- Email: [email protected]
Chorioamniotic membranes constitute a competent barrier to group b streptococcus in vitro
European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165–169
Original Article
Chorioamniotic membranes constitute a competent barrier to group b streptococcus in vitro a, d b d Niels Kjaergaard *, Rikke Bek Helmig , Henrik Carl Schønheyder , Niels Uldbjerg , c a Estrid Stæhr Hansen , Hans Madsen a
Department of Obstetrics & Gynecology, Aalborg Hospital, DK-9000 Aalborg, Denmark b Department of Clinical Microbiology, Aalborg Hospital, DK-9000 Aalborg, Denmark c Department of Pathology, Aalborg Hospital, DK-9000 Aalborg, Denmark d Department of Obstetrics & Gynecology, Aarhus University Hospital, DK-8200 Aarhus, Denmark Received 28 October 1998; received in revised form 16 December 1998; accepted 16 December 1998
Abstract Objective: To study the penetration of group B streptococcus (GBS) through human chorioamniotic membranes in vitro. Study design: Chorioamniotic membranes from seventeen healthy women were mounted onto glass cylinders and placed in tissue culture trays constituting a two-compartment system with a maternal compartment internally and a fetal compartment externally. GBS from healthy pregnant women and from newborn babies with sepsis were added to the maternal compartment at densities from 10 7 to 10 9 colony forming units (cfu) per ml. Results. Irrespective of inoculum density, GBS was not recovered from the fetal compartment within a 20 h incubation period. By histology, micro-colonies of GBS were found on the maternal surface after 8 h, but invasion of the morphologically intact membranes was not observed. A five log reduction in cfu occurred in the maternal compartment with amnion when GBS were suspended in saline. Conclusion: In this in vitro model the membranes appear to constitute an effective barrier against ascending infection. 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Preterm delivery; Group B streptococcus; In vitro model; Chorioamniotic membranes
1. Introduction Preterm birth is a problem that affects 5–10% of pregnant women, but it contributes to 80% of perinatal morbidity and mortality, excluding congenital malformations [1]. The etiology of preterm labor is multifactorial, but there is increasing evidence of infection as a cause [2]. A number of different microorganisms have been found to be associated with preterm delivery [3]. The present study focused on group B strepto-coccus (GBS) as this bac-
*Tel.: 145-99-32-12-24; fax: 145-99-32-12-40. E-mail address: [email protected] (N. Kjaergaard)
terium is the most common cause of severe neonatal sepsis, and vaginal colonisation is found in 10–20% of healthy, pregnant women [4]. GBS has been recognised as a neonatal pathogen for more than 25 years [4]. GBS has also been proved to cause maternal illness, especially postcesarean endomyometritis [5,6]. Although most reports deal with morbidity in infants and mothers carrying GBS at delivery, some researchers have followed pregnancies longitudinally for the effects of GBS colonisation [5,7]. It was shown that overall morbidity, low birth weight / preterm delivery, and preterm rupture of the membranes are all significantly more common in GBS carriers. A variety of bacteria have been found in up to 4% of amniotic fluid samples from normal pregnant women with
0301-2115 / 99 / $ – see front matter 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S0301-2115( 99 )00009-3
166
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
intact membranes near term, but GBS is rarely isolated [3,8–10]. However, there have been only a few studies of bacterial attachment to and / or migration through the chorioamniotic membranes per se [11,12]. Therefore, this study was designed to establish an in vitro model for evaluation of the capacity of bacteria to migrate through chorioamniotic membranes, and test different strains of GBS in this model.
and the epithelium facing away from the lumen. The chorion was placed with the outer aspect, with identifiable fragments of maternal decidua facing the lumen (maternal compartment). Before the glass cylinders were placed in the wells, each one was tested with physiologic saline in order to confirm that the mounted membranes were waterproof. A culture aliquot of GBS was added to the maternal compartment as specified.
2. Materials and methods
2.3. Media
2.1. Specimens
(1) Physiologic saline, (2) nutrient broth with saccharose 10%, glucose 0.1%, hemoglobin 0.05% (Statens Serum Institut, Copenhagen, Denmark) and (3) Macrodex 60mg / ml with physiologic saline (Dextran 70, MEDISAN) were used. The cell wells were coated with human plasma albumin 50 mg / ml before use (Statens Serum Institut, Copenhagen, Denmark).
To evaluate the model fresh chorioamniotic membranes were obtained under sterile conditions from four women undergoing elective cesarean section at term. Entire membranes were mounted as described below and incubated under sterile conditions. Histological examination showed separation of amnion and chorion within a few hours and therefore all subsequent experiments were carried out with amnion and chorion separately. For the study of GBS penetration membranes were obtained from 11 healthy women undergoing elective cesarean section at term. Likewise, chorioamniotic membranes were obtained from six healthy women with spontaneous vaginal delivery at term. None had been treated with antibiotics during the pregnancy.
2.2. The model The membranes were removed en-bloc from the placenta and cleaned in sterile, physiologic saline to remove blood and debris. Amnion and chorion were easily separated by blunt finger dissection. Identification of membranes covering the cervix was not possible. The amnion and chorion were mounted onto glass cylinders (cross sectional area 1.54 cm 2 ) and placed in 6.9 ml reservoirs of polystyrene trays (no. 430345, Corning Glass Works, Corning, NY) (Fig. 1). This model has previously been used in the laboratory for metabolic studies [13]. The amnion was placed with the mesenchymal layer (the maternal side of amnion) facing the lumen,
2.4. Microorganisms Five different strains of GBS were tested. The strains were isolated from a neonate who died from an early-onset infection, a neonate without sign of serious infection, but with intrapartum sign of chorioamnionitis, a hyaluronidase-positive strain, serotype III, a hyaluronidasenegative, strain serotype III, and a hyaluronidase-positive strain, serotype III from a healthy, pregnant woman. Isolates nos. 3 and 4 were obtained from M. Kilian, University of Aarhus, and in the study by Hauge et al. [14] they were included as the high-virulence clonal type 52 and low-virulence clonal type 4, respectively. One routine urinary tract isolate of Escherichia coli was tested. The inoculum was harvested from overnight cultures on 5% horse blood agar and suspended in physiologic saline to McFarland no.$4, 100 ml of which (|10 9 cfu) was added to the maternal compartment. Samples of 10 ml were taken from the maternal and fetal compartments at 1, 4, 8 and 20 h. Samples were spread on 5% horse blood agar and inoculated at 378C with 5% carbon dioxide (STERICULT 200, FORMASCIENTIFIC, Marietta, Ohio, US) for 48 h. Each cylinder was sampled only once.
2.5. Microscopy of the membranes
Fig. 1. Diagram of the fetal membrane culture system.
For morphological studies the membrane from each well was fixed in 4% buffered formalin and processed for histological examination. Paraffin sections were stained with hematoxylin and eosin. The thickness of the membranes was measured by light microscopy with a precision of 10 mm.
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
167
3. Ethics The project was approved by the regional research ethics committee. Informed written and verbal consent was obtained from each patient.
4. Statistics The two sample T-test was used for continuous variables and the paired T-test for paired data. For categorical variables, significant differences were identified by x 2 test.
5. Results In experiments with the first four placentas the chorioamniotic membranes were mounted as a unit, but histological examination showed that the amnion and chorion separated from each other within a few hours (Fig. 2). Therefore, in incubation experiments with bacteria amnion and chorion were studied separately.
5.1. Colloid osmotic pressure and thickness of the membrane barrier
Fig. 3. Amnion with colonies of group B streptococcus on the maternal surface. a: fetal surface, c: maternal surface.
5.3. Counts of gbs in maternal compartment For controls without membranes counts of GBS remained $10 7 cfu / ml. With amniom and an inoculum ¯10 9 cfu to the maternal compartment a reduction in cfu5s was observed at 4 h of incubation and counts remained low at 8 h and 20 h. Counts #10 2 cfu / ml were found in 50% of maternal compartments after 20 h. Reduction in cfu’s was also found with chorion, but the reduction was significantly less pronounced (cfu #10 2 cfu / ml in 19% after 20 h; x 2 test, p50.01).
The membrane thickness with Macrodex versus physiologic saline as medium was determined with membranes from nine women. In physiologic saline the mean thickness of amnion was 86 mm (95% confidence interval (CI) 68–110 mm) and in Macrodex 52 mm (95% CI 40–60 mm), The thickness of chorion in physiologic saline was 450 mm (95% CI 340–560 mm) and in Macrodex 330 mm (95% CI 250–410 mm).
In three experiments with one E. coli isolate amnion and chorion were penetrated within 20 h. At termination of experiments counts of E. coli were similar in maternal and fetal compartments (¯10 9 cfu / ml).
5.2. Bacterial penetration
5.5. Histopathology of membranes
Irrespective of inoculum density and strain, GBS was not recovered from the fetal compartment.
Histopathologic examination of the amnion and chorion revealed no GBS within the membranes either after 1 h, 4 h, 8 h or 20 h. However, after incubation for 20 h colonies of GBS could be seen on maternal surfaces of the amnion (Figs. 3 and 4). There was also colonies on the surface of the chorion but of less density than observed for amnion.
5.4. E.coli experiments
6. Discussion
Fig. 2. A separation occurred between amnion and chorion after a few hours. a: fetal surface, b: amnion epithelium, c: maternal surface, e: connective tissue, f: chorion.
In this in vitro model GBS did not pass from the maternal side to the fetal side of the morphologically and mechanically intact membranes. Micro-colonies of GBS were found on the membrane surface and thus intimate contact between bacteria and tissue was documented. This is in accordance with other reports [12]. Hypothetically, a GBS concentration in the order of 10 7 to 10 8 cfu / ml would suffice for covering 5–50% of the membrane if a
168
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
Fig. 4. Colonies of group B streptococcus are located on the maternal surface of the amnion and penetrate into the loose connective tissue (e), but not through the amnion epithelium. a: fetal surface, e: connective tissue, c: maternal surface.
bacterial monolayer was formed. This calculation is based on the assumption that a GBS is a spherical cell with a diameter of 1 mm. We observed an effect of colloid osmotic pressure especially on amnion, and the addition of e.g. dextran to crystalloid solutions is advisable in order to allow bacteria to migrate a defined distance through membrane tissue. Fresh membranes obtained both by cesarean section and vaginal delivery constituted a competent barrier irrespective of the patient source or the strain of GBS tested. Strains of GBS isolated from different clinical situations were tested in our model. It seems possible that differences in attachment to, and penetration of, the membranes may contribute to differences in virulence of GBS [15,16]. It is noteworthy that no difference was discernible between a hyaluronidase-positive and a hyaluronidase-negative strain, as hyaluronidase is a putative virulence factor of streptococci facilitating their spread in tissue [15]. Gyr et al. [11] have previously reported penetration of entire chorioamniotic membranes by E. coli. We corroborated the results for E. coli in three experiments, but at variance with Gyr et al. [11] the amnion and chorion were studied independently. Conversely, GBS did not penetrate in our model. Motility and chemotropism are characteristic features of E. coli, which are lacking in GBS, and thus E. coli is at a great advantage in a penetration model. In experiments with GBS suspended in saline we recorded a $5 log reduction in viable numbers of GBS and more than 50% of maternal compartments did not show any growth when sampled at 20 h. This is consistent with a bactericidal effect which was most pronounced for amnion. However, the distribution of bacteria may also have been affected by varying binding to membranes as shown by colony density on amnion and chorion. There was further circumstantial evidence to support an antibacterial effect of membranes per se as we saw no microbial contamination, neither with membranes from cesarean sections nor vaginal deliveries.
Antibacterial properties are well described for amniotic fluid [17], but only a limited number of studies have addressed effects on bacteria of membranes. It is intriguing that for decades, chorioamniotic membranes have been applied therapeutically to e.g. ulcerated skin surfaces, peritoneum and the lacerated eye [18]. Our model holds promise for ascertaining interactions of chorioamniotic membranes with established and potential pathogens. However, it should be noted that inflammatory responses in neither the mother nor fetus are represented in this model. Schoonmaker et al. has demonstrated the effects of cervicovaginal isolates of GBS and S. aureus on measures of tensile strength, elasticity, and work to rupture of chorioamniotic membranes. Thus, it is possible that e.g. bacterial attachment and formation of micro-colonies would elicit inflammatory responses in the fetomaternal unit which could affect the competence of the barrier.
Acknowledgements We are indebted to Ms. Lis Hyttel and Ms. Maiken Jensen for excellent technical assistance. Financial support was received from Stinne & Martinus Sørensen’s Fund, the Northern Jutland Fund for Medical Research, and Aalborg Stift’s Fund.
References [1] Lewis R, Mercer BM. Adjunctive care of preterm labor-the use of antibiotics. Clin Obst Gyn 1995;38(4):755–70. [2] Lamont RF. New approaches in the management of preterm labour of infective aetiology. Br J Obst Gyn 1998;105:134–7. [3] Romero R, Quintero R et al. Amniotic fluid white blood cell count: A rapid and simple test to diagnose microbial invasion of the amniotic cavity and predict preterm delivery. Am J Obstet Gynecol 1991;165:821–30. [4] Gilbert GL, editor. Infectious Diseases: Challenge for the 1990s. Bailliere’s Clinical Obstetrics and Gynaecology, Vol. 7 / Number 1, March 1993. [5] Bobitt JR, Damato JD, Sakakini J. Perinatal complications in group B streptococcal carriers: A longitudinal study of prenatal patients. Am J Obstet Gynecol 1985;151:711–7. [6] Faro S. Group B beta-hemolytic streptococci and puerperal infections. Am J Obstet Gynecol 1981;139:686–9. [7] Regan JA, Chao S, James LS. Premature rupture of membranes, preterm delivery, and group B streptococcal colonisation of mothers. Am J Obstet Gynecol 1981;141:184–6. [8] Watts DH, Krohn MA, Hillier SL, Eschenbach DA. The association of occult amniotic fluid infection with gestational age and neonatal outcome among women in preterm labor. Obstet Gynecol 1992;79:351–7. [9] Harger JH, Meyer MP, Amortegui A, Macperson TA, Kaplan L, Mueller-Heubach E. Low incidence of positive amniotic fluid cultures in preterm labor at 27–32 weeks in the absence of clinical evidence of chorioamnionitis. Obstet Gynecol 1991;77:228–34. [10] Svare J, Andersen LF et al. Oro-genital microbial colonisation and threatening preterm delivery. Acta Obstet Gynecol Scand 1994;73:460–4. [11] Gyr TN, Malek A et al. Permeation of human chorioamniotic
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169 membranes by Escherichiae coli in vitro. Am J Obstet Gynecol 1994;170:223–7. [12] Galask RP, Varner MW, Petzold R, Wilbur SL. Bacterial attachment to the chorioamniotic membranes. Am J Obstet Gynecol 1984;148:915–28. [13] Helmig R, Roseblade CK, Uldbjerg N, Elder MG, Sullivan MHF. Control of fetal membrane prostaglandin E2 production by bacteria. Acta Obstet Gynecol Scand 1990;69:405–8. [14] Hauge M, Jespersgaard C, Poulsen K, Kilian M. Population structure of Streptococcus agalactiae reveals an association between specific evolutionary lineages and putative virulence factors but not disease. Infect Immun 1996;64:919–25.
169
[15] Helmig R, Halaburt JT, Uldbjerg N, Thomsen AC, Stenderup A. Increased cell adhesion of Group B Streptococci from preterm delivered infants with neonatal sepsis. Obstet Gynecol 1990;76:825– 7. [16] Pritchard DG, Lin B. Group B Streptococcal neuraminidase is actually a hyaluronidase. Infect Immun 1993;61:3234–9. [17] Thadepalli H, Bach VT, Davidson EC. Antimicrobial effect of amniotic fluid. Obstet Gynecol 1978;52:198–204. [18] Trelford JD, Trelford-Sauder M. The amnion in surgery, past and present. Am J Obstet Gynecol 1979;134:833–45.
Original Article
Chorioamniotic membranes constitute a competent barrier to group b streptococcus in vitro a, d b d Niels Kjaergaard *, Rikke Bek Helmig , Henrik Carl Schønheyder , Niels Uldbjerg , c a Estrid Stæhr Hansen , Hans Madsen a
Department of Obstetrics & Gynecology, Aalborg Hospital, DK-9000 Aalborg, Denmark b Department of Clinical Microbiology, Aalborg Hospital, DK-9000 Aalborg, Denmark c Department of Pathology, Aalborg Hospital, DK-9000 Aalborg, Denmark d Department of Obstetrics & Gynecology, Aarhus University Hospital, DK-8200 Aarhus, Denmark Received 28 October 1998; received in revised form 16 December 1998; accepted 16 December 1998
Abstract Objective: To study the penetration of group B streptococcus (GBS) through human chorioamniotic membranes in vitro. Study design: Chorioamniotic membranes from seventeen healthy women were mounted onto glass cylinders and placed in tissue culture trays constituting a two-compartment system with a maternal compartment internally and a fetal compartment externally. GBS from healthy pregnant women and from newborn babies with sepsis were added to the maternal compartment at densities from 10 7 to 10 9 colony forming units (cfu) per ml. Results. Irrespective of inoculum density, GBS was not recovered from the fetal compartment within a 20 h incubation period. By histology, micro-colonies of GBS were found on the maternal surface after 8 h, but invasion of the morphologically intact membranes was not observed. A five log reduction in cfu occurred in the maternal compartment with amnion when GBS were suspended in saline. Conclusion: In this in vitro model the membranes appear to constitute an effective barrier against ascending infection. 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Preterm delivery; Group B streptococcus; In vitro model; Chorioamniotic membranes
1. Introduction Preterm birth is a problem that affects 5–10% of pregnant women, but it contributes to 80% of perinatal morbidity and mortality, excluding congenital malformations [1]. The etiology of preterm labor is multifactorial, but there is increasing evidence of infection as a cause [2]. A number of different microorganisms have been found to be associated with preterm delivery [3]. The present study focused on group B strepto-coccus (GBS) as this bac-
*Tel.: 145-99-32-12-24; fax: 145-99-32-12-40. E-mail address: [email protected] (N. Kjaergaard)
terium is the most common cause of severe neonatal sepsis, and vaginal colonisation is found in 10–20% of healthy, pregnant women [4]. GBS has been recognised as a neonatal pathogen for more than 25 years [4]. GBS has also been proved to cause maternal illness, especially postcesarean endomyometritis [5,6]. Although most reports deal with morbidity in infants and mothers carrying GBS at delivery, some researchers have followed pregnancies longitudinally for the effects of GBS colonisation [5,7]. It was shown that overall morbidity, low birth weight / preterm delivery, and preterm rupture of the membranes are all significantly more common in GBS carriers. A variety of bacteria have been found in up to 4% of amniotic fluid samples from normal pregnant women with
0301-2115 / 99 / $ – see front matter 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S0301-2115( 99 )00009-3
166
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
intact membranes near term, but GBS is rarely isolated [3,8–10]. However, there have been only a few studies of bacterial attachment to and / or migration through the chorioamniotic membranes per se [11,12]. Therefore, this study was designed to establish an in vitro model for evaluation of the capacity of bacteria to migrate through chorioamniotic membranes, and test different strains of GBS in this model.
and the epithelium facing away from the lumen. The chorion was placed with the outer aspect, with identifiable fragments of maternal decidua facing the lumen (maternal compartment). Before the glass cylinders were placed in the wells, each one was tested with physiologic saline in order to confirm that the mounted membranes were waterproof. A culture aliquot of GBS was added to the maternal compartment as specified.
2. Materials and methods
2.3. Media
2.1. Specimens
(1) Physiologic saline, (2) nutrient broth with saccharose 10%, glucose 0.1%, hemoglobin 0.05% (Statens Serum Institut, Copenhagen, Denmark) and (3) Macrodex 60mg / ml with physiologic saline (Dextran 70, MEDISAN) were used. The cell wells were coated with human plasma albumin 50 mg / ml before use (Statens Serum Institut, Copenhagen, Denmark).
To evaluate the model fresh chorioamniotic membranes were obtained under sterile conditions from four women undergoing elective cesarean section at term. Entire membranes were mounted as described below and incubated under sterile conditions. Histological examination showed separation of amnion and chorion within a few hours and therefore all subsequent experiments were carried out with amnion and chorion separately. For the study of GBS penetration membranes were obtained from 11 healthy women undergoing elective cesarean section at term. Likewise, chorioamniotic membranes were obtained from six healthy women with spontaneous vaginal delivery at term. None had been treated with antibiotics during the pregnancy.
2.2. The model The membranes were removed en-bloc from the placenta and cleaned in sterile, physiologic saline to remove blood and debris. Amnion and chorion were easily separated by blunt finger dissection. Identification of membranes covering the cervix was not possible. The amnion and chorion were mounted onto glass cylinders (cross sectional area 1.54 cm 2 ) and placed in 6.9 ml reservoirs of polystyrene trays (no. 430345, Corning Glass Works, Corning, NY) (Fig. 1). This model has previously been used in the laboratory for metabolic studies [13]. The amnion was placed with the mesenchymal layer (the maternal side of amnion) facing the lumen,
2.4. Microorganisms Five different strains of GBS were tested. The strains were isolated from a neonate who died from an early-onset infection, a neonate without sign of serious infection, but with intrapartum sign of chorioamnionitis, a hyaluronidase-positive strain, serotype III, a hyaluronidasenegative, strain serotype III, and a hyaluronidase-positive strain, serotype III from a healthy, pregnant woman. Isolates nos. 3 and 4 were obtained from M. Kilian, University of Aarhus, and in the study by Hauge et al. [14] they were included as the high-virulence clonal type 52 and low-virulence clonal type 4, respectively. One routine urinary tract isolate of Escherichia coli was tested. The inoculum was harvested from overnight cultures on 5% horse blood agar and suspended in physiologic saline to McFarland no.$4, 100 ml of which (|10 9 cfu) was added to the maternal compartment. Samples of 10 ml were taken from the maternal and fetal compartments at 1, 4, 8 and 20 h. Samples were spread on 5% horse blood agar and inoculated at 378C with 5% carbon dioxide (STERICULT 200, FORMASCIENTIFIC, Marietta, Ohio, US) for 48 h. Each cylinder was sampled only once.
2.5. Microscopy of the membranes
Fig. 1. Diagram of the fetal membrane culture system.
For morphological studies the membrane from each well was fixed in 4% buffered formalin and processed for histological examination. Paraffin sections were stained with hematoxylin and eosin. The thickness of the membranes was measured by light microscopy with a precision of 10 mm.
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
167
3. Ethics The project was approved by the regional research ethics committee. Informed written and verbal consent was obtained from each patient.
4. Statistics The two sample T-test was used for continuous variables and the paired T-test for paired data. For categorical variables, significant differences were identified by x 2 test.
5. Results In experiments with the first four placentas the chorioamniotic membranes were mounted as a unit, but histological examination showed that the amnion and chorion separated from each other within a few hours (Fig. 2). Therefore, in incubation experiments with bacteria amnion and chorion were studied separately.
5.1. Colloid osmotic pressure and thickness of the membrane barrier
Fig. 3. Amnion with colonies of group B streptococcus on the maternal surface. a: fetal surface, c: maternal surface.
5.3. Counts of gbs in maternal compartment For controls without membranes counts of GBS remained $10 7 cfu / ml. With amniom and an inoculum ¯10 9 cfu to the maternal compartment a reduction in cfu5s was observed at 4 h of incubation and counts remained low at 8 h and 20 h. Counts #10 2 cfu / ml were found in 50% of maternal compartments after 20 h. Reduction in cfu’s was also found with chorion, but the reduction was significantly less pronounced (cfu #10 2 cfu / ml in 19% after 20 h; x 2 test, p50.01).
The membrane thickness with Macrodex versus physiologic saline as medium was determined with membranes from nine women. In physiologic saline the mean thickness of amnion was 86 mm (95% confidence interval (CI) 68–110 mm) and in Macrodex 52 mm (95% CI 40–60 mm), The thickness of chorion in physiologic saline was 450 mm (95% CI 340–560 mm) and in Macrodex 330 mm (95% CI 250–410 mm).
In three experiments with one E. coli isolate amnion and chorion were penetrated within 20 h. At termination of experiments counts of E. coli were similar in maternal and fetal compartments (¯10 9 cfu / ml).
5.2. Bacterial penetration
5.5. Histopathology of membranes
Irrespective of inoculum density and strain, GBS was not recovered from the fetal compartment.
Histopathologic examination of the amnion and chorion revealed no GBS within the membranes either after 1 h, 4 h, 8 h or 20 h. However, after incubation for 20 h colonies of GBS could be seen on maternal surfaces of the amnion (Figs. 3 and 4). There was also colonies on the surface of the chorion but of less density than observed for amnion.
5.4. E.coli experiments
6. Discussion
Fig. 2. A separation occurred between amnion and chorion after a few hours. a: fetal surface, b: amnion epithelium, c: maternal surface, e: connective tissue, f: chorion.
In this in vitro model GBS did not pass from the maternal side to the fetal side of the morphologically and mechanically intact membranes. Micro-colonies of GBS were found on the membrane surface and thus intimate contact between bacteria and tissue was documented. This is in accordance with other reports [12]. Hypothetically, a GBS concentration in the order of 10 7 to 10 8 cfu / ml would suffice for covering 5–50% of the membrane if a
168
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169
Fig. 4. Colonies of group B streptococcus are located on the maternal surface of the amnion and penetrate into the loose connective tissue (e), but not through the amnion epithelium. a: fetal surface, e: connective tissue, c: maternal surface.
bacterial monolayer was formed. This calculation is based on the assumption that a GBS is a spherical cell with a diameter of 1 mm. We observed an effect of colloid osmotic pressure especially on amnion, and the addition of e.g. dextran to crystalloid solutions is advisable in order to allow bacteria to migrate a defined distance through membrane tissue. Fresh membranes obtained both by cesarean section and vaginal delivery constituted a competent barrier irrespective of the patient source or the strain of GBS tested. Strains of GBS isolated from different clinical situations were tested in our model. It seems possible that differences in attachment to, and penetration of, the membranes may contribute to differences in virulence of GBS [15,16]. It is noteworthy that no difference was discernible between a hyaluronidase-positive and a hyaluronidase-negative strain, as hyaluronidase is a putative virulence factor of streptococci facilitating their spread in tissue [15]. Gyr et al. [11] have previously reported penetration of entire chorioamniotic membranes by E. coli. We corroborated the results for E. coli in three experiments, but at variance with Gyr et al. [11] the amnion and chorion were studied independently. Conversely, GBS did not penetrate in our model. Motility and chemotropism are characteristic features of E. coli, which are lacking in GBS, and thus E. coli is at a great advantage in a penetration model. In experiments with GBS suspended in saline we recorded a $5 log reduction in viable numbers of GBS and more than 50% of maternal compartments did not show any growth when sampled at 20 h. This is consistent with a bactericidal effect which was most pronounced for amnion. However, the distribution of bacteria may also have been affected by varying binding to membranes as shown by colony density on amnion and chorion. There was further circumstantial evidence to support an antibacterial effect of membranes per se as we saw no microbial contamination, neither with membranes from cesarean sections nor vaginal deliveries.
Antibacterial properties are well described for amniotic fluid [17], but only a limited number of studies have addressed effects on bacteria of membranes. It is intriguing that for decades, chorioamniotic membranes have been applied therapeutically to e.g. ulcerated skin surfaces, peritoneum and the lacerated eye [18]. Our model holds promise for ascertaining interactions of chorioamniotic membranes with established and potential pathogens. However, it should be noted that inflammatory responses in neither the mother nor fetus are represented in this model. Schoonmaker et al. has demonstrated the effects of cervicovaginal isolates of GBS and S. aureus on measures of tensile strength, elasticity, and work to rupture of chorioamniotic membranes. Thus, it is possible that e.g. bacterial attachment and formation of micro-colonies would elicit inflammatory responses in the fetomaternal unit which could affect the competence of the barrier.
Acknowledgements We are indebted to Ms. Lis Hyttel and Ms. Maiken Jensen for excellent technical assistance. Financial support was received from Stinne & Martinus Sørensen’s Fund, the Northern Jutland Fund for Medical Research, and Aalborg Stift’s Fund.
References [1] Lewis R, Mercer BM. Adjunctive care of preterm labor-the use of antibiotics. Clin Obst Gyn 1995;38(4):755–70. [2] Lamont RF. New approaches in the management of preterm labour of infective aetiology. Br J Obst Gyn 1998;105:134–7. [3] Romero R, Quintero R et al. Amniotic fluid white blood cell count: A rapid and simple test to diagnose microbial invasion of the amniotic cavity and predict preterm delivery. Am J Obstet Gynecol 1991;165:821–30. [4] Gilbert GL, editor. Infectious Diseases: Challenge for the 1990s. Bailliere’s Clinical Obstetrics and Gynaecology, Vol. 7 / Number 1, March 1993. [5] Bobitt JR, Damato JD, Sakakini J. Perinatal complications in group B streptococcal carriers: A longitudinal study of prenatal patients. Am J Obstet Gynecol 1985;151:711–7. [6] Faro S. Group B beta-hemolytic streptococci and puerperal infections. Am J Obstet Gynecol 1981;139:686–9. [7] Regan JA, Chao S, James LS. Premature rupture of membranes, preterm delivery, and group B streptococcal colonisation of mothers. Am J Obstet Gynecol 1981;141:184–6. [8] Watts DH, Krohn MA, Hillier SL, Eschenbach DA. The association of occult amniotic fluid infection with gestational age and neonatal outcome among women in preterm labor. Obstet Gynecol 1992;79:351–7. [9] Harger JH, Meyer MP, Amortegui A, Macperson TA, Kaplan L, Mueller-Heubach E. Low incidence of positive amniotic fluid cultures in preterm labor at 27–32 weeks in the absence of clinical evidence of chorioamnionitis. Obstet Gynecol 1991;77:228–34. [10] Svare J, Andersen LF et al. Oro-genital microbial colonisation and threatening preterm delivery. Acta Obstet Gynecol Scand 1994;73:460–4. [11] Gyr TN, Malek A et al. Permeation of human chorioamniotic
N. Kjaergaard et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 83 (1999) 165 – 169 membranes by Escherichiae coli in vitro. Am J Obstet Gynecol 1994;170:223–7. [12] Galask RP, Varner MW, Petzold R, Wilbur SL. Bacterial attachment to the chorioamniotic membranes. Am J Obstet Gynecol 1984;148:915–28. [13] Helmig R, Roseblade CK, Uldbjerg N, Elder MG, Sullivan MHF. Control of fetal membrane prostaglandin E2 production by bacteria. Acta Obstet Gynecol Scand 1990;69:405–8. [14] Hauge M, Jespersgaard C, Poulsen K, Kilian M. Population structure of Streptococcus agalactiae reveals an association between specific evolutionary lineages and putative virulence factors but not disease. Infect Immun 1996;64:919–25.
169
[15] Helmig R, Halaburt JT, Uldbjerg N, Thomsen AC, Stenderup A. Increased cell adhesion of Group B Streptococci from preterm delivered infants with neonatal sepsis. Obstet Gynecol 1990;76:825– 7. [16] Pritchard DG, Lin B. Group B Streptococcal neuraminidase is actually a hyaluronidase. Infect Immun 1993;61:3234–9. [17] Thadepalli H, Bach VT, Davidson EC. Antimicrobial effect of amniotic fluid. Obstet Gynecol 1978;52:198–204. [18] Trelford JD, Trelford-Sauder M. The amnion in surgery, past and present. Am J Obstet Gynecol 1979;134:833–45.