hamilos DL, Waites KB 2011 Mutations in ribosomal proteins and ribosomal RNA confer macrolide resistance in human Ureaplasma spp. Int J Antimicrob Agents 37: 377-379.
85
Xie X, Zhang J (2006) Trends in the rates of resistance of Ureaplasma urealyticum to antibiotics and identification of the mutation site in the quinolone resistance-determining region in Chinese patients. FEMS Microbiol Lett 259: 181-186.
Xu C, Sun GF, Zhu YF, Wang YF (1997) The correlation of Ureaplasma urealyticum with infertility. Andrologia 29: 219-226. Yoon BH, Kim YA, Romero R, Kim JC, Park KH, Kim MH, Park JS (1999) Association of oligohydramnios in women with preterm premature rupture of membranes with an inflammatory response in fetal, amniotic, and maternal compartments. Am J Obstet Gynecol 18: 784-788. Yoon BH, Romero R, Lim JH, Shim SS, Hong JS, Shim JY, Jun JK (2003) The clinical significance of detecting Ureaplasma urealyticum by the polymerase chain reaction in the amniotic fluid of patients with preterm labor. Am J Obstet Gynecol 189: 919-924.
Yoon BH, Romero R, Park JS, Chang JW, Kim YA, Kim JC, Kim KS (1998) Microbial invasion of the amniotic cavity with Ureaplasma urealyticum is associated with a robust host response in fetal, amniotic, and maternal compartments. Am J Obstet Gynecol 179: 1254-1260.
Yoon BH, Romero R, Park JS, Kim CJ, Kim SH, Chio JH, Han TR (2000) Fetal exposure to an intra-amniotic inflammation and the development of cerebral palsy at the age of three years. Am J Obstet Gynecol 182: 675-681.
Yoshida T, Ishiko H, Yasuda M, Takahashi Y, Nomura Y, Kubota Y, Tamaki M, Maeda S, Deguchi T (2005) Polymerase chain reaction-based subtyping of Ureaplasma parvum and Ureaplasma urealyticum in first-pass urine samples from men with or without urethritis. Sex Transm Dis 32: 454-457.
Yudin MH, van Schalkwyk J, Van Eyk N, Boucher M, Castillo B, gruslin A, Money DM, Murphy K, Ogilvie G, Paquet C, Steenbeek A, Wong T, GagnonR, Hudon L, Basso M, Bos H, Delisle MF, Farine D, Grabowska K, Menticoglou S, Mundle WR, Murphy-Kaulbeck LC, Ouellet A, Pressey T, Roggensack A (2009) Antibiotic therapy in preterm premature rupture of the membranes. J Obstet Gynecol Can 31: 868-874.
Zakeri B, Wright GD (2008) Chemical biology of tetracycline antibiotics. Biochem Cell Biol 86: 124-136.
Zdrodowska-Stefanow B, Klosowska WM, Ostaszewska-Puchalska I, Bułhak-Koziol V, Kotowicz B (2006a) Ureaplasma urealyticum and Mycoplasma hominis infection in women with urogenital diseases. Adv Med Sci 51: 250-253.
Zdrodowska-Stefanow B, Klosowska WM, Ostaszewska-Puchalska I, Bułhak-Koziol V, Kotowicz B (2006b) Mycoplasma hominis and Ureaplasma urealyticum infections in male urethritis and its complications. Adv Med Sci 51: 254-257.
Zeighami H, Peerayeh SN, Yazdi RS, Sorouri R (2009) Prevalence of Ureaplasma urealyticum and Ureaplasma parvum in semen of infertile and healthy men. Int J STD AIDS 20: 387-390.
Zervomanolakis I, Ott HW, Hadziomerovic D, Mattle V, Seeber BE, Virgolini I, Heute D, Kissler S, Leyendecker G, Wildt L (2007) Physiology of upward transport in the human female genital tract. Ann N Y Acad Sci 1101: 1-20.
86
Zheng X, Lau K, Frazier M, Cassell GH, Watson HL (1996) Epitope mapping of the variable repetitive region with the MB antigen of Ureaplasma urealyticum. Clin Diag Lab Immunol 3: 774-778. Zheng X, Teng LJ, Glass JI, Blanchard A, Cao Z, Kempf MC, Watson HL, Cassell GH (1994) Size variation of a major serotype-specific antigen of Ureaplasma urealyticum. Ann N Y Acad Sci 15: 299-301.
Zheng X, Teng LJ, Watson HL, Glass JI, Blanchard A, Cassell GH (1995) Small repeating units within the Ureaplasma urealyticum MB antigen gene encode serovar specificity and are associated with antigen size variation. Infect Immun 63: 891-898.
Zheng X, Watson HL, Waites KB, Cassell GH (1992) Serotype diversity and antigen variation among invasive isolates of Ureaplasma urealyticum from neonates. Infect Immun 60: 3472-3474.
Zimmerman CR, Stiedl T, Rosengarten R, Spergser J (2009) Alternate phase variation in expression of two major surface membrane proteins (MBA and UU376) of Ureaplasma parvum serovar 3. FEMS Microbiol Lett 292: 187-193.
87
Chapter 3
PAPER 1
88
Maternal administration of erythromycin fails to eradicate intrauterine
ureaplasma infection in an ovine model
Samantha J Dando 1, Ilias Nitsos 2, John P Newnham 2, Alan H Jobe 3, Timothy
JM Moss 4, Christine L Knox 1
1 Institute of Health & Biomedical Innovation, Faculty of Science and Technology, Queensland University of Technology, Brisbane, Queensland, 4059.
2 School of Women’s and Infants’ Health, The University of Western Australia, Perth, Western Australia, 6009.
3 Department of Neonatology and Pulmonary Biology, Cincinnati Children’s Hospital Medical Center, University of Cincinnati, Cincinnati, Ohio, 45229, USA.
4 The Ritchie Centre, Monash Institute of Medical Research, Monash University, Clayton, Victoria, 3168.
Published in: Biology of Reproduction (2010) 83: 616-622.
89
Statement of Joint Authorship
Samantha J Dando (candidate):
Contributed to the experimental design and research plan; assisted with the collection of clinical samples; performed all ureaplasma cultures, histopathology analysis and prepared amniotic fluid samples for liquid chromatography-mass spectrometry. Analysed and interpreted the data, performed statistical analyses and wrote the manuscript
Ilias NItsos:
Contributed to the experimental design and research plan; performed intra-amniotic injections, administered erythromycin treatment to animals and assisted in the collection of samples. Contributed to the manuscript.
John P Newnham:
Contributed to the experimental design and research plan; performed intra-amniotic injections and assisted in the collection of samples. Contributed to the manuscript.
Alan H Jobe:
Contributed to the experimental design and research plan; performed fetal post-mortems and assisted in the collection of samples. Contributed to the manuscript.
Timothy JM Moss:
Contributed to the experimental design and research plan and made a significant contribution to the manuscript.
Christine L Knox:
Supervised the project, contributed to the experimental design and research plan and assisted in the collection of samples. Assisted in the interpretation of data and made a significant contribution to the manuscript.
90
Abstract
Erythromycin is the standard antibiotic used for treatment of Ureaplasma species
during pregnancy; however, maternally administered erythromycin may be
ineffective at eliminating intra-amniotic ureaplasma infections. We asked if
erythromycin would eradicate intra-amniotic ureaplasma infections in pregnant
sheep. At 55 days of gestation (d, term = 150 d) pregnant ewes received intra-
amniotic injections of erythromycin-sensitive U. parvum serovar 3 (n = 16) or 10B
medium (n = 16). At 100 d, amniocentesis was performed; five fetal losses
(ureaplasma group: n = 4; 10B group: n = 1) had occurred by this time. Remaining
ewes were allocated into treatment subgroups: medium only (M, n = 7); medium
and erythromycin (M/E, n = 8); ureaplasma only (Up, n = 6) or ureaplasma and
erythromycin (Up/E, n = 6). Erythromycin was administered intra-muscularly (500
mg), eight-hourly for four days (100 d – 104 d). Amniotic fluid samples were
collected at 105 d. At 125 d preterm fetuses were surgically delivered and
specimens were collected for culture and histology. Erythromycin was quantified in
amniotic fluid by liquid chromatography-mass spectrometry. Ureaplasmas were
isolated from the amniotic fluid, chorioamnion and fetal lung of animals from the Up
and Up/E groups, however, the numbers of U. parvum recovered were not different
between these groups. Inflammation in the chorioamnion, cord and fetal lung was
increased in ureaplasma-exposed animals compared to controls, but was not
different between the Up and Up/E groups. Erythromycin was detected in amniotic
fluid samples, although concentrations were low (<10 - 76 ng/mL). This study
demonstrates that maternally administered erythromycin does not eradicate chronic,
intra-amniotic ureaplasma infections or improve fetal outcomes in an ovine model,
potentially due to the poor placental passage of erythromycin.
91
Introduction
In 2008, 12.3% of babies in the United States were born preterm (less than 37
weeks of gestation). Despite the recent 3% decline in the preterm birth rate in the
United States, the preterm birth rate has risen by more than 20% between 1990 and
2006 [1]. In the developed world, preterm birth is the leading cause of neonatal
death [2] and accounts for 75% of perinatal mortality and more than half of long
term infant and childhood morbidity [3]. Approximately 30% of all preterm births are
associated with an infectious etiology [4], and the most common pathogens
associated with preterm birth are the human ureaplasmas [5].
The human ureaplasmas (Ureaplasma parvum and Ureaplasma urealyticum) are
the most frequently isolated microorganisms from the amniotic fluid (AF) [6-8], and
placentae [9, 10] of pregnant women and have been associated with adverse
pregnancy outcomes including chorioamnionitis, preterm rupture of membranes
(PROM) and preterm birth. Ureaplasmas are able to infect fetal tissues in utero and
are associated with neonatal diseases such as bronchopulmonary dysplasia,
pneumonia and meningitis. [11-13]. Experiments in rhesus macaques have
demonstrated that ureaplasmas can cause chorioamnionitis and induce preterm
labor and fetal lung injury as a sole pathogen [14].
Treatment of intra-amniotic ureaplasma infections with antimicrobials is predicted to
improve pregnancy outcomes and reduce neonatal morbidity and mortality rates
[15, 16]. In clinical practice, the standard antibiotics used to treat Ureaplasma spp.
are the macrolides, which include erythromycin, roxithromycin and azithromycin
[17]. Of these, erythromycin is the most frequently used to treat ureaplasma
infections during pregnancy, because its use is associated with a range of benefits
for the neonate after administration to pregnant women with preterm pre-labor
rupture of membranes (PPROM) in the ORACLE I trial. In this randomized trial, the
92
use of erythromycin was associated with prolongation of pregnancy, fewer major
cerebral abnormalities, fewer positive blood cultures, reduction in neonatal
surfactant treatment and decreases in oxygen dependence at ≥28 days of age.
[18].
Despite evidence of benefit to the newborn, there is little agreement within the
literature regarding the effectiveness of erythromycin for the eradication of intra-
amniotic infections. There are reports of maternally-administered erythromycin
eliminating intra-amniotic ureaplasma infections during pregnancy [19]; however,
when erythromycin was administered to women colonized with ureaplasmas in the
lower genital tract [20] or those who presented with PPROM [21], this treatment did
not prevent preterm delivery or eradicate intra-amniotic infection. Additionally,
results of a 7-year follow-up of the ORACLE II trial found that children whose
mothers had been in spontaneous preterm labour with intact membranes and
received either erythromycin or amoxicillin-clavulanate had increased rates of
cerebral palsy [22], suggesting there may be inherent risks in the use of antibiotics
during pregnancy. Despite this conflicting evidence, erythromycin still is routinely
administered to women with PPROM as standard treatment.
This study aimed to investigate the efficacy of erythromycin in eradicating chronic,
intra-amniotic ureaplasma infection in an established large animal model [23-25].
We studied the in vivo effects of erythromycin by comparing (i) ureaplasma
colonization of AF and fetal tissues; (ii) fetal inflammation; and (iii) pregnancy
outcomes between animals receiving either erythromycin treatment, or no
treatment.
93
Materials and Methods
The experiments were approved by the Animal Ethics Committees of the University
of Western Australia, Cincinnati Children’s Hospital Medical Centre and Queensland
University of Technology.
The low passage, erythromycin-sensitive U. parvum serovar 3 (isolate 442S) used
for this experiment was originally isolated from the semen of an infertile man
attending the Wesley IVF Service (Brisbane, Australia). This patient gave informed
consent for the isolate to be used for research. Ureaplasmas for injection were
prepared as previously described [24] and diluted to 2x104 colony-forming-units
(CFU) in PBS prior to injection.
At 50 days of gestation (d, term = 150 d), 32 date-mated Merino ewes bearing
single fetuses received intra-amniotic injections guided by ultrasound imaging.
Briefly, prior to injection 2 mL of AF was aspirated to confirm the injection site by
electrolyte analysis (Rapidlab 865, Bayer Diagnostics, Pymble, New South Wales).
The same needle was used to inject 2 mL of either U. parvum (2x104 CFU; n = 16)
or 10B broth (medium control; n = 16) [26].
At 100 d, ewes were examined by ultrasound, and five were no longer pregnant
(ureaplasma group: n = 4; 10B medium group: n = 1). The remaining pregnant ewes
were allocated into one of four treatment subgroups: (i) medium only (M, n = 7); (ii)
medium and erythromycin (M/E, n = 8); (iii) ureaplasma inoculum only (Up, n = 6) or
(iv) ureaplasma inoculum and erythromycin (Up/E, n = 6). The ewes that received
erythromycin were injected intra-muscularly with 500 mg of erythromycin (Abbot
Australasia, Kurnell, New South Wales), three times daily for four days (100 d – 104
d), resulting in a dosage of 30 mg/kg/day. This regimen was selected as it is similar
94
to erythromycin treatment given to pregnant women (which is variable but often 10 -
20 mg/kg/day) [19] and is an appropriate dose for administration to sheep [27].
Specimen Collection
AF samples were collected by amniocentesis at: 100 d (just prior to commencement
of erythromycin treatment); at 105 d (24 hours after completion of erythromycin
treatment); and at 125 d just prior to surgical delivery of the preterm fetus.
At 125 d ewes were anesthetised (12 mg/kg ketamine, 0.12 mg/kg metatomidine;
IM) followed by a subdural injection of 2% lignocaine (60 mg, 3 mL), and then the
uterus was incised through a maternal laparotomy. The fetal membranes were
isolated and AF was aspirated using a sterile needle and syringe. The fetus was
delivered and umbilical arterial blood was collected. The fetus was then euthanized
by injection of a lethal dose of sodium pentobarbitone (100 mg/kg, IV) and weighed.
Duplicate samples of chorioamnion and umbilical cord (cord) were aseptically
collected and either snap frozen in liquid nitrogen or fixed in 10% neutral buffered
formalin.
The lambs were exsanguinated, the chest opened, and a deflation pressure-volume
curve was performed after air inflation of the lungs to a pressure of 40 cm H2O.
Pieces of the lower lobe of the right lung were snap frozen and the upper lobe of the
right lung was fixed by airway instillation of 10% neutral buffered formalin at 30 cm
H2O pressure. Fetal cerebrospinal fluid (CSF) was collected aseptically by needle
aspiration and snap frozen for subsequent culture.
Complete blood counts (CBC) were performed on heparinised umbilical arterial
blood by a commercial laboratory (Vet Path, Ascot, Western Australia).
Ureaplasma culture and quantification
95
AF, chorioamnion, cord, fetal lung and fetal CSF specimens from all animals were
cultured for ureaplasmas to determine the number of ureaplasma CFU per mL of
fluid or per gram of tissue. Thawed chorioamnion, cord and fetal lung (0.1 grams)
were initially homogenised in 1.5 mL of 10B broth using a mini beadbeater 8-cell
disrupter (Daintree Scientific, St Helens, Tasmania). All specimens were then
inoculated into 10B broth and nine 10-fold serial dilutions were performed as
previously described [28]. For each specimen six 5 µL drops from the 10-fold
dilutions were subsequently inoculated onto A8 agar plates [29], which were
incubated in 5% CO2 for 48-72 hours. From the dilution containing countable
colonies (approximately 30 - 300 colonies per drop), ureaplasma colonies were
counted from each of the six 5 µL drops using a stereomicroscope (Leica
Microsystems, North Ryde, New South Wales). The average number of colonies per
5 µL drop was calculated and the total CFU per mL of fluid or per gram of tissue for
each specimen was determined. To eliminate bias, the microbiologist who cultured
and counted these colonies was blinded to the treatment groups of the specimens.
Specimens that were ureaplasma culture negative were tested for the presence of
ureaplasmal DNA by PCR assay. DNA was extracted from the specimen using the
QIAamp Mini Kit (Qiagen, Doncaster, Victoria) as per the DNA purification from
blood or bodily fluids protocol (provided by the manufacturer). PCR primers
targeting the multiple banded antigen gene (mba, UMS-125 5’
GTATTTGCAATCTTTATATGTTTTCG and UMA226 5’
CAGCTGATGTAAGTGCAGCATTAAATTC) [30] were used to detect and speciate
non-cultivable ureaplasmas within tissues. PCR assays were performed in a 50 µL
reaction mixture containing: 100 µM of dNTP mix (Roche Diagnostics, Castle Hill,
New South Wales), 1x PCR buffer (Tris HCl, KCl, (NH4)2SO4, pH 8.7 – Invitrogen,
Mt Waverley, Victoria), 1.5 mM of MgCl2 (Invitrogen), 0.5 µM of each primer (Sigma,
Castle Hill, New South Wales), 2.5 U of Platinum Taq Polymerase (Invitrogen) and
96
sterile distilled water. PCR cycling, in a PTC-2000 Thermal Cycler (Global Medical
Instrumentation, Ramsey, Minnesota), involved initial denaturation at 94 °C for 15
minutes; followed by 35 cycles of denaturation at 94 °C for one minute, primer
annealing at 56 °C for one minute, extension at 72 °C for one minute; and a final
extension step at 72 °C for 10 minutes. Positive PCR controls included DNA
extracted from the ureaplasma inoculum (isolate 442S) and U. parvum serovar 3
reference strain (courtesy of H. Watson, University of Alabama, Birmingham);
negative controls included master mix only and reaction mixtures with distilled water
substituted for template. All PCR products were electrophoresed through a 2%
agarose gel at 110 volts for 60 minutes. Amplicons were visualised by ethidium
bromide (10 µg/mL) staining and digitized using Grab-It Gel Dock (Ultraviolet
Products Ltd., Cambridge, United Kingdom).
Quantification of Erythromycin in AF
The detection and quantification of erythromycin in AF samples collected at 105d
and 125d from the M/E (4 out of 8 animals) and Up/E (6 out of 6 animals) groups
was performed using liquid chromatography mass spectrometry (Metabolomics
Australia, University of Queensland, St Lucia, Queensland). Samples from only 4
out of 8 animals from the M/E group were analysed due to limited sample
availability. The assay was able to detect <10 ng/mL of erythromycin, had an intra-
assay variation of 2.8%, and enabled accurate quantification of erythromycin levels
within each AF sample, calculated from a unique mass fingerprint [31, 32].
Histology and inflammatory cell counts
Formalin-fixed samples of chorioamnion, cord and fetal lung (right, upper lobe) were
embedded in paraffin and cut into 5 µm sections for staining with haematoxylin and
eosin (H & E). All stained sections were examined (in a blinded fashion) to
97
determine numbers of total white blood cells, monocytes, lymphocytes, and
neutrophils present in 20 microscopic fields per slide at x1000 total magnification.
Inflammatory cells were identified and distinguished visually based on cellular and
nuclear morphology and cell size. Chorioamnion sections were also graded for the
severity of histological chorioamnionitis. Scoring was based on a semi-quantitative
4-grade system developed within our laboratory, using a minimum score of 1
(representing minimal inflammatory cell infiltrate, no tissue fibrosis, necrosis or
abscesses), and a maximum score of 4 (heavy inflammatory cell infiltrate, severe
fibrosis, necrosis or abscesses).
Statistical analysis
Data are presented as mean values plus the standard error of the mean (SEM). The
culture data were log10 transformed to ensure normality of data, and then analysed
by independent t-tests to compare CFUs between animal groups. AF CFU/mL data
were analysed within each group across the three time points by two-way repeated
measures analysis of variance (ANOVA). Inflammatory cell counts, CBC data and
fetal physiological data were analysed by one-way ANOVA with a Tukey post hoc
test; lung compliance data were analysed by a two-way repeated measures
ANOVA. Homogeneity of variances was confirmed for each test as appropriate;
statistical significance was accepted at p < 0.05.
98
Results
Ureaplasma Culture and Quantification
At 100 d (50 days after ureaplasma injection and prior to erythromycin treatment)
100% of AF samples collected from animals inoculated with ureaplasmas tested
ureaplasma positive, with AF ureaplasma CFUs ranging from 2.0x105 to 1.76x108
CFU/mL (Figure 3.1 A). Upon retesting of AF at 105 d (24 hours after completion of
erythromycin treatment) all animals from the Up and Up/E groups remained
ureaplasma culture positive, with no change in the numbers of U. parvum
recovered. AF ureaplasma CFU/mL data were not different between Up and Up/E
groups, either before or after erythromycin treatment (p > 0.05). All control animals
tested negative for ureaplasmas by culture at these time points.
At 125 d (the time of surgical delivery of the fetus), the AF of all animals that had
been inoculated with ureaplasmas (100%) tested positive by culture for Ureaplasma
spp. Ureaplasmas were also cultured from: the chorioamnion of 4/6 (67%) animals
from the Up group and 4/6 (67%) animals from the Up/E group; the umbilical cords
of 2/6 (33%) animals from the Up group and 1/6 (17%) animals from the Up/E
group; and the fetal lung tissue from 5/6 (83%) animals from the Up group and all of
the Up/E group (100% (Figure 3.1 B)). No ureaplasmas were cultured from any fetal
CSF specimens and no ureaplasmas were cultured from any specimens from non-
infected control animals.
Ureaplasma CFU/mL at 125 d were not different between the Up and Up/E groups
(p > 0.05). AF CFU/mL at 125 d in the Up group appeared to have increased after
105 d (Figure 3.1 A); however, this increase was not statistically significant (p >
0.05). This increase was not evident in the Up/E group, in which AF ureaplasma
CFU/mL remained relatively unchanged throughout the tested time points. The
99
AF Ureaplasma Colonization
Up
Up/E
10
100
1000
10000 100d
105d
125d
A
Group
Ure
ap
lasm
a A
F C
FU
(x 1
05/m
L)
Ureaplasma Tissue Colonization at 125d
Up
Up/E
1
10
100
1000Chorioamnion
Cord
Fetal LungB
Group
Ure
ap
lasm
a C
FU
(x
10
3/m
L o
r g
)
Figure 3.1: Ureaplasma colonisation of the amniotic fluid and fetal tissues
A: Mean ureaplasma colonization of amniotic fluid samples at 100 d (d = days of gestation), 105 d and 125 d was not different between the ureaplasma (Up) group and the ureaplasma + erythromycin (Up/E) group. B: At 125 d, ureaplasmas were isolated from the chorioamnion, cord, and fetal lung; however, there were no differences in mean CFU/gram of tissue between the groups, which received intra-amniotic ureaplasma injections. AF= amniotic fluid, CFU= colony forming units. Bars represent the mean number of CFU/mL or /g + SEM.
100
number of recoverable ureaplasma CFUs were higher in the AF compared to the
chorioamnion (Up p = 0.23; Up/E p = 0.41), cord (Up p = 0.23; Up/E p = 0.47) and
fetal lung (Up p = 0.90; Up/E p = 0.14): however, these differences did not reach
statistical significance.
From those specimens that tested negative for ureaplasmas by culture, DNA was
extracted and PCR was performed to detect ureaplasmal DNA. Of these specimens:
4/4 (100%) of chorioamnion specimens, 7/9 (78%) of cord specimens and 1/1
(100%) of fetal lung specimens gave positive PCR assay results. No ureaplasmas
were detected in the fetal CSF specimens (0%, n = 27) by the PCR assay.
Chorioamnionitis and Fetal Inflammation
Histological chorioamnionitis was evident in animals from the Up and Up/E groups
when compared to controls, as demonstrated by an increase in inflammatory cells in
the chorioamnion (Figure 3.2 A). Counts of lymphocytes (p = 0.012) and neutrophils
(p = 0.03), but not monocytes (p = 0.167), were significantly higher in the Up/E
group when compared to the M/E control group. There were no differences in
inflammatory cell counts in the chorioamnion between the Up and Up/E groups (p >
0.05).
Histopathological grading of inflammation in chorioamnion tissues was consistent
with the inflammatory cell count data: animals from the control groups did not have
evidence of chorioamnionitis; whereas animals from the Up and Up/E groups had
scores indicating moderate to severe histological chorioamnionitis (Figure 3.2 B).
No differences in the severity of chorioamnionitis were found between the Up and
Up/E groups. Representative sections demonstrating the grading of histological
chorioamnionitis (Grade 1 – Grade 4) are shown in Figure 3.2 C. Inflammatory cell
counts within cord tissue were significantly increased in the Up and Up/E groups
101
Chorioamnion Inflammatory Cell Counts
MM
/E Up
Up/E
0
10
20
30
Lymphocytes
Monocytes
Neutrophils
*
*A
Group
Cell
Co
un
t(2
0 f
ield
s o
f v
iew
)
Histological Chorioamniontis Score
MM
/E Up
Up/E
0
1
2
3
4 B
Group
Ch
ori
oam
nio
n I
nfl
am
mati
on
Sco
re
Cord Inflammatory Cell Counts
MM
/E Up
Up/E
0
5
10
15
20
25
Lymphocytes
Monocytes
Neutrophils
*
*
*
*
D
*
*
Group
Cell
Co
un
t(2
0 f
ield
s o
f v
iew
)
*
**
*
*
*
Fetal Lung Inflammatory Cell Counts
MM
/E Up
Up/E
0
5
10
15
20
Lymphocytes
Monocytes
Neutrophils
#*
E
*
Group
Cell
Co
un
t(2
0 f
ield
s o
f v
iew
)
* *
C Grade 1 Grade 2 Grade 3 Grade 4
102
Figure 3.2: Inflammation of fetal tissues. A: Increased inflammatory cell counts were observed in chorioamnion sections from animals inoculated with ureaplasmas compared to controls, and the histological chorioamnionitis score (B) was also increased in these animals. C: Representative haematoxylin and eosin stained sections of chorioamnion displaying various stages of histological chorioamnionitis. Grades 1-4 are shown from left to right: Grade 1- (uninfected control) showing minimal inflammatory cell infiltrate, no tissue fibrosis, necrosis or absesses. Grade 2- mild inflammatory cell infiltrate (indicated by arrows), mild tissue fibrosis, necrosis or abscesses. Grade 3- heavy inflammatory cell infiltrate, moderate tissue fibrosis, necrosis or abscesses. Arrows represent moderate levels of scar tissue formation and fibrosis. Grade 4- heavy inflammatory cell infiltrate, severe fibrosis, necrosis or abscesses. Arrows represent severe fibrosis and disruption of normal tissue morphology. D and E: inflammatory cell counts were also increased in umbilical cord and fetal lung tissue sections in animals in the Up group and the Up/E group. Asterisk (*), probability of < 0.05 when compared to M/E control group only. Hash (#), probability of < 0.05 when compared to the M control group only. Double asterisk (**), probability of < 0.05 when compared to both M/E and M control groups. Error bars represent mean values + SEM. M= medium group; M/E= medium + erythromycin group; Up= ureaplasma group; Up/E= ureaplasma + erythromycin group. Scale bars on micrographs represent 50 µm.
103
compared to the respective control groups for monocytes (p = 0.001), lymphocytes
(p = 0.001) and neutrophils (p = 0.001) (Figure 2D). There were also increased
numbers of monocytes in fetal lung tissue in the Up and Up/E groups compared to
controls (p = 0.001). Lymphocyte numbers were significantly increased in the Up
group compared to the M group only (p = 0.041); but neutrophil numbers were not
different between groups (p = 0.119, Figure 3.2 E). In all tissues examined there
were no differences in inflammatory cell counts between the Up/E and Up groups.
Total white blood cell counts from fetal umbilical arterial blood were not different
between animals in the Up and Up/E groups compared to controls, except for
lymphocytes, which were increased in the M/E control group compared to the Up
group (p = 0.02, Table 3.1).
Fetal Growth and Wellbeing
Fetal body weight at the time of preterm delivery was not different between Up and
Up/E groups and controls (Table 3.1). Similarly, intra-amniotic ureaplasma infection
did not affect fetal lung weight relative to body weight or fetal umbilical arterial pH or
pO2 at the time of delivery (Table 3.1), when compared to control animals. Fetal
lung compliance (as indicated by lung volume at 40 cm H2O pressure) was
increased in Up and Up/E groups compared to controls (Table 3.1); however, this
increase was not statistically significant. Lung compliance was not different between
Up and Up/E animal groups
AF Erythromycin Quantification
Erythromycin was detected in 100% of AF samples from the M/E and Up/E groups
collected at 105 d and concentrations ranged from <10 ng/mL to 75.7 ng/mL (Figure
3.3). At 125 d, erythromycin was not detected within AF in 50% of animals from the
104
M M/E Up Up/E
P
Values
Fetal body weight (kg)
3.0 ± 0.1
2.7 ± 0.1
2.9 ± 0.1
3.0 ± 0.1
0.82
Fetal lung weight (g/kg body weight) 31.6 ± 0.7 33.3 ± 1.5 30.5 ± 2.1 31.7 ± 1.1 0.60
Lung volume(mL/kg) at 40 cm H2O
pressure
14.9 ± 3.7 14.6 ± 3.8 18.6 ± 4.6 19.2 ± 2.0 0.22
Umbilical arterial blood gases
pH 7.2 ± 0.0 7.2 ± 0.0 7.1 ± 0.0 7.2 ± 0.0 0.63
pO2 (mmHg) 9.6 ± 2.7 9.9 ± 2.2 9.8 ± 1.8 8.5 ± 2.0 0.97
Umbilical arterial white blood cell counts
Total (x 109/L) 3.9 ± 0.4 5.8 ± 1.3 3.2 ± 0.3 3.4 ± 0.2 0.13
Monocytes (x 109/L) 0.2 ± 0.04 0.5 ± 0.3 0.1 ± 0.0 0.1 ± 0.0 0.54
Lymphocytes (x 109/L) 2.2 ± 0.4 3.0 ± 0.5 1.3 ± 0.1 2.1 ± 0.1 0.02
Neutrophils (x 109/L) 0.6 ± 0.3 1.2 ± 0.8 1.0 ± 0.2 0.5 ± 0.1 0.70
Table 3.1: Fetal measurements at 125 days of gestation
Fetal body weight, fetal lung weight, fetal lung compliance and umbilical arterial blood measurements were not different between groups, except for umbilical arterial blood lymphocytes (p = 0.02). Data are presented as mean ± SEM. M = media group; M/E = media + erythromycin group; Up = ureaplasma group; Up/E = ureaplasma + erythromycin group.
105
M/E group, and 83% of animals from the Up/E group, and in the remaining animals
erythromycin concentrations were very low (<10 ng/mL, data not shown). Although
intra-animal variability was observed, no differences in mean AF erythromycin
concentration were found between the M/E and the Up/E groups at 105 d (p = 0.42)
and 125 d (p = 0.76). Erythromycin was not detected in any AF specimens collected
from animals that did not receive erythromycin treatment (data not shown).
Figure 3.3: Quantitation of erythromycin within the amniotic fluid of sheep, which received maternal erythromycin treatment
Amniotic fluid (AF) erythromycin concentrations at 105 days of gestation were not different between the M/E (medium + erythromycin) and Up/E (ureaplasma + erythromycin) groups (p = 0.42).
105d AF Erythromycin Quantification
M/E
Up/E
0
20
40
60
80
Group
Ery
thro
mycin
Co
ncen
trati
on
(n
g/m
L)
106
Discussion
Treatment of pregnant women with erythromycin occurs routinely in clinical
obstetrics. As a macrolide antibiotic, erythromycin has broad-spectrum activity and
is appropriate for use in pregnancy (pregnancy category A) [33]. By inhibiting
protein synthesis at the 50S ribosomal subunit [17], erythromycin is potentially
active against the Ureaplasma spp. (which lack a cell wall), unlike other broad-
spectrum beta lactam antibiotics. Our data however, clearly demonstrate that
maternal erythromycin treatment, as administered in our experiment, is ineffective at
eliminating intra-amniotic ureaplasma infections or improving fetal outcomes in an
ovine model.
Erythromycin treatment did not significantly reduce or eliminate ureaplasma
colonization within the AF, chorioamnion or fetal lung, when compared to animals
that did not receive antibiotic treatment. Remarkably, even at 105 d (24 hours after
completion of antibiotic treatment), the number of recoverable ureaplasma CFUs
within the AF remained unchanged in comparison to the AF CFU/mL prior to
erythromycin treatment (100 d). Our data showed that AF ureaplasma CFU/mL at
125 d was increased in the Up group but not in the Up/E group. This observation,
whilst not statistically significant, is consistent with the bacteriostatic, rather than
bacteriocidal action of erythromycin; suggesting that after erythromycin was
depleted the ureaplasmas were able to again replicate. Collectively, these data
indicate that the erythromycin had no significant effect on the short-term survival of
ureaplasmas or their ability to cause long-term, chronic infections. We also found
that erythromycin treatment did not reduce fetal inflammation of the chorioamnion or
fetal lungs; nor did it have any effect on the severity of histological chorioamnionitis.
Similarly, fetal outcomes (including fetal lung compliance) were not altered by the
administration of erythromycin in this experiment.
107
Our findings are in agreement with data from human clinical studies. In a study of
serial amniocenteses in which pregnant women with PPROM were treated with
ampicillin and erythromycin, and women with evidence of intra-amniotic infection
were treated with ceftriaxone, clindamycin and erythromycin [21], maternal systemic
antibiotic therapy rarely (3/18, 16.7%) eradicated intra-amniotic infection.
Furthermore, 32% (9/28) of patients who did not have intra-amniotic
infection/inflammation at admission subsequently developed intra-amniotic
inflammation, despite antibiotic treatment. In another study [20], women with lower
genital tract ureaplasma colonization received either erythromycin treatment or
placebo. Erythromycin treatment had no effect on infant birth weight, gestational
age at delivery, frequency of PROM, or neonatal outcomes compared to women
who received a placebo. Similarly, results of the ORACLE II randomized trial of
women in preterm labor with intact membranes [34] and a Cochrane review of 15
trials [35] suggest that maternally administered antibiotics do not improve pregnancy
outcomes. The literature indicates that these antibiotics have very little therapeutic
and/or prophylactic benefit. To the best of our knowledge, ours is the first animal
model to fully assess the effects of erythromycin on chronic, intra-amniotic
ureaplasma infections; and demonstrates the likely reason for the lack of efficacy in
preventing adverse pregnancy outcomes.
Conversely, some studies have reported successful eradication of human intra-
amniotic infections and improved pregnancy outcomes after antibiotic treatment
[19]. In an analysis of 2718 genetic amniocentesis specimens by Berg et al. 1999,
1.8% of AF specimens were found to be culture positive for genital mycoplasmas
(including Ureaplasma spp. and Mycoplasma hominis) [36]. Based on these results,
patients were either treated with oral erythromycin for 10 days (n = 34), or not
treated (n = 9) and pregnancy outcomes were compared. Mid-trimester pregnancy
loss was significantly decreased in the treated group (11.4%) compared to the
108
untreated group (44.4%), however, preterm delivery rates were not different
between groups. Others have reported successful eradication of Ureaplasma spp.
from the AF and prolongation of pregnancy after erythromycin treatment [37, 15,
16]. These reports should be interpreted cautiously however, as they are either
single case reports, involve different ureaplasma serovars and culture techniques,
different treatment protocols, or have large inequalities in sample sizes (evidenced
by Berg et al. 1999) [36].
To date, it is thought that the failure of erythromycin to eradicate intra-amniotic
infections is due to poor placental passage [38]. After ex vivo perfusion of single
placental cotyledons from women who delivered at term, the transplacental transfer
of erythromycin was found to be only 3.0% [38]. Our experiment shows that the
transplacental passage of erythromycin in vivo is low, and is the most likely
determinant of the ineffectiveness of erythromycin treatment during pregnancy. Our
study was not designed as a pharmacokinetic experiment, and as a result our
sample collection did not occur at times to enable us to determine the maximum
concentration of erythromycin within the AF. In humans, erythromycin has a
reported serum half-life of approximately 1.4 hours [39]; therefore, our
measurements at 105 d most likely represent the erythromycin concentration after
several half-lives have passed. However, the Merck Veterinary Manual reports that
after an intra-muscular injection of erythromycin, effective inhibitory concentrations
are maintained for 12-24 hours in sheep [40], suggesting that the half life of
erythromycin in sheep may be different from that in humans. Furthermore,
erythromycin can undergo persistent enterohepatic cycling [40], which may explain
why very low levels of erythromycin were detected with the AF of a small number of
animals at 125 d. Another limitation of our experiment was that we did not collect
serum samples for the quantification of erythromycin within maternal and fetal
plasma. Furthermore, a potential contributor towards our findings may be the
109
anatomy of the ovine placenta, which is synepitheliochorial [41, 42], with six layers
interposed between the maternal and fetal circulations [43]. In contrast, the human
placenta is haemochorial [44], with three layers separating the maternal and fetal
circulations [43]. Due to these anatomical differences, the permeability of the ovine
placenta is lower than that of humans [43, 45, 46], and may therefore provide a
greater barrier against transfer of antimicrobial agents. Despite these differences,
our study demonstrates that erythromycin can enter the AF after maternal
administration to pregnant sheep, however, it appears to be present at low
concentrations that are most likely below the minimum inhibitory concentrations
required to eradicate Ureaplasma spp. Further pharmacokinetic studies would be
helpful to better define erythromycin levels in the fetal compartment.
Based on our data and the collective reports of others, we conclude that
erythromycin has little therapeutic value for the improvement of fetal outcomes, as it
does not eradicate or reduce intra-amniotic ureaplasma infection or reduce
inflammation. Furthermore, there are potential risks associated with administering
erythromycin during pregnancy. A 7-year follow-up of the ORACLE II trial found that
there was an increase in functional impairment and cerebral palsy in children who
were exposed to erythromycin in utero [22]. One possible cause of this morbidity
arises from the concept of incomplete treatment. Partial eradication of the intra-
uterine infection may diminish the provocation of preterm labor but not remove the
infection from the fetus, thus allowing continuation of the pregnancy with the fetus
being compromised by remaining within an infected environment. Also, there are
rare cases of erythromycin causing adverse effects in newborn infants, including
cardiovascular compromise and hypertrophic pyloric stenosis [47]. As erythromycin
is a broad spectrum antibiotic, there is also the potential to target beneficial
microbes, such as the developing gut normal flora of the neonate, which is crucial in
the development of the naive immune system of newborn babies [48]. Finally, there
110
remains the issue of increasing antimicrobial resistant microorganisms due to the
unnecessary overuse of antibiotics. Recently, Krausse and Schubert (2009)
investigated the susceptibilities of clinical ureaplasma isolates collected from 1983
and from 1989-2008 to various antimicrobials [49]. It was found that ureaplasmas
showed reduced susceptibilities to all classes of antimicrobials tested throughout
the time period, and high levels of erythromycin resistance was observed from
isolates collected in 1989-2008.
The results of the present study provide further evidence that the use of
erythromycin in pregnancy is ineffective, due to its limited ability to cross the
placenta. Further research is required urgently to explore alternative antimicrobial
treatments for women at risk of preterm birth.
111
Acknowledgements
The authors would like to acknowledge Dr Shana Jacob at Metabolomics Australia
for performing liquid chromatography-mass spectrometry analysis on amniotic fluid
samples. We also acknowledge JRL Hall & Co., in particular Sara Ritchie and Fiona
Hall, who have been responsible for breeding and supplying us with the high quality
research animals necessary for this project. We also wish to thank: Drs John and
Janet Allan at Wesley Monash IVF for the research that has provided low passage
clinical ureaplasma isolates.
112
References
1. Hamilton BE, Martin JA, Ventura SJ. Births: Preliminary Data for 2008.
National Vital Statistics Reports. 2010; 58:1-17.
2. Klein LL, Gibbs RS. Use of microbial cultures and antibiotics in the
prevention of infection-associated preterm birth. Am J Obstet Gynecol. 2004;
190: 1493-502.
3. McCormick MC. The contribution of low birth weight to infant mortality and
early childhood morbidity. N Engl J Med. 1985; 312: 82-90.
4. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes
of preterm birth. Lancet. 2008; 371: 75-84.
5. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection and preterm
delivery. N Engl J Med. 2000; 342: 1500-1507.
6. Volgmann T, Ohlinger R, Panzig B. Ureaplasma urealyticum- harmless
commensal or underestimated enemy of human reproduction? A review.
Arch Gynecol Obstet. 2005; 273: 133-139.
7. Gerber S, Vial Y, Hohlfeld P, Witkin SS. Detection of Ureaplasma
urealyticum in second-trimester amniotic fluid by polymerase chain reaction
correlates with subsequent preterm labor and delivery. J Infect Dis. 2003;
187: 518-21.
8. Perni SC, Vardhana S, Korneeva I, Tuttle SL, Paraskevas LR, Chasen ST,
Kalish RB, Witkin SS. Mycoplasma hominis and Ureaplasma urealyticum in
midtrimester amniotic fluid: association with amniotic fluid cytokine levels
and pregnancy outcome. Am J Obstet Gynecol. 2004; 191: 1382-1386.
9. Kundsin RB, Driscoll SG, Monson RR, Yeh C, Biano SA, Cochran WD.
Association of Ureaplasma urealyticum in the placenta with perinatal
morbidity and mortality. N Engl J Med. 1984; 310: 941-945.
10. Pararas MV, Skevaki CL, Kafetzis DA. Preterm birth due to maternal
infection: Causative pathogens and modes of prevention. Eur J Clin
Microbiol Infect Dis. 2006; 25: 562-569.
11. Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R. Ureaplasma
urealyticum intrauterine infection: role in prematurity and disease in
newborns. Clin Microbiol Rev. 1993; 6: 69-87.
12. Waites KB, Katz B, Schelonka RL. Mycoplasmas and ureaplasmas as
neonatal pathogens. Clin Microbiol Rev. 2005; 18: 757-789.
113
13. Abele-Horn M, Wolff C, Dressel P, Pfaff F, Zimmermann A. Association of
Ureaplasma urealyticum biovars with clinical outcome for neonates, obstetric
patients, and gynecological patients with pelvic inflammatory disease. J Clin
Microbiol. 1997; 35: 1199-202.
14. Novy M, Duffy L, Axthelm M, Sadowsky D, Witkin S, Gravett M, Cassell G,
Waites K. Ureaplasma parvum or Mycoplasma hominis as sole pathogens
cause chorioamnionitis, preterm delivery, and fetal pneumonia in rhesus
macaques. Reprod Sci. 2009; 16: 56-70.
15. Romero R, Hagay Z, Jores J, Sepulveda W, Mazor M. Eradication of
Ureaplasma urealyticum from the amniotic fluid with transplacental antibiotic
treatment. Am J Obstet Gynecol. 1992; 166: 618-622.
16. Smorgick N, Frenkel E, Aaidenstein R, Lazarovitch T, Sherman DJ.
Antibiotic treatment of intra-amniotic infection with Ureaplasma urealyticum.
Fetal Diagn Ther. 2007; 22:90-93.
17. Gaynor M, Mankin AS. Macrolide antibiotics: binding site, mechanism of
action, resistance. Curr Top Med Chem. 2005; 2: 21-35.
18. Kenyon SL, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group.
Broad-spectrum antibiotics for preterm, prelabour rupture of fetal
membranes: the ORACLE I randomised trial. Lancet. 2001; 357: 979-985.
19. Hutzal CE, Boyle EM, Kenyon SL, Nash JV, Winsor S, Taylor DJ, Kirpalani
BM. Use of antibiotics for the treatment of preterm parturition and prevention
of neonatal morbidity: a metaanalysis. Am J Obstet Gynecol. 2008; 199:
620e1-e8.
20. Eschenbach EA, Nugent RP, Rao AV et al. A randomized placebo-controlled
trial of erythromycin for the treatment of Ureaplasma urealyticum to prevent
premature delivery. The Vaginal Infections and Prematurity Study Group.
Am J Obstet Gynecol. 1991; 164: 734-742.
21. Gomez R, Romero R, Nien JK et al. Antibiotic administration to patients with
preterm premature rupture of membranes does not eradicate intra-amniotic
infection. Matern Fetal Neonatal Med. 2007; 20: 167-173.
22. Kenyon S, Pike K, Jones DR, Brocklehurst P, Marlow N, Salt A, Taylor DJ.
Childhood outcomes after prescription of antibiotic to pregnant women with
spontaneous preterm labour: 7-year follow-up of the ORACLE II trial. Lancet.
2008; 372: 1319-1327.
23. Moss TJ, Nitsos I, Ikegami M, Jobe AH, Newnham JP. Experimental
intrauterine Ureaplasma infection in sheep. Am J Obstet Gynecol. 2005;
192: 1179-1186.
114
24. Moss TJ, Knox CL, Kallapur SG, Nitsos I, Theodoropoulos C, Newnham JP,
Ikegami M, Jobe AH. Experimental amniotic fluid infection in sheep: effects
of Ureaplasma parvum serovars 3 and 6 on preterm or term fetal sheep. Am
J Obstet Gynecol. 2008; 198: 122e1-e8.
25. Moss TJ, Nitsos I, Knox CL, Polglase GR, Kallapur SG, Ikegami M, Jobe
AH, Newnham JP. Ureaplasma colonization of amniotic fluid and efficacy of
antenatal corticosteroids for preterm lung maturation in sheep. Am J Obstet
Gynecol. 2009; 200: 96e1-e6.
26. Shepard MC, Lunceford CD. Serological typing of Ureaplasma urealyticum
isolates from urethritis patients by an agar growth inhibition method. J Clin
Microbiol. 1978; 8: 566-574.
27. The European Agency for the Evaluation of Medicinal Products: Veterinary
Medicines and Information Technology Unit. Erythromycin: Summary Report
(2). [Internet] January 2000 (accessed April 2010).
http://www.ema.europa.eu/pdfs/vet/mrls/082102en.pdf
28. Tully JG. Cloning and filtration techniques for mycoplasmas, Methods in
Mycoplasmology, New York: Academic Press; 1983: 173-177.
29. Shepard MC, Lunceford CD. Differential agar medium (A7) for identification
of Ureaplasma urealyticum (human T mycoplasmas) in primary cultures of
clinical material. J Clin Microbiol. 1976; 3: 613-625.
30. Teng LJ, Zheng X, Glass JI, Watson HL, Tsai J, Cassell GH. Ureaplasma
urealyticum biovar specificity and diversity are encoded in multiple-banded
antigen gene. J Clin Microbiol. 1994; 32: 1464-1469.
31. Xaio W, Chen B, Yao S, Cheng Z. Simultaneous determination of
erythromycin propionate and base in human plasma by high-performance
liquid chromatography-electrospray mass spectrometry. J Chromatogr B
Analyt Technol Biomed Life Sci. 2005; 817: 153-158.
32. Deubel A, Fandino AS, Sorgel F, Holzgrabe U. Determination of
erythromycin and related substances in commercial samples using liquid
chromatography/ion trap mass spectrometry. J Chromatogr. 2006; 1136: 39-
47.
33. Antibiotic Expert Group. Therapeutic Guidelines: Antibiotic. Version 13.
Melbourne: Therapeutic Guidelines Limited; 2006: 21, 353-362.
34. Kenyon SL, Taylor DJ, Tarnow-Mordi W. Broad-spectrum antibiotics from
spontaneous preterm labour: the ORACLE II randomised trial. Lancet. 2001;
357: 981-990.
115
35. McDonald H, Brocklehurst P, Parsons J. Antibiotics for treating bacterial
vaginosis in pregnancy. Cochrane Database Syst Rev. 2005; CD000262
36. Berg TG, Philpot KL, Welsh MS, Sanger WG, Smith CV.
Ureaplasma/mycoplasma-infected amniotic fluid: pregnancy outcome in
treated and nontreated patients. J Perinatol. 1999; 19: 275-277.
37. Mazor M, Chaim W, Horowitz S, Leiberman JR, Glezerman M. Successful
treatment of preterm labour by eradication of Ureaplasma urealyticum with
erythromycin. Arch Gynecol Obstet. 1993; 253: 215-218.
38. Heikkinen T, Laine K, Neuvonen PJ, Ekblad U. The transplacental transfer of
the macrolide antibiotics erythromycin, roxithromycin and azithromycin.
BJOG. 2000; 107: 770-775.
39. Mims Online: Erythromycin IV Prescribing Information [Internet] c2009.
http://mims.hcn.net.au
40. The Merck Veterinary Manual: Pharmacokinetic Features. [Internet]. c2009.
http://www.merckvetmanual.com
41. Sammin D, Markey B, Bassett H, Buxton D. The ovine placenta and
placentitis- A review. Vet Microbiol. 2009; 135: 90-97.
42. Wooding FBP, Flint APF. Placentation. In: Lamming GE (ed), Marshall’s
Physiology of Reproduction, Vol 3, 4th Ed. London: Chapman & Hall; 1994:
233-461.
43. Van der Aa EM, Peereboom-Stegeman JHJ, Noordhoek J, Gribnau FWJ,
Russel FGM. Mechanisms of drug transfer across the human placenta.
Pharm World Sci. 1998; 20: 139-148.
44. Benirschke K, Kaufmann P, Baergen R. Pathology of the Human Placenta.
5th Ed. New York: Springer; 2006: 30-50.
45. Enders AC, Carter AM. What can comparative studies of placental structure
tell us? – A review. Placenta. 2004; 25, Supplement A: S3-S9.
46. Carter AM. Animal models of human placentation- A review. Placenta. 2007;
21, Supplement A: S41-S47.
47. Buher C, Hoehn T, Hentschel J. Role of erythromycin for treatment of
incipient chronic lung disease in preterm infants colonised with Ureaplasma
urealyticum. Drugs. 2001; 61: 1893-1899.
48. Bedford Russell AR, Murch SH. Could peripartum antibiotics have delayed
health consequences for the infant? BJOG. 2006; 113: 758-765.
49. Krausse R, Schubert S. In vitro activities of tetracyclines, macrolides,
fluroquinolones and clindamycin against Mycoplasma hominis and
116
Ureaplasma spp. isolated in Germany over 20 years. Clin Microbiol Infect
16: 1649-1655.
117
Chapter 4
PAPER 2
118
Genetic variability and antimicrobial resistance of Ureaplasma parvum in
response to maternal erythromycin treatment: a study in pregnant sheep
Samantha J Dando 1, Ilias Nitsos 2#, Graeme R Polglase 2#, John P Newnham 2,
Alan H Jobe 2, 3, Christine L Knox 1
1 Institute of Health & Biomedical Innovation, Faculty of Science and Technology, Queensland University of Technology, Brisbane, Queensland, 4059, Australia.
2 School of Women’s and Infants’ Health, The University of Western Australia, Perth,
Western Australia, 6009, Australia. 3 Department of Neonatology and Pulmonary Biology, Cincinnati Children’s Hospital
Medical Center, University of Cincinnati, Cincinnati OH 45229, USA. # Current affiliation: The Ritchie Centre, Monash Institute of Medical Research &
Department of Obstetrics and Gynaecology, Monash University, Clayton, Victoria, 3168, Australia.
Manuscript in preparation
119
Statement of Joint Authorship
Samantha J Dando (candidate):
Contributed to the experimental design and research plan; assisted with the collection of clinical samples; performed all ureaplasma cultures, minimum inhibitory concentration testing and biofilm experiments; performed all PCRs and sequencing. Analysed and interpreted the data, performed statistical analyses and wrote the manuscript.
Ilias Nitsos:
Contributed to the experimental design and research plan; performed intra-amniotic injections, administered erythromycin treatment and assisted in the collection of samples. Contributed to the manuscript.
Graeme Polglase:
Contributed to the experimental design and research plan; assisted in the collection of samples and contributed to the manuscript.
John P Newnham:
Contributed to the experimental design and research plan; performed intra-amniotic injections and assisted in the collection of samples. Contributed to the manuscript.
Alan H Jobe:
Contributed to the experimental design and research plan; performed fetal post-mortems and assisted in the collection of samples. Contributed to the manuscript.
Christine L Knox:
Supervised the project, contributed to the experimental design and research plan and assisted in the collection of samples. Assisted in the interpretation of data and made a significant contribution to the manuscript.
120
Abstract
Ureaplasmas are the microorganisms most frequently isolated from the amniotic
fluid (AF) of pregnant women and can cause chronic infections that are difficult to
eradicate with standard macrolide treatment. We tested the effects of erythromycin
exposure on phenotypic and genotypic markers of ureaplasmal antimicrobial
resistance in a sheep model. At 55 days of gestation (term = 150 d) 12 pregnant
ewes received an intra-amniotic injection of U. parvum serovar 3 (erythromycin-
susceptible). At 100 d the ewes received erythromycin treatment (500 mg, three
times daily for 4 days, IM, n = 6) or saline (IM, n = 6). Fetuses were delivered
surgically at 125 d and AF and chorioamnion were collected for: culture, minimum
inhibitory concentration (MIC) and minimum biofilm inhibitory concentration (MBIC)
testing, 23S rRNA gene sequencing and detection of macrolide resistance genes.
MICs of erythromycin, azithromycin and roxithromycin against AF isolates were low
(range = 0.06 - 1.0 mg/L); however, chorioamnion isolates demonstrated increased
resistance to roxithromycin (0.13 - 5.33 mg/L). 62.5% of chorioamnion ureaplasmas
formed biofilms; however, MBICs were generally not higher than the MICs of
planktonic cells. Sequence variability (125 nucleotides, 29.6%) was detected in the
23S rRNA gene (domain V) of chorioamnion (but not AF) ureaplasmas, resulting in
a mosaic-like sequence. Macrolide resistance genes (erm(B), msr(C) and msr(D))
were detected in 100% of chorioamnion isolates, whereas only msr(D) was detected
in 40% of AF ureaplasmas. Sequence variability and the presence of macrolide
resistance genes occurred independently of erythromycin treatment, suggesting that
the anatomical site of isolation may exert pressures on ureaplasmas that alter the
socio-microbiological structure of the bacterial population, select for genetic
changes and alter antimicrobial susceptibility profiles.
121
Author Summary
Intra-amniotic ureaplasma infection is associated with adverse pregnancy and
neonatal outcomes. Previously, we identified that standard erythromycin treatment
of chronic, intra-amniotic ureaplasma infections was ineffective due to the minimal
placental transfer of erythromycin. As a result, ureaplasmas colonising the amniotic
fluid are potentially exposed to sub-lethal concentrations of erythromycin, which
may promote antibiotic resistance. Here, we investigated the effects of erythromycin
treatment of chronic, intra-amniotic ureaplasma infections on markers of macrolide
resistance in ureaplasmas isolated from the amniotic fluid and chorioamnion in an
ovine model. Pregnant sheep received an intra-amniotic injection of a clinical U.
parvum isolate at 55 days of gestation (d), followed by standard, maternal
erythromycin treatment at 100 d. Ureaplasmas isolated from the amniotic fluid and
chorioamnion after chronic, 70 day in utero infection demonstrated marked
differences in susceptibility to macrolide antibiotics. Furthermore, significant genetic
variability was observed in the 23S rRNA gene of chorioamnion ureaplasmas, but
not in the inoculum strain or amniotic fluid ureaplasmas. Variation of antimicrobial
susceptibilities and 23S rRNA gene sequences between amniotic fluid and
chorioamnion ureaplasmas was not induced by exposure to erythromycin.
Therefore, ureaplasma subtypes with variable antimicrobial susceptibilities were
generated in utero (even in the absence of antimicrobial treatment) and these
subtypes may selectively colonise different anatomical sites within the upper genital
tract during pregnancy.
122
Introduction
The human ureaplasmas (Ureaplasma parvum and Ureaplasma urealyticum) are
among the smallest self-replicating bacteria, typically ranging in size from 100 nm to
1 µm [1]. These microorganisms are prolific colonizers of the urogenital tract of
women, and can be isolated from the mucosal surfaces of the vagina or cervix in
40-80% of sexually active females [2]. The ureaplasmas are also the
microorganisms most frequently isolated from the amniotic fluid (AF) of pregnant
women [3-5], the placenta [6, 7] and the central nervous system and lower
respiratory tract of neonates [8, 9]. Ureaplasma infection of the AF is associated
with adverse pregnancy outcomes such as preterm birth and chorioamnionitis, and
neonatal pulmonary diseases including bronchopulmonary dysplasia and
pneumonia [9].
Eradication of intra-amniotic ureaplasma infections by antimicrobial treatment is
predicted to improve pregnancy outcomes and reduce neonatal morbidity and
mortality. As members of the class Mollicutes, the ureaplasmas lack a cell wall and
are bounded only by a plasma membrane [10], making them naturally resistant to
antimicrobials which target the bacterial cell wall, such as the β-lactams and
glycopeptides. Erythromycin (a 14-membered lactone ring macrolide) is the
standard antibiotic administered to pregnant women for the treatment of intra-
amniotic infections. However, this treatment is often ineffective [11-13] as there is
minimal placental transfer of erythromycin from the maternal circulation to the AF
and fetal circulation. Heikkinen et al. [14] report that in humans the placental
transfer of erythromycin is as low as 3%, thus microorganisms within the AF would
be exposed to low levels of the antibiotic. Recently, in an ovine model of intra-
amniotic ureaplasma infection, our group reported that standard-dose maternal
erythromycin treatment achieved low erythromycin concentrations (<10 – 76 ng/mL)
in the AF and did not eradicate the infection [15].
123
Exposure of bacteria to non-lethal concentrations of antimicrobials can promote
antimicrobial resistance, which can occur by target site modification, drug efflux
pumps, or drug inactivation mediated by short peptides [16]. Macrolide resistance in
ureaplasmas was initially thought to be uncommon [17]; however, Krausse and
Schubert [18] found that ureaplasmas demonstrated high resistance to erythromycin
over the time period 1989-2008. Mechanisms of macrolide resistance that have
been identified in ureaplasmas to date include mutations in the 23S rRNA gene and
ribosomal protein L4 and L22 genes [19-21]; and recently, erm and msr gene
subtypes (associated with post-transcriptional modification of 23S rRNA and drug
efflux activity respectively) have been identified in clinical ureaplasma isolates [22].
Bacterial biofilms are also thought to contribute to antimicrobial resistance, as the
minimum inhibitory concentration (MIC) of antibiotics for biofilm-forming bacteria
may be 10 to 1000 fold higher when compared to their planktonic counterparts [23-
25]. Biofilms are thought to increase resistance to antimicrobials by one or more of
the following mechanisms: (i) delayed antimicrobial penetration, (ii) altered growth
rate of the biofilm-forming bacteria, and (iii) other physiological and gene
transcription changes associated with biofilm formation [24]. Biofilms have been
described in several Mycoplasma species to date [26-28]; however, there are very
few studies that have described ureaplasmal biofilms and the associated
implications for antimicrobial resistance.
In this present study, with a fetal sheep model [15, 29-32], we tested whether
standard erythromycin treatment of chronic, intra-amniotic ureaplasma infections
could induce genetic markers of macrolide resistance in AF and chorioamnion
ureaplasma isolates, resulting in changes in antimicrobial susceptibility profiles. We
also investigated the ability of ureaplasma isolates to form biofilms, and the
associated effects on macrolide sensitivity in sessile and planktonic ureaplasmas.
By comparing recovered AF and chorioamnion isolates to the clinical ureaplasma
124
isolate used as the inoculum, we aimed to determine the in vivo effects of sub-lethal
exposure to erythromycin on these isolates.
125
Materials and Methods
All experimental procedures were approved by the Animal Ethics Committees of the
University of Western Australia, Cincinnati Children’s Hospital Medical Centre and
Queensland University of Technology.
Animal model and specimen collection
The inoculum used for intra-amniotic injection was a low passage, erythromycin-
susceptible, U. parvum serovar 3 isolate (442S) that was originally isolated from the
semen of an infertile man attending the Wesley IVF Service (Brisbane,
Queensland). This patient gave informed consent for the isolate to be used for
research. Ureaplasmas for injection were prepared as previously described [30] and
diluted to 2x104 CFU in PBS prior to intra-amniotic injection.
The samples analysed in this study were collected from a previous animal
experiment [15]. Briefly, at 55 days of gestation (term = 150 d) 12 date-mated
Merino ewes bearing single fetuses received a 2 mL intra-amniotic injection of U.
parvum serovar 3, as described previously [30]. Injections were guided by
ultrasound imaging, and the intra-amniotic location of the injection was confirmed by
electrolyte analysis of the aspirated fluid (Rapidlab 865, Bayer Diagnostics, Pymble,
New South Wales). At 100 d, ewes were randomly assigned into groups to receive
erythromycin treatment (Up/E group; n = 6) or saline (Up group; n = 6). Ewes that
received erythromycin treatment were injected intra-muscularly with 500 mg of
erythromycin (Abbot Australasia, Kurnell, New South Wales) three times daily for
four days (100 d - 104 d), resulting in a total dose of 30 mg/kg/day. This treatment
regimen was selected because it is similar to erythromycin treatment given to
pregnant women and is appropriate for administration to sheep. Throughout the
duration of this experiment antibiotics were not added to the supplementary feed
126
given to pregnant ewes, nor did animals receive antimicrobial treatment as part of
on-going veterinary care.
Preterm fetuses were surgically delivered at 125 d, as described previously [15].
Samples of AF and chorioamnion were aseptically collected, immediately snap
frozen in liquid nitrogen and stored at -80 ˚C for subsequent analysis.
Ureaplasma culture
Samples of AF and chorioamnion from each animal were cultured for ureaplasmas.
Briefly, 0.1 g of thawed chorioamnion was homogenized in 1.5 mL of 10B media
using a mini beadbeater 8-cell disrupter (Daintree Scientific, St Helens, Tasmania).
Homogenized chorioamnion and AF samples were then inoculated into 10B media
and nine 10-fold serial dilutions were performed [33]. Inoculated broths were
incubated aerobically at 37 ˚C for 18 - 48 hours. Positive cultures, as determined by
a colour change in the media due to an increase in pH and subsequent alkaline
shift, were frozen and stored at -80 ˚C for further analysis.
Minimum Inhibitory Concentration
The MICs of erythromycin (Sigma, Castle Hill, New South Wales), roxithromycin
(Sigma) and azithromycin (Pzifer, West Ryde, New South Wales) were determined
against ureaplasmas cultured from AF and chorioamnion samples, using a
previously described microdilution method [34, 35]. Using 96 well microtitre plates
(Nunc, Roskilde, Denmark), each of the antibiotics were serially diluted (two-fold
dilutions) in 25 µL of 10B broth to give a concentration range of 256 µg/mL to 0.008
µg/mL. Isolates to be tested were thawed, diluted to a standardised concentration of
1x104 CFU/mL (verified by serial dilutions in 10B medium and drop plate analysis on
A8 agar) and incubated at 37 °C for two hours. Wells were then inoculated with 175
µL of these pre-warmed isolates, the plates were sealed with acetate plate sealing
film (MP Biomedicals, Seven Hills, New South Wales) and then incubated
127
aerobically at 37 °C for 24 - 48 hours. Each test included an antibiotic-free positive
growth control well and media/antibiotic only negative controls. Plates were
regularly monitored for growth in the antibiotic-free growth control, as evidenced by
a colour change in the media due to an alkaline shift produced by urea hydrolysis.
The MIC for each isolate was defined as the lowest concentration in which growth of
the organism was inhibited at the time that the antibiotic-free growth control first
showed a colour change. Each isolate was tested against each of the three
antibiotics in triplicate and MIC results were expressed as the mean value of the
three experiments. Isolates were unable to be classified as either susceptible or
resistant, as breakpoints for ureaplasmas and mycoplasmas have not been
established by the Clinical and Laboratory Standards Institute (CLSI).
Minimum Biofilm Inhibitory Concentration
To determine (i) whether ureaplasmas isolated from chorioamnion samples formed
biofilms in vitro and (ii) to measure the associated minimum biofilm inhibitory
concentrations (MBIC) of erythromycin, azithromycin and roxithromycin, modified
biofilm susceptibility assays were performed [36]. All chorioamnion ureaplasma
isolates were diluted in 10B broth to a standardized concentration of 1x104 CFU/mL.
200 µL of each isolate was inoculated into the wells of a 96 well microtitre plate
(Nunc) and biofilm formation was facilitated by immersing the pegs of an Immuno
TSP Screening Lid (Nunc) into the wells. Plates were incubated at 37 °C for 24 - 48
hours, until a colour change was detected in the media, indicating ureaplasma
growth. Lids were then washed in sterile PBS three times to remove non-adherent
ureaplasmas and the lid was transferred to a new microtitre plate containing serial
dilutions of erythromycin, azithromycin and roxithromycin, as per the MIC protocol
above. Plates were then incubated until growth was observed in the antibiotic-free
well, and the biofilms were then transferred from the lids into the wells of the
microtitre plate by centrifugation at 3000 rpm for 10 minutes at 4 °C. The lid was
128
then replaced with new, sterile lids and plates were re-incubated for 24 hours.
Results were interpreted and recorded according to the criteria described above for
MIC testing. MBIC results are expressed as the average of duplicate readings.
DNA Extraction from Culture
Ureaplasmal DNA from AF and chorioamnion specimens was extracted from first
passage (P1) 10B broth cultures as described by Blanchard et al. [37]. Briefly, 500
µL of late log phase culture was centrifuged at 14,000 rpm for 20 minutes. The
supernatant was discarded and the pellet resuspended in 125 µL of Solution A (10
mM Tris HCl pH 8.5, 100 mM KCl, 2.5 mM MgCl2) and 125 µL of Solution B (10 mM
Tris HCl pH 8.5, 2.5 mM MgCl2, 1% v/v Tween 20, 1% v/v Triton X100) with 120
µg/mL proteinase K. These samples were incubated at 60 °C for one hour, then 95
°C for 10 minutes and allowed to cool before being stored at -20 °C.
23S rRNA gene PCR and Sequencing
To detect polymorphisms within genes described previously [20, 21] PCRs targeting
domain II and domain V of the 23S rRNA gene (amplifying both operons of the 23S
rRNA gene of ureaplasmas), ribosomal protein L4 gene and ribosomal protein L22
were performed on all AF and chorioamnion ureaplasma isolates. Primer sequences
are shown in Table 4.1. These PCR assays were performed in 50 µL reaction
mixtures containing: 100 µM of dNTP mix (Roche Diagnostics, Castle Hill, New
South Wales), 1x PCR buffer (Tris HCl, KCl, (NH4)2SO4, pH 8.7, Invitrogen, Mt
Waverley, Victoria), 1.5 mM of MgCl2 (Invitrogen), 0.5 µM of each primer (Sigma),
2.5 U of Platinum Taq Polymerase (Invitrogen) and sterile distilled water. PCR
cycling occurred in a PTC-2000 Thermal Cycler (Global Medical Instrumentation,
Ramsey, Minnesota) and the cycling conditions involved initial denaturation at 94 °C
for 15 minutes, followed by 35 cycles of denaturation at 94 °C for one minute, primer
annealing at 56 °C for one minute, extension at 72 °C for two minutes, plus a final
129
PRIMER TARGET AND NAME
SEQUENCE (5’ – 3’) REFERENCE
23S Domain V 23SF 23SR MH23S-11 MP23S-22 MH23S-9 MP23S-23
GTGAAATCCTGGTGAGGGTGA TTCCTACGGGCATGACAGATAG TAACTATAACGGTCCTAAGG GGCGACCGCCCCAGTCAAAC GCTCAACGGATAAAAGCTAC ACACTTAGATGCTTTCAGCG
Dongya et al. 2008 [20] Pereyre et al. 2007 [21] Pereyre et al. 2007 [21]
23S Domain II Up23S-30 Up23S-31
TGCCTTTTGAAGTATGAGCC TGGCGCCATCATAGATTCAG
Pereyre et al. 2007 [21]
Ribosomal protein L4 gene
UpL4-U UpL4-R
TCTATTGATGGTAACTTCGG GTTGAAGGTGTTTCTAAATCGC
Pereyre et al. 2007 [21]
Ribosomal protein L22 gene
UpL22-U UpL22-R
TTCGCACCGTAAAGCTTCTC GTTCTGGATCAACGTTTTCG
Pereyre et al. 2007 [21]
erm(B) GAAAAGGTACTCAACCAAATA AGTAACGGTACTTAAATTGTTTAC
Graham et al. 2009 [39]
msr(A) GGCACAATAAGAGTGTTTAA AAGTTATATCATGAATAGATTGTCCTGTT
Lina et al. 1999 [40]
msr(B) TATGATATCCATAATAATTATCCAATC AAGTTATATCATGAATAGATTGTCCTGTT
Lina et al. 1999 [40]
msr(C) AAGGAATCCTTCTCTCTCCG GTAAACAAAATCGTTCCCG
Lu et al. 2010 [22]
msr(D) TTGGACGAAGTAACTCTG GCTTGGCTCTTACGTTC
Daly et al. 2004 [38]
Table 4.1: PCR primers used for the detection of polymorphisms in the 23S rRNA gene and ribosomal protein genes; and the detection of macrolide resistance genes.
130
extension step at 72 °C for 15 minutes. Positive controls included DNA extracted
from the original ureaplasma inoculum (isolate 442S) and U. parvum serovar 3
reference strain (courtesy of H. Watson, University of Alabama, Birmingham); and
negative controls included master mix only and reaction mixtures with distilled water
substituted for template. PCR products were electrophoresed through a 2% agarose
tris-borate EDTA (TBE) gel, visualised by ethidium bromide (10 µg/mL) staining and
digitized using Grab-It Gel Dock (Ultraviolet Products Ltd., Cambridge, United
Kingdom).
PCR products from five AF isolates, four chorioamnion isolates and isolate 442S
then were selected for further analysis by sequencing. These isolates were selected
as being representative of the MIC range obtained for erythromycin, azithromycin
and roxithromycin. PCR products were purified using the High Pure PCR Product
Purification Kit (Roche) according to manufacturer’s instructions. Sequencing
reactions were performed by the Australian Genome Research Facility (AGRF;
University of Queensland, St Lucia, Queensland). The sequence data were trimmed
to obtain sequences of a uniform length and then aligned using Clustal W (Angis,
Sydney, New South Wales) to identify any sequence polymorphisms. Sequence
identity was confirmed by the Basic Local Alignment Tool (BLAST, National Center
for Biotechnology Information 2010). Partial 23S rRNA sequences from AF and
chorioamnion ureaplasma isolates have been deposited in Genbank (accession
numbers: JF521483, JF521484, JF521485, JF521486).
erm(B) and msr gene PCR
PCRs to detect selected genes associated with macrolide resistance (erm(B),
msr(A), msr(B), msr(C) and msr(D)) were performed as described previously [22,
38-40] with slight modifications. These assays were performed on the same
isolates, which were selected for 23S rRNA gene sequencing. PCRs were
131
performed in 50 µL volumes as described for PCR amplification of 23S rRNA genes,
and primer sequences are shown in Table 4.1. PCR reactions were performed using
cycling conditions described by Lu et al. [22]; however, initial denaturation and final
elongation steps were modified to 94 °C for 15 minutes and 72 °C for 15 minutes
respectively.
Statistical analysis
MIC and MBIC data were analysed by Pearson’s chi squared test, and independent
and one-sample t-tests using SPSS Version 16 (SPSS Inc., Chicago). Statistical
significance was accepted at p < 0.05.
132
Results
Ureaplasma Culture
Ureaplasmas were isolated by culture from all 12 AF samples and 8 of 12 (66.7%)
chorioamnion samples. First passage cultures were used for MIC and MBIC testing,
and DNA extraction.
Minimum Inhibitory Concentrations
The MIC values for the initial U. parvum serovar 3 inoculum (isolate 442S) injected
into the sheep were 0.13 mg/L for erythromycin and 0.5 mg/L for azithromycin and
roxithromycin. Although defined breakpoints are not available at this time, MICs of
the tested antibiotics against AF isolates were low (erythromycin MIC range = 0.08
mg/L - 0.63 mg/L; azithromycin MIC range = 0.13 mg/L - 1.0 mg/L; roxithromycin
MIC range = 0.06 mg/L- 0.83 mg/L) suggesting that all AF isolates were susceptible
to the three antibiotics (Table 4.2). All AF isolates had MIC values for erythromycin
which fell within the normal range (not associated with breakpoints, 0.02 - 4.0 mg/L)
for ureaplasmas as reported in the Cumitech 34 Manual [34]. One isolate (1 of 12,
8%) had a MIC lower than the normal range for roxithromycin (0.1 - 2.0 mg/L) and 7
isolates (7 of 12, 58%) had MICs lower than the normal range for azithromycin (0.5 -
4.0 mg/L).
MICs of erythromycin and azithromycin against chorioamnion ureaplasma isolates
were also low (erythromycin and azithromycin MIC range = 0.06 – 0.25 mg/L),
suggesting susceptibility of these isolates to these antibiotics. Conversely, the MICs
of roxithromycin against chorioamnion isolates were variable, ranging from 0.13 –
5.33 mg/L. Five chorioamnion isolates (5 of 8, 62.5%) had roxithromycin MICs that
were ≤ 0.67 mg/L, most likely indicating susceptibility to this drug; and three isolates
(3 of 8, 37.5%) had increased MIC values (isolate 226 = 4.00 mg/L; isolate 227 =
2.67 mg/L; isolate 229 = 5.33 mg/L), which could potentially represent low-level or
133
AMNIOTIC FLUID
ISOLATES
CHORIOAMNION ISOLATES
ANIMAL
NUMBER
TREATMENT
GROUP
MIC (mg/L) MIC (mg/L) MBIC (mg/L)
ERY
AZM
ROX
ERY
AZM
ROX
ERY
AZM
ROX
222 Up/E
0.17 0.33 0.50 - - - - - -
223 Up/E
0.13 0.50 0.50 - - - - - -
225 Up/E
0.13 0.13 0.13 0.06 0.13 0.13 0.06 0.06 0.50
226 Up/E
0.25 0.29 0.34 0.06 0.17 4.00 0.03 0.06 0.13
227 Up/E
0.63 0.72 0.83 0.13 0.10 2.67 NBF NBF NBF
228 Up/E
0.25 1.00 0.42 0.08 0.06 0.13 0.13 0.13 0.25
229 Up
0.33 0.50 0.17 0.25 0.25 5.33 0.13 0.13 0.50
230 Up
0.08 0.50 0.06 - - - - - -
231 Up
0.17 0.13 0.13 0.13 0.06 0.67 NBF NBF NBF
232 Up
0.08 0.33 0.50 0.06 0.06 0.25 NBF NBF NBF
233 Up
0.35 0.25 0.50 - - - - - -
234 Up
0.33 0.33 0.50 0.25 0.25 0.50 0.13 0.25 0.50
MIC50 0.17 0.33 0.42 0.08 0.10 0.50 0.13 0.13 0.5
MIC90 0.35 0.72 0.83 0.25 0.25 4.00 0.13 0.13 0.5
Table 4.2: MIC and MBIC values of erythromycin (ERY), azithromycin (AZM) and roxithromycin (ROX) against amniotic fluid and chorioamnion ureaplasma isolates
Note: MBIC analysis was only performed on chorioamnion isolates. Up/E = ureaplasma + erythromycin treatment group; Up = ureaplasma treatment group; dash (-) indicates samples which were ureaplasma culture negative; NBF = non-biofilm forming strain. Defined breakpoints are not available; however, the normal ranges (Cumitech 34 Manual [34]) are: erythromycin 0.02 - 4.0 mg/L; azithromycin 0.5 - 4.0 mg/L and roxithromycin 0.1 - 2.0 mg/L.
134
intermediate resistance to roxithromycin. All chorioamnion isolates had MICs within
the normal ranges (values indicated above) for erythromycin and all chorioamnion
isolates had MIC values for azithromycin, which were below the reported normal
ranges [34]. The three chorioamnion ureaplasma isolates that had increased
roxithromycin MICs (isolates 226, 227 and 229) also had MICs that were above the
normal reported range.
Despite all animals being injected with the same ureaplasma inoculum (isolate
442S), the MICs of ureaplasmas isolated from AF and chorioamnion samples after
chronic infection demonstrated heterogeneity between animals and between
anatomical site of isolation (AF or chorioamnion; Figure 4.1). There were no
statistical differences in average MICs for any of the antimicrobials tested between
isolates obtained from animals in the Up/E group or the Up group (erythromycin p =
0.74; azithromycin p = 0.30; roxithromycin p = 0.29). When comparing MIC values
for AF and chorioamnion ureaplasma isolates (and not taking animal group into
consideration), both erythromycin and azithromycin had significantly lower activity
(characterized by increased MICs) against AF isolates compared to chorioamnion
isolates (p = 0.005 and p = 0.001 respectively). Despite three chorioamnion
ureaplasma isolates demonstrating increased resistance to roxithromycin, there
were no statistical differences in average roxithromycin MICs between AF (mean ±
standard error = 0.38 ± 0.06 mg/L) and chorioamnion (1.71 ± 0.72 mg/L) isolates (p
= 0.11), potentially due to the large standard deviation (SD = 2.03) obtained in the
chorioamnion roxithromycin MIC data.
Minimum Biofilm Inhibitory Concentrations
Although isolate 442S (the inoculum used for the sheep experiments) was a non-
biofilm forming strain, in vitro biofilm formation was observed in five (out of eight,
62.5%) chorioamnion ureaplasma isolates. Three isolates were from animals from
135
Figure 4.1: Amniotic fluid and chorioamnion ureaplasmas demonstrate variable susceptibilities to macrolide antimicrobials.
Minimum inhibitory concentrations (MICs) of erythromycin (ERY), azithromycin (AZM) and roxithromycin (ROX) against: A, amniotic fluid (AF) ureaplasma isolates from the Up group; B, AF ureaplasma isolates from the Up/E group; C, chorioamnion ureaplasma isolates from the Up group; D, chorioamnion ureaplasma isolates from the Up/E group. MIC values were variable between isolates, particularly with respect to roxithromycin against chorioamnion isolates (C and D). MICs were not different between ureaplasmas isolated from animals from the Up group compared to the Up/E group (p > 0.05). Individual points on graphs represent the MICs of each of the tested isolates. Bars represent mean MIC ± SEM. Mean ± SEM could not be displayed for roxithromycin MICs against chorioamnion isolates due to the break in the y axis. Up = ureaplasma group; Up/E = ureaplasma + erythromycin group.
ERY
AZM
ROX
0.0
0.5
1.0
AAF MIC Up group
MIC
(m
g/L
)
ERY
AZM
ROX
0.0
0.5
1.0
BAF MIC Up/E group
MIC
(m
g/L
)
ERY
AZM
ROX
0.0
0.2
0.4
0.6
3.0
4.0
5.0
6.0
CChorioamnion MIC Up group
MIC
(m
g/L
)
ERY
AZM
ROX
0.0
0.2
0.4
0.6
3.04.05.06.0
DChorioamnion MIC Up/E group
MIC
(m
g/L
)
136
the Up/E group, and the remaining two isolates were isolated from animals from the
Up group. Unexpectedly, the MBICs for the biofilm-forming strains were generally
not higher than the MIC values of planktonic ureaplasmas isolated from the same
animal (Table 4.2). Of the three chorioamnion isolates, which had increased MICs to
roxithromycin (isolates 226, 227 and 229), two isolates formed biofilms; however,
MBIC values were much lower than the associated MICs of the planktonic cells.
There were no statistical differences in average MBIC values for the three
antimicrobials between isolates from the Up/E group and the Up group
(erythromycin p = 0.24; azithromycin p = 0.15; roxithromycin p = 0.24). The average
MBIC of roxithromycin was significantly lower than the corresponding MIC for
chorioamnion isolates (p = 0.001).
23S rRNA gene PCR and sequencing
The 442S inoculum strain shared 100% sequence identity with the U. parvum
serovar 3 reference strain (ATCC 700970; Genbank accession number AF222894).
No sequence polymorphisms were found in any ureaplasmas isolated from AF
samples across any of the targeted regions of the 23S rRNA gene and the L4 and
L22 ribosomal protein genes. All AF isolates shared 100% sequence identity with
the U. parvum serovar 3 reference strain and isolate 442S (Figures 4.2 A and 4.2
B).
Conversely, there were numerous nucleotide substitutions within all tested
chorioamnion ureaplasma isolates when compared to AF isolates, the U. parvum
serovar 3 reference strain and isolate 442S. All of the mutations were localised
within domain V of the 23S rRNA gene, specifically within the regions amplified by
PCR primers MH23S-11/MP23S-22 (Figure 4.2 A) and MH23S-9/MP23S-23 (Figure
2B). These mutations were considered non-random in nature, because the
sequence polymorphisms within these regions were identical between the four
137
A
U. parvum S3 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG 2040
442S CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
AF 225 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
AF 227 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
AF 228 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
AF 230 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
AF 234 CTCTTGACTGTCTCAACTGTAGACTCGGTGAAATCCTGGTGAGGGTGAAG
CAM 226 GGCGGCGCTGTCTCCACCCGAGACTCAGTGAAATTGAAATCGCTGTGAAG
CAM 227 GGCGGCGCTGTCTCCACCCGAGACTCAGTGAAATTGAAATCGCTGTGAAG
CAM 228 GGCGGCGCTGTCTCCACCCGAGACTCAGTGAAATTGAAATCGCTGTGAAG
CAM 229 GGCGGCGCTGTCTCCACCCGAGACTCAGTGAAATTGAAATCGCTGTGAAG
U. parvum S3 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT 2100
442S ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
AF 225 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
AF 227 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
AF 228 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
AF 230 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
AF 234 ACGCCCTCTTGGCGTGATTGGACGGAAAGACCCCATGAAGCTTTACTGTAGCTTAATATT
CAM 226 ATGCAGTGTATCCGCGGCTAGACGGAAAGACCCCGTGAACCTTTACTATAGCTTTGCACT
CAM 227 ATGCAGTGTATCCGCGGCTAGACGGAAAGACCCCGTGAACCTTTACTATAGCTTTGCACT
CAM 228 ATGCAGTGTATCCGCGGCTAGACGGAAAGACCCCGTGAACCTTTACTATAGCTTTGCACT
CAM 229 ATGCAGTGTATCCGCGGCTAGACGGAAAGACCCCGTGAACCTTTACTATAGCTTTGCACT
U. parvum S3 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA 2160
442S GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
AF 225 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
AF 227 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
AF 228 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
AF 230 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
AF 234 GGGAAATTTTATTACTTGTAGAGCATAGGTAGGAGACTGTGAAGTATACTCGCTAGGGTA
CAM 226 GGACTTTGAATTTGCTTGTGTAGGATAGGTGGGAGGCTTTGAAGCGTGGACGCCAGTTCG
CAM 227 GGACTTTGAATTTGCTTGTGTAGGATAGGTGGGAGGCTTTGAAGCGTGGACGCCAGTTCG
CAM 228 GGACTTTGAATTTGCTTGTGTAGGATAGGTGGGAGGCTTTGAAGCGTGGACGCCAGTTCG
CAM 229 GGACTTTGAATTTGCTTGTGTAGGATAGGTGGGAGGCTTTGAAGCGTGGACGCCAGTTCG
U. parvum S3 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA 2220
442S TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
AF 225 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
AF 227 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
AF 228 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
AF 230 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
AF 234 TATGGAGTCAACGTTGGAATACTACCCTTGTGATAAGATTTCTCTAACCTGCAGCCATGA
CAM 226 CGTGGAGCCAACCTTGAAATACCACCCTGGCAACTTTGAGGTTCTAACTCAGGTCCGTTA
CAM 227 CGTGGAGCCAACCTTGAAATACCACCCTGGCAACTTTGAGGTTCTAACTCAGGTCCGTTA
CAM 228 CGTGGAGCCAACCTTGAAATACCACCCTGGCAACTTTGAGGTTCTAACTCAGGTCCGTTA
CAM 229 CGTGGAGCCAACCTTGAAATACCACCCTGGCAACTTTGAGGTTCTAACTCAGGTCCGTTA
138
B
U. parvum S3 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT 2580
442S CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
AF 225 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
AF 227 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
AF 228 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
AF 230 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
AF 234 CATCTCATCCTGGAGCTGAAGCAGGTTCCAAGGGTTCGGCTGTTCGCCGATTAAAGAGAT
CAM 226 CATCACATCCTGGGGCTGAAGCCGGTCCCAAGGGTATGGCTGTTCGCCATTTAAAGTGGT
CAM 227 CATCACATCCTGGGGCTGAAGCCGGTCCCAAGGGTATGGCTGTTCGCCATTTAAAGTGGT
CAM 228 CATCACATCCTGGGGCTGAAGCCGGTCCCAAGGGTATGGCTGTTCGCCATTTAAAGTGGT
CAM 229 CATCACATCCTGGGGCTGAAGCCGGTCCCAAGGGTATGGCTGTTCGCCATTTAAAGTGGT
U. parvum S3 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG 2640
442S ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
AF 225 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
AF 227 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
AF 228 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
AF 230 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
AF 234 ACGTGAGTTGGGTTCAAACCGTCGTGAGACAGGTTGGTCCCTATCTGTCATGCCCGTAGG
CAM 226 ACGCGAGCTGGGTTTAGAACGTCGTGAGACAGTTCGGTCCCTATCTGCCGTGGACGTTTG
CAM 227 ACGCGAGCTGGGTTTAGAACGTCGTGAGACAGTTCGGTCCCTATCTGCCGTGGACGTTTG
CAM 228 ACGCGAGCTGGGTTTAGAACGTCGTGAGACAGTTCGGTCCCTATCTGCCGTGGACGTTTG
CAM 229 ACGCGAGCTGGGTTTAGAACGTCGTGAGACAGTTCGGTCCCTATCTGCCGTGGACGTTTG
U. parvum S3 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC 2700
442S AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
AF 225 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
AF 227 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
AF 228 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
AF 230 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
AF 234 AAGATTGAGAAGAGCTGTTCCTAGTACGAGAGGACCGGAATGGACACACCTCTTGTGATC
CAM 226 AGATTTGAGAGGGGCTGCTCCTAGTACGAGAGGACCGGAGTGGACGAACCTCTGGTGTTC
CAM 227 AGATTTGAGAGGGGCTGCTCCTAGTACGAGAGGACCGGAGTGGACGAACCTCTGGTGTTC
CAM 228 AGATTTGAGAGGGGCTGCTCCTAGTACGAGAGGACCGGAGTGGACGAACCTCTGGTGTTC
CAM 229 AGATTTGAGAGGGGCTGCTCCTAGTACGAGAGGACCGGAGTGGACGAACCTCTGGTGTTC
U. parvum S3 CTGTTGTCG 2709
442S CTGTTGTCG
AF 225 CTGTTGTCG
AF 227 CTGTTGTCG
AF 228 CTGTTGTCG
AF 230 CTGTTGTCG
AF 234 CTGTTGTCG
CAM 226 CGGTTGTCA
CAM 227 CGGTTGTCA
CAM 228 CGGTTGTCA
CAM 229 CGGTTGTCA
Figure 4.2: Significant genetic variability in the 23S rRNA gene of chorioamnion ureaplasmas 23S rRNA domain V sequence alignments amplified by primers MH23S-11/MP23S-22 (A) and MH23S-9/MP23S-23 (B). Sequence alignments compare amniotic fluid (AF) ureaplasma isolates (n = 5) and chorioamnion (CAM) ureaplasma isolates (n = 4) to the Ureaplasma parvum serovar 3 reference strain (U. parvum S3; ATCC 700970, Genbank Accession number AF222894) and the inoculum strain (isolate 442S). Black shading represents areas of 100% sequence homology across all isolates and grey shading represents non-random mutations found in all chorioamnion isolates, but not in amniotic fluid isolates. Numbering shown is U. parvum serovar 3 numbering.
139
sequenced chorioamnion ureaplasmas that were isolated from individual sheep.
Within the region amplified by MH23S-11/MP23S-22, there were 89 polymorphisms
(out of 230 nucleotides, 38.7%). The nucleotide similarity between the AF isolates
and chorioamnion isolates was 61.3% and the G+C content of this region was
increased in chorioamnion isolates (52% G+C) compared to AF isolates (44%
G+C). In the region amplified by MH23S-9/MP23S-23, 36 polymorphisms (out of
192 nucleotides, 18.8%) were detected, giving a nucleotide similarity of 81.2%
between the AF and chorioamnion isolates. In this region the G+C content was also
increased in chorioamnion isolates (56% G+C) compared to AF isolates (52%
G+C). Despite the large number of sequence polymorphisms, specific nucleotides
that have been previously associated with macrolide resistance in ureaplasmas and
other bacteria (nucleotides G2056, G2057 and A2058 of domain V of 23S rRNA
gene, Escherichia coli numbering, [21]; and C2243, U. urealyticum numbering [20])
remained conserved in all isolates. The non-random mutations found within the
chorioamnion isolates were localized to domain V of the 23S rRNA gene, as no
mutations were detected in domain II of the 23S rRNA gene or ribosomal proteins
L4 and L22, and sequencing of these four genes confirmed the identity of
chorioamnion isolates as U. parvum serovar 3 (data not shown).
Despite obtaining ureaplasma isolates in pure culture and having no evidence of a
polymicrobial infection, BLAST analysis of regions of domain V of the 23S rRNA
gene containing non-random mutations from chorioamnion ureaplasma isolates
revealed sequence similarity to Pseudomonas species. Specifically, P. aeruginsoa,
P. fluorescens, P. stuzeri and P. putida, had significant sequence similarities to this
fragment, suggesting that foreign DNA may have been integrated into the
ureaplasma genome. All other tested regions of the 23S rRNA gene and L4 and L22
ribosomal proteins gave positive identities for U. parvum serovar 3.
erm(B) and msr gene PCR
140
The macrolide resistance erm(B) gene was not detected in any AF ureaplasma
isolates, but was present in isolate 442S and 100% of chorioamnion ureaplasma
isolates (Figure 4.3 A). Of the four tested msr gene subtypes, msr(C) and msr(D)
were the only genes detected in the ureaplasma isolates; however, the bands were
faint (Figure 4.3 B and 4.3 C). Isolate 442S and 2 out of 5 AF isolates (40%) carried
the msr(D) gene, but not the msr(C) gene, whereas 100% of chorioamnion isolates
tested positive for both the msr(C) and msr(D) gene subtypes. Unexpectedly, the
msr(C) gene did not amplify at its expected size of 343 bp, rather, the PCR assay
produced an amplicon of >1114 bp (Figure 3B). PCR of the msr(D) gene produced
amplicons of expected sizes (370 bp). However, multiple bands of higher molecular
weight were also detected in chorioamnion isolates (Figure 4.3 C). PCR assays
were repeated to ensure all results were reproducible, and identical results were
obtained. However, DNA of sufficient purity and quantity could not be obtained for
sequence analysis and confirmation of the identity of these genes, and therefore
these results should be treated as preliminary findings only. The msr(A) and msr(B)
genes were not detected in any isolates.
141
Figure 4.3: Macrolide resistance genes were detected in chorioamnion ureaplasma isolates.
PCR detection of erm(B) (A), msr(C) (B) and msr(D) (C) resistance genes in amniotic fluid and chorioamnion ureaplasma isolates. M = Molecular weight marker VIII (Roche, Castle Hill, New South Wales); AF = amniotic fluid ureaplasma isolates; CAM = chorioamnion ureaplasma isolates; 442S = U. parvum inoculum strain; bp = base pairs.
A
B
C
142
Discussion
Exposure of microorganisms to sub-inhibitory concentrations of antimicrobials and
disinfectants promotes the emergence of antimicrobial resistance [41-43]. We
investigated whether low levels of erythromycin exposure in vivo were able to
generate phenotypic and genotypic markers of macrolide resistance in AF and
chorioamnion ureaplasma isolates. We found that exposure to erythromycin in vivo
did not have any effect on the antimicrobial resistance of ureaplasma isolates.
Rather, the anatomical site of isolation was the most important factor in determining
macrolide susceptibility patterns and changes to macrolide target sites.
MIC testing of erythromycin, azithromycin and roxithromycin suggested that
erythromycin was the most effective drug against AF and chorioamnion ureaplasma
isolates. Despite all animals in our experiment being injected with the same
ureaplasma isolate, variations in MICs were observed. This variability could not be
attributed to erythromycin exposure, as there were no differences in average MICs
between ureaplasmas isolated from animals that received erythromycin treatment
and those isolated from animals that received saline. Although we only analysed a
small number of animals (n = 12), our data suggest that the site of isolation may
have an effect on antimicrobial susceptibility, as significant differences were found
in MICs between AF and chorioamnion ureaplasma isolates. The absence of
standardized breakpoints limits our analysis in regards to the classification of
isolates as susceptible or resistant; however, the most obvious differences were
found in regards to roxithromycin, against which 37.5% of chorioamnion isolates
had MICs that could potentially represent low-level resistance. Although not
associated with breakpoints, the fact that these MICs are above the reported normal
ranges [34] supports that these isolates may demonstrate mild resistance to this
drug.
143
Our study also demonstrated that ureaplasmas isolated from the chorioamnion are
able to form biofilms in vitro. Bacterial cells growing within a biofilm are thought to
be regulated by quorum sensing mechanisms [44] and usually exhibit altered
phenotypes in terms of colonial morphology [45], growth rates [46], resistance to
stresses [26] and gene transcription [47]. Our data demonstrated that 62.5% of
chorioamnion ureaplasmas formed biofilms, although our parent isolate (442S) was
unable to form a biofilm under these conditions. It is possible that isolate 442S in its
original state may have the genes required for biofilm formation switched off, and
when introduced into the fetal model these genes may have been activated as a
survival mechanism due to pressures from the host immune response, nutrient
availability or sub-optimal pH. Similar to the MIC data we present in this report,
exposure to erythromycin had no effect on (i) the ability of ureaplasmas to form
biofilms and (ii) the associated MBIC of erythromycin, azithromycin and
roxithromycin.
Biofilm formation has been well associated with antimicrobial resistance in other
bacteria (especially P. aeruginosa), and there are many studies that demonstrate
that MBICs are often significantly higher than the MICs of associated planktonic
cells [23, 45, 25]. Our data do not support this, as average MBICs for erythromycin
and azithromycin were not different to the MICs of planktonic cells, and the average
MBICs of roxithromycin were significantly lower than the associated MICs. Similar
data have been published previously by Garcia-Castillo et al. [36], who studied in
vitro biofilm formation and antimicrobial resistance in ureaplasmas isolated from
urethral exudates or semen from patients with urethritis or chronic prostatitis (n = 9),
or urine from healthy individuals (n = 2). They concluded that biofilm formation can
protect mycoplasmas from antimicrobials, but further analysis of their data showed
that 5 out of 11 (45%) of the tested ureaplasmas had MICs for either one or several
antimicrobials (including macrolides) that were equivalent to MBICs, or in some
144
instances MICs were greater than MBICs. Similarly, McAuliffe et al. [27] reported
that biofilm formation by Mycoplasma bovis was not associated with increased
antimicrobial resistance when compared to the MIC of planktonic bacteria; however,
other phenotypic changes were observed in biofilm forming isolates. Clearly, further
studies are required to elucidate the implications of biofilm formation in
ureaplasmas.
To investigate potential mechanisms underlying differences in MICs between AF
and chorioamnion ureaplasmas we sequenced regions of the ureaplasma 23S
rRNA gene. Resistance to macrolide antibiotics can occur by target site modification
of the 23S rRNA gene, either by point mutations in the 23S rRNA gene or in
ribosomal protein L4 and L22 genes [16]. Point mutations in domain V of the 23S
rRNA gene that have been associated with macrolide resistance in ureaplasmas
include G2056U, G2057U and A2058G (E. coli numbering) [21]; and C2443 (T or C)
(U. urealyticum numbering) [20]. Additionally, a 6 bp deletion was found in the
ribosomal protein L4 gene sequence of a U. parvum serovar 1 isolate, which was
highly resistant to erythromycin (MIC >64 mg/L) [19].
Our data show high levels of sequence variability in domain V of the 23S rRNA
gene in chorioamnion ureaplasma isolates only, when compared to AF isolates and
the inoculum/parent strain (442S). Despite this sequence variability, we did not find
point mutations at any nucleotide positions which were previously associated with
macrolide resistance. As the sequence variability was identical in all chorioamnion
ureaplasmas (and not just those which had increased MICs against roxithromycin),
it appears that these non-random mutations are not associated with antimicrobial
resistance. Also, since the sequence variation was found in animals from both the
Up and Up/E animal groups, these non-random mutations were not related to sub-
lethal exposures to erythromycin. Silent mutations in macrolide target genes were
also reported in ureaplasmas by others. Dongya et al. [20] reported that
145
polymorphisms at positions A2149C and A2181T of the 23S rRNA gene (U. parvum
numbering) were found in isolates that demonstrated resistance to roxithromycin
and azithromycin; however, these polymorphisms were also present in four strains
that were susceptible to all tested macrolides. Beeton et al. [19] also reported three
species-specific conserved nucleotide polymorphisms in the ribosomal protein L4
gene at positions 309, 357 and 373, which were not associated with antimicrobial
resistance. Moreover, when investigating fluoroquinolone resistance in
ureaplasmas, Beeton et al. [48] found that there were large numbers of non-
resistance polymorphisms in GyrA (39 amino acid changes), GyrB (26 amino acid
changes), ParC (107 amino acid changes) and ParE (34 amino acid changes)
proteins.
Due to the large number of nucleotide polymorphisms found in domain V of the 23S
rRNA gene of our chorioamnion isolates, and the fact that these polymorphisms are
(i) identical between all chorioamnion ureaplasmas, and (ii) associated with an
increase in G+C content, we propose that these variable sequences may represent
a fragment transferred via horizontal gene transfer (HGT), as opposed to a
collection of unrelated nucleotide polymorphisms. Ribosomal RNA is thought to
make up part of the core genome, and it is therefore thought that the genes
encoding rRNAs are highly conserved and mutations are unlikely. However, HGT of
rRNA has been described in other bacteria, resulting in a mosaic-like structure of
rRNA genes. In a study of 708 strains of the Streptococcus anginosus group,
reverse line blot hybridisation experiments revealed 11 distinct 16S rRNA profiles.
Sequencing of these 16S rRNA genes demonstrated mosaic-like structures of the
genes and strongly suggested HGT of fragments of 16S rRNA genes [49].
Sequencing of the 23S rRNA genes from Bradyrhizobium spp (isolated from
neotropical legumes) also revealed mosaic-like structures. Specifically, an 84 bp
region within the 5’ 23S rRNA gene was found to have identical nucleotide
146
sequence to B. japonicum USDA 110, and an adjacent 288 bp sequence was found
to be identical to B. elkanii USDA 76, strongly suggesting HGT of these fragments
[50]. Similar data have also been reported in Streptomyces spp. [51] and a large
number of actinomycete species [52]. Furthermore, Asai et al. [53] reported that
inactivation of all seven E. coli chromosomal rRNA operons followed by subsequent
insertion of foreign rRNA operons derived from Salmonella typhimurium or Proteus
vulgaris had no effect on microbial fitness, and provides evidence that it is possible
to exchange entire rRNA genes between bacteria. Therefore, it is possible that our
sequence data represent the first evidence of mosaic-like rRNA structures in
ureaplasmas, although further experiments are required to confirm these results.
It is striking that the sequence variability observed in domain V of the 23S rRNA
gene was found only in chorioamnion ureaplasma isolates, and not those isolated
from the AF. Whilst the 442S inoculum strain used for intra-amniotic injection was
not 100% clonal (that is, not originating from a single CFU), we did not detect any
nucleotide polymorphisms in 23S rRNA genes within this inoculum. Therefore, these
changes may have been present within a very small sub-population of ureaplasmas
within the 442S inoculum that were undetectable by our Sanger sequencing
approach, or alternatively, the sequence variability has occurred in vivo by
acquisition of foreign DNA. Regardless, ureaplasmas containing identical, non-
random 23S rRNA gene mutations were selected for within the chorioamnion of
each of the infected ewes, potentially due to differences in selective pressures in the
microenvironment between the fetal membranes and the AF. Changes in the
bacterial environment can alter the socio-microbiological structure of the bacterial
population, driving minor subpopulations with mutant genotypes/phenotypes to
thrive [54]. This is particularly relevant for mycoplasmas and ureaplasmas, as these
microorganisms have a limited number of genes devoted to DNA repair and as
such, are associated with increased mutation rates [55]. A study of the complete
147
transcriptomes of U. urealyticum, M. genitalium and M. pneumoniae found that
selection at the codon level for these organisms is most likely to be driven by
environmental stimuli rather than phylogenetic relationships [56]. Therefore, it is
possible that the differences in microenvironment between the AF and the
chorioamnion may have driven and/or selected for sequence variability in domain V
of the 23S rRNA gene in chorioamnion ureaplasma isolates, although we are
currently unable to speculate on the mechanisms involved.
BLAST analysis of our variable sequence regions revealed sequence homology to
Pseudomonas spp., suggesting a potential source of the donor DNA. Real-time
PCR assays were performed to detect P. aeruginosa, P. fluorescens, P. putida and
P. stutzeri within chorioamnion samples; however, these assays failed to detect
pseudomonas DNA (data not shown). Additionally, there was no evidence of co-
infection within the AF or fetuses. If the animals in our experiment were exposed to
pseudomonas earlier in gestation, it is possible that lysed pseudomonas DNA or
extracellular DNA (associated with biofilm forming-bacteria) could be the source of
the foreign 23S rRNA fragment found within the chorioamnion ureaplasma isolates;
however, we believe that it is unlikely that all animals within the cohort would have
been exposed to pseudomonas. It is more likely that the acquisition of this foreign
DNA fragment occurred in the inoculum strain (442S) prior to our sheep experiment,
and was present only in a small sub-set of ureaplasmas within the population, which
were undetectable by conventional Sanger sequencing (data not shown). Whilst we
are unable to determine the mechanism by which these mutations occurred, the fact
that these mutants were selected for within the chorioamnion suggests that the
integration of this foreign DNA may confer a selective advantage at this anatomical
site.
Our study also involved the detection of an erythromycin ribosome methylase
(erm(B)) gene, which induces macrolide resistance by methylation of nucleotide
148
2058 of domain V of the 23S rRNA gene (E. coli numbering). Mono/dimethylation at
this position decreases macrolide binding activity by altering the structure of the site,
disrupting hydrogen bonding and creating steric hinderance [57, 58]. We detected
erm(B) in all tested chorioamnion ureaplasma isolates and the 442S inoculum
strain, although not in AF ureaplasma isolates. Low-level exposure to antimicrobials
is thought to induce erm(B) methylase activity; however, the gene was detected in
animals from both the Up animal group and the Up/E group, therefore was not
associated with erythromycin exposure. The presence of erm(B) could not be
attributed to increased resistance to roxithromycin in chorioamnion ureaplasma
isolates within our study, as the gene was detected in all chorioamnion isolates and
across a wide range of MICs. Similar data were reported by Lu et al. [22], in which
erm(B) detection in ureaplasmas was associated with resistance to erythromycin
and a very wide MIC range for clindamycin, azithromycin and josamycin.
In addition to erm(B), we also detected macrolide-streptogramin resistance (msr)
gene subtypes within ureaplasmas. Msr genes encode drug efflux pumps, which are
members of the ATP-binding cassette family of transporters [59]. These genes are
commonly associated with resistance to macrolide, lincosamide, streptogramin,
ketolide and oxazolidinone antimicrobials [60]. There are many msr gene subtypes;
however, our study focused on msr(A), (B), (C) and (D) as these genes have been
detected previously in ureaplasmas [22]. Unlike the work published previously by Lu
et al. [22], we did not detect msr(A) or msr(B) in any isolates. We obtained positive
PCR results for msr(C) and msr(D); however, msr(C) did not amplify at the
anticipated molecular weight, which suggests that there may be variability in
ureaplasmas isolated from different geographical locations. Whilst msr(D) did
produce an amplicon of the correct size; its presence was not associated with
macrolide resistance, nor was it induced by erythromycin exposure as again, the
gene was present in animals from both experimental groups. These findings are
149
similar to those reported by Lu et al. [22] in which msr(D) was associated with very
wide MIC ranges (0.125 - >128 mg/L) for the tested macrolides. Whilst our study is
the second to detect msr genes in ureaplasmas, we are still unable to link the
presence of these genes directly to macrolide resistance in these microorganisms
due to the variability in MIC ranges reported. Furthermore, although detected in
100% of chorioamnion ureaplasma isolates, msr(C) was not detected in the 442S
inoculum strain, which suggests that it may have been present within a small (un-
detectable) sub-population of ureaplasmas within the inoculum strain, or acquired in
utero. As we were unable to confirm the identiy of these resistance genes by
sequencing, further investigation into the identity and role of these genes in
ureaplasmas is required and these findings should be interpreted as preliminary
data.
In conclusion, injection of a single U. parvum serovar 3 clinical isolate into the AF of
pregnant sheep can generate isolates with variable MIC ranges against macrolide
antimicrobials. After extensive investigation into molecular mechanisms underlying
macrolide resistance, we found that sequence variability in domain V of the 23S
rRNA gene and detection of erm(B) and msr genes occurred independently of
erythromycin exposure. We did not detect any genetic mechanisms, which could
explain potential low-level roxithromycin resistance in chorioamnion ureaplasmas,
therefore further analysis of these strains is required. The numerous phenotypic and
genotypic changes observed in chorioamnion ureaplasma isolates compared to AF
ureaplasma isolates suggests that the anatomical site of infection and the
associated microenvironment may exert selective pressures on ureaplasmas that
result in the selection of subpopulations of mutants. Despite being unable to
demonstrate associations between sub-inhibitory levels of erythromycin and
induced macrolide resistance in ureaplasmas, our data may be the first report of
mosaic-like 23S rRNA gene sequences within ureaplasmas. Further research is
150
required to confirm these findings in amniotic fluid and chorioamnion samples
collected from larger animal studies and from pregnant women. These findings may
have significant implications for future ureaplasma research and challenge our
current understanding of these microorganisms and their so-called ‘simplistic’
minimalist genomes.
151
Acknowledgements
The authors would like to acknowledge Professor John Glass and Ms Vanya
Paralanov (J. Craig Venter Institute) for their independent analysis and advice
regarding the interpretation of 23S rRNA sequence data. We would also like to
thank JRL Hall & Co., in particular Sara Ritchie and Fiona Hall, who have been
responsible for breeding and supplying us with the high quality research animals
necessary for this project. We also acknowledge: Drs John and Janet Allan at
Wesley Monash IVF for the research that has provided low passage clinical
ureaplasma isolates, Mrs Sue Gill (Queensland University of Technology) for
providing us with Pseudomonas spp. for use as positive controls in PCR assays,
and Professor Peter Timms (Queensland University of Technology) for critically
reviewing drafts of this manuscript.
152
References
1. Shepard MC, Masover GK (1979) Special features of ureaplasmas. In: Barile
MF and Razin S, editors. The Mycoplasmas, vol. 1. New York: Academic
Press. pp. 452-494.
2. Volgmann, T, Ohlinger R, Panzig B (2005) Ureaplasma urealyticum-
harmless commensal or underestimated enemy of human reproduction? A
review. Arch Gynecol Obstet 273: 133-139.
3. Gerber S, Vial Y, Hohlfeld P, Witkin SS (2003). Detection of Ureaplasma
urealyticum in second-trimester amniotic fluid by polymerase chain reaction
correlates with subsequent preterm labor and delivery. J Infect Dis 187: 518-
521.
4. Perni SC, Vardhana S, Korneeva I, Tuttle SL, Paraskevas LR, Chasen ST,
Kalish RB, Witkin SS (2004) Mycoplasma hominis and Ureaplasma
urealyticum in midtrimester amniotic fluid: association with amniotic fluid
cytokine levels and pregnancy outcome. Am J Obstet Gynecol 191: 1382-
1386.
5. Yoon BH, Chang JW, Romero R (1998) Isolation of Ureaplasma urealyticum
from the amniotic cavity and adverse outcome in preterm labor. Obstet
Gynecol 92: 77-82.
6. Kundsin RB, Driscoll SG, Monson RR, Yeh C, Biano SA, Cochran WD
(1984) Association of Ureaplasma urealyticum in the placenta with perinatal
morbidity and mortality. N Engl J Med 310: 941-945.
7. Schelonka RL, Waites KB (2007) Ureaplasma infection and neonatal lung
disease. Semin Perinatol 31: 2-9.
8. Abele-Horn M, Wolff C, Dressel P, Pfaff F, Zimmermann A (1997)
Association of Ureaplasma urealyticum biovars with clinical outcome for
neonates, obstetric patients, and gynecological patients with pelvic
inflammatory disease. J Clin Microbiol 35: 1199-1202.
9. Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R (1993)
Ureaplasma urealyticum intrauterine infection: role in prematurity and
disease in newborns. Clin Microbiol Rev 6: 69-87.
10. Glass JI, Lefkowitz EJ, Glass JS, Heiner CR, Chen EY, Cassell GH (2000)
The complete sequence of the mucosal pathogen Ureaplasma urealyticum.
Nature 12: 757-762.
11. Gomez R, Romero R, Nien JK, Medina L, Carstens M, Kim YM, Espinoza J,
Chaiworapongsa T, Gonzalez R, Iams JD, Rojas I (2007) Antibiotic
153
administration to patients with preterm premature rupture of membranes
does not eradicate intra-amniotic infection. Matern Fetal Neonatal Med 20:
167-173.
12. Kenyon SL, Taylor DJ, Tarnow-Mordi W (2001) Broad-spectrum antibiotics
for spontaneous preterm labour: the ORACLE II randomised trial. Lancet 31:
989-994.
13. McDonald H, Brocklehurst P, Parsons J (2005) Antibiotics for treating
bacterial vaginosis in pregnancy. Cochrane Database Syst Rev CD000262.
14. Heikkinen T, Laine L, Neuvonen PJ, Ekblad U (2000) The transplacental
transfer of the macrolide antibiotics erythromycin, roxithromycin and
azithromycin. BJOG 107: 770-775.
15. Dando SJ, Nitsos I, Newnham JP, Jobe AH, Moss TJ, Knox CL (2010)
Maternal administration of erythromycin fails to eradicate intrauterine
ureaplasma infection in an ovine model. Biol Reprod 83: 616-622.
16. Gaynor M, Mankin AS (2005) Macrolide antibiotics: Binding site, mechanism
of action, resistance. Curr Top Med Chem 3: 949-961.
17. Robertson JA, Coppola JE, Heisler OR (1981) Standardized method for
determining antimicrobial susceptibility of strains of Ureaplasma urealyticum
and their response to tetracycline, erythromycin, and rosaramicin. Antimicrob
Agents Chemother 20: 53-58.
18. Krausse R, Schubert S (2010) In vitro activities of tetracyclines, macrolides,
fluoroquinolones and clindamycin against Mycoplasma hominis and
Ureaplasma spp. isolated in Germany over 20 years. Clin Microbiol Infect
16: 1649-1655.
19. Beeton ML, Chalker VL, Maxwell NC, Kotecha S, Spiller OB (2009)
Concurrent titration and determination of antibiotic resistance in Ureaplasma
species with identification of novel point mutations in genes associated with
resistance. Antimicrob Agents Chemother 53: 2020-2027.
20. Dongya M, Wencheng X, Xiaobo M, Lu W (2008) Transition mutations in
23S rRNA account for acquired resistance to macrolides in Ureaplasma
urealyticum. Microb Drug Resist 14: 183-186.
21. Pereyre S, Métifiot M, Cazanave C, Renaudin H, Charron A, Bébéar C,
Bébéar CM (2007) Characterisation of in vitro-selected mutants of
Ureaplasma parvum resistant to macrolides and related antibiotics. Int J
Antimicrob Agents 29: 207-211.
154
22. Lu C, Ye TI, Zhu Gx, Feng Py, Ma H, Lu Rb, Lai W (2010) Phenotypic and
genetic characteristics of macrolide and lincosamide resistant Ureaplasma
urealyticum isolated in Guangzhou, China. Curr Microbio 61: 44-49.
23. Anwar H, Costerton JW (1990) Enhanced activity of combination of
tobramycin and piperacillin for eradication of sessile biofilm cells of
Pseudomonas aeruginosa. Antimicrob Agents Chemother 34: 1666-1671.
24. Donlan RM, Costerton JW (2002) Biofilms: survival mechanisms of clinically
relevant microorganisms. Clin Microbiol Rev 15: 167-193.
25. Moskowitz SM, Foster JM, Emerson J, Burns JL (2004) Clinically feasible
biofilm susceptibility assay for isolates of Pseudomonas aeruginosa from
patients with cystic fibrosis. J Clin Microbiol 42: 1915-1922.
26. McAuliffe L, Ayling RD, Ellis RJ, Nicholas RAJ (2008) Biofilm-grown
Mycoplasma mycoides subsp. mycoides SC exhibit both phenotypic and
genotypic variation compared with planktonic cells. Vet Microbiol 129: 315-
324.
27. McAuliffe L, Ellis RJ, Miles K, Ayling RD, Nicholas RAJ (2006) Biofilm
formation by Mycoplasma species and its role in environmental persistence
and survival. Microbiology 152: 913-922.
28. Simmons WL, Dybvig K (2009) Mycoplasma biofilms ex vivo and in vivo.
FEMS Microbiol Lett 295: 77-81.
29. Knox CL, Dando SJ, Nitsos I, Kallapur SG, Jobe AH, Payton D, Moss TJ,
Newnham JP (2010) The severity of chorioamnionitis in pregnant sheep is
associated with in vivo variation of the surface-exposed multiple-banded
antigen/gene of Ureaplasma parvum. Biol Reprod 83: 415-426.
30. Moss TJ, Knox CL, Kallapur SG, Nitsos I, Theodoropoulos C, Newnham JP,
Ikegami M, Jobe AH (2008) Experimental amniotic fluid infection in sheep:
effects of Ureaplasma parvum serovars 3 and 6 on preterm or term fetal
sheep. Am J Obstet Gynecol 198: 122e1-e8.
31. Moss TJ, Nitsos I, Knox CL, Polglase GR, Kallapur SG, Ikegami M, Jobe
AH, Newnham JP (2009) Ureaplasma colonization of amniotic fluid and
efficacy of antenatal corticosteroids for preterm lung maturation in sheep.
Am J Obstet Gynecol 200: 96e1-e6.
32. Moss TJ, Nitsos I, Ikegami M, Jobe AH, and Newnham JP (2005)
Experimental intrauterine ureaplasma infection in sheep. Am J Obstet
Gynecol 192: 1179-1186.
155
33. Tully JG (1983) Cloning and filtration techniques for mycoplasmas. In: Razin
S, Tully JG, editors. Methods in Mycoplasmology. New York: Academic
Press pp. 173-177.
34. Waites KB, Bébéar CM, Robertson JA, Talkington DF, Kenny GE (2001) In:
Nolte FS, editor. Cumitech 34, Laboratory diagnosis of mycoplasmal
infections. Washington DC: American Society for Microbiology.
35. Waites KB, Crabb DM, Duffy LB (2003) In vitro activities of ABT-773 and
other antimicrobials against human ureaplasmas. Antimcrob Agents
Chemother 47: 39-42.
36. García-Castillo M, Morosini M, Gálvez M, Baquero F, del Campo R,
Meseguer M (2008) Differences in biofilm development and antibiotic
susceptibility among clinical Ureaplasma urealyticum and Ureaplasma
parvum isolates. J Antimicrob Chemother 62: 1027-1030.
37. Blanchard A, Gautier M, Mayau V (1991) Detection and identification of
mycoplasmas by amplification of rDNA. FEMS Microbiol Lett 65: 37-42.
38. Daly MM, Doktor S, Flamm R, Shortridge D (2004) Characterization and
prevalence of mef(A), mef(E), and the associated msr(D) gene in
Streptococcus pneumoniae clinical isolates. J Clin Microbiol 42: 3570-3574.
39. Graham JP, Price LB, Evans SL, Graczyk TK, Silbergeld EK (2009)
Antibiotic resistant enterococci and staphylococci isolated from flies
collected near confined poultry feeding operations. Sci Total Environ 407:
2701-2710.
40. Lina G, Quaglia A, Reverdy ME, Leclercq R, Vandenesch F, Etienne J
(1999) Distribution of genes encoding resistance to macrolides,
lincosamides, and streptogramins among staphylococci. Antimicrob Agents
Chemother 43: 1062-1066.
41. Pereyre S, Guyot C, Renaudin H, Charron A, Bébéar C, Bébéar CM (2004)
In vitro selection and characterization of resistance to macrolides and related
antibiotics in Mycoplasma pneumoniae. Antimicrob Agents Chemother 48:
460-465.
42. Sheldon AT Jr (2005) Antiseptic “resistance”: real or perceived threat? Clin
Infect Dis 40: 1650-1656.
43. Zhanel GG (2005) Antibacterial drivers of resistance. Treat Respir Med 4:
13-18.
44. Høiby N, Bjarnsholt T, Givskov M, Molin S, Ciofu O (2010) Antibiotic
resistance and bacterial biofilms. J Int Antimicrob Agents 35: 322-332.
156
45. Drenkard E, Ausubel FM (2002) Pseudomonas biofilm formation and
antibiotic resistance are linked to phenotypic variation. Nature 416: 740-743.
46. Del Pozo JL, Patel R (2007) The challenge of treating biofilm-associated
bacterial infections. Clin Pharmacol Ther 82: 204-209.
47. Adams JL, McLean RJ (1999) Impact of rpoS deletion on Escherichia coli
biofilms. Appl Environ Microbiol 65: 4285-4287.
48. Beeton ML, Chalker VJ, Kotecha S, Spiller OB (2009) Comparison of full
gyrA, gyrB, parC and parE gene sequences between all Ureaplasma parvum
and Ureaplasma urealyticum serovars to separate true fluoroquinolone
antibiotic resistance mutations from non-resistance mutations. J Antimicrob
Chemother 64: 529-538.
49. Schouls LM, Schot CS, Jacobs JA (2003) Horizontal transfer of segments of
the 16S rRNA genes between species of the Streptococcus anginosus
group. J Bacteriol 185: 7241-7246.
50. Parker MA (2001) Case of localized recombination in 23S rRNA genes from
divergent Bradyrhizobium lineages associated with neotropical legumes.
Appl Environ Microbiol 67: 2076-2082.
51. Ueda K, Seki T, Kudo T, Yoshida T, Kataoka M (1999) Two distinct
mechanisms cause heterogeneity of 16S rRNA. J Bacteriol 181: 78-82.
52. Asai T, Zaporojets D, Squires C, Squires CL (1999) An Escherichia coli
strain with all chromosomal rRNA operons inactivated: complete exchange
of rRNA genes between bacteria. Proc Natl Acad Sci 96: 1971-1976.
53. Wang Y, Zhang Z (2000) Comparative sequence analyses reveal frequent
occurrence of short segments containing an abnormally high number of non-
random base variations in bacterial rRNA genes. Microbiology 146: 2845-
2854.
54. Aziz RK, Kansal R, Aronow BJ, Taylor WL, Rowe SL, Kubal M, Chhatwal
GS, Walker MJ, Kotb M (2010) Microevolution of Group A Streptococci in
vivo: capturing regulatory networks engaged in sociomicrobiology, niche
adaptation, and hypervirulence. PLoS One 5: e9798.
55. Razin S, Yogev D, Naot Y (1998) Molecular biology and pathogenicity of
mycoplasmas. Microbiol Mol Biol Rev 62: 1094-1156.
56. Fadiel A, Lithwick S, Naftolin F (2005) The influence of environmental
adaptation on bacterial genome structure. Lett Appl Microbiol 40: 12-18.
57. Goldman RC, Scaglione F (2004) The macrolide-bacterium interaction and
its biological basis. Curr Drug Targets Infect Disord 4: 241-260.
157
58. Marvić G (2004) Macrolide resistance based on the Erm-mediated rRNA
methylation. Curr Drug Targets Infect Disord 4: 193-202.
59. Reynolds E, Ross JI, Cove JH (2003) Msr(A) and related
macrolide/streptogramin resistance determinants: incomplete transporters?
Int J Antimicrob Agents 22: 228-236.
60. Roberts MC (2008) Update on macrolide-lincosamide-streptogramin,
ketolide, and oxazolidinone resistance genes. FEMS Microbiol Lett 282: 147-
159.
158
Chapter 5
PAPER 3
159
The Role of the Multiple Banded Antigen of Ureaplasma parvum in Intra-
amniotic Infection: Major Virulence Factor or Decoy?
Samantha J Dando 1, Ilias Nitsos 2, 3#, Suhas G Kallapur 2,4, John P Newnham 2,
Graeme R Polglase 2, 3#, J Jane Pillow 2, Alan H Jobe 2, 4, Peter Timms 1,
Christine L Knox 1
1 Institute of Health & Biomedical Innovation, Faculty of Science and Technology, Queensland University of Technology, Brisbane, Queensland, 4059, Australia.
2 School of Women’s and Infants’ Health, The University of Western Australia,
Perth, Western Australia, 6009, Australia. 3 The Ritchie Centre, Monash Institute of Medical Research, Monash University,
Clayton, Victoria, 3168, Australia. 4 Department of Neonatology and Pulmonary Biology, Cincinnati Children’s Hospital
Medical Center, University of Cincinnati, Cincinnati OH 45229, USA. # Current affiliation
Published in: PLoS One (2012) 7: e29856
160
Statement of Joint Authorship
Samantha J Dando (candidate):
Contributed to the experimental design and research plan; performed all ureaplasma cultures, western blots and PCR assays. Analysed tissue samples for histopathology; performed quantitative real time PCR; analysed and interpreted the data and wrote the manuscript.
Ilias Nitsos:
Contributed to the experimental design and research plan; performed intra-amniotic injections, and assisted in the collection of samples. Contributed to the manuscript.
Suhas G Kallapur:
Contributed to the experimental design and research plan; assisted in the collection of samples and contributed to the manuscript.
Graeme Polglase:
Contributed to the experimental design and research plan; assisted in the collection of samples and contributed to the manuscript.
J. Jane Pillow:
Contributed to the experimental design and research plan; assisted in the collection of samples and contributed to the manuscript.
John P Newnham:
Contributed to the experimental design and research plan; performed intra-amniotic injections and assisted in the collection of samples. Contributed to the manuscript.
Alan H Jobe:
Contributed to the experimental design and research plan; performed fetal post-mortems and assisted in the collection of samples, and made a significant contribution to the manuscript.
Peter Timms:
Assisted in research design and the interpretation of data. Contributed to the manuscript.
Christine L Knox:
Supervised the project, contributed to the experimental design and research plan and assisted in the collection of samples. Assisted in the interpretation of data and made a significant contribution to the manuscript.
161
Abstract
The multiple banded antigen (MBA) is a predicted virulence factor of Ureaplasma
species. Antigenic variation of the MBA is a potential mechanism by which
ureaplasmas avoid immune recognition and cause chronic infections of the upper
genital tract of pregnant women. We tested whether the MBA is involved in the
pathogenesis of intra-amniotic infection and chorioamnionitis by injecting virulent or
avirulent-derived ureaplasma clones (expressing single MBA variants) into the
amniotic fluid of pregnant sheep. At 55 days of gestation pregnant ewes (n=20),
received intra-amniotic injections of virulent-derived or avirulent-derived U. parvum
serovar 6 strains (2x104 CFU), or 10B medium (n=5). Amniotic fluid was collected
every two weeks post-infection and fetal tissues were collected at the time of
surgical delivery of the fetus (140 days of gestation). Whilst chronic colonisation
was established in the amniotic fluid of animals infected with avirulent-derived and
virulent-derived ureaplasmas, the severity of chorioamnionitis and fetal inflammation
was not different between these groups (p>0.05). MBA size variants (32-170 kDa)
were generated in vivo in amniotic fluid samples from both the avirulent and virulent
groups, whereas in vitro antibody selection experiments led to the emergence of
MBA-negative escape variants in both strains. Anti-ureaplasma IgG antibodies were
detected in the maternal serum of animals from the avirulent (40%) and virulent
(55%) groups, and these antibodies correlated with increased IL-1β, IL-6 and IL-8
expression in chorioamnion tissue (p<0.05). We demonstrate that ureaplasmas are
capable of MBA phase variation in vitro; however, ureaplasmas undergo MBA size
variation in vivo, to potentially prevent eradication by the immune response. Size
variation of the MBA did not correlate with the severity of chorioamnionitis.
Nonetheless, the correlation between a maternal humoral response and the
expression of chorioamnion cytokines is a novel finding. This host response may be
162
important in the pathogenesis of inflammation-mediated adverse pregnancy
outcomes.
163
Introduction
The two Ureaplasma species, which cause infections in humans are Ureaplasma
parvum (serovars 1, 3, 6 and 14) and Ureaplasma urealyticum (serovars 2, 4, 5, 7-
13) [1]. Phenotypically the ureaplasmas are distinguished from the closely related
Mycoplasma species by their ability to hydrolyse urea to generate 95% of their ATP
[2, 3]. The ureaplasmas are generally regarded as commensals of the lower genital
tract in both males and females and can be isolated from the vagina or cervix in 40-
80% of sexually active females [4, 5]. However, ureaplasma infection of the upper
genital tract during pregnancy is associated with adverse pregnancy outcomes
including preterm birth and chorioamnionitis [5, 6].
Ureaplasmas are hypothesized to gain access to the upper genital tract of pregnant
women by various mechanisms including (i) ascending invasive infection from the
lower genital tract; (ii) transplacental or haematogenous spread; or (iii) iatrogenic
needle contamination at the time of amniocentesis or chorionic villous sampling [7].
Although ureaplasmas are the bacteria most frequently isolated from infected
amniotic fluid (AF) in pregnant women [8-10], the pathogenic role of these
microorganisms during pregnancy is unclear, as ureaplasmas have also been
isolated from the AF of women with apparently normal pregnancy outcomes after
delivery at term [8, 9, 11]. These discrepancies demonstrate that a causal
relationship has not been established between intra-amniotic ureaplasma infection
and adverse pregnancy outcomes.
Initial serotyping studies of invasive ureaplasmas isolated from CSF and blood
cultures of neonates demonstrated that no one serovar was more associated with
disease, and that invasiveness was not likely to be limited to one particular serotype
[12]. Rather, it was hypothesised that the virulence of individual ureaplasma strains
may be determined by antigenic variation and/or host factors [5, 12]. The multiple
164
banded antigen (MBA) is a surface exposed lipoprotein, which can undergo size
and phase variation in vitro and in vivo [13-17]. The MBA gene (mba) consists of a
5’ conserved region, which encodes a signal peptide and membrane anchor and a
3’ repetitive region, which consists of multiple tandem repeat units [14]. The MBA is
predicted to be a major ureaplasmal virulence factor and is the predominant antigen
recognised by sera during infections in humans [18]. Recently, our group
demonstrated that MBA size variation was associated with the severity of
histological chorioamnionitis in a pregnant sheep model of intra-amniotic
ureaplasma infection [17]. From this previous work, we cultured a clonal U. parvum
serovar 6 virulent-derived strain (associated with severe histological
chorioamnionitis) and a clonal avirulent-derived strain (associated with no signs of
histological chorioamnionitis), which we aimed to characterize further in vivo.
Whilst it is evident that certain ureaplasma isolates are more associated with severe
disease than others, it is not known if the invasive properties associated with these
isolates are determined primarily by bacterial factors (such as size/phase variation
of the MBA) or host factors (including the immune response). Using our established
sheep model of chronic intra-amniotic ureaplasma infection, we tested whether
clonal ureaplasma isolates with defined MBA profiles (derived from virulent or
avirulent parent strains, associated with either severe, or no chorioamnionitis) are
intrinsically virulent or avirulent. For this study we defined virulence as the extent of
damage to the host during infection with a pathogen, as defined by Brown et al. [19].
We hypothesized that virulence is not likely to be associated with specific
ureaplasma isolates, but rather that the severity of disease may be determined by
the host immune response generated against intra-amniotic ureaplasma infection.
We predicted that interactions between ureaplasmas and the host immune
response may be mediated by size/phase variation of the MBA and that variable
expression of the MBA would enable ureaplasmas to avoid eradication by host
165
immune factors. By measuring pregnancy/fetal outcomes, ureaplasma colonization
of fetal tissues, MBA in vivo expression profiles and the host immune response we
aimed to provide insight into the role of the MBA during microbial invasion of the
amniotic cavity and the chorioamnion.
166
Materials and Methods
Ethics Statement
This study was carried out in accordance with the NHMRC ‘Australian code of
practice for the care and use of animals for scientific purposes’ and approved by the
UWA Animal Ethics Committee (Approval No. RA/ 3/100/619). Preterm lambs were
surgically delivered by Caesarean section. Ewes were pre-medicated with an intra-
venous injection of ketamine (10 mg/kg bodyweight) and medetomidine (0.02
mg/kg) and a subdural injection of 2% lignocaine (60 mg). The fetus was delivered
and then euthanized using sodium pentobarbitone, 100 mg/kg. The ewe was killed
by sodium pentobarbitone, 100 mg/kg.
Source of ureaplasma isolates
The U. parvum serovar 6 strains used for intra-amniotic injection were expanded
from single colony forming units (CFUs) derived from ‘virulent’ and ‘avirulent’
ureaplasmas isolated from the AF of pregnant sheep from our previous experiment
[17]. The parent strains (E24 and E22) were classified as virulent and avirulent
respectively, based on the severity of chorioamnionitis associated with intra-
amniotic infection and the number of MBA variants detected within the AF at the
time of preterm delivery. Infection with isolate E24 was associated with severe
chorioamnionitis (resulting in fibrosis and tissue lesions) and a low number of AF
MBA size variants (n = 5); whereas infection with isolate E22 did not result in
histological chorioamnionitis and a greater number of MBA size variants were
detected within the AF (n = 14, data presented in reference [17]). For this current
experiment, to obtain ureaplasmas expressing single MBA size variants derived
from single CFUs, E22 and E24 cultures were cloned and filtered three times, as
described previously [20]. Cloned ureaplasmas were designated E22 5.8.1
167
(originating from the avirulent parent strain) and E24 3.2.1 (originating from the
virulent parent strain).
Animal model
At 55 days of gestation (d, term = 150 d) 25 pregnant Merino ewes were
randomized to receive ultrasound-guided intra-amniotic injections (virulent-derived
clone E24 3.2.1: n = 10; avirulent-derived clone E22 5.8.1: n = 10; 10B medium
control: n = 5). Prior to the injection of ureaplasmas, AF was aspirated to verify: (i)
the injection site by electrolyte analysis (Rapidlab 865, Bayer Diagnostics, Pymble,
New South Wales); and (ii) to test that animals did not have pre-existing intra-
amniotic ureaplasma infections. Ureaplasma isolates and 10B medium were
prepared for injection in 2 mL volumes, which consisted of 2x104 CFU or 10B
medium diluted in PBS [21]. AF was sampled from each animal by ultrasound-
guided amniocentesis approximately every two weeks (at 73, 87, 101, 115 and 126
days of gestation) post intra-amniotic injection and was tested by culture and
western blot.
Near-term fetuses were delivered surgically at 140 d [17, 22] and samples of AF,
chorioamnion, cord, fetal lung and fetal CSF were aseptically collected. Complete
blood counts were performed on umbilical arterial blood.
Ureaplasma culture
Ureaplasmas were cultured from AF, chorioamnion, cord, fetal lung and fetal CSF
specimens in 10B broth medium [17, 20, 22]. Ureaplasmas within specimens were
quantified by standard drop plate analysis, and are reported as the number of
CFU/mL of fluid, or CFU/gram of tissue.
mba PCR
168
To determine the size and number of mba variants within clonal ureaplasma isolates
E22 5.8.1 and E24 3.2.1, the downstream repeat region of the mba was amplified
by PCR, as described previously [17].
MBA western blots
Ureaplasma MBA size variants were detected in un-cultured AF specimens
collected at 73 d, 87 d, 101 d, 115 d, 126 d and 140 d by western blot. Western
blots were performed directly on centrifuged AF rather than ureaplasmas cultured
from AF, as culture can select for sub-populations of MBA variants, and therefore
results are not representative of the pool of size variants originally present within the
AF (data not shown). 10 mL of thawed AF was centrifuged at 5250 x g for 20
minutes at 4 °C. The supernatant was discarded and the pellet then was
resuspended in 100 µL of PBS. SDS-PAGE and western blots were performed as
described previously [17]. The primary antibody used for detection of the MBA was
rabbit polyclonal antisera raised against U. parvum serovar 6 (courtesy of Emeritus
Dr Patricia Quinn) diluted 1/5000 in blocking solution (5% skim milk, 150 mM NaCl,
50 mM Tris). This antibody has been previously shown to be reactive only against
the MBA [17]. Membranes were further probed with a goat anti-rabbit IgG-HRP
secondary antibody (Sigma Aldrich, Castle Hill, New South Wales) diluted 1/5000;
and MBA protein bands were detected by 3’, 3’-diaminobenzidine tetrahydrochloride
(DAB) staining with cobalt chloride (Sigma). Proteins extracted from ureaplasma
isolates E22 5.8.1 and E24 3.2.1 were included in each western blot as positive
controls.
Histopathology
Formalin-fixed paraffin-embedded tissues were cut into 5 µm sections and stained
with haematoxylin and eosin (H & E, all tissue types) and Masson’s trichrome stain
(chorioamnion samples only). Sectioning and staining was performed by QML
169
Vetnostics (Murarrie, Queensland). Inflammatory cell counts were performed on H &
E stained sections to determine the numbers of monocytes, macrophages,
lymphocytes, band neutrophils and polymorphonuclear neutrophils (PMNs) in 20
fields of view at x 1000 total magnification. Masson’s trichrome-stained
chorioamnion sections were graded using our previously described scoring system
[22] to determine the severity of histological chorioamnionitis. Histopathological
analysis of H & E and Masson’s trichrome stained tissues was performed blinded to
animal treatment groups.
RNA extraction and RT-PCR
Total RNA was extracted from 20 µg of freshly thawed chorioamnion and fetal lung
samples collected at 140 d using the Qiagen RNeasy Mini Kit (Qiagen, Doncaster,
Victoria). The quantity of eluted RNA was measured using the NanoDrop 1000
spectrophotometer (Thermo Fischer Scientific Australia Pty Ltd, Scoresby, Victoria)
and 1.5 µg of cDNA was synthesised from RNA template by RT-PCR using
Invitrogen’s Super Script III First-Strand Synthesis Supermix for q-PCR (Invitrogen,
Mulgrave, Victoria). RT-PCR conditions consisted of: 25 ˚C for 10 minutes, 50 ˚C for
30 minutes, and 85 ˚C for 5 minutes. cDNA samples were then chilled on ice, and
contaminating RNA was degraded by the addition of RNase H followed by
incubation at 37 ˚C for 20 minutes. cDNA was stored at -20 ˚C until use.
Quantitative Real time PCR
Quantitative real time PCR was performed on chorioamnion and fetal lung cDNA
samples to determine the expression of: Toll-like receptor (TLR) 1, TLR2, TLR6,
interleukin (IL)-1β, IL-6, IL-8, IL-10 and tumor necrosis factor-α (TNF-α), relative to
the expression of GAPDH. Previously published PCR primers [23-25] (Table 5.1)
were used to amplify sheep-specific sequences. PCR assays incorporated 1x
Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen); either 0.4 µM of each
170
PRIMER SEQUENCE (5’ – 3’) ANNEALING TEMPERATURE
SOURCE
GAPDH (F) GAPDH (R)
GTCCGTTGTGGATCTGACCT TGCTGTAGCCGAATTCATTG
58 °C
Chang et al. 2009
[25]
TLR1 (F) TLR1 (R)
TTGCACATCAGCAAGGTTTT CACTGTGGTGCTGACTGACA
58 °C
Chang et al. 2009
[25]
TLR2 (F) TLR2 (R)
GGCTGTAATCAGCGTGTTCA GATCTCGTTGTCGGACAGGT
58 °C
Chang et al. 2009
[25]
TLR6 (F) TLR6 (R)
TTTGTCCTCAGGAACCAAGC TCATATTCCAAAGAATTCCAGCTA
58 °C
Chang et al. 2009
[25]
IL-1β (F) IL-1β (R)
CCTTGGGTATCAGGGACAA TGCGTATGGCTTTCTTTAGG
57 °C
McNeilly et al. 2008
[24]
IL-6 (F) IL-6 (R)
TCCAGAACGAGTTTGAGG CATCCGAATAGCTCTCAG
52 °C
Egan et al. 1996
[26]
IL-8 (F) IL-8 (R)
ATCAGTACAGAACTTCGA TCATGGATCTTGCTTCTC
55 °C
McNeilly et al. 2008
[24]
IL-10 (F) IL-10 (R)
TGAAGGACCAACTGAACAGC TTCACGTGCTCCTTGATGTC
55 °C
Egan et al. 1996
[26]
TNF-α (F) TNF-α (R)
GAATACCTGGACTATGCCGA CCTCACTTCCCTACATCCCT
58 °C
McNeilly et al. 2008
[24]
Table 5.1: PCR primers
These primers were used for the amplification of selected Toll like receptor (TLR) and cytokine genes from chorioamnion and fetal lung cDNA. IL = interleukin, TNF = tumor necrosis factor, F = forward primer, R = reverse primer.
171
primer (GAPDH, TLR1, TLR2 and TLR6) or 0.5 µM of each primer (IL-1β IL-6, IL-8,
IL-10, TNF-α); 25 ng of cDNA and sterile distilled H2O to a final volume of 20 µL.
Real time PCR cycling was performed in a Qiagen Rotor-Gene Q thermocycler
(Qiagen, Doncaster, Victoria), and included initial incubation at 50 ˚C for 10 minutes
and an initial denaturation step at 95 ˚C for 10 minutes. Cycling then consisted of 40
cycles of: denaturation at 94 ˚C for 15 seconds (GAPDH, TLR1, TLR2 and TLR6) or
20 seconds (IL-1β IL-6, IL-8, IL-10 and TNF-α); annealing at 52 ˚C to 58 ˚C (specific
annealing temperatures for each primer pair are shown in Table 5.1); and extension
at 72 ˚C for 20 seconds (IL-1β IL-6, IL-8, IL-10 and TNF-α) or 40 seconds (GAPDH,
TLR1, TLR2 and TLR6). Cycle thresholds (CT) were calculated using Rotor-Gene Q
series software version 1.7 (Qiagen) and PCR product specificity was confirmed by
standard melt curve analysis. Prior to the testing of chorioamnion and fetal lung
cDNA samples, the amplification efficiency of each of the primer pairs was validated
(data not shown), as per the protocol published by Schmittgen and Livak [26]. All
primer efficiencies were found to be within +/- 10% of the efficiency of the reference
gene (GAPDH, 95% efficiency).
All chorioamnion and fetal lung samples were tested in triplicate and mean CT
values from animals infected with virulent-derived (E24 3.2.1) and avirulent-derived
(E22 5.8.1) ureaplasmas were used to calculate the expression of target genes
relative to GAPDH and normalised against the expression of target genes in non-
infected control animals. Relative expression was determined by the equation
published by Pfaffl [27], where the relative expression ratio = (Efficiencytarget
gene)ΔCT(control – sample) / (Efficiencyreference gene)
ΔCT(control – sample). This method was selected
as opposed to the ΔΔCT method of relative expression, as the equation presented
by Pfaffl [27] does not assume that PCR efficiencies are equal between the
reference and target genes.
Detection of anti-ureaplasma IgG antibodies
172
Western blots were performed to determine if anti-ureaplasma IgG antibodies were
present in maternal and/or fetal sera, as indicators of a humoral immune response.
Standardised protein extracts of whole ureaplasma isolates E22 5.8.1 and E24 3.2.1
were loaded into the wells of 10% SDS-PAGE gels, electrophoresed, transferred
onto nitrocellulose membrane and blocked as described above. Maternal or fetal
serum (diluted 1/100 in blocking solution) was used as the primary antibody, and
membranes were incubated with these sera overnight at 4 °C. Serum from the
avirulent group was probed against the E22 5.8.1 whole ureaplasma protein extract,
whereas serum from the virulent group was probed against the E24 3.2.1 whole
ureaplasma protein extract. Membranes were washed and then probed with
secondary antibody (anti-sheep IgG (whole molecule)-HRP, raised in donkey
(Sigma)) diluted 1/1000. The presence of protein bands (detected by DAB staining
with cobalt chloride (Sigma)) indicated binding of antibodies within the serum to
ureaplasmal proteins from either the virulent-derived or avirulent-derived
ureaplasma strain. All samples were tested in duplicate.
Serial passage of virulent and avirulent-derived clonal ureaplasmas
Serial passage experiments were performed by inoculating 2x104 CFU/mL of
ureaplasma strains E22 5.8.1 and E24 3.2.1 into 1.8 mL of 10B medium containing
rabbit polyclonal U. parvum serovar 6 antiserum (at a final dilution of 1/500).
Inoculated broths were incubated at 37 °C aerobically until a colour change was
evident within the media (usually occurring between 12 and 18 hours). Ureaplasmas
were then transferred into fresh 10B medium (at a concentration of approximately
104 to 105 colour changing units per mL) containing antibodies, and were again
incubated. Each ureaplasma isolate was serially passaged 20 times in culture
media containing antibodies. After each passage, samples were collected for
western blot. As a control, ureaplasma isolates E22 5.8.1 and E24 3.2.1 were also
serially transferred in 10B medium without antibodies 20 times.
173
Statistical analysis
Pregnancy outcomes, complete blood count data, inflammatory cell counts and
ureaplasma tissue colonization data were initially analysed for homogeneity of
variances by Levene’s test. Those data for which homogeneity of variances were
confirmed were subsequently analysed by one-way analysis of variance (ANOVA)
with a Tukey post hoc test. If the assumption of homogeneity of variance was
violated, the Welch statistic was alternatively reported. Fetal lung compliance, AF
colonization and MBA size variant data were analysed by a two-way repeated
measures ANOVA, and degrees of freedom were corrected using Greenhouse-
Geisser estimates if the assumption of sphericity was violated. Independent t-tests
were used to analyse humoral immune response and qRT-PCR data. Data are
presented as mean ± standard error of the mean (SEM) and statistical significance
was accepted at p < 0.05.
174
Results
Pregnancy outcomes
Of the 25 pregnant ewes that received intra-amniotic injections, fetuses were
spontaneously aborted from three ewes (virulent group n = 2; avirulent group n = 1)
at approximately 82 d, 115 d and 131 d (Table 5.2). One ewe (avirulent group) also
delivered a stillborn fetus at 131 d after preterm labor. Oligohydramnios was
observed at least once during the amniocentesis sampling period in three animals
from the avirulent group, in two animals from the virulent group, but not in the
control group. Meconium-stained AF was present in four animals from the virulent
group at least once throughout the sampling period; however, meconium was not
present in AF from animals from the avirulent group or the media control group.
Pregnancy loss and oligohydramnios occurred independently of the animal group (p
= 0.59 and p = 0.42 respectively). The presence of meconium-stained AF was
significantly increased in the virulent group compared to the avirulent group and the
media control group (p = 0.01).
Fetal birth weight, fetal lung weight and umbilical arterial cord blood pH, pO2 and
white blood cell counts were not different between animal groups (p > 0.05, Table
5.2). Chronic intra-amniotic infection with the avirulent-derived (E22 5.8.1)
ureaplasma strain tended to increase lung compliance in near-term fetuses (as
determined by a deflation pressure-volume curve, Table 5.2), when compared to the
virulent and control groups. However this observed increase was not statistically
significant (p = 0.06).
Ureaplasmas can chronically colonise the amniotic fluid
AF collected from all ewes prior to intra-amniotic injection at 55 d tested negative for
ureaplasmas. Following intra-amniotic injection, the AF from all ewes inoculated
with ureaplasmas (either isolate E22 5.8.1 or isolate E24 3.2.1) tested positive for
175
Table 5.2: Pregnancy outcomes and fetal measurements at the time of
delivery (140 d)
VIRULENT
GROUP (n = 10)
AVIRULENT GROUP (n = 10)
CONTROL GROUP (n = 5)
P VALUE
Pregnancy outcomes
Abortion/stillborn fetus
2 (20%)
2 (20%)
0 (0%)
0.59
Oligohydramnios
2 (20%)
3 (30%)
0 (0%)
0.42
Meconium-stained amniotic fluid
4 (40%)
0 (0%)
0 (0%)
0.01
Gender (female : male)
5:3
5:3
3:2
0.10
Fetal birth weight (kg)
4.8 ± 0.3
5.3 ± 0.2
5.3 ± 0.3
0.36
Fetal lung weight (g/kg body weight)
28.3 ± 1.4
31.4 ± 1.9
30.4 ± 2.8
0.50
Lung volume (mL/kg) at 40 cm H2O pressure
36.2 ± 4.0
42.1 ± 1.7
37.7 ± 3.6
0.06
Umbilical arterial cord blood gases
pH
7.2 ± 0.03 7.21 ± 0.03 7.19 ± 0.05 0.09
pO2 (mmHg) 9.6 ± 0.6 9.9 ± 0.9 12.8 ± 1.6 0.08
Umbilical arterial white blood cell counts
Total (x10
9/L)
4.2 ± 0.9 3.9 ± 0.5 3.6 ± 0.6 0.87
Monocytes (x10
9/L)
0.3 ± 0.1 0.1 ± 0.04 0.2 ± 0.1 0.54
Lymphocytes (x10
9/L)
2.2 ± 0.3 2.0 ± 0.2 2.1 ± 0.4 0.86
Neutrophils (x10
9/L)
1.2 ± 0.6 0.9 ± 0.1 0.8 ± 0.2 0.75
176
ureaplasmas at all time points (Figure 5.1A). The peak of infection occurred at 87 d
for the avirulent group (7.2 ± 3.1 x 107 CFU/mL) and at 101 d for the virulent group
(5.3 ± 2.1 x 107 CFU/mL). Ureaplasma AF colonization at 87 d was significantly
increased in the avirulent group compared to the virulent group (p = 0.002);
however, there were no differences in colonization at any other time points between
these groups. High numbers of ureaplasmas were recovered from the AF in both
the avirulent (9.6 ± 6.4 x 106 CFU/mL) and virulent (1.6 ± 0.8 x 107 CFU/mL) groups
at 140 d. For each experimental animal group differences in the AF ureaplasma
CFU/mL were observed: (i) between 87 d and 101 d for the virulent group only (p =
0.04); and (ii) between 87 d and 126 d for both the avirulent and virulent groups (p =
0.02). Ureaplasmas were not detected in the AF of non-infected controls.
All animals from the avirulent and virulent groups tested culture-positive for
ureaplasmas within the chorioamnion at 140 d (Figure 5.1B). Ureaplasmas were
cultured from the cords of 4 out of 8 (50%) animals from both the virulent and
avirulent groups; and from the fetal lungs of 5 out of 8 (62.5%) lambs from the
virulent group and 6 out of 8 (75%) lambs from the avirulent group. No ureaplasmas
were detected in fetal CSF specimens (as determined by both culture and PCR,
data not shown), or from any tissue specimens from non-infected control animals.
Ureaplasma colonization in chorioamnion (p = 0.38), cord (p = 0.66) and fetal lung
(p = 0.49) tissues were not different between treatment groups.
Intra-amniotic ureaplasma infection is associated with fetal inflammation
Inflammatory cell counts within chorioamnion tissue were higher in animals injected
with virulent-derived and avirulent-derived ureaplasmas when compared to controls.
However, inflammatory cell counts were not different between the virulent and
avirulent groups for any of the cell types (p > 0.05, Figure 5.2A). The number of
177
Figure 5.1: Ureaplasma colonization of amniotic fluid and fetal tissues
(A) Chronic infection of the amniotic fluid was observed in all ewes experimentally infected with ureaplasmas from the time of inoculation (55 d) until fetuses were delivered at 140 d. Amniotic fluid ureaplasma colonization was significantly increased in the avirulent group when compared to the virulent group at 87 d (p < 0.05, denoted by #). Statistically significant differences in amniotic fluid ureaplasma colonization within animal groups occurred between 87 d and 101 d; and 87 d and 126 d (p < 0.05). (B) Ureaplasmas were isolated from the chorioamnion, cord and fetal lung; however, recovered ureaplasma CFU/g was not different between animal groups for the tested tissue types. * = statistically significant difference between time points in the virulent group only; ** = statistically significant difference between time points in both groups. AF = amniotic fluid; CFU = colony forming units; d = days of gestation. Data are presented as mean ± SEM.
Amniotic Fluid Colonization
50 d
73 d
87 d
101
d
115
d
126
d
140
d
1000
10000
100000 Virulent group
Avirulent group
Days of Gestation
Ure
ap
lasm
a A
F C
FU
(x 1
04/m
L)
* *
*
A
#
Ureaplasma Tissue Colonization
Virul
ent g
roup
Avi
rule
nt g
roup
1
10
100
1000
10000 Chorioamnion
Cord
Fetal lung
Ure
ap
lasm
a C
olo
niz
ati
on
(x 1
05 C
FU
/gra
m)
B
178
179
Figure 5.2: Fetal inflammation induced by intra-amniotic ureaplasma infection
Inflammatory cell infiltrates within chorioamnion tissue (A) and the severity of histological chorioamnionitis (B) were increased in animals from the virulent and avirulent groups when compared to the control group. Representative chorioamnion sections ((C) stained with haematoxylin and eosin (top row) or Masson’s trichrome stain (bottom row), photographed at x 200 total magnification) demonstrate the 4 stages of histological chorioamnionitis. From left to right: Grade 1 (uninfected control), minimal inflammatory cell infiltrate and no tissue fibrosis, necrosis or abscesses; Grade 2, mild inflammatory cell infiltrate and mild tissue fibrosis, necrosis or abscesses; Grade 3, heavy inflammatory cell infiltrate and moderate tissue fibrosis, necrosis or abscesses; Grade 4, heavy inflammatory cell infiltrate and sever fibrosis, necrosis or abscesses. Stars on haematoxylin and eosin stained sections indicate localized inflammatory cell influx. Arrows on Masson’s trichrome stained sections represent tissue fibrosis and disruption of normal tissue morphology. Size bars represent 50 µm. Inflammatory cell infiltrates within cord tissue (D) and fetal lung tissue (E) were not statistically different between treatment groups. Data are presented as mean + SEM. * p < 0.05 when compared to the control group.
180
macrophages and PMNs within chorioamnion sections were significantly increased
in both the virulent and avirulent groups when compared to controls (p = 0.02 and p
= 0.03 respectively). Grading of chorioamnion sections demonstrated that animals
in the 10B medium control group had no evidence of histological chorioamnionitis,
but moderate to severe histological chorioamnionitis was evident in both the
avirulent and virulent groups (p = 0.001, Figure 5.2B). Representative chorioamnion
sections stained with H & E and Masson’s trichrome stain (Figure 5.2C)
demonstrated the various grades of histological chorioamnionitis. Non-infected
chorioamnion tissues were characterized by minimal/no inflammatory cell influx and
a well-defined structure consisting of a thin layer of fibroblasts bordering intact
epithelial cells. In contrast, histological chorioamnionitis was associated with
increased localized inflammatory cell influx, fibrosis and/or scar tissue (as indicated
by thickening of fibroblast layers and lesion formation) and an irregular epithelial
layer. Umbilical cord and fetal lung inflammatory cell counts were not different
between groups (Figures 5.2D and 5.2E, p > 0.05).
MBA size variation occurs in vivo
Ureaplasma MBA size variants were detected by western blot in AF specimens
collected at 73 d, 87 d, 101 d, 115 d, 126 d and 140 d. The avirulent-derived and
virulent-derived ureaplasma clones used for intra-amniotic injection each expressed
a single mba size variant as determined by PCR (Figure 5.3A); however, these
appeared as double bands in western blots (Figure 5.3B). MBA bands of
approximately 45 kDa and 50 kDa (E22 5.8.1); and 50 kDa and 55 kDa (E24 3.2.1)
were detected in the avirulent-derived and virulent-derived strains respectively.
Antigenic size variation was detected within all tested AF samples from the avirulent
group (Figure 5.3C) and the virulent group (Figure 5.3D), and MBA bands ranged in
size from 32 kDa to 170 kDa. In the avirulent group all detected MBA proteins were
equal in size to, or had an increased molecular weight when compared to the E22
181
182
Figure 5.3: Size variation of the MBA was observed in amniotic fluid samples
PCR of the repeat region of the mba from ureaplasma isolates E22 5.8.1 and E24 3.2.1 produced single amplicons prior to injection into pregnant sheep (A), indicating that these clonal isolates contain only one mba size variant; however, western blots of the same isolates (B) demonstrated that the MBA appears as a double band. MBA size variants were detected in vivo within the amniotic fluid of animals from the avirulent group (C) and the virulent group (D) by western blot. Each western blot demonstrates the ureaplasma MBA size variants generated at 73, 87, 101, 115, 126 and 140 days of gestation in each animal. Note: samples were not collected at some time points for certain animals due to oligohydramnios or other complicating factors. Protein preparations from the avirulent-derived (E22 5.8.1) and virulent-derived (E24 3.2.1) ureaplasma clones used for intra-amniotic injection were included in each western blot (in the last lane) as a positive control and for size comparison. The MBA was detected by anti-ureaplasma polyclonal antibodies, which are specific for the MBA. M = DNA molecular weight marker VIII (Roche, Castle Hill, New South Wales) or protein marker (BioRad, Gladesville, New South Wales). The average number of MBA size variants generated over time (E) was not different between animal groups.
183
5.8.1 inoculum. In the virulent group, the range of MBA size variants was greater, as
MBA bands of lower, equal and higher molecular weight were detected.
From these AF specimens, the total numbers of MBA size variants detected ranged
from one to eight. The average number of AF MBA size variants in the avirulent
group increased over the first three time points and reached a maximum of four
variants at 101 d (Figure 5.3E). Following this, the number of MBA size variants
decreased over the last three time points. Within the virulent group the average
number of MBA size variants within the AF initially increased at 87 d, peaked at 126
d, and then decreased at 140 d. The number of MBA size variants detected within
AF specimens was not different between animal groups at any of the tested time
points (p = 0.87), nor were there differences in the number of MBA size variants
within the two groups for the entire gestation (p = 0.32).
Intra-amniotic ureaplasma infection stimulates a maternal and fetal humoral
response
Anti-ureaplasma IgG antibodies were detected in the maternal serum of four ewes
from the avirulent group and five ewes from the virulent group (Figure 5.4A). In the
avirulent group, the maternal antibodies from each positive ewe reacted with
ureaplasmal proteins (ranging in size from approximately 45 kDa to 87 kDa), which
were different in size to the MBA variants detected in AF specimens (as determined
by molecular weight comparison). Within this avirulent group, the maternal serum of
one ewe (animal 2) reacted with two ureaplasmal proteins, whilst the sera from the
other three animals reacted with only one protein. Positive sera were also probed
against protein extracts from MBA-negative avirulent-derived ureaplasma clones
(generated from serial in vitro transfer experiments). Comparison of protein bands
recognised by maternal serum when probed against whole ureaplasma protein
extract or MBA-negative ureaplasma protein extract (Table 5.3) demonstrated that
184
Figure 5.4: Demonstration of maternal and fetal anti-ureaplasma humoral responses
Anti-ureaplasma IgG antibodies were detected in the maternal serum (A) and the fetal serum (B) of animals in both the avirulent and virulent groups. Antibodies in serum samples reacted with ≥ 1 ureaplasmal protein, over a wide molecular weight range (approximately 45 kDa to 80 kDa). Numbers above western blots indicate the animal number from which the serum was obtained. M = protein marker (BioRad).
185
Table 5.3: Comparison of the protein bands detected by anti-ureaplasma IgG antibodies in maternal serum when probed against whole ureaplasma protein extract or MBA-negative ureaplasma protein extract.
Those proteins indicated in bold were detected in western blots using both whole ureaplasma and MBA negative ureaplasma protein extracts. kDa = kilodaltons.
1 Serum from the avirulent group was probed against the E22 5.8.1 whole ureaplasma protein extract, whereas serum from the virulent group was probed against the E24 3.2.1 whole ureaplasma protein extract.
2 MBA-negative protein extracts were generated from serial in vitro transfer experiments. Serum from ewes from the avirulent group was probed against protein extracts from MBA-negative avirulent ureaplasma clones and serum from the virulent group was probed against protein extracts from MBA-negative virulent ureaplasma clones.
ANIMAL # TREATMENT
GROUP
PROTEIN BANDS RECOGNISED BY MATERNAL SERUM WHEN PROBED AGAINST:
Whole ureaplasma protein
extract1
MBA-negative ureaplasma protein extract
2
2 Avirulent 45 kDa, 87 kDa 45 kDa
29 Avirulent 80 kDa N/A
91 Avirulent 45 kDa 45 kDa
102 Avirulent 70 kDa N/A
42 Virulent 50 kDa 45 kDa
66 Virulent 55 kDa, 62 kDa, 87 kDa N/A
83 Virulent 45 kDa, 50 kDa 45 kDa
85 Virulent 45 kDa, 50 kDa, 55 kDa, 60 kDa N/A
96 Virulent 50 kDa, 55 kDa, 60 kDa 45 kDa
186
two (out of four) serum samples from the avirulent group detected a protein of
approximately 45 kDa, which was expressed in both whole ureaplasma and MBA-
negative ureaplasma preparations.
In the virulent group, one ewe (animal 42) produced antibodies only against a 50
kDa protein, whereas the other four ewes produced antibodies against more than
one ureaplasmal protein (ranging in size from approximately 45 kDa to 80 kDa),
including (but not limited to) MBA size variants that were detected within the AF of
each animal. When positive sera from the virulent group were probed against MBA-
negative virulent-derived ureaplasma protein extracts, three (out of five) serum
samples also detected a protein of approximately 45 kDa (Table 5.3).
Maternal serum collected from non-infected controls did not react with ureaplasmal
proteins and 10B medium, demonstrating that antibodies were not generated
against ureaplasmas or 10B medium components. The number of ewes which
generated anti-ureaplasma IgG antibodies was not different between the avirulent
and virulent groups (p = 0.35).
Anti-ureaplasma IgG was present within the fetal serum of one fetus delivered from
the avirulent group, and one fetus delivered from the virulent group (Figure 5.4B).
Fetal serum from both positive fetuses reacted with only one ureaplasmal protein
(animal 48, avirulent group: 60 kDa; and animal 83, virulent group: 80 kDa), both of
which were not detected within the pool of MBA size variants from the
corresponding AF sample. Interestingly, anti-ureaplasma IgG antibodies were
produced in both the maternal and fetal serum of animal 83; however, these sera
reacted with different ureaplasmal proteins (maternal serum: approximately 45 kDa
and 55 kDa; fetal serum: approximately 80 kDa).
Relative expression of ovine Toll-like receptors and cytokines
187
Within chorioamnion tissue there were no statistically significant increases or
decreases in the expression of ovine Toll-like receptors or cytokines between the
avirulent and virulent groups (Table 5.4). The expression of IL-1β, IL-6 and IL-8
tended to be up-regulated in both the avirulent and virulent groups, relative to the
expression of GAPDH, although high levels of intra-animal variation were observed,
suggesting significant variability in the host immune response. Within both
ureaplasma-infected groups, the relative expression levels of TLR1, TLR2 and
TLR6 were similar to expression levels in media control animals, whereas the
expression of TNF-α and IL-10 were slightly down-regulated (albeit not statistically
significant).
Similarly, within fetal lung tissue, high levels of intra-animal variation were observed
and there were no statistically significant differences in the expression of Toll-like
receptors or cytokines between the avirulent and virulent groups (Table 5.4). Within
the avirulent group, the expression of IL-1β, IL-6 and IL-8 tended to be down-
regulated relative to GAPDH, whereas in the virulent group some animals
demonstrated up-regulation of TNF-α and IL-10.
Interestingly, when these gene expression data were grouped based on the
presence of anti-ureaplasma IgG antibodies within maternal serum (as opposed to
treatment group) significant differences were found. Specifically, the relative
expression of IL-1β, IL-6 and IL-8 within the chorioamnion was significantly
increased in animals which tested positive for anti-ureaplasma IgG antibodies, when
compared to those animals which tested negative for the presence of these
antibodies (IL-1β: p = 0.04; IL-6: p = 0.02; IL-8: p = 0.04, Figure 5.5A). Furthermore,
the relative expression of TNF-α (p = 0.02) and IL-10 (p = 0.04) within chorioamnion
tissue was significantly decreased in IgG negative animals, when compared to
those animals which were positive for anti-ureaplasma IgG antibodies (Figure 5.5B).
The relative expression of TLR1, TLR2 and TLR6 also tended to be decreased in
188
CHORIOAMNION FETAL LUNG
AVIRULENT
GROUP
VIRULENT
GROUP
P VALUE
AVIRULENT
GROUP
VIRULENT
GROUP
P VALUE
TLR1 -2.0 ± 3.1 -2.7 ± 1.9 0.9 0.2 ± 0.7 12.3 ± 10.8 0.3
TLR2 -3.1 ± 2.9 -3.1 ± 2.9 0.9 -2.7 ± 0.4 12.3 ± 14.6 0.3
TLR6 -3.1 ± 2.6 -2.8 ± 1.7 0.9 -0.5 ± 0.5 18.3 ± 17.4 0.3
IL-1β 5.4 ± 2.9 48.2 ± 45.5 0.4 -13.2 ± 3.7 -9.2 ± 11.9 0.7
TNF-α -7.4 ± 4.9 -4.0 ± 3.3 0.6 -0.7 ± 0.7 81.3 ± 81.1 0.4
IL-6 47.5 ± 44.1 71.8 ± 39.7 0.7 -30.9 ± 12.8 -7.6 ± 9.3 0.2
IL-8 87.3 ± 71.4 98.6 ± 81.4 0.9 -38.3 ± 11.6 -32.1 ± 22.2 0.8
IL-10 -5.8 ± 4.3 -8.5 ± 5.9 0.7 -1.8 ± 0.7 230.6 ± 229.5 0.4
Table 5.4: Expression of Toll-like receptors and selected cytokines within the chorioamnion tissue and fetal lung tissue
The relative expression levels of IL-1β, IL-6 and IL-8 were up-regulated within the chorioamnion of animals from the avirulent and virulent groups when compared to the expression of the housekeeping gene GAPDH and after normalisation against 10B medium control animals. In the fetal lung tissue, up-regulation of TLR1, TLR2, TLR6, TNF-α and IL-10 was observed in the virulent group only, and IL-1β, IL-6 and IL-8 were down-regulated in both groups. However, due to the large standard errors associated with high levels of intra-animal variation, no statistical differences were observed. Data are presented as mean ± SEM. TLR = Toll like receptor, IL= interleukin, TNF = tumor necrosis factor.
189
Figure 5.5: Pro-inflammatory cytokines were up-regulated in animals, which produced anti-ureaplasma IgG antibodies
The relative expression of IL-1β, IL-6 and IL-8 within the chorioamnion was significantly increased in ewes that tested positive for anti-ureaplasma IgG antibodies (IgG positive) within serum samples, when compared to animals in which these antibodies were not generated (IgG negative) (A). Conversely, the relative expression of TNF-α and IL-10 (B); and TLR1, TLR2 and TLR6 (C) within the chorioamnion were decreased in IgG negative animals when compared to IgG positive animals. Data are presented as mean fold change ± SEM. * p < 0.05. Expression of genes is determined relative to the expression of GAPDH after normalisation against 10B medium control animals.
190
IgG negative animals, when compared to IgG positive animals; however, these data
were not statistically significant (p = 0.20, p = 0.08, p = 0.16 respectively, Figure
5.5C).
Serial passage of virulent and avirulent ureaplasmas
Serial passage of avirulent-derived E22 5.8.1 and virulent-derived E24 3.2.1
ureaplasma strains in 10B medium without the presence of anti-ureaplasma
polyclonal rabbit sera did not lead to the emergence of MBA escape variants or
MBA size variants after 20 passages (Figure 5.6). Conversely, when these strains
were serially transferred in 10B medium containing rabbit anti-ureaplasma
antibodies, MBA escape variants were generated in both the avirulent-derived and
virulent-derived ureaplasma isolates (Figure 5.6). MBA-negative ureaplasmas were
generated after three serial transfers for isolate E22 5.8.1, and after four serial
transfers for isolate E24 3.2.1. MBA size variation was not observed in any of these
isolates.
191
Figure 5.6: Ureaplasmas are phase variable in vitro
Western blot analysis demonstrated that MBA expression was not affected in avirulent-derived (E22 5.8.1) and virulent-derived (E24 3.2.1) ureaplasmas, which were serially transferred in 10B medium without the presence of polyclonal antibodies (top panel). MBA negative escape variants were generated for both avirulent-derived and virulent-derived ureaplasma strains after serial transfer in 10B medium containing anti-ureaplasma polyclonal antibodies (α-Up, bottom panel). P = passage number.
192
Discussion
The pathogenic role of Ureaplasma spp. in adverse pregnancy outcomes is
controversial. The isolation of these microorganisms from the upper genital tract of
pregnant women who deliver at term with no evidence of chorioamnionitis [8, 9, 11]
suggests that a causal relationship between intra-amniotic ureaplasma infection and
adverse pregnancy outcomes does not always exist. However, advances in our
understanding of the complexities of disease pathogenesis have changed how we
define microbial pathogens and highlighted the importance of host-pathogen
interactions in predicting disease [28]. In this study, we investigated the role of the
MBA from virulent-derived and avirulent-derived ureaplasma strains in a fetal sheep
model of chronic intra-amniotic infection. Our data suggest that ureaplasmas are not
intrinsically virulent/avirulent, as size variation of the MBA did not directly contribute
to fetal inflammation and chorioamnionitis. However, variation of this surface-
exposed antigen may prevent the eradication of ureaplasmas by the host immune
response. For the first time, we have demonstrated a significant association
between the up-regulation of chorioamnion pro-inflammatory cytokines and the
presence of maternal serum anti-ureaplasma antibodies. We predict that a strong
host immune response may be an important determinant in distinguishing
asymptomatic intra-amniotic ureaplasma infections from those resulting in adverse
outcome.
Our data demonstrate that the incidences of fetal abortion and oligohydramnios
were not different between animals that received intra-amniotic injections of virulent-
derived ureaplasmas, avirulent-derived ureaplasmas or 10B medium. Lung
compliance appeared to be increased in the avirulent group when compared to the
virulent and control groups; however, the number of animals studied was too few to
demonstrate statistical significance. Previous data published by our group showed
that long term intra-amniotic ureaplasma infection induced fetal lung compliance in
193
preterm lambs (125 d) when compared to controls [29, 30]. However, fetuses in the
current experiment were delivered at 140 d (near-term), at which time fetal lungs are
in the late saccular/early alveolar developmental phase [31] and are mature. We
observed a significant increase in the presence of meconium-stained AF in the
virulent group compared to both the avirulent and control groups. Meconium-stained
AF is reported to be a sign of fetal stress and was associated previously with
increased incidences of intra-amniotic infection, chorioamnionitis and
intraventricular haemorrhage [32-36]. However, in our experiment fetal outcomes
were not different between animals with/without meconium-stained AF, suggesting
that the presence of meconium within the AF may be an indirect indicator of
infection in some animals.
To the best of our knowledge, this is the first study to demonstrate the progress of
chronic intra-amniotic ureaplasma infections over time. Remarkably, after 85 days of
in utero infection, ureaplasma colonization within the AF remained elevated
throughout gestation, as determined by amniocentesis sampling at two weekly
intervals. Even at 140 d, the AF ureaplasma CFU/mL in both the avirulent and
virulent animal groups was still two-to-three logs higher than the original inoculum
dose of 2x104 CFU, demonstrating that AF is an excellent growth medium for these
microorganisms. It is well documented that AF has bacteriostatic/bacteriocidal
activity against numerous bacteria such as Listeria monocytogenes, Escherichia
coli, Staphylococcus aureus and group B streptococci [37-39] due to various
components within the AF including zinc and phosphate [40, 41]. Therefore, the
ability of the ureaplasmas to thrive within the AF suggests that this may be a niche
environment for these microorganisms.
Our results indicate that ureaplasma colonization of the AF was not affected by
treatment group, as there were no differences in CFU/mL (except at 87 d) between
the avirulent and virulent animal groups. Previously, we demonstrated that
194
ureaplasma colonization of the AF and fetal tissues occurred independently of the
serovar as well as the dose (high dose or low dose) used for intra-amniotic injection
[17]. We predict that a major regulator of ureaplasmal growth within the AF may be
pH, as ureaplasmas require a pH of 6.0 for optimal growth [42] and the pH of ovine
AF ranges from 8.4 (at 70 d) to 7.4 (at 145 d) [43]. Similarly in humans, the pH of AF
from third trimester pregnancies is usually 7.1 ± 0.08 [44], which may have a limiting
effect on ureaplasmal growth. Long-term colonization of the AF may be facilitated by
biofilm formation at the chorioamnion-AF interface, or within ‘amniotic fluid sludge’
(as reported by Romero et al. [45]). Garcia-Castillo et al. [46] demonstrated that
82% of ureaplasmas isolated from urine specimens collected from males diagnosed
with urethritis, chronic prostatitis or healthy individuals formed biofilms in vitro. We
also have preliminary data to suggest that ureaplasmas isolated from the
chorioamnion of experimentally-infected sheep are capable of forming biofilms
(Dando et al. 2010, unpublished), indicating that this may be an important survival
mechanism for these microorganisms in vivo.
We did not observe differences in the severity of histological chorioamnionitis
between animals infected with the avirulent-derived ureaplasma clone or the
virulent-derived ureaplasma clone. Both ureaplasma-infected groups had moderate
to severe chorioamnionitis, as determined by tissue scoring and inflammatory cell
counts. This is in contrast with our previous findings, in which infection with the
parent strains of avirulent (E22 5.8.1) and virulent (E24 3.2.1) clones resulted in
either no histological evidence of chorioamnionitis, or severe chorioamnionitis
respectively [17]. We hypothesise that these differences may be attributed to the
elaboration of different MBA size variants in utero between these two studies. In this
experiment, both the avirulent-derived and virulent-derived ureaplasma clones
produced low numbers of MBA variants in vivo (average = 4.2 and 4.6 size variants
respectively). Whereas previously, intra-amniotic infection with a clinical ureaplasma
195
isolate elaborated the avirulent parent strain (associated with 14 MBA variants) and
the virulent parent strain (associated with 5 MBA variants). Despite differences in
the numbers of MBA size variants generated between these two studies, the
present data do confirm our previous observation that low numbers of MBA size
variants (≤ 5) are associated with severe histological chorioamnionitis.
At the present time, we do not know the mechanisms that drive MBA size variation
in vivo, and are unable to explain the differences in the number of MBA variants
produced by the clonal strains in comparison to the parent strains. However, an
important distinction between these two studies is that the inoculum strain used for
the previous experiment [17] was a non-clonal clinical ureaplasma isolate, which
may have comprised a mixture of MBA subtypes. In contrast, the inocula used for
the current experiment were clonal strains, each expressing a single mba variant.
Our data demonstrate that clonal ureaplasma strains, unlike non-clonal mixtures,
may have a limited ability to generate MBA size variants in vivo. Furthermore, clonal
selection based on MBA antigenic variation may have also selected for clones with
altered expression of other proposed ureaplasmal virulence factors, which include
urease, and phospholipase A and C [3]. Therefore, experiments to characterize the
expression of these additional virulence factors throughout gestation may elucidate
differences between clonally-derived ureaplasma strains.
Antigenic variation, as defined by Deitsch et al. [47], refers to the capacity of a
microorganism to alter the proteins exposed to the host immune system, such that
the host is confronted with a continually changing antigenic population that is
difficult to eliminate. Antigenic variation can refer to either phase variation (on/off
switching) or expression of alternate forms of an antigen (such as size variation).
Perhaps the best characterised examples of antigenic variation are flagella phase
variation in Salmonella spp. [48] and Opa phase variation in Neisseria spp. [49].
However, high frequency antigenic variation is also prominent in numerous
196
Mycoplasma spp. To date, the mechanisms of antigenic variation described in
mycoplasmas include: (i) slipped strand mispairing and/or nucleotide
insertions/deletions in simple sequence repeats; and (ii) DNA rearrangements via
site specific recombination and promoter inversions [50]. Zimmerman et al. [16]
demonstrated that the MBA of U. parvum underwent alternate phase variation with
an adjacent gene, UU376. By in vitro selection using antibody pressure, it was
demonstrated that alternate expression of the MBA/UU376 was associated with a
DNA inversion event in which the 5’ conserved region of the mba and its putative
promoter were opposed to either the 3’ repeat region of the mba or UU376.
Furthermore, phase variation of MBA N-terminal paralogs (UU171 and UU172) was
recently described in both U. parvum and U. urealyticum [51]. Similarly to the
alternate expression of MBA/UU376, phase variation of UU171/UU172 is predicted
to occur via DNA inversion and rearrangement of potential promoter sequences.
Whilst Zimmerman et al. [16] demonstrated that in vitro antibody selection led to the
emergence of MBA-negative escape variants (via phase variation), our in vivo data
demonstrated size variability of the MBA after injection of clonal ureaplasmas
expressing a single mba variant. PCR of the repeat region of the mba confirmed
that isolates E22 5.8.1 and E24 3.2.1 each had one mba size variant prior to intra-
amniotic injection; however, the MBA appeared as a double band in western blots.
These findings are similar to those reported by Zheng et al. [14], who suggested
that MBA doublets may be due to inefficient signal peptide cleavage. However, it is
also possible that this may be a result of post-translational modification, such as
glycosylation, but this remains to be confirmed.
We have previously reported that ureaplasmas isolated from the AF of pregnant
sheep undergo MBA size variation [17]. As yet, MBA-negative ureaplasmas have
not been isolated from patients or generated in vivo, therefore it is possible that
phase variation of the MBA is induced only by strong selection using antibodies
197
directed against the repetitive region of the MBA. To select for MBA-negative
escape variants, Zimmerman et al. [16] incubated low numbers of clonal
ureaplasmas with hyperimmune antisera diluted 1/100. It is highly unlikely that these
conditions would be mimicked in vivo, which is perhaps why we have not observed
MBA phase variation in our sheep model. To determine whether isolates E22 5.8.1
and E24 3.2.1 were capable of MBA phase variation, we applied antibody pressure
to these clonal ureaplasmas in vitro and found that MBA-negative variants were
elaborated after either three or four serial transfers. Similarly, we were only able to
achieve this using a high concentration of rabbit antisera, which further suggests
that MBA phase variation may only be inducible in vitro using concentrated
antibodies.
A primary function of antigenic variation is to evade the adaptive immune response
and to a lesser extent, the innate immune response [47]. Our data demonstrated
that MBA size variants were generated in all animals, not just those in which anti-
ureaplasma IgG antibodies were detected. This suggests that MBA size variation
was not driven by the development of a host humoral response, nor did it prevent
recognition by host pattern recognition receptors. However, continual size variation
of the MBA may prevent the eradication of ureaplasmas due to changes to epitopes
within the repeat region of the protein. As IgG is unable to cross the placenta in
sheep, the ureaplasma-specific antibodies detected in maternal sera were most
likely generated in response to ureaplasma invasion of maternal tissues, such as
the decidua. As expected, the sheep anti-ureaplasma IgG antibodies were
predominantly reactive against MBA size variants that were produced in vivo
throughout gestation. However, not all proteins that were detected by sheep sera
within the whole ureaplasma protein extract matched the size of MBA variants,
suggesting that non-MBA proteins may also be immunogenic in these animals.
Alternatively, it is also possible that these proteins represent MBA size variants that
198
were not detected during our two weekly sampling intervals. To test this, sheep
serum samples were also probed against MBA-negative ureaplasma protein
extracts. These results demonstrated that a 45kDa non-MBA protein within these
extracts was immunogenic, suggesting that antibodies can be produced against
proteins other than the MBA in vivo. Although the identity of this 45kDa protein was
not determined, Figure 5.6 confirms that MBA expression was not detected in the
MBA-negative ureaplasma isolates, thus providing evidence that the 45kDa protein
is a non-MBA protein. Previous studies have demonstrated that the presence of
serum antibodies against ureaplasmas was more associated with preterm birth, low
birth weight, stillbirth, neonatal respiratory disease and fetal death when compared
to patients without anti-ureaplasma antibodies [52, 53]. Quinn [52] reported that a
fetal antibody response to U. urealyticum occurred in 77.3% of stillbirths, 58.3% of
respiratory disease cases, 69.3% of neonatal deaths, 80.4% of term neonates with
complications, but only in 6.5% of healthy, term neonates (p ≤ 0.001). Additionally,
Horowitz et al. [53] reported that the rates of preterm birth and fetal death were
significantly higher in women with antibodies against U. urealyticum compared to
those without these antibodies (90% vs. 43% (p = 0.006); and 85% vs. 28% (p =
0.001) respectively). In our experiment we did not observe any differences in fetal
outcomes in animals with/with-out anti-ureaplasma IgG, except for the presence of
meconium-stained AF, which was increased in those animals with IgG antibodies
(data not presented). Interestingly, only two fetuses developed IgG antibodies in
response to chronic ureaplasma infection.
It is well documented that intra-uterine infection/ inflammation is associated with
adverse pregnancy outcomes (especially preterm birth [54]), via mechanisms
reviewed elsewhere [7, 55, 56]. Specifically, elevation of pro-inflammatory
cytokines/chemokines such as IL-1β, Il-6, IL-8 and TNF-α within the AF and fetal
membranes have been associated with preterm delivery and chorioamnionitis [57-
199
62]. In this study, intra-amniotic infection with either the avirulent-derived or virulent-
derived ureaplasma clone tended to result in the increased expression of IL-1β, IL-6
and IL-8 within chorioamnion tissue; however, there were high levels of intra-animal
variation, suggesting variability in the host immune response. Surprisingly, we did
not observe any increases in TLR1, TLR2 or TLR6 expression, which are the
pattern recognition receptors through which the MBA activates nuclear factor
kappaB [63]. Reyes et al. [64] also noted variability in the innate immune response
after inoculation of U. parvum into the bladder of Fischer 344 rats. They
demonstrated that the severity of urinary tract infection was associated with distinct
urine cytokine profiles. Asymptomatic urinary tract infection was associated with
elevation of interferon-γ, IL-18 and monocyte chemotactic protein-1, whereas
complicated urinary tract infection with struvite formation was characterised by
increased IL-1α, IL-1β and growth related oncogene/keratinocyte chemoattractant
(analogous to human IL-8). Kasper et al. [65] also found differences in the
expression of IL-8 within human AF infected with U. parvum and determined that
bacterial load significantly influenced the levels of AF IL-8. In our study, differences
in IL-1β, IL-6 and IL-8 expression did not correlate with ureaplasma CFU/mL or
gram of tissue, or the severity of histological chorioamnionitis. Others have
demonstrated that intra-amniotic ureaplasma infection did not result in increased
levels of any tested cytokines (IL-1β, IL-1 receptor antagonist, IL-4, IL-6 and TNF-α
[9]). Also, in vitro stimulation experiments, in which human choriodecidua or fetal
membrane tissues were stimulated with ureaplasmas, demonstrated a T helper-2
dominant cytokine response (characterized by IL-10 production, [66, 67]). Taken
together, these data suggest that the host innate immune response to intra-amniotic
ureaplasma infection is not uniform, and this could account for the variety of
outcomes associated with in-utero ureaplasma infection in humans.
200
We found that the animals with maternal anti-ureaplasma IgG antibodies had
significantly higher levels of IL-1β, IL-6 and IL-8 within chorioamnion tissue, when
compared to IgG negative animals. We predict that a strong pro-inflammatory innate
response within the chorioamnion may have induced a humoral immune response
in these animals. This is the first time a correlation between the innate and adaptive
immune response during intra-amniotic ureaplasma infection has been described
and is a finding unique to this animal model.
In conclusion, we have demonstrated that U. parvum avirulent-derived and virulent-
derived clones are able to chronically colonise the AF of pregnant sheep and cause
histological chorioamnionitis. Our data suggest that ureaplasmas may not be
intrinsically virulent or avirulent; and it appears to be more likely that the host
immune response generated against intra-amniotic ureaplasma infection is a key
determinant of adverse pregnancy outcomes. Size variation of the MBA did not
correlate with different histological outcomes, and MBA size variation occurred in all
animals, regardless of the intensity of the innate and adaptive immune responses.
This suggests that MBA size variability did not prevent recognition by host pattern
recognition receptors. However, it may prevent the host immune response from
eradicating ureaplasmas from the amniotic cavity and thus play a role in the
virulence of these microorganisms.
201
Acknowledgements
The authors wish to thank JRL Hall & Co., in particular Sara Ritchie and Fiona Hall,
who have been responsible for breeding and supplying us with the high quality
research animals necessary for this project.
202
References
1. Robertson JA, Stemke GW, Davis JW, Jr, Harasawa R, Thirkell D, et al.
(2002) Proposal of Ureaplasma parvum sp. nov. and emended description of
Ureaplasma urealyticum (Shepard et al. 1974) Robertson et al. 2001. Int J
Syst Evol Microbiol 52: 587-597.
2. Smith DGE, Russell WC, Ingledew WJ, Thirkell D (1993) Hydrolysis of urea
by Ureaplasma urealyticum generates a transmembrane potential with
resultant ATP synthesis. J Bacteriol 175: 3253-3258.
3. Glass JI, Lefkowitz EJ, Glass JS, Heiner CR, Chen EY, et al. (2000) The
complete sequence of the mucosal pathogen Ureaplasma urealyticum.
Nature 407: 757-762.
4. Volgmann T, Ohlinger R, Panzig B (2005) Ureaplasma urealyticum-harmless
commensal or underestimated enemy of human reproduction? A review.
Arch Gynecol Obstet 273: 133-139.
5. Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R (1993)
Ureaplasma urealyticum intrauterine infection: role in prematurity and
disease in newborns. Clin Microbiol Rev 6: 69-87.
6. Waites KB, Schelonka RL, Xiao L, Grigsby PL, Novy MJ (2009) Congenital
and opportunistic infections: Ureaplasma species and Mycoplasma hominis.
Semin Fetal Neonatal Med 14: 190-199.
7. Goldenberg RL, Hauth JC, Andrews WW (2000) Intrauterine infection and
preterm delivery. New Engl J Med 342: 1500-1507.
8. Gerber S, Vial Y, Hohlfeld P, Witkin SS (2003) Detection of Ureaplasma
urealyticum in second-trimester amniotic fluid by polymerase chain reaction
correlates with subsequent preterm labor and delivery. J Infect Dis 187: 518-
521.
203
9. Perni SC, Vardhana S, Korneeva I, Tuttle SL, Paraskevas L, et al. (2004)
Mycoplasma hominis and Ureaplasma urealyticum in midtrimester amniotic
fluid: association with amniotic fluid cytokine levels and pregnancy outcome.
Am J Obstet Gynecol 191: 1382-1386.
10. Yoon BH, Chang JW, Romero R (1998) Isolation of Ureaplasma urealyticum
from the amniotic cavity and adverse outcome in preterm labor. Obstet
Gynecol 92: 77-82.
11. Horowitz S, Mazor M, RomeroR, Horowitz J, Glezerman M (1995) Infection
of the amniotic cavity with Ureaplasma urealyticum in the midtrimester of
pregnancy. J Reprod Med 40: 375-379.
12. Zheng X, Watson HL, Waites KB, Cassell GH (1992) Serotype diversity and
antigen variation among invasive isolates of Ureaplasma urealyticum from
neonates. Infect Immun 60: 3472-3474.
13. Zheng X, Teng LJ, Glass JI, Blanchard A, Cao Z, et al. (1994) Size variation
of a major serotype-specific antigen of Ureaplasma urealyticum. Ann N Y
Acad Sci 15: 299-301.
14. Zheng X, Teng LJ, Watson HL, Glass JI, Blanchard A, et al. (1995) Small
repeating units within the Ureaplasma urealyticum MB antigen gene encode
serovar specificity and are associated with antigen size variation. Infect
Immun 63: 891-898.
15. Monecke S, Helbig JH, Jacobs E (2003) Phase variation of the multiple
banded protein in Ureaplasma urealyticum and Ureaplasma parvum. Int J
Med Microbiol 293: 203-211.
16. Zimmerman CR, Stiedl T, Rosengarten R, Spergser J (2009) Alternate
phase variation in expression of two major surface membrane proteins (MBA
and UU376) of Ureaplasma parvum serovar 3. FEMS Microbiol Lett 292:
187-193.
204
17. Knox CL, Dando SJ, Nitsos I, Kallapur SG, Jobe AH, et al. (2010) The
severity of chorioamnionitis in pregnant sheep is associated with in vivo
variation of the surface-exposed multiple-banded antigen/gene of
Ureaplasma parvum. Biol Reprod 83: 415-426.
18. Watson HL, Blalock DK, Cassell GH (1990) Variable antigens of Ureaplasma
urealyticum containing both serovar-specific and serovar-cross-reactive
epitopes. Infect Immun 58: 3679-3688.
19. Brown NF, Wickham ME, Coombes BK, Finlay BB (2006) Crossing the line:
selection and evolution of virulence traits. PLoS Pathog 2: 346-353.
20. Tully JG (1983) Cloning and filtration techniques for mycoplasmas. In: Razin
S, Tully G, editors. Methods in Mycoplasmology. New York: Academic
Press. pp. 173-177.
21. Moss TJ, Knox CL, Kallapur SG, Nitsos I, Theodoropoulos C, et al. (2008)
Experimental amniotic fluid infection in sheep: effects of Ureaplasma parvum
serovars 3 and 6 on preterm or term fetal sheep. Am J Obstet Gynecol 198:
122e1-e8.
22. Dando SJ, Nitsos I, Newnham JP, Jobe AH, Moss TJM, et al. (2010)
Maternal administration of erythromycin fails to eradicate intrauterine
ureaplasma infection in an ovine model. Biol Reprod 83: 616-622.
23. McNeilly TN, Baker A, Brown JK, Collie D, MacLachlan G, et al. (2008) Role
of alveolar macrophages in respiratory transmission of visna/maedi virus. J
Virol 82: 1526-1536.
24. Chang J, Russell GC, Jann O, Glass EJ, Werling D, et al. (2009) Molecular
cloning and characterization of Toll-like receptors 1-10 in sheep. Vet
Immunol Immunopathol 127: 94-105.
25. Egan PJ, Kimpton W, Seow H-F, Bowles VM, Brandon MR, et al. (1996)
Inflammation-induced changes in the phenotype and cytokine profile of cells
migrating through skin and afferent lymph. Immunology 89: 539-546.
205
26. Schmittgen TD, Livak KJ (2008) Analyzing real-time PCR data by the
comparative CT method. Nat Protoc 3: 1101-1108.
27. Pfaffl MW (2001) A new mathematical model for relative quantification in
real-time RT-PCR. Nucleic Acids Res 29: 2002-2007.
28. Falkow S (2004) Molecular Koch’s postulates applied to bacterial
pathogenicity- a personal recollection 15 years later. Nat Rev Microbiol 2:
67-72.
29. Moss TJM, Nitsos I, Knox CL, Polglase GR, Kallapur SG, et al. (2009)
Ureaplasma colonization of amniotic fluid and efficacy of antenatal
corticosteroids fore preterm lung maturation in sheep. Am J Obstet Gynecol
200: 96e1-e6.
30. Polglase GR, Dalton RGB, Nitsos I, Knox CL, Pillow JJ, et al. (2010)
Pulmonary vascular and alveolar development in preterm lambs chronically
colonized with Ureaplasma parvum. Am J Physiol Lung Cell Mol Physiol
299: L232-L241.
31. Gnanalingham MG, Mostyn A, Dandrea J, Yakubu DP, Symonds ME, et al.
(2005) Ontogeny and nutritional programming of uncoupling protein-2 and
glucocorticoid receptor mRNA in the ovine lung. J Physiol 565.1: 159-169.
32. Romero R, Hanaoka S, Mazor M, Athanassiadis AP, Callahan R, et al.
(1991) Meconium-stained amniotic fluid: a risk factor for microbial invasion of
the amniotic cavity. Am J Obstet Gynecol 164: 859-862.
33. Mazor M, Furman B, Wiznitzer A, Shoham-Vardi I, Cohen J, et al. (1995)
Maternal and perinatal outcome of patients with preterm labor and
meconium-stained amniotic fluid. Obstet Gynecol 86: 830-833.
34. Chapman S, Duff P (1995) Incidence of chorioamnionitis in patients with
meconium-stained amniotic fluid. Infect Dis Obstet Gynecol 2: 210-212.
206
35. Rao S, Pavlova Z, Incerpi MH, Ramanathan R (2001) Meconium-stained
amniotic fluid and neonatal morbidity in near-term and term deliveries with
acute histologic chorioamnionitis and/or funisitis. J Perinatol 21: 537-540.
36. Tybulewicz AT, Clegg SK, Fonfé GJ, Stenson BJ (2004) Preterm meconium
staining of the amniotic fluid: associated findings and risk of adverse clinical
outcome. Arch Dis Child Fetal Neonatal Ed 89: F328-F330.
37. Schlievert P, Larsen B, Johnson W, Galask RP (1975) Bacterial growth
inhibition by amniotic fluid. III. Demonstration of the variability of bacterial
growth inhibition by amniotic fluid with a new plate-count technique. Am J
Obstet Gynecol 122: 809-819.
38. Evans HE, Levy E, Glass L (1977) Effect of amniotic fluid on bacterial
growth. Obstet Gynecol 49: 35-37.
39. Eidelman AI, Nevet A, Rudensky B, Rabinowitz R, Hammerman C, et al.
(2002) The effect of meconium staining of amniotic fluid on the growth of
Escherichia coli and Group B Streptococcus. J Perinatol 22: 467-471.
40. Schlievert P, Johnson W, Galask RP (1976) Bacterial growth inhibition by
amniotic fluid. V. Phosphate-to-zinc ratio as a predictor of bacterial growth-
inhibitory activity. Am J Obstet Gynecol 125: 899-905.
41. Schlievert PS, Johnson W, Galask RP (1976) Bacterial growth inhibition by
amniotic fluid. VI. Evidence for a zinc-peptide antibacterial system. Am J
Obstet Gynecol 125: 906-910.
42. Shepard MC, Lunceford CD (1965) Effect of pH on human Mycoplasma
strains. J Bacteriol 89: 265-270.
43. Prestes NC, Chalhoub MCL, Lopes MD, Takahira RK (2001) Amniocentesis
and biochemical evaluation of amniotic fluid in ewes at 70, 100 and 145 days
of pregnancy. Small Rumin Res 39: 277-281.
207
44. Fadel HE, Northrop G, Misenhimer HR, Harp RJ (1979) Acid-base
determinations in amniotic fluid and blood of normal late pregnancy. Obstet
Gynecol 53: 99-104.
45. Romero R, Schaudinn C, Kusanovic JP, Gorur A, Gotsch F, et al. (2008)
Detection of a microbial biofilm in intraamniotic infection. Am J Obstet
Gynecol 198: 135 e1-5.
46. García-Castillo M, Morosini MI, Gàlvez M, Baquero F, del Campo R, et al.
(2008) Differences in biofilm development and antibiotic susceptibility among
clinical Ureaplasma urealyticum and Ureaplasma parvum isolates. J
Antimicrob Chemother 62: 1027-1030.
47. Deitsch KW, Lukehart SA, Stringer JR (2009) Common strategies for
antigenic variation by bacterial, fungal and protozoan pathogens. Nat Rev
Microbiol 7: 493-503.
48. Silverman M, Zieg J, Hilmen M, Simon M (1979) Phase variation in
Salmonella: genetic analysis of a recombinational switch. Proc Natl Acad Sci
USA 76: 391-395.
49. Cannon JG, Sparling PF (1984) The genetics of the gonococcus. Ann Rev
Microbiol 38: 111-133.
50. Citti C, Nouvel L-X, Baranowski E (2010) Phase and antigenic variation in
mycoplasmas. Future Microbiol 5: 1073-1085.
51. Zimmerman CR, Rosengarten R, Spergser J (2011) Ureaplasma antigenic
variation beyond MBA phase variation: DNA inversions generating chimeric
structures and switching in expression of the MBA N-terminal paralogue
UU172. Mol Microbiol 79: 663-676.
52. Quinn PA (1986) Evidence of an immune response to Ureaplasma
urealyticum in perinatal morbidity and mortality. Pediatr Infect Dis 5: S282-
S287.
208
53. Horowitz S, Mazor M, Horowitz J, Porath A, Glezerman M (1995) Antibodies
to Ureaplasma urealyticum in women with intraamniotic infection and
adverse pregnancy outcome. Acta Obstet Gynecol Scand 74: 132-136.
54. Romero R, Espinoza J, Gonçalves LF, Kusanovic JP, Friel LA, et al. (2006)
Inflammation in preterm and term labour and delivery. Semin Fetal Neonatal
Med 11: 317-326.
55. Goldenberg RL, Culhane JF, Iams JD, Romero R (2008) Epidemiology and
causes of preterm birth. Lancet 371: 75-84.
56. Peltier MR (2003) Immunology of term and preterm labor. Reprod Biol
Endocrinol 1: 1-11.
57. Hillier SL, Witkin SS, Krohn MA, Watts DH, Kiviat NB, et al. (1993) The
relationship of amniotic fluid cytokines and preterm delivery, amniotic fluid
infection, histologic chorioamnionitis, and chorioamnion infection. Obstet
Gynecol 81: 947-948.
58. Arntzen KJ, Kjøllesdal AM, Halgunset J, Vatten L, Austgulen R (1998) TNF,
IL-1, IL-6, IL-8 and soluble TNF receptors in relation to chorioamnionitis and
premature labor. J Perinatal Med 26: 17-26.
59. Saji F, Samejima Y, Kamiura S, Sawai K, Shimoya K, et al. (2000) Cytokine
production in chorioamnionitis. J Reprod Immunol 47: 185-196.
60. Jacobsson B, Mattsby-Baltzer I, Hagberg H (2005) Interleukin-6 and
interleukin-8 in cervical and amniotic fluid: relationship to microbial invasion
of the chorioamniotic membranes. BJOG 112: 719-724.
61. Thomakos N, Daskalakis G, Papapanagiotou A, Papantoniou N, Mesogitis
S, et al. (2010) Amniotic fluid interleukin-6 and tumor necrosis factor-α at
mid-trimester genetic amniocentesis: relationship to intra-amniotic microbial
invasion and preterm delivery. Eur J Obstet Gynecol Reprod Biol 148: 147-
151.
209
62. Marconi C, Ramos BRA, Peraçoli JC, Donders GGG, Silva MG (2011)
Amniotic fluid interleukin-1 beta and interleukin-6, but not interleukin-8
correlate with microbial invasion of the amniotic cavity in preterm labor. Am J
Reprod Immunol. In press.
63. Shimizu T, Kida Y, Kuwano K (2008) Ureaplasma parvum lipoproteins,
including MB antigen, activate NF-κB through TLR1, TLR2 and TLR6.
Microbiology 154: 1318-1325.
64. Reyes L, Reinhard M, Brown MB (2009) Different inflammatory responses
are associated with Ureaplasma parvum-induced UTI and urolith formation.
BMC Infect Dis 9: 1-12.
65. Kasper DC, Mechtler TP, Reischer GH, Witt A, Langgartner M, et al. (2010)
The bacterial load of Ureaplasma parvum in amniotic fluid is correlated with
an increased intrauterine inflammatory response. Diagn Microbiol Infect Dis
67: 117-121.
66. Aaltonen R, Heikkinen J, Vahlberg T, Jensen JS, Alanen A (2007) Local
inflammatory response in choriodecidua induced by Ureaplasma
urealyticum. BJOG 114: 1432-1435.
67. Menon R, Peltier MR, Eckardt J, Fortunato SJ (2009) Diversity in cytokine
response to bacteria associated with preterm birth by fetal membranes. Am
J Obstet Gynecol 201: 306.e1-6.
210
Chapter 6
GENERAL DISCUSSION
211
6.1 DISCUSSION
Despite being commensals of the lower genital tract of sexually active females, the
Ureaplasma spp. are the most prevalent potentially pathogenic bacteria isolated
from the upper genital tract of pregnant women. Intra-uterine infection and
inflammation of the fetal membranes is a leading cause of preterm birth
(Goldenberg et al. 2000), which is associated with numerous adverse sequelae for
the newborn infant, including respiratory distress, neurodevelopmental impairment
and death (Saigal and Doyle 2008; Simmons et al. 2010). Ureaplasmas are the
microorganisms most frequently isolated from infected amniotic fluids and placentas
and are associated with preterm birth and other adverse pregnancy outcomes
(Cassell et al. 1993). Therefore, research investigating the pathogenesis of intra-
amniotic ureaplasma infections and potentially informing treatment options may lead
to a reduction in preventable preterm birth and the associated neonatal morbidity
and mortality.
Intra-amniotic ureaplasma infections are often clinically asymptomatic, chronic and
difficult to treat (Romero et al. 2003; Waites et al. 2009). To date, there has been
very little research investigating the host and microbial factors contributing to this.
The overall hypothesis of this PhD project was that intra-amniotic ureaplasma
infections are able to persist in utero due to (i) the ineffectiveness of current
antimicrobial treatment protocols; and (ii) the inability of the host immune system to
eliminate ureaplasmas due to variable expression of the MBA. We tested this
hypothesis using a sheep model of intra-uterine infection, by establishing chronic
intra-amniotic infections with either U. parvum serovar 3 or serovar 6, which are the
two most prevalent ureaplasma serovars isolated from both men and women
(Cassell et al. 1993; Knox and Timms 1998; Knox et al. 2003). The data obtained
from this PhD project confirmed that maternal erythromycin treatment of intra-
amniotic ureaplasma infections is ineffective, potentially due to the minimal transfer
212
of erythromycin across the placental barrier. Furthermore, ureaplasmas isolated
from the amniotic fluid and the chorioamnion of pregnant sheep demonstrated
marked differences in their sensitivity to macrolide antibiotics (especially
roxithromycin) and in the nucleotide sequences of macrolide binding sites (ie.
domain V of the 23S rRNA gene). It was also determined that intra-amniotic
ureaplasma infection induced a host immune response; however, this response was
not capable of neutralising in utero infection. Within the ovine model, ureaplasmas
demonstrated MBA size variation, although in vivo variation of the MBA did not
mediate inflammation or the severity of chorioamnionitis. Combined, the data
presented in this thesis suggest that the host immune response alone is insufficient
to eradicate intra-amniotic ureaplasma infection and this confirms the need for
antimicrobial treatment. However, this study has highlighted the short-comings of
current treatment protocols and has demonstrated that ureaplasmas can undergo
high rates of selection in vivo, resulting in sub-populations of ureaplasmas
colonising different anatomical sites.
This study further validated the sheep model of intra-amniotic ureaplasma infection.
Whilst the sheep model has been used extensively to study the effects of
Escherichia coli lipopolysaccharide on preterm fetuses, the model of injection of live
U. parvum into the amniotic cavity of pregnant sheep has been less well
characterised. Similar to the effects of intra-amniotic ureaplasma infection on human
fetuses, increased lung compliance was observed in preterm, but not term, ovine
fetuses. Also, in utero ureaplasma exposure resulted in chorioamnionitis, funisitis,
fetal lung inflammation and innate and adaptive immune responses similar to those
observed in humans. Using this ovine model, we were able to establish chronic
intra-amniotic ureaplasma infections from 55 days of gestation to either 125
(preterm) or 140 (near term) days of gestation. The timing and duration of these
ureaplasma infections are similar to clinically silent, intra-amniotic infections
213
established early in pregnancy. Although the natural infective dose of ureaplasmas
required for initiation of intra-amniotic infection (in both humans and sheep) is
unknown, a low inoculum dose of 2x104 CFU was used in all experiments to
potentially represent an ascending invasive infection. However, the number of
ureaplasmas required to establish an in vivo infection may be significantly lower, as
Reyes et al. (2009) demonstrated that inoculation of 101 CFU of U. parvum into the
bladder of Fisher 344 rats resulted in successful colonisation of the bladder two
weeks post infection in 29% of animals. Based on the similarity between intra-
amniotic ureaplasma infection in sheep and humans, the data obtained from this
PhD project are clinically relevant and contribute towards our understanding of
chronic intra-uterine ureaplasma infection and inflammation in pregnant women.
The pathogenic role of ureaplasmas has been strongly debated within the literature,
but the data presented in this thesis and from others investigating the role of intra-
amniotic ureaplasma infection in a rhesus macaque model (Novy et al. 2009) have
demonstrated that U. parvum, as a sole pathogen, causes chorioamnionitis and
fetal lung injury. Given that these data support a causal association between intra-
amniotic ureaplasma infection and adverse pregnancy/fetal outcomes, antimicrobial
treatment of these infections may potentially reduce morbidity and mortality rates in
newborn infants. Erythromycin is the standard antimicrobial recommended for the
treatment of human intra-amniotic infections and preterm prelabour rupture of
membranes (Kenyon et al. 2001); however, there is significant controversy
regarding the efficacy of this treatment. In this PhD program of study we have
demonstrated that standard maternal erythromycin treatment is ineffective at
eradicating intra-amniotic ureaplasma infections in pregnant sheep. Following intra-
amniotic injection of U. parvum serovar 3 at 55 days of gestation, ewes were
administered intra-muscular erythromycin (500 mg, three times daily for four days;
30 mg/kg/day) at 100 days of gestation. Despite receiving erythromycin treatment,
214
the amniotic fluid of pregnant ewes remained colonised until the time of preterm
delivery of the fetus (at 125 days of gestation); and chorioamnion, cord and fetal
lung tissues were also found to be heavily colonised, demonstrating the
ineffectiveness of this treatment protocol. Liquid chromatography-mass
spectrometry analysis of amniotic fluid demonstrated that erythromycin levels were
low, suggesting limited transfer of erythromycin across the placenta.
Kiefer et al. (1955) first reported that the placental transfer of erythromycin was low,
after analysis of maternal and fetal plasma at the time of therapeutic abortion.
Similar findings have been reported by others (Philipson et al. 1973; Heikkinen et al.
2000); however, erythromycin still remains first-line therapy for administration to
pregnant women. According to the Australian Antibiotic Therapeutic Guidelines
(Antibiotic Expert Group, 2010) there are 16 antibacterial drugs, which are classified
as Category A, and therefore have been well studied and are appropriate for use
during pregnancy. Of these, the antibiotics with the most efficient placental transfer
are the β-lactams and the anti-mycobacterial drugs, which ureaplasmas are
inherently resistant to as they lack a cell wall. Comparison of the Category A
antibiotics (Table 6.1) revealed that only erythromycin and azithromycin have
consistent bacteriostatic activity (or bacteriocidal activity, in high concentrations)
against ureaplasmas. In this study, it was demonstrated that erythromycin and
azithromycin are effective at eliminating both planktonic and sessile ureaplasma
populations in vitro. However, these two macrolides, potentially due to their large
molecular weight, demonstrate the poorest placental transfer.
The macrolides, which inhibit bacterial protein synthesis, have a wide spectrum of
activity that effectively targets Gram positive and Gram negative aerobes and
anaerobes, as well as Mycoplasma spp., Ureaplasma spp., and Chlamydia spp.
(Antibiotic Expert Group, 2010). Therefore, the macrolides are the best available
215
Table 6.1: Comparison of the placental transfer and anti-ureaplasmal activity of Category A antibiotics
Placental transfer: + incomplete or low level transfer across the placenta, often achieves only sub-therapeutic concentrations; ++ moderate transfer across the placenta; +++ complete transfer across
the placenta associated with therapeutic concentrations.
Anti-ureaplasma activity: - no anti-ureaplasmal activity; + some demonstrated activity, although results are highly variable; ++ moderate activity, although results are variable; +++ high levels of
bacteriostatic/bacteriocidal activity.
Adapted from Mylonas 2011.
ANTIBIOTIC CLASSIFICATION MODE OF
ACTION
PLACENTAL
TRANSFER
ANTI-
UREAPLASMAL
ACTIVITY
Amoxycillin Beta lactam Inhibit cell wall
synthesis ++ -
Ampicillin Beta lactam Inhibit cell wall
synthesis +++ -
Benzathine
penicillin Beta lactam
Inhibit cell wall
synthesis +++ -
Benzylpenicillin Beta lactam Inhibit cell wall
synthesis +++ -
Cefalotin Beta lactam Inhibit cell wall
synthesis + -
Cephalexin Beta lactam Inhibit cell wall
synthesis +++ -
Phenoxymethyl
penicillin Beta lactam
Inhibit cell wall
synthesis unknown -
Procaine penicillin Beta lactam Inhibit cell wall
synthesis ++ -
Chloramphenicol Phenicol
Inhibit bacterial
protein
synthesis
+++ ++
Ethambutol Anti-mycobacterial
Inhibit
arabinogalactan
synthesis
+++ -
Isoniazid Anti-mycobacterial Inhibit mycolic
acid synthesis +++ -
Clindamycin Lincosamide
Inhibit bacterial
protein
synthesis
++ +
Lincomycin Lincosamide
Inhibit bacterial
protein
synthesis
+ +
Azithromycin Macrolide
Inhibit bacterial
protein
synthesis
+ +++
Erythromycin Macrolide
Inhibit bacterial
protein
synthesis
+ +++
Nitrofurantoin Nitrofurantoin Degrade
bacterial DNA unknown +
216
treatment option for human intra-amniotic infections as they are active against the
majority of bacteria that are capable of invading the amniotic cavity. The poor
placental transfer of these antimicrobials represents a significant challenge in
clinical obstetrics, and the data presented in this thesis confirms that the placental
barrier drastically limits the effectiveness of erythromycin treatment of intra-amniotic
ureaplasma infections. Direct injection of antimicrobials into the amniotic fluid may
be required to achieve therapeutic concentrations within the fetal compartment. Our
group has recently compared the concentrations of erythromycin and azithromycin
in maternal plasma, fetal plasma and amniotic fluid after maternal intra-muscular
injection, maternal intra-venous injection, or direct intra-amniotic injection of
antimicrobials (Keelan et al. 2011). Similar to the results reported here, these data
confirmed that maternal administration of either erythromycin or azithromycin
resulted in sub-therapeutic concentrations within the amniotic fluid. However, direct
injection of antimicrobials into the amniotic cavity achieved high amniotic fluid
concentrations (erythromycin Cmax = 8.7 µg/mL; azithromycin Cmax = 18.9 µg/mL)
and therapeutic levels were maintained for 48 hours after a bolus injection. These
findings are promising, although further data regarding the safety of intra-amniotic
antimicrobial injections in pregnant women is required. Similar techniques such as
amniocentesis and chorionic villus sampling are associated with a procedure-related
miscarriage rate of 0.5 - 1.0% (Tabor and Alfirevic 2010), therefore further studies to
determine the risk of intra-amniotic antimicrobial injections in pregnant women are
warranted.
Currently, the implications of incomplete placental transfer of erythromycin have
only been considered in terms of effects on the fetus arising from prolonged
exposure to intra-uterine infection/inflammation. The effects of sub-lethal exposures
of erythromycin on microbial populations within the amniotic fluid and chorioamnion
have not been determined. Antimicrobials are thought to exhibit hormetic
217
behaviour, which is defined as the phenomenon by which an agent has different and
even opposite effects at high and low concentrations (Couce and Blázquez 2009).
Therefore, exposure of microorganisms to sub-inhibitory concentrations of
antimicrobials can not only select for pre-existing resistant phenotypes, but can also
result in changes to bacterial physiology, morphology, virulence and also increase
rates of genetic mutability (Davies et al. 2006; Davies and Davies 2010).
The effects of sub-inhibitory concentrations of erythromycin on ureaplasmas
isolated from the amniotic fluid and chorioamnion of pregnant ewes were
determined in this study. High levels of variability were observed in macrolide MICs
between amniotic fluid and chorioamnion ureaplasma isolates; however, in vivo
erythromycin exposure was not found to be a contributing factor to this variability.
This experiment is the first to demonstrate that injection of a single U. parvum
clinical isolate (isolated originally from the semen of an infertile man and containing
a mixture of ureaplasmas that were adherent and non-adherent to spermatozoa)
into the amniotic fluid of pregnant sheep can generate ureaplasmas with variable
MICs. These findings alone have significant implications for antimicrobial treatment
of intra-amniotic ureaplasma infections, as populations of sensitive and resistant
ureaplasmas may be generated in utero. As there are currently no standardised
methods or breakpoints for in vitro antimicrobial susceptibility testing of
ureaplasmas, reporting of antimicrobial resistance is often not uniform between
laboratories. The microbroth dilution method used in Chapter 4 is the most widely
used method for testing Ureaplasma spp. and enables the accurate determination of
MICs using a standardised bacterial inoculum. In this study MIC testing was
performed in triplicate to ensure that differences in macrolide sensitivities between
amniotic fluid and chorioamnion ureaplasmas were truly representative of variability
generated in vivo, as opposed to intra-assay variation. Whilst the MICs of
erythromycin and azithromycin demonstrated variability between cultured isolates,
218
roxithromycin MICs were notably higher in ureaplasmas isolated from the
chorioamnion, when compared to ureaplasmas isolated from the amniotic fluid.
Previous data have demonstrated that exposure of microorganisms to one
antimicrobial can result in resistance to other related compounds. Exposure of E.
coli to ketolide antimicrobials (erythromycin derivatives) resulted in the induction of
erm(C) resistance genes and subsequent resistance to macrolides (Bailey et al.
2008). However, in the data presented in this current study, roxithromycin MICs
were increased in chorioamnion ureaplasma isolates irrespective of whether they
were exposed to erythromycin or not.
Increased roxithromycin resistance in chorioamnion ureaplasma isolates could not
be attributed to nucleotide polymorphisms at positions previously described in the
23S rRNA gene, or due to the presence of macrolide resistance genes (erm(B),
msr(C), msr(D)). According to the most recent comprehensive review of the subject
(Roberts 2008), erm(B) has been described as the rRNA methylase gene
responsible for macrolide resistance in 33 bacterial genera. Recently, erm(B) was
detected in 17 Streptococcus agalactiae (group B streptococcus) isolates from
pregnant women in the third trimester (Brzychczy-Włoch et al. 2010). In these
isolates, erm(B) conferred high levels of macrolide resistance resulting in MICs
>256 µg/mL. Macrolide resistance due to the activity of drug efflux pumps encoded
by msr(C) and msr(D) has also been described in numerous microorganisms. Of
macrolide-resistant Enterococcus faecium environmental isolates, 92.2% were
found to carry the msr(C) gene (Diarra et al. 2010); and transformation of
Staphylococcus aureus RN4220 with a plasmid containing the msr(D) gene and its
promoter resulted in a 16 fold increase in erythromycin MICs (Reynolds and Cove
2005). The role of these macrolide resistance genes in ureaplasmas currently
remains unknown. Lu et al. (2010) detected erm(B), msr(B) and msr(D) in U.
parvum isolated from either cervical or urethral specimens; however, the presence
219
of these genes was associated with wide MIC ranges, indicating that they did not
confer macrolide resistance. Furthermore, the data presented in Chapter 4 of this
thesis demonstrated the presence of erm(B), msr(C) and msr(D) in all tested
chorioamnion ureaplasma isolates (not just in those with increased resistance to
roxithromycin), and therefore provides further evidence that these genes may not be
functional, or perhaps are not activated in ureaplasmas.
In this study, the molecular investigations into macrolide resistance in amniotic fluid
and chorioamnion ureaplasmas failed to elucidate mechanisms, which may have
resulted in increased roxithromycin resistance in chorioamnion ureaplasmas.
Similarly, Xiao et al. (2011) recently reported two human clinical U. parvum isolates
with erythromycin MICs of 8 µg/mL; however, neither of these isolates were found to
contain previously described genetic alterations or macrolide resistance genes that
could explain these elevated MICs. Therefore, it is possible that isolates with
reported low-level resistance (≤ 8 µg/mL) may not represent true macrolide
resistance in the absence of specific genetic markers. However, given the
significant differences in genome size and gene arrangement between ureaplasmas
and other microorganisms in which macrolide resistance was initially characterised,
it is possible that previously undescribed mechanisms may be responsible for low-
level ureaplasmal macrolide resistance. Since the secondary structure of 23S rRNA
can influence macrolide binding, mutations in other regions of the 23S rRNA gene
or in ribosomal proteins could alter the secondary structure and subsequently inhibit
macrolide binding by steric hindrance or by blocking access to key binding sites.
Until comparative sequencing of the full 23S rRNA gene of macrolide sensitive,
macrolide intermediate-resistant and macrolide resistant ureaplasmas occurs, it is
difficult to conclude what defines true macrolide resistance in these microorganisms.
A novel finding of this project was the large number of non-resistance associated
polymorphisms found in domain V of the 23S rRNA gene in chorioamnion
220
ureaplasmas, but not amniotic fluid ureaplasmas. The region of variable 23S rRNA
sequence demonstrated high nucleotide similarity to Pseudomonas spp. and it was
proposed this may represent a previous genetic recombination event in which a
fragment of pseudomonas 23S rRNA was integrated into the ureaplasma 23S rRNA
gene via horizontal gene transfer (HGT). Whilst it was not within the scope of this
PhD, further sequencing of these isolates to identify the positions at which foreign
DNA was inserted into the ureaplasma chromosome would provide evidence of
genetic transfer. As real time PCR failed to detect Pseudomonas spp. DNA within
chorioamnion specimens, it was proposed that these genetic events did not occur in
vivo in the sheep model. Rather, these non-random mutations may have already
been present within a small population of ureaplasma strains within the clinical U.
parvum isolate that was used for intra-amniotic injection, and this population may
have been strongly selected for within the chorioamnion.
There were numerous differences between ureaplasmas isolated from the amniotic
fluid and those isolated from the chorioamnion including: (i) macrolide MICs; (ii) the
presence of macrolide resistance genes and (iii) the nucleotide sequence of domain
V of the 23S rRNA gene. These significant differences demonstrate that different
ureaplasma sub-populations can be selected for at different anatomical sites. It is
remarkable that this selection resulted in the expansion of identical ureaplasma
populations isolated from the chorioamnion from different experimentally-infected
animals. However, similar findings have been previously reported using a model of
E. coli evolution. Cooper et al. (2003) sampled two E.coli populations (derived from
a common ancestor) after propagation in glucose-limited minimal medium for 20,
000 generations (approximately 3076 days, based on a doubling time of 3.7 hours
for these strains). DNA expression arrays of these populations revealed that the
expression of 59 genes had changed significantly in both populations. The
expression of all 59 genes was changed in the same direction relative to the
221
ancestral strain, indicating that two independently propagated E. coli populations
evolved in parallel. The data presented in this thesis also demonstrate in vivo
parallel selection of ureaplasmas within the chorioamnion of different animals;
however, these events have occurred over a very rapid time scale (after 70 days of
in utero infection).
There are significant differences in the microenvironment of the amniotic fluid, when
compared to that of the chorioamnion. Whilst these two anatomical sites are in
contact with each other, they are very different in terms of cellular composition,
protein expression, metabolites and presence of innate immune factors (Michaels et
al. 2007; Calvin and Oyen 2007). Therefore, it is possible that the
microenvironments associated with the amniotic fluid and chorioamnion could exert
different selective pressures resulting in the colonisation of these sites with specific
ureaplasma populations. The importance of the local microenvironment has been
discussed previously with respect to colonisation of the female lower genital tract
with sexually transmitted bacteria and viruses. Innate immune factors, the
expression of Toll-like receptors and the female sex hormones estrogen and
progesterone have all been determined to play a key role in whether sexually
transmitted pathogens are capable of establishing an infection, and whether
infections are exacerbated or down-modulated (Brabin 2002; Kaushic 2009;
Kaushic et al. 2010). Furthermore, the recent genome sequencing of a field strain of
Mycoplasma agalactiae revealed large differences to the genome of the previously
sequenced type strain (Nouvel et al. 2010). The two genomes were found to differ
by 130 kbp, with the field strain having increased mobile genetic elements and
expanded gene families that encode surface proteins. It was suggested that M.
agalactiae possesses a very dynamic genome, which is influenced by gene flow
between ruminant mycoplasmas and driven by localised genetic micro-events. The
data presented in this thesis suggests that the local microenvironment may also be
222
a key factor influencing the dynamics of microbial infections in the female upper
genital tract.
The variable sequences detected in chorioamnion ureaplasma isolates may
represent the first report of mosaic-like 23S rRNA structures in these
microorganisms, and may also be the first evidence of significant genetic variability
within the ureaplasmal core genome. As ribosomal RNA is thought to make up part
of the core genome, it is predicted that the genes encoding rRNAs are highly
conserved and mutations are unlikely to occur. Jain et al. (1999) postulated the
‘complexity hypothesis’, which states that the complexity of gene interactions is a
significant factor that influences HGT, therefore RNA genes (which have numerous
interactions with other genes/gene products) are highly unlikely to be involved in
HGT. However, Wang and Yang (2000) have since postulated the ‘simplified
complexity hypothesis’, which takes into account that the gene is not the smallest
unit of transfer in HGT, and that gene segments are readily transferred. Therefore, it
is possible that the variable 23S rRNA sequences found in chorioamnion
ureaplasma isolates represent fragments of 23S rRNA transferred via HGT,
resulting in a mosaic-like structure.
The implications of these findings are significant. Of particular concern, the ability of
ureaplasmas to integrate foreign DNA into their chromosome indicates a significant
potential for acquired antimicrobial resistance. The increase in ureaplasmal
macrolide and ciprofloxacin resistance reported over a 20 year period (Krausse and
Shubert 2010) supports this notion. Whilst this study is the first to report the rapid
selection of 23S rRNA ureaplasma variants in vivo, previous analysis of
mycoplasma rRNA sequences by Woese et al. (1985) determined that mycoplasma
rRNA has unique evolutionary characteristics and exhibits significantly higher
amounts of variation, when compared to phylogenetically related clostridia and other
low G+C Gram positive bacteria. This was recently confirmed by Marques et al.
223
(2011), who sequenced 34 field isolates of Ureaplasma diversum, a bovine
pathogen, and demonstrated a large number of single nucleotide polymorphisms
within the 16S rRNA gene of these isolates. Although researchers have been
unable to genetically manipulate ureaplasmas and mycoplasmas in vitro, these
microorganisms are thought to undergo rapid selection and genetic hypermutability
(Razin et al. 1998; Mrázek 2006). Ureaplasmas (and mycoplasmas) lack the methyl-
directed mismatch repair pathway encoded by the mutS, mutL and mutH genes
(Razin et al. 1998; Sachadyn 2010), which are responsible for the prevention of
point and frameshift mutations as well as recombinational processes (Metzgar and
Wills 2000). Defects in any one of these genes (or the absence of these genes) can
result in increased mutation rates and increased genetic recombination. The
importance of mutS in DNA repair was illustrated in natural populations of E. coli
and Salmonella enterica (LeClerc et al. 1996). 2.6% of isolates demonstrated a
hypermutable phenotype due to a single defect in the mutS allele. Ureaplasmas
also lack key regulators of the SOS pathway and the activity of uracil-DNA
glycosylase, which is responsible for the removal of uracil residues from DNA
arising by spontaneous deamination of cytosine residues, is also absent (Razin et
al. 1998). Therefore, it is not surprising that ureaplasmas can exhibit a
hypermutable phenotype and undergo genetic recombination.
The concept of genetic variability in ureaplasmas is not new. A major research focus
thus far has been genetic variability and recombination of surface exposed antigens
(such as the MBA and UU376) resulting in size and/or phase variable expression
(Zheng et al. 1995; Zimmerman et al. 2009). This PhD project investigated antigenic
variation of the MBA to determine: (i) if the MBA contributes to the severity of intra-
uterine inflammation; (ii) if variability of the MBA is a predictor of ureaplasmal
virulence; and (iii) the immune response generated against ureaplasma clones
expressing different MBA profiles. Previous studies have demonstrated that the
224
MBA undergoes size variation in vivo (Zheng et al. 1994; Zheng et al. 1995; Knox et
al. 2010) and phase variation in vitro (Monecke et al. 2003; Zimmerman et al. 2009).
It is hypothesised that size/phase variation of the MBA occurs as a mechanism to
avoid recognition by the host immune response, thus enabling these
microorganisms to cause chronic, intra-amniotic infections. However, these
phenomena have not been previously investigated.
In a severe combined immunodeficient mouse model, infection with two clonal
Borrelia turicatae isolates (each expressing a different sized variable major protein
(Vmp)), resulted in differences in arthritis, myocarditis and neurological disease
between animal groups (Cadavid et al. 1994). As the only discernable difference
between the two isolates was the size of Vmp, it was concluded that specific Vmp
size variants can be more virulent than others. Additionally, Talkington et al. (1989)
demonstrated that size variation of the variable antigen-1 (V-1) of Mycoplasma
pulmonis was associated with virulence in mice. These authors demonstrated a
significant correlation between the severity of lung lesions and the percentage of V-
1 size variants recovered from the respiratory tracts of mice intra-nasally infected
with M. pulmonis. In this PhD project, we investigated whether clonal U. parvum
serovar 6 isolates expressing different sized MBA proteins demonstrated
differences in virulence, specifically with respect to the severity of histological
chorioamnionitis. No differences were observed in pregnancy outcomes, the
number of ureaplasmas colonising individual tissues, chorioamnion histopathology,
host immune responses and the elaboration of MBA size variants in vivo between
animals infected with these clonal isolates. However, subtle differences were
observed in lung compliance and birth weight between animals infected with virulent
and avirulent ureaplasma strains. Intra-amniotic infection with the virulent
ureaplasma strain was also significantly associated with the presence of meconium-
stained amniotic fluid, which is thought to be an indicator of fetal distress. Our
225
findings were consistent with previous observations that low numbers of MBA size
variants produced within the amniotic fluid were associated with severe histological
chorioamnionitis (Knox et al. 2010); however, different MBA size variants were not
more associated with fetal inflammation and adverse fetal outcomes than others.
Microbial virulence factors are traditionally categorised as adhesins/invasins, toxins
or secretion systems (Casadevall and Pirofski 2001; Rasko and Sperandio 2010).
Monecke et al. (2003) published limited data suggesting that the MBA may function
in cytadherence to host cells; however, conclusive data were not presented to
support this hypothesis. Recent definitions of bacterial virulence factors have been
expanded to include microbial factors that are involved in evasion of the host
immune response via numerous mechanisms (reviewed by Finlay and McFadden
2006). This PhD project investigated the role of the MBA in immune evasion. We
demonstrated that MBA size variants were produced in vivo; however, strong innate
and adaptive immune responses were generated in some animals, suggesting that
size variability of the MBA does not prevent recognition by the host. Whilst antigenic
variation is a key microbial strategy for avoiding recognition by the host immune
response, variation of surface-exposed antigens can also interfere with antibody
function and binding due to changes in B cell epitopes (Deitsch et al. 2009). In utero
ureaplasma colonisation persisted in all animals, despite the up-regulation of pro-
inflammatory cytokines in chorioamnion tissue and the production of anti-
ureaplasma IgG antibodies. Therefore, MBA variation did not prevent recognition by
the host immune response, but may have prevented the eradication of ureaplasmas
by these immune factors.
This study has demonstrated for the first time that the MBA of clonal ureaplasma
isolates (originating from single CFUs) is phase variable in vitro, but only size
variation was observed in vivo. It is unknown what stimulates in vivo MBA size
variation in ureaplasmas; however, this study has clearly demonstrated that the
226
MBA responds very differently to selective pressures within in vitro and in vivo
environments. Zimmerman et al. (2009) recently proposed the molecular
mechanisms of MBA phase variation in U. parvum; however, the mechanisms of
MBA size variation have not been elucidated. The most common mechanism of
antigenic size variation is due to slipped strand mispairing of DNA polymerases in
repeat sequences, resulting in a misalignment between mother and daughter DNA
strands during replication. This misalignment could occur on either the leading or
lagging DNA strand, thus causing either an increase or decrease in the number of
repeat units (Levinson and Gutman 1987; van Belkum et al. 1998; van Belkum et al.
1999; van der Woude and Bäumler 2004). Size variation of surface-exposed
antigens via slipped strand mispairing occurs in several Mycoplasma spp. including
the variable adherence-associated antigen of M. hominis (Zhang and Wise 1996)
and the M. arthriditis T-cell mitogen superantigen (Tu et al. 2005). Therefore it is
also likely that size variation of the MBA occurs due to slipped strand mispairing of
repeat units in the 3’ region of the mba gene.
Another novel finding of this project was that the presence of maternal anti-
ureaplasma IgG antibodies was correlated with up-regulation of IL-1β, IL-6 and IL-8
in chorioamnion tissue. Up-regulation of these key pro-inflammatory cytokines in the
amniotic fluid and chorioamnion of pregnant women is associated with inflammatory
cascades, which lead to preterm birth (Goldenberg et al. 2000). Previous data have
also demonstrated that women with intra-amniotic ureaplasma infections are more
likely to deliver preterm if they have produced anti-ureaplasma antibodies (Quinn
1986; Horowitz et al. 1995). The data presented here are the first to demonstrate a
relationship between the innate and adaptive immune response in intra-amniotic
infection, and pregnant women who develop a similar systemic immune response
may be at high-risk for preterm delivery. The development of an innate and adaptive
immune response against intra-amniotic ureaplasma infection may be an important
227
predictor of adverse pregnancy outcomes. The host immune response has been
implicated as a key factor in the pathogenesis of Chlamydia trachomatis genital tract
infection in females (Darville and Hiltke 2010; Rusconi and Greub 2011). Hvid et al.
(2007) reported that IL-1 production by epithelial cells after C. trachomatis infection
resulted in severe tissue destruction in a human Fallopian tube organ culture model.
The addition of IL-1 receptor antagonist to human Fallopian tubes prior to C.
trachomatis infection prevented tissue damage, demonstrating that IL-1 plays an
important role in the sequelae associated with chlamydial infection. Therefore, it is
possible that the host immune response may be important in determining whether
intra-amniotic ureaplasma infections are asymptomatic or result in adverse
outcomes.
6.2 CONCLUSIONS
The data from this PhD project have demonstrated that U. parvum can chronically
colonise the amniotic fluid and fetus, despite antimicrobial treatment and the
development of a host immune response. These data have highlighted the
variability of U. parvum in vivo, with respect to sensitivity to macrolide
antimicrobials, 23S rRNA gene sequences, the presence of macrolide resistance
genes and the expression of MBA size variants within the amniotic fluid. Similarly,
the host immune response generated against intra-amniotic ureaplasma infection
was highly variable between animals; although a significant association was found
between increased IL-1β, IL-6 and IL-8 expression and the presence of anti-
ureaplasma IgG antibodies within maternal serum.
To address the hypothesis of this project, it was confirmed that maternal
erythromycin treatment is unable to eradicate intra-amniotic ureaplasma infection. It
appears that the ineffectiveness of this treatment is due to the placental barrier
preventing transfer of erythromycin from the maternal circulation to the amniotic fluid
228
and fetal circulation. This scenario is further complicated by the ability of
ureaplasmas to undergo rapid selection and genetic variation in vivo, resulting in U.
parvum isolates with variable MICs to macrolide antimicrobials. Furthermore,
continuous size variation of the MBA by ureaplasma populations may prevent the
eradication of these microorganisms by the host immune response within the
amniotic cavity.
A model of chronic, intra-amniotic ureaplasma infections is presented in Figure 6.1
After colonisation of the amniotic fluid and fetus, MBA antigenic variants are
generated and the local microenvironment may select for sub-populations of
ureaplasmas at various anatomical sites. If the host mounts an immune response
against intra-amniotic ureaplasma infection, it is unlikely that ureaplasmas will be
eradicated due to continual MBA size variation. However, the immune response
may lead to clinical chorioamnionitis, due to the increase in inflammatory cells and
pro-inflammatory cytokines in the chorioamnion. In those women who present with
clinical chorioamnionitis (ie. symptomatic chorioamnionitis), maternal erythromycin
treatment is unlikely to be effective due to the minimal placental transfer of
macrolides, and also potentially due to populations of ureaplasmas with mixed
antimicrobial sensitivities colonising the amniotic fluid and the chorioamnion. Intra-
amniotic ureaplasma infection will therefore persist; however, the increase in pro-
inflammatory cytokines (IL-1β, IL-6 and IL-8) will stimulate the production of matrix
metalloproteases and prostaglandins, which will result in weakening (and
subsequent rupture) of the fetal membranes and preterm birth. The preterm infant is
also likely to develop BPD due to a pro-inflammatory immune response within the
fetal lung. Alternatively, if a strong host immune response is not generated, intra-
amniotic ureaplasma infections may persist for lengthy periods without any clinical
evidence of chorioamnionitis. Even in the absence of clinical chorioamnionitis,
ureaplasma invasion of the amniotic cavity and chorioamnion may result in
229
Figure 6.1: Model of chronic intra-amniotic infection
ASYMPTOMATIC
CHORIOAMNIONITIS
Preterm rupture of
membranes
PRETERM BIRTH
NEONATAL BPD
Attachment to
sperm
Endometrium
colonisation
Haematogenous
spread Iatrogenic Ascending invasive
infection
Amniotic fluid colonisation and fetal infection
Elaboration of MBA variants & selection of
ureaplasma sub-populations by the local
microenvironment
HOST
IMMUNE
RESPONSE
NO HOST
IMMUNE
RESPONSE
Ureaplasmas not eradicated by immune
response due to MBA size variation
SYMPTOMATIC
CLINICAL
CHORIOAMNIONITIS
Maternal erythromycin
treatment
Ineffective due to
minimal placental
transfer and variable
MICs of ureaplasmas
Intra-amniotic infection
persists
Intra-amniotic infection
persists
Preterm rupture
of membranes
PRETERM BIRTH
± NEONATAL BPD
TERM BIRTH
± SEQUELAE
230
inflammatory cell influx and weak to moderate increases in pro-inflammatory
cytokines, leading to preterm birth. Women with asymptomatic intra-amniotic
ureaplasma infection may also deliver at term if ureaplasmas successfully avoid
recognition by elements of the innate immune response. In cases of clinically
asymptomatic, intra-amniotic ureaplasma infections, the neonate may or may not
develop long term sequelae, depending on the extent of inflammation and injury to
vulnerable organs (such as the lungs and brain).
The ureaplasmas are important pathogens of the upper genital tract during
pregnancy. Results from this PhD project suggest that the ineffectiveness of current
antimicrobial treatment protocols and the genetic variability of ureaplasmal core
genes (such as the 23S rRNA gene) and surface-exposed antigens (such as the
MBA) are key factors, which contribute to the ability of ureaplasmas to cause
chronic, intra-amniotic infections. As the host immune response alone is unable to
eliminate ureaplasmas from the amniotic cavity, more effective treatment options
are required to eradicate and/or prevent these opportunistic in utero infections.
6.3 FUTURE DIRECTIONS
Based on the results of this study, alternative methods of antimicrobial
administration to pregnant women must be explored. Direct injection or implantation
of antimicrobials into the amniotic fluid should be further investigated as a potential
method to bypass the placental barrier. Multi-drug therapies may also be required to
eradicate ureaplasma populations with variable MICs, and specific antimicrobial
susceptibility testing of amniotic fluid isolates should inform drug selection to
prevent the emergence of antimicrobial resistance. Given the often asymptomatic
nature of intra-amniotic ureaplasma infections, amniotic fluid collected from
pregnant women undergoing amniocentesis should be routinely tested for
ureaplasmas. As culture of these microorganisms is technically difficult and time
231
consuming, the development of rapid detection and serotyping methods are urgently
required.
Furthermore, the role of the MBA should be further investigated to determine if it is
essential for the establishment of chronic intra-amniotic infections. Experiments to
compare the virulence of ureaplasma isolates expressing the MBA or not
expressing the MBA would provide insight into the role of this surface-exposed
antigen in intra-amniotic ureaplasma infections. In vitro, MBA phase variation is
reversible after the removal of selective pressure (Monecke et al. 2003; Zimmerman
et al. 2009). However, it is not known if MBA expression and/or size variation is
essential for the establishment of in utero infections or colonisation. As the MBA is
the predominant antigen recognised by the host immune response, a MBA vaccine
may be an efficacious method to prevent intra-amniotic ureaplasma infections.
However, there are inherent problems with this approach due to the genetic
variability, hypermutability and rapid selection of ureaplasmas in vivo. Although
ureaplasma vaccines have not been trialled to date, it is unlikely that this approach
would be successful due to the continuous size variability of B cell epitopes within
the repeat region of the MBA. Furthermore, the presence of anti-ureaplasma
antibodies in maternal serum appears to be associated with adverse pregnancy
outcomes and thus may result in increased rates of preterm birth and spontaneous
abortion. Additionally, a vaccine would most likely eradicate commensal
ureaplasmas from the lower genital tract of females, which could potentially alter the
balance of vaginal flora and lead to bacterial vaginosis or other reproductive health
problems. Therefore, it is unlikely that a vaccine approach is feasible, and future
research should be focused on further characterisation of ureaplasma pathogenesis
and the development of effective screening and treatment options.
232
6.4 LITERATURE CITED
Antibiotic Expert Group (2010) Therapeutic Guidelines: Antibiotic Version 14. Melbourne: Therapeutic Guidelines Limited pp. 374-380.
Bailey M, Chettiath T, Mankin AS (2008) Induction of erm(C) expression by noninducing antibiotics. Antimicrob Agents Chemother 52: 866-874.
Brabin L (2002) Interactions of the female hormonal environment, susceptibility to viral infections, and disease progression. AIDS Patient Care STDs 16: 211-221.
Brzychczy-Włoch M, Gosiewski T, Bodaszewska M, Pabian W, Bulanda M, Kochan P, Strus M, Heczko PB (2010) Genetic characterization and diversity of Streptococcus agalactiae isolates with macrolide resistance. J Med Microbiol 59: 780-786.
Cadavid D, Thomas DD, Crawley R, Barbour AG (1994) Variability of a bacterial surface protein and disease expression in a possible mouse model of systemic Lyme Borreliosis. J Exp Med 179: 631-642.
Calvin SE, Oyen ML (2007) Microstructure and mechanics of the chorioamnion membrane with an emphasis on fracture properties. Ann N Y Acad Sci 1101: 166-185.
Casadevall A, Pirofski L (2001) Host-pathogen interactions: the attributes of virulence. J Infect Dis 184: 337-344.
Cassell GH, Waites KB, Watson HL, Crouse DT, Harasawa R (1993) Ureaplasma urealyticum intrauterine infection: role in prematurity and disease in newborns. Clin Microbiol Rev 6: 69-87.
Cooper TF, Rozen DE, Lenski RE (2003) Parallel changes in gene expression after 20 000 generations of evolution in Escherichia coli. Proc Natl Acad 100: 1072-1077.
Couce A, Blázquez J (2009) Side effects of antibiotics on genetic variability. FEMS Microbiol Rev 33: 531-538.
Darville T, Hiltke TJ (2010) Pathogenesis of genital tract disease due to Chlamydia trachomatis. J Infect Dis 201(S2): S114-S125.
Davies J, Davies D (2010) Origins and evolution of antibiotic resistance. Microbiol Mol Biol Rev 74: 417-433.
Davies J, Spiegelman GB, Yim G (2006) The world of subinhibitory antibiotic concentrations. Curr Opin Microbiol 9: 445-453.
Deitsch KW, Lukehart SA, Stringer JR (2009) Common strategies for antigenic variation by bacterial, fungal and protozoan pathogens. Nat Rev Microbiol 7: 493-503.
Diarra MS, Rempel H, Champagne J, MassonL, Pritchard J, Topp E (2010) Distribution of antimicrobial resistance and virulence genes in Enterococcus spp. and characterization of isolates from broiler chickens. Appl Environ Microbiol 76: 8033-8043.
233
Finlay BB, McFadden G (2006) Anti-immunology: evasion of the host immune system by bacterial and viral pathogens. Cell 124: 767-782.
Goldenberg RL, Hauth JC, Andrews WW (2000) Intrauterine infection and preterm delivery. New Engl J Med 342: 1500-1507.
Heikkinen T, Laine K, Neuvonen PJ, Ekblad U (2000) The transplacental transfer of the macrolide antibiotics erythromycin, roxithromycin and azithromycin. BJOG 107: 770-755.
Horowitz S, Mazor M, Horowitz J, Porath A, Glezerman M (1995) Antibodies to Ureaplasma urealyticum in women with intraamniotic infection and adverse pregnancy outcome. Acta Obstet Gynecol Scand 74: 132-136.
Hvid M, Baczynska A, Deleuran B, Fedder J, Knudsen HJ, Christiansen G, Birkelund S (2007) Interleukin-1 is the initiator of Fallopian tube destruction during Chlamydia trachomatis infection. Cell Microbiol 9: 2795-2803.
Jain R, Rivera MC, Lake JA (1999) Horizontal gene transfer among genomes: the complexity hypothesis. Proc Natl Acad Sci 96: 3801-3806.
Kaushic C (2009) The role of the local microenvironment in regulating susceptibility and immune responses to sexually transmitted viruses in the female genital tract. J Reprod Immunol 83: 168-172.
Kaushic C, Ferreira VH, Kafka JK, Nazli A (2010) HIV infection in the female genital tract: discrete influence of the local mucosal microenvironment. Am J Reprod Immunol 63: 566-575.
Keelan JA, Nitsos I, Saito M, Musk GC, Kemp MW, Timminis M, Li S, Yaegashi N, Newnham JP (2011) Maternal-amniotic-fetal distribution of macrolide antibiotics following intravenous, intramuscular, and intraamniotic administration in late pregnant sheep. Am J Obstet Gynecol 204: 1.e1-1.e8.
Kenyon SL, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group (2001) Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLE I randomised trial. Lancet 357: 979-988.
Kiefer L, Rubin A, McCoy JB, Foltz EL (1955) The placental transfer of erythromycin. Am J Obstet Gynecol 69: 147-177.
Knox CL, Allan JA, Allan JM, Edirisinghe WR, Stenzel D, Lawrence FA, Purdie D.M, Timms P (2003) Ureaplasma parvum and Ureaplasma urealyticum are detected in semen after washing before assisted reproductive technology procedures. Fertil Steril 80:921-929. Knox CL, Dando SJ, Nitsos I, Kallapur SG, Jobe AH, Payton D, Moss TJ, Newnham JP (2010) The severity of chorioamnionitis in pregnant sheep is associated with in vivo variation of the surface-exposed multiple-banded antigen/gene of Ureaplasma parvum. Biol Reprod 83: 415-426.
Knox CL, Timms P (1998). Comparison of PCR, nested PCR, and random amplified polymorphic DNA PCR for detection and typing of Ureaplasma urealyticum in specimens from pregnant women. J Clin Microbiol 36:3032-3039.
234
Krausse R, Schubert S (2010) In vitro activities of tetracyclines, macrolides, fluoroquinolones and clindamycin against Mycoplasma hominis and Ureaplasma spp. isolated in Germany over 20 years. Clin Microbiol Infect 16: 1649-1655. LeClerc JE, Li B, Payne WL, Cebula TA (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274: 1208-1211. Levinson G, Gutman GA (1987) Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Mol Biol Evol 4: 203-221.
Lu C, Ye TI, Zhu Gx, Feng Py, Ma H, Lu Rb, Lai W (2010) Phenotypic and genetic characteristics of macrolide and lincosamide resistant Ureaplasma urealyticum isolated in Guangzhou, China. Curr Microbio 61: 44-49.
Marques LM, Buzinhani M, Guimaraes AMS, Marques RCP, Farias ST, Neto RL, Yamaguti M, Oliveria RC, Timenetsky J (2011) Intraspecific sequence variation in 16S rRNA gene of Ureaplasma diversum isolates. Vet Microbiol (article in press).
Metzgar D, Wills C (2000) Evolutionary changes in mutation rates and spectra and their influence on the adaptation of pathogens. Microbes Infect 2: 1513-1522.
Michaels JE, Dasari S, Pereira L, Reddy AP, Lapidus JA, Lu X, Jacob T, Thomas A, Rodland M, Roberts CT Jr, Gravett MG, Nagalla SR (2007) Comprehensive proteomic analysis of the human amniotic fluid proteome: gestational age-dependent changes. J Proteone Res 6: 1244-1285.
Monecke S, Helbig JH, Jacobs E (2003) Phase variation of the multiple banded protein in Ureaplasma urealyticum and Ureaplasma parvum. Int J Med Microbiol 293: 203-211.
Mrázek J (2006) Analysis of distribution indicates diverse functions of simple sequence repeats in Mycoplasma genomes. Mol Biol Evol 23: 1370-1385.
Mylonas I (2011) Antibiotic chemotherapy during pregnancy and lactation period: aspects for consideration. Arch Gynecol Obstet 283: 7-18.
Nouvel LX, Sirand-Pugnet P, Marenda MS, Sagné E, Barbe V, Mangenot S, Schenowitz C, Jacob D, Barré A, Claverol S, Blanchard A, Citti C (2010) Comparative genomic and proteomic analyses of two Mycoplasma agalactiae strains: clues to the macro- and micro-events that are shaping mycoplasma diversity. BMC Genomics. 11: 86-105.
Novy MJ, Duffy L, Axthelm MK, Sadowsky DW, Witkin SS, Gravett MG, Cassell GH, Waites KB (2009) Ureaplasma parvum or Mycoplasma hominis as sole pathogens cause chorioamnionitis, preterm delivery, and fetal pneumonia in rhesus macaques. Reprod Sci 16: 56-70.
Philipson A, Sabath LD, Charles D (1973) Transplacental passage of erythromycin and clindamycin. New Engl J Med 288: 1219-1221.
Quinn PA (1986) Evidence of an immune response to Ureaplasma urealyticum in perinatal morbidity and mortality. Pediatr Infect Dis 5: S282-S287.
235
Rasko DA, Sperandio V (2010) Anti-virulence strategies to combat bacteria-mediated disease. Nat Rev Drug Discov 9: 117-128.
Razin S, Yogev D, Naot Y (1998) Molecular biology and pathogenicity of mycoplasmas. Microbiol Mol Biol Rev 62: 1094-1156.
Reyes L, Reinhard M, Brown MB (2009) Different inflammatory responses are associated with Ureaplasma parvum-induced UTI and urolith formation. BMC Infect Dis 9: 1-12.
Reynolds ED, Cove JH (2005) Resistance to telithromycin is conferred by msr(A), msr(C) and msr(D) in Staphylococcus aureus. J Antimicrob Chemother 56: 117-1182.
Roberts MC (2008) Update on macrolide-lincosamide-streptogramin, ketolide, and oxazolidinone resistance genes. FEMS Microbiol Lett 282: 147-159.
Romero R, Chaiworapongsa T, Espinoza J (2003) Micronutrients and intrauterine infection, preterm birth and the fetal inflammatory response syndrome. J Nutr 133: 1668S-1673S.
Rusconi B, Greub G (2011) Chlamydiales and the innate immune response: friend or foe? FEMS Immunol Med Microbiol 61: 231-244.
Sachadyn P (2010) Conservation and diversity of MutS proteins. Mutat Res 694: 20-30.
Saigal S, Doyle LW (2008) An overview of mortality and sequelae of preterm birth from infancy to adulthood. Lancet 371: 261-269.
Simmons LE, Rubens CE, Darmstadt GL, Gravett MG (2010) Preventing preterm birth and neonatal mortality: exploring the epidemiology, causes and interventions. Semin Perinatol 34: 408-415.
Tabor A, Alfirevic Z (2010) Update on procedure-related risks for prenatal diagnosis techniques. Fetal Diagn Ther 27: 1-7.
Talkington DF, Fallon MT, Watson HL, Thorp RK, Cassell GH (1989) Mycoplasma pulmonis V-1 surface protein variation: occurrence in vivo and association with lung lesions. Microb Pathog 7: 429-436.
Tu T, Clapper B, Schoeb TR, Elgavish A, Zhang J, Liu L, Yu H, Dybvig K (2005) Association of a major protein antigen of Mycoplasma arthriditis with virulence. Infect Immun 73: 245-249.
Van Belkum A, Scherer S, Van Alphen L, Verbrugh H (1998) Short-sequence DNA repeats in prokaryotic genomes. Microbiol Mol Biol Rev 62: 275-293.
Van Belkum A, van Leeuwen W, Scherer S, Verbrugh H (1999) Occurrence and structure-function relationship of pentameric short sequence repeats in microbial genomes. Res Microbiol 150: 617-626.
Van der Woude MW, Bäumler AJ (2004) Phase and antigenic variation in bacteria. Clin Microbiol Rev 17: 581-611.
236
Waites KB, Schelonka RL, Xiao L, Grigsby PL, Novy MJ (2009) Congenital and opportunistic infections: Ureaplasma species and Mycoplasma hominis. Semin Fetal Neonatal Med 14: 190-199.
Wang Y, Zhang Z (2000) Comparative sequence analyses reveal frequent occurrence of short segments containing an abnormally high number of non-random base variations in bacterial rRNA genes. Microbiology. 146: 2845-2854.
Woese CR, Stackebrandt E, Ludwig W (1985) What are mycoplasmas: the relationship of tempo and mode in bacterial evolution. J Mol Evol 21: 305-316.
Xiao L, Crabb DM, Duffy LB, Paralanov V, Glass JI, Hamilos DL, Waites KB (2011) Mutations in ribosomal proteins and ribosomal RNA confer macrolide resistance in human Ureaplasma spp. Int J Antimicrob Agents 37: 377-379.
Zhang Q, Wise KS (1996) Molecular basis of size and antigenic variation of a Mycoplasma hominis adhesis encoded by divergent vaa genes. Infect Immun 64: 2737-2744.
Zheng X, Teng LJ, Glass JI, Blanchard A, Cao Z, Kempf MC, Watson HL, Cassell GH (1994) Size variation of a major serotype-specific antigen of Ureaplasma urealyticum. Ann N Y Acad Sci 15: 299-301.
Zheng X, Teng LJ, Watson HL, Glass JI, Blanchard A, Cassell GH (1995) Small repeating units within the Ureaplasma urealyticum MB antigen gene encode serovar specificity and are associated with antigen size variation. Infect Immun 63: 891-898.
Zimmerman CR, Stiedl T, Rosengarten R, Spergser J (2009) Alternate phase variation in expression of two major surface membrane proteins (MBA and UU376) of Ureaplasma parvum serovar 3. FEMS Microbiol Lett 292: 187-193.