- Email: [email protected]
Cytokines in normal and abnormal parturition: Elevated amniotic fluid interleukin-6 levels in women with premature rupture of membranes associated with intrauterine infection
CYTOKINES IN NORMAL AND ABNORMAL PARTURITION: ELEVATED AMNIOTIC FLUID INTERLEUKIN-6 LEVELS IN WOMEN WITH PREMATURE RUPTURE OF MEMBRANES ASSOCIATED WITH INTRAUTERINE INFECTION Uma Santhanam,l Cecilia Avila,’ Roberto Romero,’ Huguette Viguet,’ Nobuo Ida,3 Shingou Sakurai,4 Pravinkumar B. Sehgall,” The participation of interleukin-6 (IL-6) in the pathophysiology of normal and abnormal human parturition was evaluated by determining IL-6 concentrations in amniotic fluid (AF). Biologically active IL-6 was determined (in U/ml) using the B9 hybridoma growth factor assay, while the concentrations of immunoreactive IL-6 species (in pglml) were assessed using a monoclonal antibody (moAb)-based ELISA. Two hundred and twenty-seven AF samples from women in normal labor and from those presenting with a clinical diagnosis of premature rupture of membranes (PROM) were assayed. In selected instances, IL-6 levels were evaluated simultaneously in AF and in maternal and fetal plasma. Women with a normal pregnancy had low titers of biologically active IL-6 in AF both at midtrimester (group 1, n = 27; median IL-6 concentration = 16 U/ml) and at term (group 2, n = 33; median = 15 U/ml). There was an increase in the IL-6 bioactivity in AP from women in normal labor at term (group 3, n = 40; median = 74 U/ml; p < 0.001). In order to distinguish between the relative contributions of parturition per se and of intrauterine infection to the elevation of biologically active IL-6 levels in AF, IL-6 titers were compared in four different groups of women with PROM. Median AP IL-6 titers were 67 U/ml in women with microbial culture-negative AP who were not in active labor at the time of admission (group 4a, n = 17); 109 U/ml in those with culture-negative AP who were in active labor (group 4b, n = 25); 336 U/ml in those with culture-positive AF who were not in labor (group 4c, n = 21) and 4,330 U/ml in women with culture-positive AP who were in labor (group 4d, n = 33). The IL-6 ELISA provided data consistent with the B9 bioassay results. The median concentrations of AF IL-6 in women at midtrimester (n = 16) and in those at term who were not in labor (n = 15) were both 200 pg/ml; those in women with PROM were 800, 1,000,2,320, and 18,800 pg./ml in groups 4a (n = 12), 4b (n = 23), 4c (n = 14), and 4d (n = 31), respectively. Taken together, these data indicate that: (1) PROM per se may be associated with a modest elevation of AF IL-6 titers; (2) in the absence of infection, preterm parturition is not associated with a further elevation of AF IL-6; and most strikingly, (3) intraamniotic microbial infection is associated with marked elevations of AF IL-6 that are particularly high
‘The Rockefeller University, New York, NY 10021. *The Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, CT 06510. ‘Basic Research Laboratories, Toray Industries, Kamakura 248, Japan. 4Mishima Plants, Toray Industries, Mishima 411, Japan. *To whom correspondence should be addressed at The Rockefeller University, 1230 York Avenue, New York, NY 10021. Copyright o 1991 by W.B. Saunders Company 1043-4666/91/0302-0011$5.00/O KEY WORDS: amniotic fluid/B9 hybridoma IL-6 ELI&A/pregnancy and labor/premature
CYTOKINE,
Vol. 3, No. 2 (March),
growth factor assay/ rupture of membranes
1991: pp 155-163
Recent studies suggest that immunologically active cytokines participate in the pathophysiology of normal and abnormal pregnancy and parturition. Colonystimulating factor-l (CSF-1) has been implicated in the process of implantation,’ and granulocyte-macrophage colony-stimulating factor (GM-CSF) has been shown to stimulate placental growth.233 Interleukin-1 (IL-l) and tumor necrosis factor (TNF) have been implicated in the initiation of parturition in the setting of intrauterine infection.4-8 Indeed, IL-l and TNF have been detected in the amniotic fluid (AF) of women in 155
156 / Santhanam
et al.
CYTOKINE,
Vol.
3, No. 2 (March
1991: 155-163)
when associated with active parturition (p < 0.001). A marked elevation in the AF IL-6 level was sometimes, but not always, accompanied by an elevation of the IL-6 level in the fetal circulation and vice versa, suggesting that separate pathologic events trigger IL-6 accumulation in these two intrauterine compartments. At best, only trace levels of IL-6 biological activity and immunoreactivity were observed in the maternal circulation, even in women with markedly elevated amniotic fluid and/or fetal plasma IL-6. The present data clearly implicate IL-6 in the physiology of normal human parturition and also point to a strong association between markedly elevated AF or fetal plasma IL-6 levels and abnormal labor that accompanies intrauterine infection. Copyright o 1991 by W.B. Saunders Company
preterm labor (PTL) with intraamniotic infection.4” These cytokines are produced by human decidua in response to endotoxin,6 and IL-l and TNF can stimulate the production of prostaglandins by human amnion and decidua.‘,’ These observations prompted us to investigate the participation of IL-6 in parturition, as this cytokine is a major mediator of the host response to infection and tissue injury. IL-6 is produced by endometrial stromal cells in response to IL-l and interferon-y (IFN-Y)~ and by decidual explants in response to endotoxin.” Moreover, we have previously reported that IL-6 is normally present in human amniotic fluid and that its concentrations increase during spontaneous labor at term and in women with preterm labor and intraamniotic infecti0n.l’ From our previous study, we were unable to dissect the relative contributions of parturition per se and intraamniotic infection to elevations in AF IL-6 concentrations, due to constraints of the study design. All patients in that study who had intraamniotic infection were in preterm labor. The current study addresses this issue by exploiting a naturally occurring pregnancy complication, preterm premature rupture of membranes (PROM), which is also the leading identifiable cause of prematurity.‘l We have evaluated AF IL-6 levels in patients with PROM in the presence or absence of intraamniotic infection. Since patients with PROM may present with or without active labor at the time of admission, these patients afford a unique opportunity to distinguish between the individual contributions of intraamniotic infection and labor to changes in AF IL-6 levels. The presented data suggest that while active labor and PROM each coiltribute a modest increase to AF IL-6 levels, intratiterine microbial infection is the dominant factor. In some instances, we were able to sample the AF and the maternal and fetal circulation in the same patient. We report marked elevations in IL-6 levels in fetal plasma associated with PTL and intrauterine infection. The.available data suggest that the amniotic cavity, the fetal circulation, and the maternal circulation behave as three separate compartments in their
IL-6 response to infection of intrauterine tissues. The marked elevations observed in AF IL-6 levels in the presence of microbial infection strongly suggest that, in principle, this cytokine is likely to represent a reporter cytokine of considerable prognostic value. We have used both a bioassay (the B9 hybridoma growth factor assay”) and a two-antibody sandwich ELISA (based on high affinity anti-IL-6 moAb’3z13a)to evaluate AF IL-6 levels. Overall, the two assays were comparable in their ability to distinguish between infected and non-infected patients. RESULTS We have previously reported the detection of low levels of IL-6 bioactivity in AF from women in the second trimester of pregnancy and at term, in the absence of labor, and a modest increase in IL-6 levels during spontaneous labor at term using the hepatocytestimulating factor assay in Hep3B cel1s.l’ In the present study, we employed the more convenient B9 assay for determination of titers of biologically active IL-6. The AF samples in these previous three groups were retested using the B9 assay. (Fig. 1). AF from women in the midtrimester (group 1) and third trimester at term (group 2) of normal pregnancy contained IL-6 bioactivity that could be detected in the B9 assay (median = 16 U/ml in group 1 and 15 U/ml in group 2). Also, AF IL-6 titers were higher in women in spontaneous labor at term (group 3) compared to women at term who were not in labor (group 2) (median = 74 U/ml, range: 5 to 1,383 U/ml for women in active labor versus median = 15 U/ml, range: < 3 to 105 U/ml for women not in labor; p < 0.05). The validity of determining AF IL-6 concentrations with an ELISA was tested by determining the concentration of IL-6 antigen in undiluted AF samples collected in midtrimester and at term in the presence or absence of labor (n = 16, 15, and 15, respectively; samples different from those depicted in Fig. 1). Women in active labor at term had significantly higher AF concentrations of IL-6 antigen than women at term
Amniotic fluid IL-6 in parturition
1200 1100 i
IL-6 U/ml
.
600
400 t
loo0
I ‘, .*
Midtrimester (n = 27)
Figure 1. Biologically and at term.
active
IL-6
i
t I GA. -.* A
Term No Labor (n = 33)
Term Labor (n = 40)
in AF of women
/ 157
the midtrimester (group 1) or at term (group 2) (group 4a, median = 67 U/ml; group 1, median = 16 U/ml; group 2, median = 15 U/ml; p < 0.05; for comparisons between group 4a and group 1 and for group 4a and group 2, see Fig. 3). A similar set of results was obtained when determining immunoreactive IL-6 (Fig. 4). AF IL-6 antigen concentrations were significantly higher in women with PROM and without infection and labor (group 4a) than in women in the midtrimester (group 1) at term who had intact membranes (group 2) (group 4a-median = 800 pg/ml, range: 200 to 1,240 pg/ml, Fig. 4; group l-median = 200 pg/ml, range: 60 to 1,600 pgiml; group 2-median = 200 pg/ml, range: 150 to 850 pg/ml; see Fig. 4). These data indicate that rupture of the membranes per se (in the absence of labor or intraamniotic infection) is associated with increased bioactive and immunoreactive IL-6 concentrations in the amniotic cavity. The group of patients with PROM provided us with an opportunity to dissect the relative contributions of parturition and intraamniotic infection on AF IL-6 titers. In these patients, in the absence of intraam-
in midtrimester
Scatter diagram illustrates three different groups of patients (AF samples used in this analysis are the same as in Fig. 5 in reference 10): (A) midtrimester (group 1: n = 27; median = 16 U/ml; range, <3 (0 166 U/ml); (B)‘at term but not in labor (group 2: n = 33; median = 1.5 U/ml: range. <5 to 105 U/ml): and CC) at term and spontaneous labor ‘(groip’3: n = 40; median = 74 c/r& range, 5 to 1,383 U/ml). Kruskal-Wallis H = 22.30; p = 0.0000144; A compared to B: p > 0.05; B compared to C: p < 0.05; C compared to A: p < 0.05 (Dunn’s test).
without labor or in the midtrimester (Fig. 2) (midtrimester-median = 200 pg/ml, range: 60 to 1600 pg/ml; term without labor-median = 200 pg/ml, range: 1.50 to 800 pg/ml; term in active labor-median = 1,600 pgiml, range: 200 to 15,600 pg/ml; Kruskal-Wallis H = 25.24; p = 0.0000033; Dunn’s test for multiple comparisons between term labor and the other two groups p < 0.05). These data indicate that both immunoreactive and biologically active IL-6 concentrations are elevated in the amniotic fluid of women in active labor at term. The pathophysiologic events that lead to PROM in the absence of active labor are complex and poorly understood. Inflammation of chorioamniotic membranes secondary to an antecedent extraamniotic infection has been postulated to play a role.14,15Therefore, it is important to determine whether PROM per se (in the absence of intraamniotic infection) is associated with alterations in AF IL-6 titers. Women with preterm PROM in the absence of active labor and without demonstrable intraamniotic infection (group 4a, Fig. 3) had higher AF IL-6 biologic activity than those with intact membranes without labor or infection, either in
12500 1
10000 t
IL-6 pgiml
50°0 2500
t
2000 1500 i
-
-
‘::;Ii ;- ; Midtrimester (n = 16)
Figure 2. Immunoreactive and at term.
Term No Labor (n = 15)
IL-6
in AF of women
Term Labor (n = 15)
in midtrimester
Scatter diagram illustrates three different groups of patients (AF samples used in this analysis are different from those in Fig. 1 and were all obtained transabdominally): (A) midtrimester (n = 16; median = 200 pgiml; range, 60 to 1,600 pg/ml; mean = 331 pgiml; SD = 386 pgiml); (B) term but not in labor (n = 15; median = 200 pg/ml; range, < 100 to 850 pgiml; mean = 296 pgiml; SD = 243 pgiml); and (C) term active labor (n = 15; median = 1,600 pg/ml; range, 200 to 15,600 pgiml; mean = 3,300 pg/ml; SD = 4,185 pgiml). Kruskal-Wallis H = 25.24; p = 0.0000033; A compared to B: p > 0.05; B compared to C: p < 0.05; C compared to A: p < 0.05 (Dunn’s test).
158 I Santhanam
CYTOKINE,
et al.
75000
. ..
. 45000 F
tt V
15000
.
11000 i
.
9000 7000 5000 IL-6 U/ml
. :
i Y-
t
3000
.t.
f*
1250
.
.
1000
”
Figure
3.
No Labor No Infection (n = 17)
Biologically
active
Labor No Infection (n = 25)
IL-6
No Labor Infection (n = 21)
in Al7 of women
Vol. 3, No. 2 (March 1991: 155-163)
pg/ml, p < 0.05). These data indicate that in the setting of microbial invasion of the amniotic cavity, parturition is associated with a marked increase in AF IL-6 bioactivity and immunoreactivity. The possibility that IL-6 produced by infected intrauterine tissues may enter the maternal circulation, leading to systemic effects in the mother, or that IL-6 may be present in the fetal circulation in such patients, was evaluated by comparing biologically active IL-6 titers in AF and in maternal plasma of patients who presented with PTL and fever (Table 1). In some of these instances, we were also able to simultaneously evaluate IL-6 titers in fetal plasma and, depending upon availability of material, to verify the presence of IL-6 using the ELBA. Table 1 shows that many of the fetal plasma samples tested contained high levels of IL-6. Only low levels of IL-6, however, were observed in the maternal circulation even in the presence of high AF and fetal plasma IL-6 titers. An inspection of the
Labor Infection (n = 33)
with
2. 27000
PROM.
Scatter diagram illustrates four different groups of patients: (A) no infection and no delivery (group 4a; n = 17~;median = 67‘U)ml; range. 24 to 479 U/ml: mean = 101 U/ml: SD = 103 U/ml): (B) no infeition but active labor (group 4b: n = 25; median = lb;9‘U~ml; range, 10 to 967 U/ml; mean = 238 U/ml; SD = 260 U/ml); (C) infection but no delivery (group 4c: n = 21; median 336 U/ml; range, 8 to 56,978 U/ml; mean = 4,142 U/ml; SD = 12,367 U/ml); and (D) infection and labor (group 4d: n = 33; median = 4,330 U/ml; range, 27 to 67,140 U/ml; mean = 12,238 U/ml; SD = 17,435 U/ml). After logarithmic transformation, ANOVA indicated F = 18.7, p < 0.00001. Multiple comparisons were made with the Scheffe’s test. A p value of <0.05 was considered significant. IL-6 in 4d was significantly higher than in all other groups; IL-6 in 4c was higher than in 4a but not higher than in 4b. No differences were detected between groups 4a and 4b.
i
c
I
22000
I .
-
17000 -
i . .
. . .
12000 -
. ;;j:,
7000 I I . . 1500 .
i ’
1000
,
. .
.
.
.
.
-
-
l . :
-
.
. !
niotic infection, preterm parturition was not associated with a further elevation in AF IL-6 titers (Figs. 3 and 4, group 4b versus 4a; p > 0.05). Women with intraamniotic infection (Figs. 3 and 4, pooled data from groups 4c and 4d) had significantly higher AF IL-6 levels than women without infection (pooled data from groups 4a and 4b) (bioactive IL-6, intraamniotic infectionmedian = 1,557 U/ml; range: 8 to 67, 140 U/ml versus non-infected women-median = 90 U/ml; range: 10 to 967 U/ml; p < 0.00001, Mann-Whitney U test; immunoreactive IL-6, intraamniotic infection-median = 14,400 pg/ml; range: 200 to 28,800 pg/ml versus non-infected women-median = 800 pg/ml; range: 20026,800 pg/ml; p < 0.0001, Mann-Whitney U test). Among women with intraamniotic infections, those with spontaneous PTL (group 4d) had significantly higher AF IL-6 biological activity and IL-6 antigen (median = 4,330 U/ml and 18,800 pg/ml) than those without labor (group 4c, median = 336 U/ml and 2,320
500
1
:
2.
.. .
.. .
01 No
Labor No Infection (n = 12)
Figure
4.
Immunoreactive
.
:.
Labor No Infection (n = 23)
IL-6
No Labor Infection (n = 14)
in AF’ of women
Labor Infection (n =31)
with PROM.
Scatter diagram illustrates four different groups of patients (the AF samples analyzed were from within the groups described in Fig. 3): (A) no infection and no delivery (n = 12; median = 800 pg/ml; range, 5200 to 1,240 pgiml; mean = 703 pg./ml; SD = 401 pg/ml); (B) no infection but in labor (n = 23; median = 1,000 pg/ml; range, 200 to 45,000 pg/ml; mean = 2,923 pgiml; SD = 5,638 is/ml); CC) infection but no delivery (n = 14, median = 2,320; range, 5 200 to 16,000 pg/ml; mean = 5,020 pgiml; SD = 5,407 pgiml); and (D) infection and labor (n = 31: median = 18.800 Da/ml: range. 1.280 to 92,500 pgiml; mean ‘= 16,904 pgiml; SD 2 9,34
Amniotic
&ml
1 2 3 4 5 6
< 10 117 Cl0 6,819
150 1,200 < 100 5,900 < 100 2,700
7
676
500
/ 159
The present data confirm that IL-6 is a physiologic constituent of AF during the second and third trimesters of pregnancy. Additionally, there is no change in the concentration of biologically active or immunoreactive IL-6 with advancing gestational age. Spontaneous term parturition is associated with modest elevations of AF IL-6 levels. Furthermore, preterm PROM is also associated with a modest, albeit significant, elevation of Al! IL-6 levels. However, intraamniotic infection in this setting is associated with a dramatic elevation of AF IL-6 levels. The highest AF IL-6 levels are seen in women with PROM in whom intraamniotic infection is associated with spontaneous preterm parturition. Thus, although intramniotic infection is the major contributor to elevations of AF IL-6 levels, parturition also contributes to increases in AF IL-6 in such patients. Markedly elevated AF IL-6 in patients with PROM may be a good diagnostic indicator of intraamniotic infection and be prognostic of impending spontaneous labor. The presence of IL-6 in normal amniotic fluid has now been confirmed with two different bioassays (the hepatocyte stimulating assay”’ and the B9 hybridoma growth factor assay [this report]) and two different immunoassays (immunoblot analysis” and ELISA [this report]). It is of considerable interest that there is a correlation between immunoreactive and biologically active IL-6 in AF over a wide range of concentrations (r = 0.92, p < 0.0001). This observation is of considerable importance because the B9 bioassay takes 4 days to perform, while the ELISA allows quantitation in a matter of hours. Therefore, IL-6 determinations can now be carried out in the clinical setting in a timely fashion and their results used in patient management. Further studies are required to determine whether this
Fetal Plasma U/ml
in parturition
DISCUSSION
Table 1. IL-6 in Fetal Plasma.’
Patient number
fluid IL-6
AF (U/ml)
Comments
42 24,682 6,862 12 1,4532 14,0023 1O,6544 162
PTL PTL PTL PTL Term PROM PROM
‘IL-6 in maternal plasma samples was I 10 U/ml in the B9 assay and 5 110 pg/ml by ELISA in all seven patients. *3,750 pg/ml by ELISA 3192.500 p&n1 by ELISA 4117,500 pgiml by ELISA
data in Table 1 suggests that the amniotic cavity, the fetal circulation, and the maternal circulation appear to behave essentially as three separate compartments. The appearance of a high level of IL-6 bioactivity and of IL-6 antigen in fetal plasma without an associated increase in AF IL-6 (patient #4) may be indicative of microbial infection of the fetus itself. These data demonstrate, for the first time, the detection of high levels of IL-6 in the fetal circulation using both the B9 assay and an IL-6 ELISA. Figure 5 illustrates a comparison of IL-6 B9 assay titers and of IL-6 ELISA concentrations in 77 AF samples from PROM patients in this study that we were able to analyze using both assays. The correlation between the two assays is good (r = 0.92, p < 0.0000001; Spearman correlation). This overall correlation between the two assays is reflected in the ability of both assays to distinguish between the PROM groups that were with and without microbial infection (Figs. 3 and 4). 25.0 22.5 I 20.0 17.5
Figure 5. munologic
Correlation between assays for IL-6.
the biologic
and im-
The small figure on the right illustrates the relationship between AP IL-6 concentrations determined with B9 assay and with ELISA in all 77 AF samples. Spearman correlation r = 0.92, p < 0.00001. Since most of the results are grouped below 10,000 U/ml (horizontal axis) and 25 ngiml (vertical axis), this part of the graph has been magnified to illustrate the correlation between assay results at these concentrations (left graph).
a y
15.0
w E
12.5
B 3
10.0
-
I
125.0
t
100.0 i *
I
.
10000 20000 IL-6 U/ml (B9)
IL-6 U/ml (B9)
30000
160 I Santhanam
et al.
ELISA detects all of the different biologically active human IL-6 isoforms and aggregates with equal efficiency. The immunoassay employed in this study, however, provides results qualitatively similar to the B9 bioassay when applied to the study of parturition and intraamniotic infection. The emerging picture from our studies is that while certain cytokines (e.g., IL-6 and CSF-1) are physiologic constituents of amniotic fluid,‘,” others appear only under pathologic conditions (e.g., TNF)? The biologic role and source of AF IL-6 in normal gestation remains to be elucidated. Since the production of this cytokine is a general feature of processes in which there is tissue disruption,16 its presence in normal amniotic fluid may be an indicator of the extensive fetal and maternal tissue remodeling associated with fetal growth. In situ nucleic acid hybridization analysis for cells containing IL-6 mRNA and immunohistochemical localization of IL-6 protein in tissue sections are procedures currently available to address directly questions concerning the sources of amniotic fluid IL-6. Our data indicate that spontaneous parturition at term is associated with increased concentrations of biologically active and immunoreactive IL-6 in amniotic fluid. In contrast to our previous study, in which amniotic fluid samples from women in active labor at term were obtained by transvaginal amniotomy. AF was obtained transabdominally in the current study (Figs. 2-4). AF collected transvaginally might be contaminated with endocervical or vaginal flora, and therefore, cultures of this material may not reflect the microbiologic state of the amniotic cavity. The elevation of AF IL-6 in women in term labor described in our previous study could have been due to the presence of a subclinical intraamniotic infection. In the current study, patients with spontaneous labor at term had negative AF cultures for microorganisms. Thus, the elevation in AF IL-6 in these women is attributable to parturition per se rather than to a subclinical intraamniotic infection. IL-6 secretion during spontaneous labor at term may reflect tissue disruption associated with cervical ripening and dilatation and myometrial contractility. Indeed, cervical ripening has been likened to an inflammatory reaction that is under endocrine contr01.15 Recently, smooth muscle has been identified as a rich source of IL-6,17 and this cytokine may be part of the stress-related phenomenon associated with dramatic changes in myometrial contractility required for normal parturition. Patients with PROM and without labor or intraamniotic infection (group 4a) had a modest but significant increase in AF IL-6 concentrations in comparison to women with intact membranes (groups 1 and 2). These data suggest that the pathologic process responsible for
CYTOKINE,
Vol. 3, No. 2 (March
1991: 155-163)
rupture of the membranes also leads to an elevation of AF IL-6. Alternatively, the tissue trauma associated with membrane disruption and breakage may be sufficient for stimulation of IL-6 production by fibroblasts and other IL-6-producing cells in the chorioamniotic membranes. In a previous study,l’ we reported dramatic elevations of AF IL-6 concentrations (the highest measured in our laboratory in any biological fluid) in women with intraamniotic infection and preterm labor with intact membranes. In that study, the relative contribution of preterm parturition and intraamniotic infection could not be determined because patients were ascertained by the diagnosis of preterm labor, and consequently, there was not a group who had intraamniotic infection without labor. In the current study, we have addressed this question in women with preterm PROM. The results indicate that infection is the most important contributor to elevations in AF IL-6. Indeed, in the absence of intraamniotic infection, we could not demonstrate an elevation in AF IL-6 with preterm labor alone in the setting of PROM (comparisons between groups 4a and 4b). These data are consistent with similar studies of AF IL-l and TNF that collectively indicate that preterm parturition in the absence of infection is not associated with increased secretion of cytokines into the AF compartment.“’ In the case of K-6, it is possible that the elevation associated with membrane rupture per se may obscure the effect of preterm parturition. This view is supported by our previous observation that preterm parturition with intact membranes is associated with elevated levels of AF IL-6, even in the absence of a demonstrable intraamniotic infection.” In contrast to preterm parturition in the absence of intraamniotic infection, preterm labor in the presence of intraamniotic infection is associated with a significant increase in AF IL-6 concentrations (comparisons of groups 4c and 4d). Similar observations have been made with IL-1.4 These data indicate that the onset of parturition in the setting of infection is temporally associated with the secretion of cytokines into the amniotic cavity and support the view that the initiation of parturition may be viewed as part of the repertoire of host-defense mechanisms elicited in response to intraamniotic infection. In comparing IL-6 levels in maternal plasma, fetal plasma, and AF in women with fever and preterm labor, it is striking that some of the fetal plasma samples had unusually high levels of IL-6. At best, only low levels of IL-6 were detected in maternal plasma, suggesting that sampling the maternal circulation for IL-6 content is not likely to be a useful test in the management of these patients. The ease of sampling of the amniotic fluid compartment and the marked elevations of IL-6 observed in this compartment suggest that
Amniotic fluid IL-6 in parturition /
AF IL-6 measurements are likely to prove of greater value. It is apparent from the data in Table 1 that AF IL-6 levels do not always correlate with fetal plasma IL-6 levels. The ease of sampling the amniotic cavity versus that of sampling the fetal circulation in utero makes AF IL-6 measurement a more generally applicable procedure. The biologic role of IL-6 in maternal and fetal physiology remains to be clearly defined. Elevation of CRP levels in fetal plasma appears to be of protective value because the survival rate of neonates with sepsis and high serum CRP plasma is greater than those with no detectable serum CRP.‘*,19 It is likely that IL-6 in the fetal plasma enhances not only acute phase plasma protein synthesis in the liver but also the ability of the fetal immune system to fight off microbial infection of the fetal tissues. To summarize, our data suggest that IL-6 is an excellent reporter cytokine whose AF levels are markedly elevated in the presence of intraamniotic infection. The present study justifies a prospective evaluation of the diagnostic and prognostic value of AF IL-6 measurements in patients with preterm PROM.
within 24 h of amniocentesis. Pelvic examinations were not performed at the time of amniocentesis, since this could result in an increased risk of infection.14 Tocolytic agents were not used in patients with preterm PROM. After defining the criteria for entry into the study, AF samples were obtained from the AF bank of the Department of Obstetrics and Gynecology, Yale University School of Medicine. This bank consistsof AF samples that remain after clinical studies. AF samples were centrifuged at 400g for 10 min at 4°C immediately after collection, and the fluid was separated from the cell pellet and stored frozen at -70°C. AF samples analyzed in this study were from consecutive patients admitted to the hospital who fulfilled the criteria of each of the study groups. Informed consent was obtained from all patients following the guidelines of the Yale University Human Investigations Committee.
Collection of Maternal
Patient Population
and Collection of Amniotic
Fluid
AF was collected transabdominally from women in four different populations: group 1, women in the midtrimester of pregnancy (gestational age: 16 to 18 weeks) who underwent
amniocenteses for genetic indications (maternal age > 35 years) (n = 43); group 2, women in the third trimester (gestational age: 37 to 40 weeks) who had amniocenteses for the assessmentof fetal lung maturity before elective cesarean sections (n = 48); group 3, women in spontaneous active labor (cervical dilatation of at least 4 cm) who underwent a primary or repeat cesarean section or women who were suspected to have preterm labor and who subsequently delivered ( < 24 h) a term newborn (by Dubowitz evaluation) with birthweight above 2,800 g (n = 40); and group 4, women admitted with preterm PROM on whom amniocenteses were performed for microbiologic assessment of the amniotic cavity and for fetal lung maturity studies (n = 96). To study the relationship between intraamniotic infection, preterm labor, and AF IL-6 levels, a cross-sectional study of patients in group 4 (PROM) was constructed. These patients were subdivided into four subgroups according to the results of their AF culture and the presence or absence of labor at the time of amniocentesis, as follows: group 4a, women without labor or infection (n = 17); group 4b, women with labor but without an intraamniotic infection (n = 25); group 4c, women with an intraamniotic infection but without labor (n = 21); and group 4d, women with both intraamniotic infection and labor (n = 33). An intraamniotic infection was defined as the presence of a positive amniotic fluid culture. Labor was defined as the presence of regular uterine contractions that led to cervical dilatation and resulted in delivery
and Fetal (Neonatal) Plasma
Maternal blood was collected before amniocentesis, and fetal umbilical cord blood, at the time of delivery. Blood was collected in vacutainer tubes containing EDTA (BectonDickinson, Rutherford New Jersey) centrifuged at 4008 for 10 min at 4”C, and the plasma was separated and stored frozen at -70°C.
Microbiologic
MATERIALS AND METHODS
161
Studies
AF in a capped plastic syringe was transported to the microbiology laboratory immediately after collection and plated within 30 min of amniocentesis. AF was cultured for aerobic and anaerobic bacteria as well as for Mycoplasma hominis and Ureaplasma urealyticum, as described previ0us1y.~~AF samples from women in groups 1 and 2 were sterile.
IL-6 Bioassay The IL-6 content of body fluids was assayedby monitoring the proliferation of the IL-bdependent murine hybridoma cell line B9.” Briefly, duplicate sets of serial threefold dilutions, starting at a 1:20 dilution of the test body fluid, were added to B9 cells in 96-well plates (final volume 200 ~1 containing 5,000 cells). The cultures were incubated for 92 to 96 h, and the growth of B9 cells was then quantitated by the uptake of MTT (3-[4,5 Dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide; Sigma, St. Louis, MO) and measured at 5701690 nm wavelengths using the EL312 Biotek microplate reader.‘l Neutralizing rabbit anti-rIL-6 antiserum confirmed that the proliferative response of B9 cells to representative samples of human body fluids was specific for IL-6.” Prior to use in the assay,the plasma or serum samples, but not the AF samples, were heated at 56°C for 30 min and then clarified by centrifugation. Dilutions of unheated human plasma/serum, but not of AF, inhibited B9 cell proliferation in response to exogenous IL-6. Interassay variation was a maximum of one dilution step (threefold). IL-6 titers, expressed as laboratory reference units (U/ml) in body fluids, are the reciprocals of the highest dilutions that produced a half-maximal proliferative response in the B9 assay. Each assay is calibrated in terms of
162 / Santhanam et al.
CYTOKINE,
the biologic activity of an IL-6 laboratory reference standard (purified IL-l-induced, human fibroblast-derived, natural IL-6 preparation 060888 or 071989). These reference preparations contain multiple differentially modified forms of IL-6 in the size range from 23 to 30 kD as characterized by SDS polyacrylamide gel electrophoresis (PAGE) and immunoblot analyses, using a rabbit antiserum to Escherichia coli-derived human IL-6.2’ The mass concentration of IL-6 in these laboratory reference preparations has been estimated by immunoblot comparisons (using the polyclonal anti-rIL-6 antiserum) of these IL-6 preparations with preparations of electrophoretically homogeneous, E. coli-derived IL-6, which have, in turn, been quantitated by Coomassie blue staining and by silver staining. In our hands, 1 U/ml of B9 activity of natural glycosylated IL-6 corresponds to a concentration of 15 pg/ml. In this assay, the E. coli-derived IL-6 preparation 881514, distributed by the NIH as an interim human IL-6 reference standard, titers at least threefold higher than its arbitrarily assigned potency (36,296 U/ml vs. 12,500 arbitrarily assigned U/ml designation). Furthermore, one U/ml of E. coli-derived IL-6 881514in the B9 assaycorresponds to 70
pg/ml. It is not yet clear how titers of biologically active IL-6 in human body fluids that are determined using the B9 assay and expressed as units per milliliter relate numerically to those obtained using other biologic and immunologic assays for IL-6?3-26The difficulties encountered in comparing quantitative
data between
different
bioassays include
the inhibi-
tory effect of human body fluids in the B9 assayz6and the observation that the slope of the log concentration-response plot in the B9 assay for IL-6 in body fluids is often quite different from that of the reference IL-6 preparation. Apparent IL-6 titers in human serum in the B9 assay have been reported to be considerably lower than would have been expected on the basis of the hepatocyte-stimulating factor assay.23,26Our experience comparing AF IL-6 levels using the B9 assay and the hepatocyte (Hep3B) assay indicates that although
there
is an overall
correlation
between
the IL-6
concentrations estimated using the two bioassays, the B9 assay registers a lower level of IL-6 biologic activity in AF than that detected as IL-6-specific biologic activity in Hep3B cells.1o,26
IL-4 ELBA To evaluate a faster assayfor IL-6 that could be used in patient
management,
we have assayed AF samples using a
sandwich ELISA developed recently.‘3z’3”Immunoreactive IL-6 detected in AF and plasma samples using this assay is expressed as picograms per milliliter of antigen calibrated in terms of a particular preparation of purified natural fibroblastderived IL-6; interassay variation was < 10%. The absolute amount of IL-6 in this preparation
was determined
by amino
acid analysis. This assaywas validated using AF from women in normal midtrimester and third-trimester pregnancies. Undiluted AF did not interfere with the detection of exogenously added IL-6 (Toray IL-6 standard preparation or preparation NIH 88/514). The relationship between IL-6 determinations with the B9 assay and the ELISA was evaluated by comparing the results of the two assayson a set of 77 AF samples from patients with PROM.
Vol. 3, No. 2 (March 1991: 155-163)
Statistical Analysis A Wilk-Shapiro test was employed to test the hypothesis that AF IL-6 concentrations were normally distributed. Logarithmic transformation was required for normality. Oneway analysis of variance (ANOVA) was used to compare AF IL-6 concentrations among patients with preterm PROM. Multiple comparisons among the four subgroups were performed
with a Duncan’s
multiple
range test. A p value of
< 0.05 was considered statistically significant. A KruskalWallis test was used to compare AF IL-6 concentrations between multiple
groups 1,2, and 3. A Dunn’s test was employed for comparisons (True Epistat, Epistat Services, Rich-
ardson, TX). A logarithmic transformation was not used for these values since several patients had less than detectable levels of AF IL-6 bioactivity. The correlation between the B9 assayand ELISA was tested using the Spearman correlation rank test.
Acknowledgments We thank Dr. Igor Tamm for his enthusiastic support and Carol Brekus, C.N.M., M.S.N., and Valerie Kraemer, M.F.A., for their editorial assistance. Supported by Research Grant CA-44365 (P.B.S.) and a Physician-Scientist Award from the National Institutes of Health (R.R.) and a contract from the National Foundation for Cancer Research.
REFERENCES 1. Pollard JW, Bartocci A, Arceci R, Orlofsky A, Ladner MD, Stanley ER (1987) Apparent role of the macrophage growth factor, CSF-1, in placental development. Nature 330:484-486. 2. Athanassakis I, Bleackly RC, Paetkau V, Guilbert L, Barr PJ, Wegmann TG (1987) The immunostimulatory effects of T cells and T cell lymphokines on murine fetally derived placental cells. J Immunol138:37-44. 3. Wegmann TG, Athanassakis I, Guilbert L, Branch D, Dy M, Menu E, Chaouat G (1989) Placental immunotrophism: maternal T cells enhance placental growth. Am J Reprod Immunol15:566-569. 4. Romero R, Brody DT, Oyarzun E, Mazor M, Wu YK, Hobbins JC, Durum SK (1989a) Infection and labor. III. Interleukin-l: a signal for the onset of parturition. Am J Obstet Gynecol 160:1117-1123. 5. Romero R, Manogue KR, Mitchell MD, Wu YK, Oyarzun E, Hobbins JC, Cerami A (1989) Infection and labor. IV. Cachectintumor necrosis factor in the amniotic fluid of women with intraamniotic infection and preterm labor. Am J Obstet Gynecol 161:336-341. 6. Romero R, Wu YK, Brody DT, Oyarzun E, Duff GW, Durum SK (1989) Human decidua: a source of interleukin-1. Am J Obstet Gynecol73:31-34. 7. Romero R, Durum S, Dinarello CA, Oyarzun E, Mitchell MD (1989) Interleukin-1 stimulates prostaglandin biosynthesis by human amnion. Prostaglandins 37:13-22. 8. Romero R, Mazor M, Wu YK, Avila C, Oyarzun E, Mitchell MD (1989) Bacterial endotoxin and tumor necrosis factor stimulate prostaglandin production by human decidua. Prostaglandins, Leukotrienes & Essential Fatty Acids 37:X33-186. 9. Tabibzadeh SS, Santhanam U, Sehgal PB, May LT (1989) Cytokine-induced production of interferonp,/interleukin-6 by freshlyexplanted human endometrial stromal cells. J Immunol 142:31343139. 10. Romero R, Avila C, Santhanam U, Sehgal PB (1990)
Amniotic fluid IL-6 in parturition Amniotic fluid interleukin-6 in preterm labor: association with infection. J Clin Invest 85:1392-1400. 11. Arias F, Tomich P (1982) Etiology and outcome of low birth weight and preterm infants. Obstet Gynecol60:277-282. 12. Brakenhoff JPJ, De Groot ER, Evers RF, Pannekoek H, Aarden LA (1987) Molecular cloning and expression of hybridoma growth factor in E. coli. J Immunol139:4116-4121. 13. Ida N, Hosaka T, Hosoi K, Kunitomo T, Shimazu T, Maruyama T, Matsuura Y, Kohase M (1989) Establishment of strongly neutralizing monoclonal antibody to human interleukin-6 and its epitope analysis. Biochem Biophys Res Commun 165:728734. 13a. Ida N, Sakurai S, Hosaka T, Hosoi K, Kunitomo T, Matsuura Y, Kohase M (1990) An enzyme-linked immunosorbent assay for the measurement of human interleukin-6. J Immunol Methods 133:279-284. 34. Schutte MD, Treffers PE, Kloosterman GJ (1983) Management of premature rupture of membranes: the risk of vaginal examination to the infant. Am J Obstet Gynecol144:395-399. 15. Liggans GC (1981) Cervical ripening as an inflammatory reaction. In Ellwood DA, Anderson ABM, Embrey MP (eds) The Cervix in Pregnancy and Labour: Clinical and Biochemical Investigations, Churchill Livingstone, Edinburgh, pp 1. 16. Sehgal PB, Grieninger G, Tosato G (1989) Regulation of the acute phase and immune responses: interleukin-6. Ann NY Acad Sci 557:1-583. 17. Loppnow H, Libby P (1990) Proliferating or interleukin-lactivated human vascular smooth muscle cells secrete copious interleukin-6. J Clin Invest 85:731-738. 18. Philip AGS (1985) Response of C-reactive protein in neonatal group B streptococcal infection. Pediatr Infect Dis 4:145148.
/ 163
19. Philip AGS (1979) The protective effect of acute phase response reactants in neonatal sepsis. Acta Paediatr Scan 68:481483. 20. Romero R, Scharf K, Mazor M, Emamian M, Ryan J, Hobbins JC (1988) The clinical value of aas liauid chromatoaraohv in the detection of intraamniotic microbiil in&ion. Obstet”G&col 72:44-50. 21. Mosmann T (1983) Rapid calorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays.J Immunol Methods 65:55-63. 22. May LT, Ghrayeb J, Santhanam U, Tatter SB, Sthoeger Z, Helfgott DC, Chiorazzi N, Grieninger G, Sehgal PB (1988) Synthesis and secretion of multiple forms of &-interferon/B cell differentiation factor-Uhepatocyte stimulating factor by human fibroblasts and monocytes. J Biol Chem 263:7760-7766. 23. Jablons DM, Mule JJ, McIntosh JK, Sehgal PB, May LT, Huang CM, Rosenberg SA, Lotze MT (1982) Interleukin-hl interferon-B, as a circulating hormone: induction by cytokine administration in man. J Immunol 142:1542-1547. 24. Helfgott DC, Tatter SB, Santhanam U, Clarick RH, Bhardwaj N, May LT, Sehgal PB (1989) Multiple forms of IFN-&/IL-6 in serum and body fluids during acute bacterial infection. J Immunol 143:948-953. 25. Fong Y, Moldawer LL, Marano M, Wei H, Tatter SB, Clarick RM, Santhanam U, Sherris D, May LT, Sehgal PB, Lowry SF (1989) Endotoxemia elicits increasing circulating B,-IFN/ILd in man. J Immunol142:2321-2324. 26. Grossman RM, Krueger J, Yourish D, Granelli-Piperno A, Murphy DP, May LT, Kupper TS, Sehgal PB, Gottlieb AB (1989) Interleukin-6 (IL-6) is expressed in high levels in psoriatic skin and stimulates proliferation of cultured human keratinocytes. Proc Nat1 Acad Sci USA 86:6367-6371.
‘The Rockefeller University, New York, NY 10021. *The Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, CT 06510. ‘Basic Research Laboratories, Toray Industries, Kamakura 248, Japan. 4Mishima Plants, Toray Industries, Mishima 411, Japan. *To whom correspondence should be addressed at The Rockefeller University, 1230 York Avenue, New York, NY 10021. Copyright o 1991 by W.B. Saunders Company 1043-4666/91/0302-0011$5.00/O KEY WORDS: amniotic fluid/B9 hybridoma IL-6 ELI&A/pregnancy and labor/premature
CYTOKINE,
Vol. 3, No. 2 (March),
growth factor assay/ rupture of membranes
1991: pp 155-163
Recent studies suggest that immunologically active cytokines participate in the pathophysiology of normal and abnormal pregnancy and parturition. Colonystimulating factor-l (CSF-1) has been implicated in the process of implantation,’ and granulocyte-macrophage colony-stimulating factor (GM-CSF) has been shown to stimulate placental growth.233 Interleukin-1 (IL-l) and tumor necrosis factor (TNF) have been implicated in the initiation of parturition in the setting of intrauterine infection.4-8 Indeed, IL-l and TNF have been detected in the amniotic fluid (AF) of women in 155
156 / Santhanam
et al.
CYTOKINE,
Vol.
3, No. 2 (March
1991: 155-163)
when associated with active parturition (p < 0.001). A marked elevation in the AF IL-6 level was sometimes, but not always, accompanied by an elevation of the IL-6 level in the fetal circulation and vice versa, suggesting that separate pathologic events trigger IL-6 accumulation in these two intrauterine compartments. At best, only trace levels of IL-6 biological activity and immunoreactivity were observed in the maternal circulation, even in women with markedly elevated amniotic fluid and/or fetal plasma IL-6. The present data clearly implicate IL-6 in the physiology of normal human parturition and also point to a strong association between markedly elevated AF or fetal plasma IL-6 levels and abnormal labor that accompanies intrauterine infection. Copyright o 1991 by W.B. Saunders Company
preterm labor (PTL) with intraamniotic infection.4” These cytokines are produced by human decidua in response to endotoxin,6 and IL-l and TNF can stimulate the production of prostaglandins by human amnion and decidua.‘,’ These observations prompted us to investigate the participation of IL-6 in parturition, as this cytokine is a major mediator of the host response to infection and tissue injury. IL-6 is produced by endometrial stromal cells in response to IL-l and interferon-y (IFN-Y)~ and by decidual explants in response to endotoxin.” Moreover, we have previously reported that IL-6 is normally present in human amniotic fluid and that its concentrations increase during spontaneous labor at term and in women with preterm labor and intraamniotic infecti0n.l’ From our previous study, we were unable to dissect the relative contributions of parturition per se and intraamniotic infection to elevations in AF IL-6 concentrations, due to constraints of the study design. All patients in that study who had intraamniotic infection were in preterm labor. The current study addresses this issue by exploiting a naturally occurring pregnancy complication, preterm premature rupture of membranes (PROM), which is also the leading identifiable cause of prematurity.‘l We have evaluated AF IL-6 levels in patients with PROM in the presence or absence of intraamniotic infection. Since patients with PROM may present with or without active labor at the time of admission, these patients afford a unique opportunity to distinguish between the individual contributions of intraamniotic infection and labor to changes in AF IL-6 levels. The presented data suggest that while active labor and PROM each coiltribute a modest increase to AF IL-6 levels, intratiterine microbial infection is the dominant factor. In some instances, we were able to sample the AF and the maternal and fetal circulation in the same patient. We report marked elevations in IL-6 levels in fetal plasma associated with PTL and intrauterine infection. The.available data suggest that the amniotic cavity, the fetal circulation, and the maternal circulation behave as three separate compartments in their
IL-6 response to infection of intrauterine tissues. The marked elevations observed in AF IL-6 levels in the presence of microbial infection strongly suggest that, in principle, this cytokine is likely to represent a reporter cytokine of considerable prognostic value. We have used both a bioassay (the B9 hybridoma growth factor assay”) and a two-antibody sandwich ELISA (based on high affinity anti-IL-6 moAb’3z13a)to evaluate AF IL-6 levels. Overall, the two assays were comparable in their ability to distinguish between infected and non-infected patients. RESULTS We have previously reported the detection of low levels of IL-6 bioactivity in AF from women in the second trimester of pregnancy and at term, in the absence of labor, and a modest increase in IL-6 levels during spontaneous labor at term using the hepatocytestimulating factor assay in Hep3B cel1s.l’ In the present study, we employed the more convenient B9 assay for determination of titers of biologically active IL-6. The AF samples in these previous three groups were retested using the B9 assay. (Fig. 1). AF from women in the midtrimester (group 1) and third trimester at term (group 2) of normal pregnancy contained IL-6 bioactivity that could be detected in the B9 assay (median = 16 U/ml in group 1 and 15 U/ml in group 2). Also, AF IL-6 titers were higher in women in spontaneous labor at term (group 3) compared to women at term who were not in labor (group 2) (median = 74 U/ml, range: 5 to 1,383 U/ml for women in active labor versus median = 15 U/ml, range: < 3 to 105 U/ml for women not in labor; p < 0.05). The validity of determining AF IL-6 concentrations with an ELISA was tested by determining the concentration of IL-6 antigen in undiluted AF samples collected in midtrimester and at term in the presence or absence of labor (n = 16, 15, and 15, respectively; samples different from those depicted in Fig. 1). Women in active labor at term had significantly higher AF concentrations of IL-6 antigen than women at term
Amniotic fluid IL-6 in parturition
1200 1100 i
IL-6 U/ml
.
600
400 t
loo0
I ‘, .*
Midtrimester (n = 27)
Figure 1. Biologically and at term.
active
IL-6
i
t I GA. -.* A
Term No Labor (n = 33)
Term Labor (n = 40)
in AF of women
/ 157
the midtrimester (group 1) or at term (group 2) (group 4a, median = 67 U/ml; group 1, median = 16 U/ml; group 2, median = 15 U/ml; p < 0.05; for comparisons between group 4a and group 1 and for group 4a and group 2, see Fig. 3). A similar set of results was obtained when determining immunoreactive IL-6 (Fig. 4). AF IL-6 antigen concentrations were significantly higher in women with PROM and without infection and labor (group 4a) than in women in the midtrimester (group 1) at term who had intact membranes (group 2) (group 4a-median = 800 pg/ml, range: 200 to 1,240 pg/ml, Fig. 4; group l-median = 200 pg/ml, range: 60 to 1,600 pgiml; group 2-median = 200 pg/ml, range: 150 to 850 pg/ml; see Fig. 4). These data indicate that rupture of the membranes per se (in the absence of labor or intraamniotic infection) is associated with increased bioactive and immunoreactive IL-6 concentrations in the amniotic cavity. The group of patients with PROM provided us with an opportunity to dissect the relative contributions of parturition and intraamniotic infection on AF IL-6 titers. In these patients, in the absence of intraam-
in midtrimester
Scatter diagram illustrates three different groups of patients (AF samples used in this analysis are the same as in Fig. 5 in reference 10): (A) midtrimester (group 1: n = 27; median = 16 U/ml; range, <3 (0 166 U/ml); (B)‘at term but not in labor (group 2: n = 33; median = 1.5 U/ml: range. <5 to 105 U/ml): and CC) at term and spontaneous labor ‘(groip’3: n = 40; median = 74 c/r& range, 5 to 1,383 U/ml). Kruskal-Wallis H = 22.30; p = 0.0000144; A compared to B: p > 0.05; B compared to C: p < 0.05; C compared to A: p < 0.05 (Dunn’s test).
without labor or in the midtrimester (Fig. 2) (midtrimester-median = 200 pg/ml, range: 60 to 1600 pg/ml; term without labor-median = 200 pg/ml, range: 1.50 to 800 pg/ml; term in active labor-median = 1,600 pgiml, range: 200 to 15,600 pg/ml; Kruskal-Wallis H = 25.24; p = 0.0000033; Dunn’s test for multiple comparisons between term labor and the other two groups p < 0.05). These data indicate that both immunoreactive and biologically active IL-6 concentrations are elevated in the amniotic fluid of women in active labor at term. The pathophysiologic events that lead to PROM in the absence of active labor are complex and poorly understood. Inflammation of chorioamniotic membranes secondary to an antecedent extraamniotic infection has been postulated to play a role.14,15Therefore, it is important to determine whether PROM per se (in the absence of intraamniotic infection) is associated with alterations in AF IL-6 titers. Women with preterm PROM in the absence of active labor and without demonstrable intraamniotic infection (group 4a, Fig. 3) had higher AF IL-6 biologic activity than those with intact membranes without labor or infection, either in
12500 1
10000 t
IL-6 pgiml
50°0 2500
t
2000 1500 i
-
-
‘::;Ii ;- ; Midtrimester (n = 16)
Figure 2. Immunoreactive and at term.
Term No Labor (n = 15)
IL-6
in AF of women
Term Labor (n = 15)
in midtrimester
Scatter diagram illustrates three different groups of patients (AF samples used in this analysis are different from those in Fig. 1 and were all obtained transabdominally): (A) midtrimester (n = 16; median = 200 pgiml; range, 60 to 1,600 pg/ml; mean = 331 pgiml; SD = 386 pgiml); (B) term but not in labor (n = 15; median = 200 pg/ml; range, < 100 to 850 pgiml; mean = 296 pgiml; SD = 243 pgiml); and (C) term active labor (n = 15; median = 1,600 pg/ml; range, 200 to 15,600 pgiml; mean = 3,300 pg/ml; SD = 4,185 pgiml). Kruskal-Wallis H = 25.24; p = 0.0000033; A compared to B: p > 0.05; B compared to C: p < 0.05; C compared to A: p < 0.05 (Dunn’s test).
158 I Santhanam
CYTOKINE,
et al.
75000
. ..
. 45000 F
tt V
15000
.
11000 i
.
9000 7000 5000 IL-6 U/ml
. :
i Y-
t
3000
.t.
f*
1250
.
.
1000
”
Figure
3.
No Labor No Infection (n = 17)
Biologically
active
Labor No Infection (n = 25)
IL-6
No Labor Infection (n = 21)
in Al7 of women
Vol. 3, No. 2 (March 1991: 155-163)
pg/ml, p < 0.05). These data indicate that in the setting of microbial invasion of the amniotic cavity, parturition is associated with a marked increase in AF IL-6 bioactivity and immunoreactivity. The possibility that IL-6 produced by infected intrauterine tissues may enter the maternal circulation, leading to systemic effects in the mother, or that IL-6 may be present in the fetal circulation in such patients, was evaluated by comparing biologically active IL-6 titers in AF and in maternal plasma of patients who presented with PTL and fever (Table 1). In some of these instances, we were also able to simultaneously evaluate IL-6 titers in fetal plasma and, depending upon availability of material, to verify the presence of IL-6 using the ELBA. Table 1 shows that many of the fetal plasma samples tested contained high levels of IL-6. Only low levels of IL-6, however, were observed in the maternal circulation even in the presence of high AF and fetal plasma IL-6 titers. An inspection of the
Labor Infection (n = 33)
with
2. 27000
PROM.
Scatter diagram illustrates four different groups of patients: (A) no infection and no delivery (group 4a; n = 17~;median = 67‘U)ml; range. 24 to 479 U/ml: mean = 101 U/ml: SD = 103 U/ml): (B) no infeition but active labor (group 4b: n = 25; median = lb;9‘U~ml; range, 10 to 967 U/ml; mean = 238 U/ml; SD = 260 U/ml); (C) infection but no delivery (group 4c: n = 21; median 336 U/ml; range, 8 to 56,978 U/ml; mean = 4,142 U/ml; SD = 12,367 U/ml); and (D) infection and labor (group 4d: n = 33; median = 4,330 U/ml; range, 27 to 67,140 U/ml; mean = 12,238 U/ml; SD = 17,435 U/ml). After logarithmic transformation, ANOVA indicated F = 18.7, p < 0.00001. Multiple comparisons were made with the Scheffe’s test. A p value of <0.05 was considered significant. IL-6 in 4d was significantly higher than in all other groups; IL-6 in 4c was higher than in 4a but not higher than in 4b. No differences were detected between groups 4a and 4b.
i
c
I
22000
I .
-
17000 -
i . .
. . .
12000 -
. ;;j:,
7000 I I . . 1500 .
i ’
1000
,
. .
.
.
.
.
-
-
l . :
-
.
. !
niotic infection, preterm parturition was not associated with a further elevation in AF IL-6 titers (Figs. 3 and 4, group 4b versus 4a; p > 0.05). Women with intraamniotic infection (Figs. 3 and 4, pooled data from groups 4c and 4d) had significantly higher AF IL-6 levels than women without infection (pooled data from groups 4a and 4b) (bioactive IL-6, intraamniotic infectionmedian = 1,557 U/ml; range: 8 to 67, 140 U/ml versus non-infected women-median = 90 U/ml; range: 10 to 967 U/ml; p < 0.00001, Mann-Whitney U test; immunoreactive IL-6, intraamniotic infection-median = 14,400 pg/ml; range: 200 to 28,800 pg/ml versus non-infected women-median = 800 pg/ml; range: 20026,800 pg/ml; p < 0.0001, Mann-Whitney U test). Among women with intraamniotic infections, those with spontaneous PTL (group 4d) had significantly higher AF IL-6 biological activity and IL-6 antigen (median = 4,330 U/ml and 18,800 pg/ml) than those without labor (group 4c, median = 336 U/ml and 2,320
500
1
:
2.
.. .
.. .
01 No
Labor No Infection (n = 12)
Figure
4.
Immunoreactive
.
:.
Labor No Infection (n = 23)
IL-6
No Labor Infection (n = 14)
in AF’ of women
Labor Infection (n =31)
with PROM.
Scatter diagram illustrates four different groups of patients (the AF samples analyzed were from within the groups described in Fig. 3): (A) no infection and no delivery (n = 12; median = 800 pg/ml; range, 5200 to 1,240 pgiml; mean = 703 pg./ml; SD = 401 pg/ml); (B) no infection but in labor (n = 23; median = 1,000 pg/ml; range, 200 to 45,000 pg/ml; mean = 2,923 pgiml; SD = 5,638 is/ml); CC) infection but no delivery (n = 14, median = 2,320; range, 5 200 to 16,000 pg/ml; mean = 5,020 pgiml; SD = 5,407 pgiml); and (D) infection and labor (n = 31: median = 18.800 Da/ml: range. 1.280 to 92,500 pgiml; mean ‘= 16,904 pgiml; SD 2 9,34
Amniotic
&ml
1 2 3 4 5 6
< 10 117 Cl0 6,819
150 1,200 < 100 5,900 < 100 2,700
7
676
500
/ 159
The present data confirm that IL-6 is a physiologic constituent of AF during the second and third trimesters of pregnancy. Additionally, there is no change in the concentration of biologically active or immunoreactive IL-6 with advancing gestational age. Spontaneous term parturition is associated with modest elevations of AF IL-6 levels. Furthermore, preterm PROM is also associated with a modest, albeit significant, elevation of Al! IL-6 levels. However, intraamniotic infection in this setting is associated with a dramatic elevation of AF IL-6 levels. The highest AF IL-6 levels are seen in women with PROM in whom intraamniotic infection is associated with spontaneous preterm parturition. Thus, although intramniotic infection is the major contributor to elevations of AF IL-6 levels, parturition also contributes to increases in AF IL-6 in such patients. Markedly elevated AF IL-6 in patients with PROM may be a good diagnostic indicator of intraamniotic infection and be prognostic of impending spontaneous labor. The presence of IL-6 in normal amniotic fluid has now been confirmed with two different bioassays (the hepatocyte stimulating assay”’ and the B9 hybridoma growth factor assay [this report]) and two different immunoassays (immunoblot analysis” and ELISA [this report]). It is of considerable interest that there is a correlation between immunoreactive and biologically active IL-6 in AF over a wide range of concentrations (r = 0.92, p < 0.0001). This observation is of considerable importance because the B9 bioassay takes 4 days to perform, while the ELISA allows quantitation in a matter of hours. Therefore, IL-6 determinations can now be carried out in the clinical setting in a timely fashion and their results used in patient management. Further studies are required to determine whether this
Fetal Plasma U/ml
in parturition
DISCUSSION
Table 1. IL-6 in Fetal Plasma.’
Patient number
fluid IL-6
AF (U/ml)
Comments
42 24,682 6,862 12 1,4532 14,0023 1O,6544 162
PTL PTL PTL PTL Term PROM PROM
‘IL-6 in maternal plasma samples was I 10 U/ml in the B9 assay and 5 110 pg/ml by ELISA in all seven patients. *3,750 pg/ml by ELISA 3192.500 p&n1 by ELISA 4117,500 pgiml by ELISA
data in Table 1 suggests that the amniotic cavity, the fetal circulation, and the maternal circulation appear to behave essentially as three separate compartments. The appearance of a high level of IL-6 bioactivity and of IL-6 antigen in fetal plasma without an associated increase in AF IL-6 (patient #4) may be indicative of microbial infection of the fetus itself. These data demonstrate, for the first time, the detection of high levels of IL-6 in the fetal circulation using both the B9 assay and an IL-6 ELISA. Figure 5 illustrates a comparison of IL-6 B9 assay titers and of IL-6 ELISA concentrations in 77 AF samples from PROM patients in this study that we were able to analyze using both assays. The correlation between the two assays is good (r = 0.92, p < 0.0000001; Spearman correlation). This overall correlation between the two assays is reflected in the ability of both assays to distinguish between the PROM groups that were with and without microbial infection (Figs. 3 and 4). 25.0 22.5 I 20.0 17.5
Figure 5. munologic
Correlation between assays for IL-6.
the biologic
and im-
The small figure on the right illustrates the relationship between AP IL-6 concentrations determined with B9 assay and with ELISA in all 77 AF samples. Spearman correlation r = 0.92, p < 0.00001. Since most of the results are grouped below 10,000 U/ml (horizontal axis) and 25 ngiml (vertical axis), this part of the graph has been magnified to illustrate the correlation between assay results at these concentrations (left graph).
a y
15.0
w E
12.5
B 3
10.0
-
I
125.0
t
100.0 i *
I
.
10000 20000 IL-6 U/ml (B9)
IL-6 U/ml (B9)
30000
160 I Santhanam
et al.
ELISA detects all of the different biologically active human IL-6 isoforms and aggregates with equal efficiency. The immunoassay employed in this study, however, provides results qualitatively similar to the B9 bioassay when applied to the study of parturition and intraamniotic infection. The emerging picture from our studies is that while certain cytokines (e.g., IL-6 and CSF-1) are physiologic constituents of amniotic fluid,‘,” others appear only under pathologic conditions (e.g., TNF)? The biologic role and source of AF IL-6 in normal gestation remains to be elucidated. Since the production of this cytokine is a general feature of processes in which there is tissue disruption,16 its presence in normal amniotic fluid may be an indicator of the extensive fetal and maternal tissue remodeling associated with fetal growth. In situ nucleic acid hybridization analysis for cells containing IL-6 mRNA and immunohistochemical localization of IL-6 protein in tissue sections are procedures currently available to address directly questions concerning the sources of amniotic fluid IL-6. Our data indicate that spontaneous parturition at term is associated with increased concentrations of biologically active and immunoreactive IL-6 in amniotic fluid. In contrast to our previous study, in which amniotic fluid samples from women in active labor at term were obtained by transvaginal amniotomy. AF was obtained transabdominally in the current study (Figs. 2-4). AF collected transvaginally might be contaminated with endocervical or vaginal flora, and therefore, cultures of this material may not reflect the microbiologic state of the amniotic cavity. The elevation of AF IL-6 in women in term labor described in our previous study could have been due to the presence of a subclinical intraamniotic infection. In the current study, patients with spontaneous labor at term had negative AF cultures for microorganisms. Thus, the elevation in AF IL-6 in these women is attributable to parturition per se rather than to a subclinical intraamniotic infection. IL-6 secretion during spontaneous labor at term may reflect tissue disruption associated with cervical ripening and dilatation and myometrial contractility. Indeed, cervical ripening has been likened to an inflammatory reaction that is under endocrine contr01.15 Recently, smooth muscle has been identified as a rich source of IL-6,17 and this cytokine may be part of the stress-related phenomenon associated with dramatic changes in myometrial contractility required for normal parturition. Patients with PROM and without labor or intraamniotic infection (group 4a) had a modest but significant increase in AF IL-6 concentrations in comparison to women with intact membranes (groups 1 and 2). These data suggest that the pathologic process responsible for
CYTOKINE,
Vol. 3, No. 2 (March
1991: 155-163)
rupture of the membranes also leads to an elevation of AF IL-6. Alternatively, the tissue trauma associated with membrane disruption and breakage may be sufficient for stimulation of IL-6 production by fibroblasts and other IL-6-producing cells in the chorioamniotic membranes. In a previous study,l’ we reported dramatic elevations of AF IL-6 concentrations (the highest measured in our laboratory in any biological fluid) in women with intraamniotic infection and preterm labor with intact membranes. In that study, the relative contribution of preterm parturition and intraamniotic infection could not be determined because patients were ascertained by the diagnosis of preterm labor, and consequently, there was not a group who had intraamniotic infection without labor. In the current study, we have addressed this question in women with preterm PROM. The results indicate that infection is the most important contributor to elevations in AF IL-6. Indeed, in the absence of intraamniotic infection, we could not demonstrate an elevation in AF IL-6 with preterm labor alone in the setting of PROM (comparisons between groups 4a and 4b). These data are consistent with similar studies of AF IL-l and TNF that collectively indicate that preterm parturition in the absence of infection is not associated with increased secretion of cytokines into the AF compartment.“’ In the case of K-6, it is possible that the elevation associated with membrane rupture per se may obscure the effect of preterm parturition. This view is supported by our previous observation that preterm parturition with intact membranes is associated with elevated levels of AF IL-6, even in the absence of a demonstrable intraamniotic infection.” In contrast to preterm parturition in the absence of intraamniotic infection, preterm labor in the presence of intraamniotic infection is associated with a significant increase in AF IL-6 concentrations (comparisons of groups 4c and 4d). Similar observations have been made with IL-1.4 These data indicate that the onset of parturition in the setting of infection is temporally associated with the secretion of cytokines into the amniotic cavity and support the view that the initiation of parturition may be viewed as part of the repertoire of host-defense mechanisms elicited in response to intraamniotic infection. In comparing IL-6 levels in maternal plasma, fetal plasma, and AF in women with fever and preterm labor, it is striking that some of the fetal plasma samples had unusually high levels of IL-6. At best, only low levels of IL-6 were detected in maternal plasma, suggesting that sampling the maternal circulation for IL-6 content is not likely to be a useful test in the management of these patients. The ease of sampling of the amniotic fluid compartment and the marked elevations of IL-6 observed in this compartment suggest that
Amniotic fluid IL-6 in parturition /
AF IL-6 measurements are likely to prove of greater value. It is apparent from the data in Table 1 that AF IL-6 levels do not always correlate with fetal plasma IL-6 levels. The ease of sampling the amniotic cavity versus that of sampling the fetal circulation in utero makes AF IL-6 measurement a more generally applicable procedure. The biologic role of IL-6 in maternal and fetal physiology remains to be clearly defined. Elevation of CRP levels in fetal plasma appears to be of protective value because the survival rate of neonates with sepsis and high serum CRP plasma is greater than those with no detectable serum CRP.‘*,19 It is likely that IL-6 in the fetal plasma enhances not only acute phase plasma protein synthesis in the liver but also the ability of the fetal immune system to fight off microbial infection of the fetal tissues. To summarize, our data suggest that IL-6 is an excellent reporter cytokine whose AF levels are markedly elevated in the presence of intraamniotic infection. The present study justifies a prospective evaluation of the diagnostic and prognostic value of AF IL-6 measurements in patients with preterm PROM.
within 24 h of amniocentesis. Pelvic examinations were not performed at the time of amniocentesis, since this could result in an increased risk of infection.14 Tocolytic agents were not used in patients with preterm PROM. After defining the criteria for entry into the study, AF samples were obtained from the AF bank of the Department of Obstetrics and Gynecology, Yale University School of Medicine. This bank consistsof AF samples that remain after clinical studies. AF samples were centrifuged at 400g for 10 min at 4°C immediately after collection, and the fluid was separated from the cell pellet and stored frozen at -70°C. AF samples analyzed in this study were from consecutive patients admitted to the hospital who fulfilled the criteria of each of the study groups. Informed consent was obtained from all patients following the guidelines of the Yale University Human Investigations Committee.
Collection of Maternal
Patient Population
and Collection of Amniotic
Fluid
AF was collected transabdominally from women in four different populations: group 1, women in the midtrimester of pregnancy (gestational age: 16 to 18 weeks) who underwent
amniocenteses for genetic indications (maternal age > 35 years) (n = 43); group 2, women in the third trimester (gestational age: 37 to 40 weeks) who had amniocenteses for the assessmentof fetal lung maturity before elective cesarean sections (n = 48); group 3, women in spontaneous active labor (cervical dilatation of at least 4 cm) who underwent a primary or repeat cesarean section or women who were suspected to have preterm labor and who subsequently delivered ( < 24 h) a term newborn (by Dubowitz evaluation) with birthweight above 2,800 g (n = 40); and group 4, women admitted with preterm PROM on whom amniocenteses were performed for microbiologic assessment of the amniotic cavity and for fetal lung maturity studies (n = 96). To study the relationship between intraamniotic infection, preterm labor, and AF IL-6 levels, a cross-sectional study of patients in group 4 (PROM) was constructed. These patients were subdivided into four subgroups according to the results of their AF culture and the presence or absence of labor at the time of amniocentesis, as follows: group 4a, women without labor or infection (n = 17); group 4b, women with labor but without an intraamniotic infection (n = 25); group 4c, women with an intraamniotic infection but without labor (n = 21); and group 4d, women with both intraamniotic infection and labor (n = 33). An intraamniotic infection was defined as the presence of a positive amniotic fluid culture. Labor was defined as the presence of regular uterine contractions that led to cervical dilatation and resulted in delivery
and Fetal (Neonatal) Plasma
Maternal blood was collected before amniocentesis, and fetal umbilical cord blood, at the time of delivery. Blood was collected in vacutainer tubes containing EDTA (BectonDickinson, Rutherford New Jersey) centrifuged at 4008 for 10 min at 4”C, and the plasma was separated and stored frozen at -70°C.
Microbiologic
MATERIALS AND METHODS
161
Studies
AF in a capped plastic syringe was transported to the microbiology laboratory immediately after collection and plated within 30 min of amniocentesis. AF was cultured for aerobic and anaerobic bacteria as well as for Mycoplasma hominis and Ureaplasma urealyticum, as described previ0us1y.~~AF samples from women in groups 1 and 2 were sterile.
IL-6 Bioassay The IL-6 content of body fluids was assayedby monitoring the proliferation of the IL-bdependent murine hybridoma cell line B9.” Briefly, duplicate sets of serial threefold dilutions, starting at a 1:20 dilution of the test body fluid, were added to B9 cells in 96-well plates (final volume 200 ~1 containing 5,000 cells). The cultures were incubated for 92 to 96 h, and the growth of B9 cells was then quantitated by the uptake of MTT (3-[4,5 Dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide; Sigma, St. Louis, MO) and measured at 5701690 nm wavelengths using the EL312 Biotek microplate reader.‘l Neutralizing rabbit anti-rIL-6 antiserum confirmed that the proliferative response of B9 cells to representative samples of human body fluids was specific for IL-6.” Prior to use in the assay,the plasma or serum samples, but not the AF samples, were heated at 56°C for 30 min and then clarified by centrifugation. Dilutions of unheated human plasma/serum, but not of AF, inhibited B9 cell proliferation in response to exogenous IL-6. Interassay variation was a maximum of one dilution step (threefold). IL-6 titers, expressed as laboratory reference units (U/ml) in body fluids, are the reciprocals of the highest dilutions that produced a half-maximal proliferative response in the B9 assay. Each assay is calibrated in terms of
162 / Santhanam et al.
CYTOKINE,
the biologic activity of an IL-6 laboratory reference standard (purified IL-l-induced, human fibroblast-derived, natural IL-6 preparation 060888 or 071989). These reference preparations contain multiple differentially modified forms of IL-6 in the size range from 23 to 30 kD as characterized by SDS polyacrylamide gel electrophoresis (PAGE) and immunoblot analyses, using a rabbit antiserum to Escherichia coli-derived human IL-6.2’ The mass concentration of IL-6 in these laboratory reference preparations has been estimated by immunoblot comparisons (using the polyclonal anti-rIL-6 antiserum) of these IL-6 preparations with preparations of electrophoretically homogeneous, E. coli-derived IL-6, which have, in turn, been quantitated by Coomassie blue staining and by silver staining. In our hands, 1 U/ml of B9 activity of natural glycosylated IL-6 corresponds to a concentration of 15 pg/ml. In this assay, the E. coli-derived IL-6 preparation 881514, distributed by the NIH as an interim human IL-6 reference standard, titers at least threefold higher than its arbitrarily assigned potency (36,296 U/ml vs. 12,500 arbitrarily assigned U/ml designation). Furthermore, one U/ml of E. coli-derived IL-6 881514in the B9 assaycorresponds to 70
pg/ml. It is not yet clear how titers of biologically active IL-6 in human body fluids that are determined using the B9 assay and expressed as units per milliliter relate numerically to those obtained using other biologic and immunologic assays for IL-6?3-26The difficulties encountered in comparing quantitative
data between
different
bioassays include
the inhibi-
tory effect of human body fluids in the B9 assayz6and the observation that the slope of the log concentration-response plot in the B9 assay for IL-6 in body fluids is often quite different from that of the reference IL-6 preparation. Apparent IL-6 titers in human serum in the B9 assay have been reported to be considerably lower than would have been expected on the basis of the hepatocyte-stimulating factor assay.23,26Our experience comparing AF IL-6 levels using the B9 assay and the hepatocyte (Hep3B) assay indicates that although
there
is an overall
correlation
between
the IL-6
concentrations estimated using the two bioassays, the B9 assay registers a lower level of IL-6 biologic activity in AF than that detected as IL-6-specific biologic activity in Hep3B cells.1o,26
IL-4 ELBA To evaluate a faster assayfor IL-6 that could be used in patient
management,
we have assayed AF samples using a
sandwich ELISA developed recently.‘3z’3”Immunoreactive IL-6 detected in AF and plasma samples using this assay is expressed as picograms per milliliter of antigen calibrated in terms of a particular preparation of purified natural fibroblastderived IL-6; interassay variation was < 10%. The absolute amount of IL-6 in this preparation
was determined
by amino
acid analysis. This assaywas validated using AF from women in normal midtrimester and third-trimester pregnancies. Undiluted AF did not interfere with the detection of exogenously added IL-6 (Toray IL-6 standard preparation or preparation NIH 88/514). The relationship between IL-6 determinations with the B9 assay and the ELISA was evaluated by comparing the results of the two assayson a set of 77 AF samples from patients with PROM.
Vol. 3, No. 2 (March 1991: 155-163)
Statistical Analysis A Wilk-Shapiro test was employed to test the hypothesis that AF IL-6 concentrations were normally distributed. Logarithmic transformation was required for normality. Oneway analysis of variance (ANOVA) was used to compare AF IL-6 concentrations among patients with preterm PROM. Multiple comparisons among the four subgroups were performed
with a Duncan’s
multiple
range test. A p value of
< 0.05 was considered statistically significant. A KruskalWallis test was used to compare AF IL-6 concentrations between multiple
groups 1,2, and 3. A Dunn’s test was employed for comparisons (True Epistat, Epistat Services, Rich-
ardson, TX). A logarithmic transformation was not used for these values since several patients had less than detectable levels of AF IL-6 bioactivity. The correlation between the B9 assayand ELISA was tested using the Spearman correlation rank test.
Acknowledgments We thank Dr. Igor Tamm for his enthusiastic support and Carol Brekus, C.N.M., M.S.N., and Valerie Kraemer, M.F.A., for their editorial assistance. Supported by Research Grant CA-44365 (P.B.S.) and a Physician-Scientist Award from the National Institutes of Health (R.R.) and a contract from the National Foundation for Cancer Research.
REFERENCES 1. Pollard JW, Bartocci A, Arceci R, Orlofsky A, Ladner MD, Stanley ER (1987) Apparent role of the macrophage growth factor, CSF-1, in placental development. Nature 330:484-486. 2. Athanassakis I, Bleackly RC, Paetkau V, Guilbert L, Barr PJ, Wegmann TG (1987) The immunostimulatory effects of T cells and T cell lymphokines on murine fetally derived placental cells. J Immunol138:37-44. 3. Wegmann TG, Athanassakis I, Guilbert L, Branch D, Dy M, Menu E, Chaouat G (1989) Placental immunotrophism: maternal T cells enhance placental growth. Am J Reprod Immunol15:566-569. 4. Romero R, Brody DT, Oyarzun E, Mazor M, Wu YK, Hobbins JC, Durum SK (1989a) Infection and labor. III. Interleukin-l: a signal for the onset of parturition. Am J Obstet Gynecol 160:1117-1123. 5. Romero R, Manogue KR, Mitchell MD, Wu YK, Oyarzun E, Hobbins JC, Cerami A (1989) Infection and labor. IV. Cachectintumor necrosis factor in the amniotic fluid of women with intraamniotic infection and preterm labor. Am J Obstet Gynecol 161:336-341. 6. Romero R, Wu YK, Brody DT, Oyarzun E, Duff GW, Durum SK (1989) Human decidua: a source of interleukin-1. Am J Obstet Gynecol73:31-34. 7. Romero R, Durum S, Dinarello CA, Oyarzun E, Mitchell MD (1989) Interleukin-1 stimulates prostaglandin biosynthesis by human amnion. Prostaglandins 37:13-22. 8. Romero R, Mazor M, Wu YK, Avila C, Oyarzun E, Mitchell MD (1989) Bacterial endotoxin and tumor necrosis factor stimulate prostaglandin production by human decidua. Prostaglandins, Leukotrienes & Essential Fatty Acids 37:X33-186. 9. Tabibzadeh SS, Santhanam U, Sehgal PB, May LT (1989) Cytokine-induced production of interferonp,/interleukin-6 by freshlyexplanted human endometrial stromal cells. J Immunol 142:31343139. 10. Romero R, Avila C, Santhanam U, Sehgal PB (1990)
Amniotic fluid IL-6 in parturition Amniotic fluid interleukin-6 in preterm labor: association with infection. J Clin Invest 85:1392-1400. 11. Arias F, Tomich P (1982) Etiology and outcome of low birth weight and preterm infants. Obstet Gynecol60:277-282. 12. Brakenhoff JPJ, De Groot ER, Evers RF, Pannekoek H, Aarden LA (1987) Molecular cloning and expression of hybridoma growth factor in E. coli. J Immunol139:4116-4121. 13. Ida N, Hosaka T, Hosoi K, Kunitomo T, Shimazu T, Maruyama T, Matsuura Y, Kohase M (1989) Establishment of strongly neutralizing monoclonal antibody to human interleukin-6 and its epitope analysis. Biochem Biophys Res Commun 165:728734. 13a. Ida N, Sakurai S, Hosaka T, Hosoi K, Kunitomo T, Matsuura Y, Kohase M (1990) An enzyme-linked immunosorbent assay for the measurement of human interleukin-6. J Immunol Methods 133:279-284. 34. Schutte MD, Treffers PE, Kloosterman GJ (1983) Management of premature rupture of membranes: the risk of vaginal examination to the infant. Am J Obstet Gynecol144:395-399. 15. Liggans GC (1981) Cervical ripening as an inflammatory reaction. In Ellwood DA, Anderson ABM, Embrey MP (eds) The Cervix in Pregnancy and Labour: Clinical and Biochemical Investigations, Churchill Livingstone, Edinburgh, pp 1. 16. Sehgal PB, Grieninger G, Tosato G (1989) Regulation of the acute phase and immune responses: interleukin-6. Ann NY Acad Sci 557:1-583. 17. Loppnow H, Libby P (1990) Proliferating or interleukin-lactivated human vascular smooth muscle cells secrete copious interleukin-6. J Clin Invest 85:731-738. 18. Philip AGS (1985) Response of C-reactive protein in neonatal group B streptococcal infection. Pediatr Infect Dis 4:145148.
/ 163
19. Philip AGS (1979) The protective effect of acute phase response reactants in neonatal sepsis. Acta Paediatr Scan 68:481483. 20. Romero R, Scharf K, Mazor M, Emamian M, Ryan J, Hobbins JC (1988) The clinical value of aas liauid chromatoaraohv in the detection of intraamniotic microbiil in&ion. Obstet”G&col 72:44-50. 21. Mosmann T (1983) Rapid calorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays.J Immunol Methods 65:55-63. 22. May LT, Ghrayeb J, Santhanam U, Tatter SB, Sthoeger Z, Helfgott DC, Chiorazzi N, Grieninger G, Sehgal PB (1988) Synthesis and secretion of multiple forms of &-interferon/B cell differentiation factor-Uhepatocyte stimulating factor by human fibroblasts and monocytes. J Biol Chem 263:7760-7766. 23. Jablons DM, Mule JJ, McIntosh JK, Sehgal PB, May LT, Huang CM, Rosenberg SA, Lotze MT (1982) Interleukin-hl interferon-B, as a circulating hormone: induction by cytokine administration in man. J Immunol 142:1542-1547. 24. Helfgott DC, Tatter SB, Santhanam U, Clarick RH, Bhardwaj N, May LT, Sehgal PB (1989) Multiple forms of IFN-&/IL-6 in serum and body fluids during acute bacterial infection. J Immunol 143:948-953. 25. Fong Y, Moldawer LL, Marano M, Wei H, Tatter SB, Clarick RM, Santhanam U, Sherris D, May LT, Sehgal PB, Lowry SF (1989) Endotoxemia elicits increasing circulating B,-IFN/ILd in man. J Immunol142:2321-2324. 26. Grossman RM, Krueger J, Yourish D, Granelli-Piperno A, Murphy DP, May LT, Kupper TS, Sehgal PB, Gottlieb AB (1989) Interleukin-6 (IL-6) is expressed in high levels in psoriatic skin and stimulates proliferation of cultured human keratinocytes. Proc Nat1 Acad Sci USA 86:6367-6371.