- Email: [email protected]
Cytokine abundance in placental tissues: Evidence of inflammatory activation in gestational membranes with term and preterm parturition
Cytokine abundance in placental tissues: Evidence of inflammatory activation in gestational membranes with term and preterm parturition Jeffrey A. Keelan, PhD,a Keith W. Marvin, PhD,a Timothy A. Sato, MSc,a Matthew Coleman, MD,b Lesley M.E. McCowan, MD,b and Murray D. Mitchell, DSca Auckland, New Zealand OBJECTIVES: This study of the changes in cytokine concentrations in gestational tissues from women with term and preterm labor was undertaken to assess the extent of inflammatory activation associated with spontaneous labor and delivery. STUDY DESIGN: Extracts of amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (n = 15), at term after labor (n = 15), and preterm (n = 31) were assayed for interleukin 1β, interleukin 6, and interleukin 8. RESULTS: In amniotic tissues of women delivered by spontaneous labor at term the median interleukin-6, interleukin-8, and interleukin-1β concentrations were 3.8 to 5.4 times those of tissues from women delivered at term without labor (P < .05, Mann-Whitney U test). Interleukin-6 and interleukin-8 concentrations were also significantly increased (3.3-4 times) in chorionic-decidual tissues. Marked increases (approximately 3-6 times) in the concentrations of all 3 cytokines were observed in both amniotic and chorionic-decidual tissues from women with preterm deliveries with respect to those from women with term deliveries after labor. Cytokine concentrations were significantly correlated within amniotic tissues from both women with term delivery after labor and women with preterm delivery and also in preterm chorionic-decidual tissues but not preterm placental tissues. Concentrations of cytokines in the tissues of women delivered preterm were not significantly affected by mode of delivery, treatment with antibiotics, or twin birth. In preterm tissues with evidence of intrauterine infection only amniotic interleukin-1β concentrations were significantly elevated (P < .05). Little or no labor-related change in cytokine concentrations was seen within placental tissues. CONCLUSIONS: Increased cytokine abundance in gestational membranes associated with labor supports the view that an inflammatory process is involved in both term and preterm labor. This process does not, however, appear to be evident in the villous placenta. (Am J Obstet Gynecol 1999;181:1530-6.)
Key words: Cytokines, inflammation, parturition, preterm labor
Human parturition is associated with altered rates of production of numerous paracrine and autocrine factors within gestational tissues, many of which have been proposed to play roles in regulating the onset and maintenance of parturition. Among these are inflammatory cytokines such as tumor necrosis factor α, interleukin 1β (IL-1β), interleukin 6 (IL-6), and interleukin 8 (IL-8).1-3 Although these proteins are expressed in the placenta
From the Departments of Pharmacology and Clinical Pharmacologya and Obstetrics and Gynaecology,b University of Auckland School of Medicine. Supported by the Auckland University Research Committee, New Zealand Lottery Health Grants Board, the Health Research Council of New Zealand, and the Research Centre in Reproductive Medicine. Received for publication December 4, 1998; revised April 21, 1999; accepted June 18, 1999. Reprint requests: Jeffrey A. Keelan, PhD, Department of Pharmacology and Clinical Pharmacology, University of Auckland School of Medicine, Private Bag 92019, Auckland, New Zealand. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/1/101143
1530
and extraplacental membranes in normal pregnancies during late gestation,3-12 attention has previously been focused primarily on their production and role in intrauterine infection–associated preterm labor.10-13 Several studies have reported findings of elevated concentrations of these cytokines in amniotic fluid taken from women who were delivered preterm, particularly those with an amniotic fluid culture positive for pathogens or with chorioamnionitis.12-14 Changes in cytokine production also occur after the onset of spontaneous labor at term. Increased amniotic fluid cytokine concentrations,6-9, 15, 16 increased production of IL-1β, IL-6, and IL-8 by gestational tissue explants,9, 16 and increased levels of IL-1β and IL-8 in gestational tissues have been reported with term labor.3, 5 These findings suggest that an increase in cytokine production in gestational tissues occurs with term labor, although the data are not entirely consistent.10, 11, 16 Although an increase in villous placental cytokine production after labor was noted in 1 study,16 this finding
Volume 181, Number 6 Am J Obstet Gynecol
contrasts with several other reports to the contrary.4, 9-11, 17 It remains unclear whether intrauterine cytokines play a role in the onset of labor or whether enhanced cytokine production is a sequela of the process of labor and delivery.18 The aims of this study were to characterize the changes in cytokine concentrations in gestational tissues seen with term and preterm labor to assess the extent of inflammatory activation in these tissues associated with spontaneous labor and delivery. The measurement of production of inflammatory mediators is notoriously susceptible to modulation by factors associated with cell dispersal, isolation, and culture.19, 20 Therefore, in an attempt to address the issue of labor-associated changes in cytokine production in gestational tissues, we measured cytokine abundance in tissue segments frozen shortly after delivery. A similar approach has been used in many classic endocrine studies to estimate changes in hormone production in source tissues. Although we acknowledge that this approach cannot differentiate between the production of a secreted substance and its accumulation from adjacent or distal sites, it can provide an accurate “snapshot” of concentrations of a substance in tissues under various physiologic conditions. It also overcomes the limitation of measurement of messenger ribonucleic acid abundance, which may not detect changes in protein production that are independent of altered gene expression. Material and methods Cytokines and the anticytokine antisera were obtained from R&D Systems (Minneapolis, Minn). Molecular Probes, Inc (Eugene, Ore), supplied 4-methylumbelliferyl phosphate. Pepstatin-A, phenylmethylsulfonylfluoride, and bicinchoninic acid were supplied by Sigma (St Louis, Mo). Bovine serum albumin free of fatty acids and protease inhibitors were supplied by SERVA Electrophoresis GmbH (Heidelberg, Germany). Tissue isolation and extraction. Maternal consent for collection of placental tissues was obtained from all women according to the guidelines of the Auckland Human Ethics Committee. Placentas were collected and processed within 60 minutes of delivery. Placentas were obtained from women who (1) were delivered by the cesarean route for the indications of previous cesarean delivery or malpresentation at term (37-42 weeks’ gestation) before labor (term no labor, n = 15), (2) had uncomplicated vaginal delivery after spontaneous labor at term (term spontaneous labor, n = 15), and (3) were delivered after spontaneous preterm (≤34 weeks’ gestation) labor (preterm delivery, n = 31). The maternal and neonatal clinical records were assessed for evidence of intrauterine infection as indicated by maternal fever (temperature >37.8°C), maternal or fetal tachycardia (pulse rates >100 and >160 beats/min, respectively), uterine
Keelan et al 1531
tenderness, foul-smelling liquor, histologically evident chorioamnionitis, or neonatal sepsis.21 The final assessment was graded as definite infection, probable infection, or no infection present. After manual stripping of the amnion from the chorion-decidua, sections of the membranes (approximately 2 cm2) were cut, washed briefly in cold phosphate-buffered sodium chloride solution to remove excessive amounts of blood, and rapidly frozen in liquid nitrogen. At the same time a small piece of villous placental tissue (approximately 1 cm3) was dissected free of maternal decidua from the placental bed and similarly frozen after a brief wash in phosphate-buffered sodium chloride solution. When they were ready for processing, samples were thawed in 2 mL lysis buffer (0.05-mol/L tris[hydroxymethyl]aminomethane hydrochloride, pH 6.8; 2-mmol/L ethylenediaminetetraacetic acid; 100 g/L sucrose; 50 µg/mL pepstatin-A; 0.1 g/L phenylmethylsulfonylfluoride; and 0.1 g/L sodium azide) at 4°C. They were initially processed with a Polytron (Brinkmann Instruments, Inc, Westbury, NY) homogenizer for 15 seconds on ice, followed by further homogenization with a Branson (Branson Ultrasonics Corporation, Danbury, Conn) sonicator for 30 seconds. A 1-mL aliquot of each homogenate was taken and centrifuged for 7 minutes at 10,000g; the supernatants were removed and stored frozen at –70°C before assay. A small aliquot of each supernatant was taken and diluted 20-fold in 0.1N sodium hydroxide for protein assay by the bicinchoninic acid method.22 The assay had a working range of approximately 10 to 1000 µg/mL calibrated against bovine serum albumin. Cytokine enzyme-linked immunosorbent assays. IL-6 and IL-8 were measured by sandwich enzyme-linked immunosorbent assays as described previously.23 The sensitivity of the IL-6 assay was 20 pg/mL; that of the IL-8 assay was 6 pg/mL. The interassay and intra-assay precisions (coefficients of variation) were 9.0% and 8.6% for IL-6 and 7.6% and 8.0% for IL-8. The IL-1β assay also used antisera from R&D Systems but used a streptavidin–alkaline phosphatase signal generation system combined with fluorescence detection (4-methylumbelliferyl phosphate) to enhance sensitivity. The IL-1β assay had a sensitivity of 6 pg/mL; it performed with interassay and intra-assay precisions of 11.7% and 3.9%, respectively. Recovery from tissue homogenates ranged from 99.4% to 144.5%. All enzyme-linked immunosorbent assays were calibrated against recombinant human cytokine standards. Tissue homogenates diluted linearly in all 3 enzyme-linked immunosorbent assays. Statistical analysis and representation of data. Tissue concentrations of cytokines were expressed in picograms per milligram of protein. The Mann-Whitney U test was used for the analysis of all nonparametric data. Spearman rank correlations were performed to assess sig-
1532 Keelan et al
December 1999 Am J Obstet Gynecol
Fig 1. Concentrations of IL-1β in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
Table I. Summary of cytokine concentrations in gestational tissue samples IL-1β
Amnion Term no labor Term spontaneous labor Preterm delivery Chorion-decidua Term no labor Term spontaneous labor Preterm delivery Placenta Term no labor Term spontaneous labor Preterm delivery
No.
Median
15 15 31
21.7 83.2* 423.9†
15 15 31
270.2 292.7 1468.5†
15 15 31
42.8 50.5 70.9
IL-6 10th-90th percentiles
IL-8
Median
10th-90th percentiles
Median
10th-90th percentiles
5-172 21-2697 112-6713
84.6 330.4* 2030.6†
47-174 179-899 439-5853
93.0 497.8* 2241.5†
0-321 325-1911 626-4722
51-559 70-696 175-11,022
84.9 278.1* 1220.4†
53-132 177-1547 193-3953
193.3 766.4* 2152.2†
85-1387 337-1847 494-3544
63.8 65.1 102.2
17-125 44-130 29-216
69.8 48.8 88.9
42-179 21-238 32-354
22-142 28-95 28-179
Data are all given in picograms per milligram of protein. *Significantly different from term no labor. †Significantly different from term spontaneous labor and term no labor.
nificant relationships between the concentrations of various cytokines in the tissues studied. P <.05 was considered to be significant. Results The gestational age at delivery for both term groups (term no labor and term spontaneous labor) ranged from 37 to 42 weeks. All women had singleton pregnancies, with the exception of 1 twin birth in the term no labor group. There was no clinical evidence of intrauterine infection in any woman in either term group. In the preterm delivery group 65% of the women (20/31) were delivered vaginally. Of the remainder 10 underwent emergency cesarean delivery in labor (duration of labor, 2-84 hours) and 1 underwent cesarean delivery before labor after preterm premature rupture of membranes. The gestational age at delivery in the preterm delivery group ranged from 24 to 34 weeks (median, 31 weeks).
Thirty-two percent of the women in the preterm delivery group (10/31) had twin pregnancies. Intravenous albuterol (Salbutamol) was administered to 52% of the women with preterm deliveries (16/31) and 55% (17/31) received antibiotics. All received antenatal corticosteroids. Spontaneous rupture of membranes occurred in 55% of the preterm delivery group (17/31); 35% of the preterm delivery group tissue specimens (10/31) were from pregnancies that were scored as “infection positive” (definite or probable intrauterine infection). Placental histologic information was available in only 6 of 31 cases; 2 of these had evidence of chorioamnionitis. IL-1β concentrations were increased after term labor in amniotic tissues (P < .05, term spontaneous labor vs term no labor) but not in chorionic-decidual or placental tissues (Fig 1 and Table I). More marked increases were observed in preterm deliveries, with further significant
Keelan et al 1533
Volume 181, Number 6 Am J Obstet Gynecol
Fig 2. Concentrations of IL-6 in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
increases detected in both amniotic and chorionic-decidual tissues (preterm delivery vs term spontaneous labor, P < .001 and P < .005 for amnion and chorion-decidua, respectively). There were no significant differences in placental IL-1β concentrations between those women who had spontaneous labor at term and those who were delivered before labor (term spontaneous labor vs term no labor), nor were there any differences between those women who were delivered preterm with evidence of intrauterine infection and those without such evidence. Concentrations of IL-1β in the amniotic, chorionic-decidual, and placental tissues of the term no labor group showed significant correlation with each other (Table II). In the term spontaneous labor group IL-1β concentrations were significantly correlated only between chorionic-decidual and placental tissues, whereas in the preterm group only amniotic and chorionic-decidual concentrations were significantly correlated (Table II). Tissue concentrations of IL-6 in the amnion and the chorion-decidua were also increased after term labor (term spontaneous labor vs term no labor, P = .0001 for both amnion and chorion-decidua; Fig 2 and Table I) and were further increased after preterm pregnancies (preterm delivery vs term spontaneous labor, P = .0001 and P < .005 for amnion and chorion-decidua, respectively). In contrast, IL-6 concentrations in placental tissues taken after labor were not significantly increased with respect to those from women who were delivered before labor either at term or preterm. IL-6 concentrations in the various tissues showed little or no correlation with each other in the 2 term groups (Table II). In the preterm delivery group, however, IL-6 concentrations in the amnion, the chorion-decidua, and the placenta were all significantly correlated. IL-8 concentrations were increased after term labor in the extraplacental membranes (term spontaneous labor
Table II. Spearman ranked correlation coefficients (rs) for cytokine concentrations between amnion and chorion-decidua, chorion-decidua and placenta, and placenta and amnion
Term no labor IL-1β IL-6 IL-8 Term spontaneous labor IL-1β IL-6 IL-8 Preterm delivery IL-1β IL-6 IL-8
Amnion and choriondecidua
Choriondecidua and placenta
Placenta and amnion
0.682* 0.484* NS
0.697† NS NS
0.699† NS 0.581*
NS NS NS
0.768‡ NS NS
NS NS NS
0.633‡ 0.525‡ 0.389*
NS 0.501† NS
NS 0.602† 0.421*
In the interests of clarity only significant correlations are displayed. NS, Not significant. *P < .05. †P < .01. ‡P < .005.
vs term no labor, P = .0001 and P < .005 in amnion and chorion-decidua, respectively; Fig 3 and Table I). More marked increases were observed in amniotic and chorionic-decidual tissues from preterm pregnancies (preterm delivery versus term spontaneous labor, P < .0005 and P < .005 for amnion and chorion-decidua, respectively). Again the IL-8 concentrations in placental tissues were not significantly increased after labor at term or preterm. There was a significant correlation between IL-8 concentrations in amniotic and placental tissues in the term no labor group and also between placental and amniotic or chorionic-decidual concentrations in the preterm delivery group (Table II).
1534 Keelan et al
December 1999 Am J Obstet Gynecol
Fig 3. Concentrations of IL-8 in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
Table III. Spearman ranked correlation coefficients (rs) between cytokine concentrations in amniotic, chorionicdecidual, and placental tissue homogenates
Amnion Term no labor Term spontaneous labor Preterm delivery Chorion-decidua Term no labor Term spontaneous labor Preterm delivery Placenta Term no labor Term spontaneous labor Preterm delivery
IL-1β and IL-6
IL-1β and IL-8
IL-6 and IL-8
NS 0.818* 0.516*
NS 0.532† 0.794*
NS NS 0.727*
NS NS 0.595‡
NS NS 0.511‡
0.575† NS 0.598‡
NS NS NS
NS NS 0.378†
NS 0.586† NS
In the interests of clarity only significant correlations are displayed. NS, Not significant. *P < .005. †P < .05. ‡P < .01.
Concentrations of the 3 cytokines showed little intratissue correlation with each other in the term no labor group (Table III). In the term spontaneous labor group, however, amniotic IL-1β concentrations were significantly correlated with those of IL-6 and IL-8, whereas in the preterm delivery group all 3 cytokines showed significant correlations with each other within amniotic and chorionic-decidual tissues. Overall, cytokine concentrations showed the poorest correlations within placental tissues. Concentrations of cytokines in the amniotic, chorionic-decidual, and placental tissues were not significantly affected by subdivision of the preterm delivery tissues on
the basis of mode of delivery, treatment with antibiotics, or twin birth. In the amnion the IL-6 and IL-8 concentrations were significantly lower in pregnancies with spontaneous rupture of membranes than in those without (IL6, median, 1346 pg/mg protein; 10th-90th percentiles, 346-3851 pg/mg protein with spontaneous rupture of membranes; vs median, 2674 pg/mg protein; 10th-90th percentiles, 1352-7118 pg/mg protein without spontaneous rupture of membranes; IL-8, median, 1654 pg/mg protein; 10th-90th percentiles, 568-3745 pg/mg protein; vs median, 2862 pg/mg protein; 10th-90th percentiles, 1933-5382 pg/mg protein; P < .02 for both). Unexpectedly, when the tissues were subdivided on the basis of evidence of intrauterine infection, only IL-1β concentrations in the amnion were significantly elevated in the infection-positive subgroup (median, 4083 pg/mg protein; 10th-90th percentiles, 133-11,582 pg/mg protein infected; vs median, 251 pg/mg protein; 10th-90th percentiles, 82-1878 pg/mg protein noninfected; P < .05). The IL-6 concentrations in the infection-positive group were also modestly elevated compared with those in the noninfected group (median, 2621 pg/mg; 10th-90th percentiles, 656-2793 pg/mg infected; vs median, 1214 pg/mg; 10th-90th percentiles, 297-2021 pg/mg noninfected), but this difference did not reach statistical significance. Similar trends were seen in the chorionic-decidual samples, but again these did not reach statistical significance (IL-1β, median, 2609 pg/mg; 10th-90th percentiles, 350-11,252 pg/mg infected; vs median, 1465 pg/mg; 10th-90th percentiles, 124-10,399 pg/mg noninfected). Amniotic and chorionic-decidual cytokine concentrations in the infection-negative preterm delivery subgroup remained significantly elevated with respect to those of the term spontaneous labor group (P < 0.05, P < .005, and P < .005 for IL-1β, IL-6, and IL-8, respec-
Volume 181, Number 6 Am J Obstet Gynecol
tively; both tissues). No significant differences or apparent trends were observed in cytokine concentrations in the placental tissues with or without infection or spontaneous rupture of membranes. Comment This is the first comprehensive study of cytokine concentrations in gestational tissues from women delivered at term and preterm with and without labor. The results demonstrate increased levels of inflammatory cytokines in the gestational membranes of women with term or preterm labor. These findings support the view of labor as an inflammatory process and are consistent with the findings of increased cytokine production by gestational tissues after labor at term6, 7, 17 and preterm.10-12 Few previous studies have attempted to measure cytokine abundance in gestational tissues. In this study concentrations of IL-1β in decidua and placenta were higher than those found by Ammala et al6 but close to those reported by Kauma et al.1 Although concentrations of IL-6 and IL-8 have not been reported previously in these tissues in comparable units, we expected IL-8 concentrations to be in excess of those of IL-6 because these tissues produce more IL-8 than IL-6.23 However, concentrations of IL-6 in the amnion and the placenta were similar to those of IL-8, although their comparative levels in the chorion-decidua were closer to those anticipated. IL1β protein has previously been detected in the amnion by immunohistochemical methods7 but not quantified. However, IL-1β secretion and messenger ribonucleic acid expression by the amnion have been reported to be low.3, 5 In this study amniotic IL-1β concentrations were readily detectable, even in tissues from women who did not have labor. A possible explanation for this finding is that the amnion might accumulate cytokines originating from the chorion-decidua. The evidence for this hypothesis is equivocal, however, because, although amniotic and chorionic-decidual concentrations of IL-1β in tissues from the term no labor and preterm delivery groups were significantly correlated, those in tissues from the term spontaneous labor group were not. Our data indicate that the placenta is excluded from the inflammatory processes that appear to occur with labor in the gestational membranes. This finding is in agreement with those of Opsjøn et al,9 Gunn et al,4 and Matsuzaki et al,11 who also found no change in placental IL-6 content or production after labor at term. We have expanded on their findings to include determinations of IL-1β and IL-8 concentrations, and we can report that these cytokines also exhibit no significant changes after labor at term or preterm. Our findings contrast, however, with 2 reports of elevated IL-6 and IL-8 production in placentas from women with intrauterine infection at delivery.10, 11 Histologic studies have shown that sporadic neutrophilic invasion of the chorionic plate is apparent in a
Keelan et al 1535
proportion of placentas delivered at term.8 Our samples were of villous placenta and thus may not have been deep enough into the plate to sample regions where inflammatory changes might have taken place. Alternatively, the glucocorticoids administered to all women who were delivered preterm may have inhibited placental cytokine production, thereby reducing cytokine concentrations in this tissue. These considerations notwithstanding, the lack of changes in cytokine concentrations found in the placenta is in striking contrast to the changes found in the membranes. The marked increases in concentrations of cytokines in the extraplacental membranes after normal spontaneous labor suggest that upward regulation of cytokine production in these tissues is a parturition-dependent event that is independent of a pathogen-driven response or gestational maturity. Whether this process plays a causal role in the onset of parturition or is merely a response to stresses or processes during parturition remains unanswered by these studies. The possibility also cannot be excluded that the inflammatory response that we detected at term is a result of subclinical infection or is a response to exposure of the membranes to vaginal flora during labor.18 Our findings do not support these hypotheses, however, because under those circumstances the maternal face of the membranes would be expected to show the biggest response, which was not the case (median cytokine concentrations in the preterm delivery group were 4.5- to 6-fold those of the term spontaneous labor group in the amnion and 2.8- to 5-fold in the chorion-decidua). Several reports have documented infiltration of the cervix and lower uterine tract with neutrophils and macrophages before labor,24, 25 possibly in response to chemotactic agents such as IL-8,26 suggesting that some form of inflammatory activation may be involved in the mechanism of normal parturition. In our view the evidence to date is supportive of the hypothesis that cytokines are involved in the mechanism of normal parturition. The findings in this study of elevated levels of all 3 cytokines in the tissues from women with preterm delivery, irrespective of apparent infection status, support the contention that most preterm deliveries are associated with an inflammatory response in the extraplacental membranes unaccompanied by overt clinical signs of infection or chorioamnionitis. The fact that cytokine concentrations in the tissues from women who were delivered preterm were higher than those observed after term labor (despite glucocorticoid therapy) indicates that the inflammatory responses observed are part of the pathogenesis of preterm labor rather than merely a response to parturition. Several reports, including those based on cytokine measurements in amniotic fluid, have suggested that subclinical infection is a significant cause, if not the major cause, of preterm labor and delivery.2, 13, 14, 21, 27
1536 Keelan et al
Our findings are consistent with this view. In our group of preterm tissues categorized as infection-positive, only IL-1β concentrations in the amnion were significantly increased with respect to those in the infection-negative group. Our ability to accurately diagnose pregnancies with intrauterine infection was limited, however, and this deficiency may well have obscured some differences in the data. Other studies have highlighted the difficulty in comprehensively diagnosing intrauterine infection or chorioamnionitis, even with the aid of extensive histologic studies.21, 27, 28 This study supports the view that an inflammatory reaction takes place in the gestational membranes in a significant proportion of women who are delivered preterm. Our findings also support the hope that the development of effective strategies for the diagnosis and treatment of intrauterine inflammatory activation, with its strong association with preterm labor, will significantly improve the neonatal morbidity and mortality rates associated with preterm delivery. We gratefully acknowledge the assistance of the nursing and theater staff at National Women’s Hospital with the collection of the placental tissues and the excellent technical assistance of Ren-Li Zhou. We also thank Dr M. Battin for screening records for neonatal sepsis.
December 1999 Am J Obstet Gynecol
11.
12.
13.
14.
15.
16.
17.
18.
REFERENCES
1. Kauma S, Matt D, Strom S, Eierman D, Turner T. Interleukin-1 beta, human leukocyte antigen HLA-DR alpha, and transforming growth factor-beta expression in endometrium, placenta, and placental membranes. Am J Obstet Gynecol 1990;163:14307. 2. Dudley DJ, Trautman MS. Infection, inflammation, and contractions: the role of cytokines in the pathophysiology of preterm labor. Semin Reprod Endocrinol 1994;12:263-73. 3. Dudley DJ, Colmer D, Mitchell MD, Trautman MS. Inflammatory cytokine mRNA in human gestational tissues: implications for term and preterm labor. J Soc Gynecol Invest 1996;3:328-35. 4. Gunn L, Hardiman P, Tharmaratnam S, Lowe D, Chard T. Measurement of interleukin-1 alpha and interleukin-6 in pregnancy-associated tissues. Reprod Fertil Dev 1996;8:1069-73. 5. Menon R, Swan KF, Leyden TW, Rote NS, Fortunato SJ. Expression of inflammatory cytokines (interleukin-1 beta and interleukin-6) in amniochorionic membranes. Am J Obstet Gynecol 1995;172:493-500. 6. Ammala M, Nyman T, Salmi A, Rutanen EM. The interleukin-1 system in gestational tissues at term: effect of labour. Placenta 1997;18:717-23. 7. Osmers RGW, Blasr J, Kuhn W, Tschesche T. Interleukin-8 synthesis and the onset of labor. Obstet Gynecol 1995;86:223-9. 8. Halgunset J, Johnsen H, Kjøllesdal AM, Qvigstad E, Espevik T, Austgulen R. Cytokine levels in amniotic fluid and inflammatory changes in the placenta from normal deliveries at term. Eur J Obstet Gynecol Reprod Biol 1994;56:153-60. 9. Opsjøn SL, Wathen NC, Tingulstad S, Wiedswang G, Sundan A, Waage A, et al. Tumor necrosis factor, interleukin-1, and interleukin-6 in normal human parturition. Am J Obstet Gynecol 1993;169:397-404. 10. Shimoya K, Matsuzaki N, Taniguchi T, Kameda T, Koyama M, Neki R, et al. Human placenta constitutively produces inter-
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
leukin-8 during pregnancy and enhances its production in intrauterine infection. Biol Reprod 1992;47:220-6. Matsuzaki N, Taniguchi T, Shimoya K, Neki R, Okada T, Saji F, et al. Placental interleukin-6 production is enhanced in intrauterine infection but not in labor. Am J Obstet Gynecol 1993;168:94-7. Stallmach T, Hebisch G, Joller-Jemelka HI, Orban P, Schwaller J, Engelmann M. Cytokine production and visualized effects in the feto-maternal unit: quantitative and topographic data on cytokines during intrauterine disease. Lab Invest 1995;73:384-92. Romero R, Yoon BH, Mazor M, Gomez R, Gonzalez R, Diamond MP, et al. A comparative study of the diagnostic performance of amniotic fluid glucose, white blood cell count, interleukin-6, and Gram stain in the detection of microbial invasion in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 1993;169:839-51. Coultrip LL, Lien JM, Gomez R, Kapernick P, Khoury A, Grossman JH. The value of amniotic fluid interleukin-6 determination in patients with preterm labor and intact membranes in the detection of microbial invasion of the amniotic cavity. Am J Obstet Gynecol 1994;171:901-11. Laham N, Rice GE, Bishop GJ, Ransome C, Brennecke SP. Interleukin-8 concentrations in amniotic fluid and peripheral venous plasma during human pregnancy and parturition. Acta Endocrinol 1993;129:220-4. Steinborn A, Günes H, Halberstadt E. Signal for term parturition is of trophoblast and therefore of fetal origin. Prostaglandins 1995;50:237-52. Laham N, Brennecke SP, Bendtzen K, Rice GE. Differential release of interleukin-6 from human gestational tissues in association with labour and in vitro endotoxin treatment. J Endocrinol 1996;149:431-9. Casey ML, MacDonald PC. Cytokines in the human placenta, fetal membranes, uterine decidua, and amniotic fluid. In: Rice GE, Brennecke SP, editors. Molecular aspects of placental and fetal membrane autacoids. Boca Raton (FL): CRC Press; 1993. p. 361-94. Kauma SW, Walsh SW, Nestler JE, Turner TT. Interleukin-1 is induced in the human placenta by endotoxin and isolation procedures for trophoblasts. J Clin Endocrinol Metab 1992;75:951-5. Lonsdale LB, Elder MG, Sullivan MH. A comparison of cytokine and hormone production by decidual cells and tissue explants. J Endocrinol 1996;151:309-13. Gibbs RS, Romero R, Hillier SL, Eschenbach DA, Sweet RL. A review of premature birth and subclinical infection. Am J Obstet Gynecol 1992;166:1515-28. Redinbaugh MG, Turley RB. Adaption of the bicinchoninic acid protein assay for use with microtiter plates and sucrose gradient fractions. Anal Biochem 1986;153:267-71. Keelan JA, Sato T, Mitchell MD. Interleukin (IL)-6 and IL-8 production by human amnion cells in vitro: regulation by cytokines, growth factors, glucocorticoids, phorbol esters and bacterial lipopolysaccharide. Biol Reprod 1997;57:1438-44. Junqueira LC, Zugaib M, Montes GS, Toledo OM, Krisztán RM, Shigihara KM. Morphologic and histochemical evidence for the occurrence of collagenolysis and for the role of neutrophilic polymorphonuclear leukocytes during cervical dilation. Am J Obstet Gynecol 1980;138:273-81. Bokström H, Brännström M, Alexandersson M, Norström A. Leukocyte subpopulations in the human uterine cervical stroma at early and term pregnancy. Hum Reprod 1997:12;586-90. Osmers RG, Adelmann-Grill BC, Rath W, Stuhlsatz HW, Tschesche H, Kuhn W. Biochemical events in cervical ripening and dilatation during pregnancy and parturition. J Obstet Gynecol 1995;21:185-94. Zlatnik FJ, Gellhause TM, Benda JA, Koonitz FP, Burmeister LF. Histologic chorioamnionitis, microbial infection, and premature labor. Obstet Gynecol 1990;76:335-59. Quinn PA, Butany J, Taylor J, Hannah W. Chorioamnionitis: its association with pregnancy outcome and microbial infection. Am J Obstet Gynecol 1987;156:379-87.
Key words: Cytokines, inflammation, parturition, preterm labor
Human parturition is associated with altered rates of production of numerous paracrine and autocrine factors within gestational tissues, many of which have been proposed to play roles in regulating the onset and maintenance of parturition. Among these are inflammatory cytokines such as tumor necrosis factor α, interleukin 1β (IL-1β), interleukin 6 (IL-6), and interleukin 8 (IL-8).1-3 Although these proteins are expressed in the placenta
From the Departments of Pharmacology and Clinical Pharmacologya and Obstetrics and Gynaecology,b University of Auckland School of Medicine. Supported by the Auckland University Research Committee, New Zealand Lottery Health Grants Board, the Health Research Council of New Zealand, and the Research Centre in Reproductive Medicine. Received for publication December 4, 1998; revised April 21, 1999; accepted June 18, 1999. Reprint requests: Jeffrey A. Keelan, PhD, Department of Pharmacology and Clinical Pharmacology, University of Auckland School of Medicine, Private Bag 92019, Auckland, New Zealand. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/1/101143
1530
and extraplacental membranes in normal pregnancies during late gestation,3-12 attention has previously been focused primarily on their production and role in intrauterine infection–associated preterm labor.10-13 Several studies have reported findings of elevated concentrations of these cytokines in amniotic fluid taken from women who were delivered preterm, particularly those with an amniotic fluid culture positive for pathogens or with chorioamnionitis.12-14 Changes in cytokine production also occur after the onset of spontaneous labor at term. Increased amniotic fluid cytokine concentrations,6-9, 15, 16 increased production of IL-1β, IL-6, and IL-8 by gestational tissue explants,9, 16 and increased levels of IL-1β and IL-8 in gestational tissues have been reported with term labor.3, 5 These findings suggest that an increase in cytokine production in gestational tissues occurs with term labor, although the data are not entirely consistent.10, 11, 16 Although an increase in villous placental cytokine production after labor was noted in 1 study,16 this finding
Volume 181, Number 6 Am J Obstet Gynecol
contrasts with several other reports to the contrary.4, 9-11, 17 It remains unclear whether intrauterine cytokines play a role in the onset of labor or whether enhanced cytokine production is a sequela of the process of labor and delivery.18 The aims of this study were to characterize the changes in cytokine concentrations in gestational tissues seen with term and preterm labor to assess the extent of inflammatory activation in these tissues associated with spontaneous labor and delivery. The measurement of production of inflammatory mediators is notoriously susceptible to modulation by factors associated with cell dispersal, isolation, and culture.19, 20 Therefore, in an attempt to address the issue of labor-associated changes in cytokine production in gestational tissues, we measured cytokine abundance in tissue segments frozen shortly after delivery. A similar approach has been used in many classic endocrine studies to estimate changes in hormone production in source tissues. Although we acknowledge that this approach cannot differentiate between the production of a secreted substance and its accumulation from adjacent or distal sites, it can provide an accurate “snapshot” of concentrations of a substance in tissues under various physiologic conditions. It also overcomes the limitation of measurement of messenger ribonucleic acid abundance, which may not detect changes in protein production that are independent of altered gene expression. Material and methods Cytokines and the anticytokine antisera were obtained from R&D Systems (Minneapolis, Minn). Molecular Probes, Inc (Eugene, Ore), supplied 4-methylumbelliferyl phosphate. Pepstatin-A, phenylmethylsulfonylfluoride, and bicinchoninic acid were supplied by Sigma (St Louis, Mo). Bovine serum albumin free of fatty acids and protease inhibitors were supplied by SERVA Electrophoresis GmbH (Heidelberg, Germany). Tissue isolation and extraction. Maternal consent for collection of placental tissues was obtained from all women according to the guidelines of the Auckland Human Ethics Committee. Placentas were collected and processed within 60 minutes of delivery. Placentas were obtained from women who (1) were delivered by the cesarean route for the indications of previous cesarean delivery or malpresentation at term (37-42 weeks’ gestation) before labor (term no labor, n = 15), (2) had uncomplicated vaginal delivery after spontaneous labor at term (term spontaneous labor, n = 15), and (3) were delivered after spontaneous preterm (≤34 weeks’ gestation) labor (preterm delivery, n = 31). The maternal and neonatal clinical records were assessed for evidence of intrauterine infection as indicated by maternal fever (temperature >37.8°C), maternal or fetal tachycardia (pulse rates >100 and >160 beats/min, respectively), uterine
Keelan et al 1531
tenderness, foul-smelling liquor, histologically evident chorioamnionitis, or neonatal sepsis.21 The final assessment was graded as definite infection, probable infection, or no infection present. After manual stripping of the amnion from the chorion-decidua, sections of the membranes (approximately 2 cm2) were cut, washed briefly in cold phosphate-buffered sodium chloride solution to remove excessive amounts of blood, and rapidly frozen in liquid nitrogen. At the same time a small piece of villous placental tissue (approximately 1 cm3) was dissected free of maternal decidua from the placental bed and similarly frozen after a brief wash in phosphate-buffered sodium chloride solution. When they were ready for processing, samples were thawed in 2 mL lysis buffer (0.05-mol/L tris[hydroxymethyl]aminomethane hydrochloride, pH 6.8; 2-mmol/L ethylenediaminetetraacetic acid; 100 g/L sucrose; 50 µg/mL pepstatin-A; 0.1 g/L phenylmethylsulfonylfluoride; and 0.1 g/L sodium azide) at 4°C. They were initially processed with a Polytron (Brinkmann Instruments, Inc, Westbury, NY) homogenizer for 15 seconds on ice, followed by further homogenization with a Branson (Branson Ultrasonics Corporation, Danbury, Conn) sonicator for 30 seconds. A 1-mL aliquot of each homogenate was taken and centrifuged for 7 minutes at 10,000g; the supernatants were removed and stored frozen at –70°C before assay. A small aliquot of each supernatant was taken and diluted 20-fold in 0.1N sodium hydroxide for protein assay by the bicinchoninic acid method.22 The assay had a working range of approximately 10 to 1000 µg/mL calibrated against bovine serum albumin. Cytokine enzyme-linked immunosorbent assays. IL-6 and IL-8 were measured by sandwich enzyme-linked immunosorbent assays as described previously.23 The sensitivity of the IL-6 assay was 20 pg/mL; that of the IL-8 assay was 6 pg/mL. The interassay and intra-assay precisions (coefficients of variation) were 9.0% and 8.6% for IL-6 and 7.6% and 8.0% for IL-8. The IL-1β assay also used antisera from R&D Systems but used a streptavidin–alkaline phosphatase signal generation system combined with fluorescence detection (4-methylumbelliferyl phosphate) to enhance sensitivity. The IL-1β assay had a sensitivity of 6 pg/mL; it performed with interassay and intra-assay precisions of 11.7% and 3.9%, respectively. Recovery from tissue homogenates ranged from 99.4% to 144.5%. All enzyme-linked immunosorbent assays were calibrated against recombinant human cytokine standards. Tissue homogenates diluted linearly in all 3 enzyme-linked immunosorbent assays. Statistical analysis and representation of data. Tissue concentrations of cytokines were expressed in picograms per milligram of protein. The Mann-Whitney U test was used for the analysis of all nonparametric data. Spearman rank correlations were performed to assess sig-
1532 Keelan et al
December 1999 Am J Obstet Gynecol
Fig 1. Concentrations of IL-1β in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
Table I. Summary of cytokine concentrations in gestational tissue samples IL-1β
Amnion Term no labor Term spontaneous labor Preterm delivery Chorion-decidua Term no labor Term spontaneous labor Preterm delivery Placenta Term no labor Term spontaneous labor Preterm delivery
No.
Median
15 15 31
21.7 83.2* 423.9†
15 15 31
270.2 292.7 1468.5†
15 15 31
42.8 50.5 70.9
IL-6 10th-90th percentiles
IL-8
Median
10th-90th percentiles
Median
10th-90th percentiles
5-172 21-2697 112-6713
84.6 330.4* 2030.6†
47-174 179-899 439-5853
93.0 497.8* 2241.5†
0-321 325-1911 626-4722
51-559 70-696 175-11,022
84.9 278.1* 1220.4†
53-132 177-1547 193-3953
193.3 766.4* 2152.2†
85-1387 337-1847 494-3544
63.8 65.1 102.2
17-125 44-130 29-216
69.8 48.8 88.9
42-179 21-238 32-354
22-142 28-95 28-179
Data are all given in picograms per milligram of protein. *Significantly different from term no labor. †Significantly different from term spontaneous labor and term no labor.
nificant relationships between the concentrations of various cytokines in the tissues studied. P <.05 was considered to be significant. Results The gestational age at delivery for both term groups (term no labor and term spontaneous labor) ranged from 37 to 42 weeks. All women had singleton pregnancies, with the exception of 1 twin birth in the term no labor group. There was no clinical evidence of intrauterine infection in any woman in either term group. In the preterm delivery group 65% of the women (20/31) were delivered vaginally. Of the remainder 10 underwent emergency cesarean delivery in labor (duration of labor, 2-84 hours) and 1 underwent cesarean delivery before labor after preterm premature rupture of membranes. The gestational age at delivery in the preterm delivery group ranged from 24 to 34 weeks (median, 31 weeks).
Thirty-two percent of the women in the preterm delivery group (10/31) had twin pregnancies. Intravenous albuterol (Salbutamol) was administered to 52% of the women with preterm deliveries (16/31) and 55% (17/31) received antibiotics. All received antenatal corticosteroids. Spontaneous rupture of membranes occurred in 55% of the preterm delivery group (17/31); 35% of the preterm delivery group tissue specimens (10/31) were from pregnancies that were scored as “infection positive” (definite or probable intrauterine infection). Placental histologic information was available in only 6 of 31 cases; 2 of these had evidence of chorioamnionitis. IL-1β concentrations were increased after term labor in amniotic tissues (P < .05, term spontaneous labor vs term no labor) but not in chorionic-decidual or placental tissues (Fig 1 and Table I). More marked increases were observed in preterm deliveries, with further significant
Keelan et al 1533
Volume 181, Number 6 Am J Obstet Gynecol
Fig 2. Concentrations of IL-6 in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
increases detected in both amniotic and chorionic-decidual tissues (preterm delivery vs term spontaneous labor, P < .001 and P < .005 for amnion and chorion-decidua, respectively). There were no significant differences in placental IL-1β concentrations between those women who had spontaneous labor at term and those who were delivered before labor (term spontaneous labor vs term no labor), nor were there any differences between those women who were delivered preterm with evidence of intrauterine infection and those without such evidence. Concentrations of IL-1β in the amniotic, chorionic-decidual, and placental tissues of the term no labor group showed significant correlation with each other (Table II). In the term spontaneous labor group IL-1β concentrations were significantly correlated only between chorionic-decidual and placental tissues, whereas in the preterm group only amniotic and chorionic-decidual concentrations were significantly correlated (Table II). Tissue concentrations of IL-6 in the amnion and the chorion-decidua were also increased after term labor (term spontaneous labor vs term no labor, P = .0001 for both amnion and chorion-decidua; Fig 2 and Table I) and were further increased after preterm pregnancies (preterm delivery vs term spontaneous labor, P = .0001 and P < .005 for amnion and chorion-decidua, respectively). In contrast, IL-6 concentrations in placental tissues taken after labor were not significantly increased with respect to those from women who were delivered before labor either at term or preterm. IL-6 concentrations in the various tissues showed little or no correlation with each other in the 2 term groups (Table II). In the preterm delivery group, however, IL-6 concentrations in the amnion, the chorion-decidua, and the placenta were all significantly correlated. IL-8 concentrations were increased after term labor in the extraplacental membranes (term spontaneous labor
Table II. Spearman ranked correlation coefficients (rs) for cytokine concentrations between amnion and chorion-decidua, chorion-decidua and placenta, and placenta and amnion
Term no labor IL-1β IL-6 IL-8 Term spontaneous labor IL-1β IL-6 IL-8 Preterm delivery IL-1β IL-6 IL-8
Amnion and choriondecidua
Choriondecidua and placenta
Placenta and amnion
0.682* 0.484* NS
0.697† NS NS
0.699† NS 0.581*
NS NS NS
0.768‡ NS NS
NS NS NS
0.633‡ 0.525‡ 0.389*
NS 0.501† NS
NS 0.602† 0.421*
In the interests of clarity only significant correlations are displayed. NS, Not significant. *P < .05. †P < .01. ‡P < .005.
vs term no labor, P = .0001 and P < .005 in amnion and chorion-decidua, respectively; Fig 3 and Table I). More marked increases were observed in amniotic and chorionic-decidual tissues from preterm pregnancies (preterm delivery versus term spontaneous labor, P < .0005 and P < .005 for amnion and chorion-decidua, respectively). Again the IL-8 concentrations in placental tissues were not significantly increased after labor at term or preterm. There was a significant correlation between IL-8 concentrations in amniotic and placental tissues in the term no labor group and also between placental and amniotic or chorionic-decidual concentrations in the preterm delivery group (Table II).
1534 Keelan et al
December 1999 Am J Obstet Gynecol
Fig 3. Concentrations of IL-8 in amniotic, chorionic-decidual, and placental tissues from women delivered at term before labor (TNL, n = 15), after spontaneous labor (TSL, n = 15), and preterm (PTD, n = 31). Median values are indicated by horizontal lines. Asterisk, P < .05; Mann-Whitney U test.
Table III. Spearman ranked correlation coefficients (rs) between cytokine concentrations in amniotic, chorionicdecidual, and placental tissue homogenates
Amnion Term no labor Term spontaneous labor Preterm delivery Chorion-decidua Term no labor Term spontaneous labor Preterm delivery Placenta Term no labor Term spontaneous labor Preterm delivery
IL-1β and IL-6
IL-1β and IL-8
IL-6 and IL-8
NS 0.818* 0.516*
NS 0.532† 0.794*
NS NS 0.727*
NS NS 0.595‡
NS NS 0.511‡
0.575† NS 0.598‡
NS NS NS
NS NS 0.378†
NS 0.586† NS
In the interests of clarity only significant correlations are displayed. NS, Not significant. *P < .005. †P < .05. ‡P < .01.
Concentrations of the 3 cytokines showed little intratissue correlation with each other in the term no labor group (Table III). In the term spontaneous labor group, however, amniotic IL-1β concentrations were significantly correlated with those of IL-6 and IL-8, whereas in the preterm delivery group all 3 cytokines showed significant correlations with each other within amniotic and chorionic-decidual tissues. Overall, cytokine concentrations showed the poorest correlations within placental tissues. Concentrations of cytokines in the amniotic, chorionic-decidual, and placental tissues were not significantly affected by subdivision of the preterm delivery tissues on
the basis of mode of delivery, treatment with antibiotics, or twin birth. In the amnion the IL-6 and IL-8 concentrations were significantly lower in pregnancies with spontaneous rupture of membranes than in those without (IL6, median, 1346 pg/mg protein; 10th-90th percentiles, 346-3851 pg/mg protein with spontaneous rupture of membranes; vs median, 2674 pg/mg protein; 10th-90th percentiles, 1352-7118 pg/mg protein without spontaneous rupture of membranes; IL-8, median, 1654 pg/mg protein; 10th-90th percentiles, 568-3745 pg/mg protein; vs median, 2862 pg/mg protein; 10th-90th percentiles, 1933-5382 pg/mg protein; P < .02 for both). Unexpectedly, when the tissues were subdivided on the basis of evidence of intrauterine infection, only IL-1β concentrations in the amnion were significantly elevated in the infection-positive subgroup (median, 4083 pg/mg protein; 10th-90th percentiles, 133-11,582 pg/mg protein infected; vs median, 251 pg/mg protein; 10th-90th percentiles, 82-1878 pg/mg protein noninfected; P < .05). The IL-6 concentrations in the infection-positive group were also modestly elevated compared with those in the noninfected group (median, 2621 pg/mg; 10th-90th percentiles, 656-2793 pg/mg infected; vs median, 1214 pg/mg; 10th-90th percentiles, 297-2021 pg/mg noninfected), but this difference did not reach statistical significance. Similar trends were seen in the chorionic-decidual samples, but again these did not reach statistical significance (IL-1β, median, 2609 pg/mg; 10th-90th percentiles, 350-11,252 pg/mg infected; vs median, 1465 pg/mg; 10th-90th percentiles, 124-10,399 pg/mg noninfected). Amniotic and chorionic-decidual cytokine concentrations in the infection-negative preterm delivery subgroup remained significantly elevated with respect to those of the term spontaneous labor group (P < 0.05, P < .005, and P < .005 for IL-1β, IL-6, and IL-8, respec-
Volume 181, Number 6 Am J Obstet Gynecol
tively; both tissues). No significant differences or apparent trends were observed in cytokine concentrations in the placental tissues with or without infection or spontaneous rupture of membranes. Comment This is the first comprehensive study of cytokine concentrations in gestational tissues from women delivered at term and preterm with and without labor. The results demonstrate increased levels of inflammatory cytokines in the gestational membranes of women with term or preterm labor. These findings support the view of labor as an inflammatory process and are consistent with the findings of increased cytokine production by gestational tissues after labor at term6, 7, 17 and preterm.10-12 Few previous studies have attempted to measure cytokine abundance in gestational tissues. In this study concentrations of IL-1β in decidua and placenta were higher than those found by Ammala et al6 but close to those reported by Kauma et al.1 Although concentrations of IL-6 and IL-8 have not been reported previously in these tissues in comparable units, we expected IL-8 concentrations to be in excess of those of IL-6 because these tissues produce more IL-8 than IL-6.23 However, concentrations of IL-6 in the amnion and the placenta were similar to those of IL-8, although their comparative levels in the chorion-decidua were closer to those anticipated. IL1β protein has previously been detected in the amnion by immunohistochemical methods7 but not quantified. However, IL-1β secretion and messenger ribonucleic acid expression by the amnion have been reported to be low.3, 5 In this study amniotic IL-1β concentrations were readily detectable, even in tissues from women who did not have labor. A possible explanation for this finding is that the amnion might accumulate cytokines originating from the chorion-decidua. The evidence for this hypothesis is equivocal, however, because, although amniotic and chorionic-decidual concentrations of IL-1β in tissues from the term no labor and preterm delivery groups were significantly correlated, those in tissues from the term spontaneous labor group were not. Our data indicate that the placenta is excluded from the inflammatory processes that appear to occur with labor in the gestational membranes. This finding is in agreement with those of Opsjøn et al,9 Gunn et al,4 and Matsuzaki et al,11 who also found no change in placental IL-6 content or production after labor at term. We have expanded on their findings to include determinations of IL-1β and IL-8 concentrations, and we can report that these cytokines also exhibit no significant changes after labor at term or preterm. Our findings contrast, however, with 2 reports of elevated IL-6 and IL-8 production in placentas from women with intrauterine infection at delivery.10, 11 Histologic studies have shown that sporadic neutrophilic invasion of the chorionic plate is apparent in a
Keelan et al 1535
proportion of placentas delivered at term.8 Our samples were of villous placenta and thus may not have been deep enough into the plate to sample regions where inflammatory changes might have taken place. Alternatively, the glucocorticoids administered to all women who were delivered preterm may have inhibited placental cytokine production, thereby reducing cytokine concentrations in this tissue. These considerations notwithstanding, the lack of changes in cytokine concentrations found in the placenta is in striking contrast to the changes found in the membranes. The marked increases in concentrations of cytokines in the extraplacental membranes after normal spontaneous labor suggest that upward regulation of cytokine production in these tissues is a parturition-dependent event that is independent of a pathogen-driven response or gestational maturity. Whether this process plays a causal role in the onset of parturition or is merely a response to stresses or processes during parturition remains unanswered by these studies. The possibility also cannot be excluded that the inflammatory response that we detected at term is a result of subclinical infection or is a response to exposure of the membranes to vaginal flora during labor.18 Our findings do not support these hypotheses, however, because under those circumstances the maternal face of the membranes would be expected to show the biggest response, which was not the case (median cytokine concentrations in the preterm delivery group were 4.5- to 6-fold those of the term spontaneous labor group in the amnion and 2.8- to 5-fold in the chorion-decidua). Several reports have documented infiltration of the cervix and lower uterine tract with neutrophils and macrophages before labor,24, 25 possibly in response to chemotactic agents such as IL-8,26 suggesting that some form of inflammatory activation may be involved in the mechanism of normal parturition. In our view the evidence to date is supportive of the hypothesis that cytokines are involved in the mechanism of normal parturition. The findings in this study of elevated levels of all 3 cytokines in the tissues from women with preterm delivery, irrespective of apparent infection status, support the contention that most preterm deliveries are associated with an inflammatory response in the extraplacental membranes unaccompanied by overt clinical signs of infection or chorioamnionitis. The fact that cytokine concentrations in the tissues from women who were delivered preterm were higher than those observed after term labor (despite glucocorticoid therapy) indicates that the inflammatory responses observed are part of the pathogenesis of preterm labor rather than merely a response to parturition. Several reports, including those based on cytokine measurements in amniotic fluid, have suggested that subclinical infection is a significant cause, if not the major cause, of preterm labor and delivery.2, 13, 14, 21, 27
1536 Keelan et al
Our findings are consistent with this view. In our group of preterm tissues categorized as infection-positive, only IL-1β concentrations in the amnion were significantly increased with respect to those in the infection-negative group. Our ability to accurately diagnose pregnancies with intrauterine infection was limited, however, and this deficiency may well have obscured some differences in the data. Other studies have highlighted the difficulty in comprehensively diagnosing intrauterine infection or chorioamnionitis, even with the aid of extensive histologic studies.21, 27, 28 This study supports the view that an inflammatory reaction takes place in the gestational membranes in a significant proportion of women who are delivered preterm. Our findings also support the hope that the development of effective strategies for the diagnosis and treatment of intrauterine inflammatory activation, with its strong association with preterm labor, will significantly improve the neonatal morbidity and mortality rates associated with preterm delivery. We gratefully acknowledge the assistance of the nursing and theater staff at National Women’s Hospital with the collection of the placental tissues and the excellent technical assistance of Ren-Li Zhou. We also thank Dr M. Battin for screening records for neonatal sepsis.
December 1999 Am J Obstet Gynecol
11.
12.
13.
14.
15.
16.
17.
18.
REFERENCES
1. Kauma S, Matt D, Strom S, Eierman D, Turner T. Interleukin-1 beta, human leukocyte antigen HLA-DR alpha, and transforming growth factor-beta expression in endometrium, placenta, and placental membranes. Am J Obstet Gynecol 1990;163:14307. 2. Dudley DJ, Trautman MS. Infection, inflammation, and contractions: the role of cytokines in the pathophysiology of preterm labor. Semin Reprod Endocrinol 1994;12:263-73. 3. Dudley DJ, Colmer D, Mitchell MD, Trautman MS. Inflammatory cytokine mRNA in human gestational tissues: implications for term and preterm labor. J Soc Gynecol Invest 1996;3:328-35. 4. Gunn L, Hardiman P, Tharmaratnam S, Lowe D, Chard T. Measurement of interleukin-1 alpha and interleukin-6 in pregnancy-associated tissues. Reprod Fertil Dev 1996;8:1069-73. 5. Menon R, Swan KF, Leyden TW, Rote NS, Fortunato SJ. Expression of inflammatory cytokines (interleukin-1 beta and interleukin-6) in amniochorionic membranes. Am J Obstet Gynecol 1995;172:493-500. 6. Ammala M, Nyman T, Salmi A, Rutanen EM. The interleukin-1 system in gestational tissues at term: effect of labour. Placenta 1997;18:717-23. 7. Osmers RGW, Blasr J, Kuhn W, Tschesche T. Interleukin-8 synthesis and the onset of labor. Obstet Gynecol 1995;86:223-9. 8. Halgunset J, Johnsen H, Kjøllesdal AM, Qvigstad E, Espevik T, Austgulen R. Cytokine levels in amniotic fluid and inflammatory changes in the placenta from normal deliveries at term. Eur J Obstet Gynecol Reprod Biol 1994;56:153-60. 9. Opsjøn SL, Wathen NC, Tingulstad S, Wiedswang G, Sundan A, Waage A, et al. Tumor necrosis factor, interleukin-1, and interleukin-6 in normal human parturition. Am J Obstet Gynecol 1993;169:397-404. 10. Shimoya K, Matsuzaki N, Taniguchi T, Kameda T, Koyama M, Neki R, et al. Human placenta constitutively produces inter-
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
leukin-8 during pregnancy and enhances its production in intrauterine infection. Biol Reprod 1992;47:220-6. Matsuzaki N, Taniguchi T, Shimoya K, Neki R, Okada T, Saji F, et al. Placental interleukin-6 production is enhanced in intrauterine infection but not in labor. Am J Obstet Gynecol 1993;168:94-7. Stallmach T, Hebisch G, Joller-Jemelka HI, Orban P, Schwaller J, Engelmann M. Cytokine production and visualized effects in the feto-maternal unit: quantitative and topographic data on cytokines during intrauterine disease. Lab Invest 1995;73:384-92. Romero R, Yoon BH, Mazor M, Gomez R, Gonzalez R, Diamond MP, et al. A comparative study of the diagnostic performance of amniotic fluid glucose, white blood cell count, interleukin-6, and Gram stain in the detection of microbial invasion in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 1993;169:839-51. Coultrip LL, Lien JM, Gomez R, Kapernick P, Khoury A, Grossman JH. The value of amniotic fluid interleukin-6 determination in patients with preterm labor and intact membranes in the detection of microbial invasion of the amniotic cavity. Am J Obstet Gynecol 1994;171:901-11. Laham N, Rice GE, Bishop GJ, Ransome C, Brennecke SP. Interleukin-8 concentrations in amniotic fluid and peripheral venous plasma during human pregnancy and parturition. Acta Endocrinol 1993;129:220-4. Steinborn A, Günes H, Halberstadt E. Signal for term parturition is of trophoblast and therefore of fetal origin. Prostaglandins 1995;50:237-52. Laham N, Brennecke SP, Bendtzen K, Rice GE. Differential release of interleukin-6 from human gestational tissues in association with labour and in vitro endotoxin treatment. J Endocrinol 1996;149:431-9. Casey ML, MacDonald PC. Cytokines in the human placenta, fetal membranes, uterine decidua, and amniotic fluid. In: Rice GE, Brennecke SP, editors. Molecular aspects of placental and fetal membrane autacoids. Boca Raton (FL): CRC Press; 1993. p. 361-94. Kauma SW, Walsh SW, Nestler JE, Turner TT. Interleukin-1 is induced in the human placenta by endotoxin and isolation procedures for trophoblasts. J Clin Endocrinol Metab 1992;75:951-5. Lonsdale LB, Elder MG, Sullivan MH. A comparison of cytokine and hormone production by decidual cells and tissue explants. J Endocrinol 1996;151:309-13. Gibbs RS, Romero R, Hillier SL, Eschenbach DA, Sweet RL. A review of premature birth and subclinical infection. Am J Obstet Gynecol 1992;166:1515-28. Redinbaugh MG, Turley RB. Adaption of the bicinchoninic acid protein assay for use with microtiter plates and sucrose gradient fractions. Anal Biochem 1986;153:267-71. Keelan JA, Sato T, Mitchell MD. Interleukin (IL)-6 and IL-8 production by human amnion cells in vitro: regulation by cytokines, growth factors, glucocorticoids, phorbol esters and bacterial lipopolysaccharide. Biol Reprod 1997;57:1438-44. Junqueira LC, Zugaib M, Montes GS, Toledo OM, Krisztán RM, Shigihara KM. Morphologic and histochemical evidence for the occurrence of collagenolysis and for the role of neutrophilic polymorphonuclear leukocytes during cervical dilation. Am J Obstet Gynecol 1980;138:273-81. Bokström H, Brännström M, Alexandersson M, Norström A. Leukocyte subpopulations in the human uterine cervical stroma at early and term pregnancy. Hum Reprod 1997:12;586-90. Osmers RG, Adelmann-Grill BC, Rath W, Stuhlsatz HW, Tschesche H, Kuhn W. Biochemical events in cervical ripening and dilatation during pregnancy and parturition. J Obstet Gynecol 1995;21:185-94. Zlatnik FJ, Gellhause TM, Benda JA, Koonitz FP, Burmeister LF. Histologic chorioamnionitis, microbial infection, and premature labor. Obstet Gynecol 1990;76:335-59. Quinn PA, Butany J, Taylor J, Hannah W. Chorioamnionitis: its association with pregnancy outcome and microbial infection. Am J Obstet Gynecol 1987;156:379-87.