Resistin in amniotic fluid and its association with intra-amniotic infection and inflammation

Abstract

Objective. Intra-amniotic infection/inflammation (IAI) is one of the most important mechanisms of disease in preterm birth. Resistin is an adipocytokine that has been linked to insulin resistance, diabetes, obesity and inflammation. The objective of this study was to determine if resistin is present in amniotic fluid (AF) and if its concentration changes with gestational age, in the presence of labour, and in IAI in patients with spontaneous preterm labour (PTL) and intact membranes, preterm prelabour rupture of membranes (PPROM) and clinical chorioamnionitis.

Study design. This cross-sectional study included 648 patients in the following groups: (1) women in the mid-trimester of pregnancy (14–18 weeks) who underwent amniocentesis for genetic indications and delivered a normal neonate at term (n = 61); (2) normal pregnant women at term with (n = 49) and without (n = 50) spontaneous labour; (3) patients with an episode of PTL and intact membranes who were classified into: (a) PTL who delivered at term (n = 153); (b) PTL who delivered preterm (<37 weeks gestation) without IAI (n = 108); and (c) PTL with IAI (n = 84); (4) women with PPROM with (n = 47) and without (n = 44) IAI; and (5) patients with clinical chorioamnionitis at term with (n = 22) and without (n = 30) microbial invasion of the amniotic cavity. Resistin concentration in AF was determined by enzyme-linked immunoassay. Non-parametric statistics were used for analyses.

Results. (1) Resistin was detected in all AF samples; (2) the median AF resistin concentration at term was significantly higher than in the mid-trimester (23.6 ng/mL vs. 10 ng/mL; p < 0.001); (3) among patients with PTL, the median AF resistin concentration was significantly higher in patients with IAI than in those without IAI (144.9 ng/mL vs. 18.7 ng/mL; p < 0.001) and those with PTL and intact membranes who delivered at term (144.9 ng/mL vs. 16.3 ng/mL; p < 0.001); (4) patients with PPROM with IAI had a significantly higher median AF resistin concentration than those without IAI (132.6 ng/mL vs. 13 ng/mL; p < 0.001); (5) no significant differences were observed in the median AF resistin concentration between patients with spontaneous labour at term and those at term not in labour (28.7 ng/mL vs. 23.6 ng/mL; p = 0.07); and (6) AF resistin concentration ≥37 ng/mL (derived from a receiver-operating characteristic curve) had a sensitivity of 85.4% and a specificity of 94.3% for the diagnosis of intra-amniotic inflammation.

Conclusions. Resistin is a physiologic constituent of the AF, and its concentrations in AF: (1) are significantly elevated in the presence of IAI; (2) increase with advancing gestation; and (3) do not change in the presence of spontaneous labour at term. We propose that resistin may play a role in the innate immune response against intra-amniotic infection.

Keywords:

• Preterm labour
• preterm delivery
• preterm prelabour rupture of membranes
• PPROM
• pregnancy
• amniocentesis
• microbial invasion of the amniotic cavity
• MIAC
• adipokines
• cytokines
• chorioamnionitis

Introduction

Intrauterine infection is one of the most important mechanisms of disease in preterm birth and is the single pathologic process for which a causal association with prematurity has been recognised [1-11]. Microbial invasion of the amniotic cavity (MIAC), elevated pro-inflammatory cytokines in the amniotic fluid (AF) and histologic chorioamnionitis are part of the ‘inflammatory cluster’ related to patients with preterm parturition [12]. Intra-amniotic infection and/or inflammation (IAI) is present in approximately one third of the patients with spontaneous preterm labour (PTL) [10],[13],[14], is associated with development of foetal inflammatory response syndrome [15],[16], and is considered a risk factor for foetal injury [17-28]. Moreover, the outcome of patients with spontaneous PTL with intra-amniotic inflammation and a negative AF culture is similar to that of patients with microbiologically proven intra-amniotic infection [14].

A solid body of evidence supports a role for pro-inflammatory cytokines in the mechanisms of preterm parturition: (1) interleukin (IL)-1β and tumour necrosis factor (TNF)-α can induce preterm parturition when administered systemically to pregnant animals [29-31]; (2) IL-1β and TNF-α stimulate prostaglandin production by amnion, decidua and myometrium [4],[32]; (3) human decidua can produce IL-1β and TNF-α in response to bacterial products [33-35]; and (4) AF IL-1β and TNF-α concentrations and bioactivity are elevated in women with PTL and intra-amniotic infection [36],[37]. Other cytokines, such as IL-6 [38-42], IL-16 [43], IL-18 [44] and IL-10 [45] have also been implicated in preterm parturition.

Adipocytokines are soluble mediators mainly produced by adipose tissue and include adiponectin, leptin, visfatin, resistin, as well as other cytokines such as TNF-α, IL-6 and IL-1 [46]. Resistin belongs to the resistin-like molecule (RELM) family of cysteine-rich proteins. In mice, resistin is exclusively expressed and secreted by adipocytes, its serum concentration is increased in both diet-induced and genetic obesity models and administration of resistin impairs glucose tolerance and insulin action [47]. These findings led to the suggestion that resistin may play an important role in glucose homeostasis and obesity-induced insulin resistance [48],[49]. In contrast to adiponectin and leptin, which are produced abundantly by adipocytes [46], resistin is mainly expressed in humans by circulating monocytes [50],[51] and macrophages [52]. Consistent with these findings, compelling evidence support a role for resistin in inflammation: (1) resistin mRNA expression is higher in human peripheral blood mononuclear cells (PBMC) after exposure to IL-1β, IL-6, TNF-α and lipopolysaccharide (LPS) [53]; (2) resistin stimulation of PBMC leads to increased expression of TNF-α, IL-1β and IL-6 mRNA, and the concentrations of cytokines released after stimulation with the highest resistin concentration are similar to those induced by LPS [54]; and (3) resistin induces nuclear factor (NF)-κB activity and its translocation from the cytoplasm to the nucleus [54]. In addition, serum resistin concentration has been reported to be significantly elevated in patients with severe sepsis or septic shock when compared with healthy controls [55], as well as in chronic inflammatory processes such as alcoholic liver disease, hepatitis C virus-induced chronic hepatitis [56], rheumatoid arthritis and osteoarthritis [54],[57],[58] and chronic kidney disease [59].

In normal pregnancy, the maternal plasma resistin concentration is significantly higher than in the non-pregnant state [60-63], its concentration increased with gestational age [61], and contradictory results have been reported regarding the maternal plasma/serum resistin concentrations in patients with preeclampsia [62],[64-66] and in those with gestational diabetes [60],[67],[68]. Of note, current data concerning resistin in pregnancy is restricted to maternal circulating concentrations of this adipocytokine and there is a paucity in the literature about its concentrations in AF. The objective of this study was to determine if resistin is present in AF, if its concentration changes with advancing gestation, spontaneous labour at term and in the presence of IAI in patients with spontaneous PTL and intact membranes, preterm prelabour rupture of membranes (PPROM) and clinical chorioamnionitis.

Materials and methods

Study design and population

A cross-sectional study was designed by searching our clinical database and bank of biological samples and included 648 patients in the following groups: (1) women in the mid-trimester of pregnancy (14–18 weeks) who underwent amniocentesis for genetic indications and delivered a normal neonate at term (n = 61); (2) normal pregnant women at term with (n = 49) and without (n = 50) spontaneous labour; (3) patients with an episode of PTL and intact membranes who were classified into: (a) PTL who delivered at term (n = 153); (b) PTL who delivered preterm (<37 weeks gestation) without IAI (n = 108); and (c) PTL with IAI (n = 84); (4) women with PPROM with (n = 47) and without (n = 44) IAI; and (5) patients with clinical chorioamnionitis at term with (n = 22) and without (n = 30) MIAC.

All women provided written informed consent prior to the collection of AF. The collection and utilisation of AF for research purposes was approved by the Institutional Review Boards of the participant institutions and the Eunice Kennedy ShriverNational Institute of Child Health and Human Development, NIH, DHHS. Many of these samples have been previously used to study the biology of inflammation, haemostasis and growth factor concentrations in normal pregnant women and those with pregnancy complications.

Definitions

Patients were considered to have a normal pregnancy outcome if they did not have any medical, obstetrical or surgical complication, and delivered a term neonate (≥37 weeks) of appropriate birth weight for gestational age [69],[70] without complications. Spontaneous PTL was defined by the presence of regular uterine contractions occurring at a frequency of at least two every 10 min associated with cervical change before 37 completed weeks of gestation that required hospitalisation. Preterm PROM was diagnosed by sterile speculum examination confirming pooling of AF in the vagina in association with nitrazine and ferning tests when necessary, before 37 weeks of gestation and in the absence of labour. Clinical chorioamnionitis was diagnosed according to the criteria proposed by Gibbs et al. [71] including maternal temperature of ≥37.8°C and two or more of the following criteria: uterine tenderness, malodorous vaginal discharge, maternal leukocytosis (≥15,000 cells/mm³), maternal tachycardia (>100 beats/min) and foetal tachycardia (>160 beats/min). Microbial invasion of the amniotic cavity was defined as a positive AF culture for micro-organisms. Intra-amniotic inflammation was diagnosed by an AF IL-6 concentration ≥2.6 ng/mL [14]. Histologic chorioamnionitis was diagnosed based on the presence of inflammatory cells in the chorionic plate and/or chorioamniotic membranes. Acute funisitis was diagnosed by the presence of neutrophils in the wall of the umbilical vessels and/or Wharton's jelly using the criteria previously described [72].

Sample collection

AF samples were obtained from transabdominal amniocentesis performed for genetic indication, evaluation of microbial status of the amniotic cavity and/or assessment of foetal lung maturity in patients approaching term. Women at term in labour consisted of women who were admitted for suspected PTL because of uncertain dates and had an amniocentesis for the assessment of foetal lung maturity. The criteria for considering that these patients were at term in labour was derived retrospectively, if the following criteria were met: (1) spontaneous labour; (2) delivery within 24 h from amniocentesis; (3) analysis of AF consistent with maturity; (4) birthweight >2500 g; (5) absence of respiratory distress syndrome or other complications of prematurity; and (6) physical examination of the newborn by paediatricians consistent with a term neonate. Samples of AF were transported to the laboratory in a sterile capped syringe and cultured for aerobic/anaerobic bacteria and genital mycoplasmas. White blood cell (WBC) count, glucose concentration and Gram-stain were also performed shortly after collection as previously described [73-75]. The results of these tests were used for clinical management. AF IL-6 concentrations were used only for research purposes. AF not required for clinical assessment was centrifuged for 10 min at 4°C and the supernatant was aliquoted and stored at −70°C until analysis. Among patients with spontaneous PTL with intact membranes who delivered within 72 h of amniocentesis, placenta, umbilical cord and chorioamniotic membranes were collected and the presence or absence of histologic chorioamnionitis and/or funisitis was assessed. The 72-h interval was chosen to preserve a meaningful temporal relationship between AF resistin concentration and placental histopathologic findings.

Determination of human resistin concentration in AF

Specific and sensitive enzyme-linked immunoassays were used to determine concentrations of resistin in human AF. Immunoassays for human resistin were purchased from Linco Research (St. Charles, MO, USA). Resistin assays were validated for use in human AF in our laboratory prior to their use in this study. The calculated inter-assay and intra-assay coefficients of variation for resistin in our laboratory were 3.4% and 4.5%, respectively. The sensitivity was 0.05 ng/mL.

Statistical analysis

The normality of the data was tested using the Shapiro-Wilk and Kolmogorov-Smirnov tests. Because AF resistin concentrations were not normally distributed, non-parametric tests were used for analyses. Comparisons between proportions were performed with the Chi-square test. Kruskal-Wallis with post-hoc analysis and Mann–Whitney U tests were used for continuous variables. Adjustment for multiple comparisons was performed using the Bonferroni method [76]. Analysis of covariance (ANCOVA) was used to investigate the association between the PTL and PPROM subgroups, AF resistin concentration and storage time. Spearman rank correlation was utilised to assess correlations between AF concentration of resistin, WBC count and IL-6. Among patients with PTL and intact membranes, receiver-operating characteristic (ROC) curve analyses were performed to determine AF resistin concentration cutoffs for the identification of patients who had MIAC and intra-amniotic inflammation. A Kaplan-Meier survival analysis was conducted to assess the amniocentesis-to-delivery interval, using an AF resistin concentration cutoff derived from the ROC curve for the presence of intra-amniotic inflammation. Spontaneous labour was entered in the analysis as the event of interest, and patients who delivered due to foetal or maternal indications were treated as censored observations with a censoring time equal to the amniocentesis-to-delivery interval. Cox proportional hazard modelling was performed to examine the differences in amniocentesis-to-delivery interval, according to resistin concentration in AF while controlling for other confounding factors [maternal body mass index (BMI), AF culture result, cervical dilatation and gestational age at amniocentesis]. A p-value of <0.05 was considered statistically significant. The statistical package used was SPSS v.15.0 (SPSS Inc., Chicago, IL, USA).

Results

Demographic and clinical characteristics of the study population

Table I presents the demographic and clinical characteristics of patients in the mid-trimester, term not in labour and term in labour groups. Tables II and III display the demographic and clinical characteristics of patients with spontaneous PTL and intact membranes and those with PPROM, respectively. Among patients with PTL with intact membranes, those with IAI had a significantly lower median gestational age at amniocentesis than those without IAI who delivered preterm and those who delivered at term (Table I). Similar results were observed between patients with PPROM with IAI than in those without IAI (Table II).

Table I.  Demographic and clinical characteristics of patients in the mid-trimester and those at term with and without spontaneous labour.

	Mid-trimester (n = 61)	p^a	Term no labour (n = 50)	Term in labor (n = 49)	p^b
Maternal age (years)	36 (35–38)	<0.01	27 (21–32)	22 (19–27)	<0.01
Gestational age at amniocentesis (weeks)	16 (16–17)	<0.01	39 (38–39)	38.5 (37.7–39.3)	NS
Gestational age at delivery (weeks)	39 (38–40)	NS	39 (38–39)	38.5 (37.7–39.3)	NS
Birthweight (g)	3,344 (3,145–3,627)	NS	3,260 (3,080–3,630)	3,360 (3,085–3,550)	NS

Values are expressed as percentage (proportion) or median (interquartile range).

NS, not significant.

^ap, comparison between patients in the mid-trimester and those at term not in labour.

^bp, comparison between patients at term not in labor and those at term in labour.

Table II.  Demographic and clinical characteristics of patients presenting with spontaneous preterm labour with intact membranes.

	PTL without IAI term delivery (n = 153)	p	PTL without IAI preterm delivery (n = 108)	p^a	PTL with IAI preterm delivery (n = 84)	p^b
Maternal age (years)	22 (19–29.8)	NS	22 (19–30)	NS	23 (20–28)	NS
GA at amniocentesis (weeks)	31.9 (29.4–33.3)	NS	31.9 (29.8–33.1)	<0.01	28.7 (25.1–33)	<0.01
GA at delivery (weeks)	38.7 (38–39.7)	<0.01	34.6 (33.3–35.6)	<0.01	29.8 (25.6–33.3)	<0.01
Birthweight (g)	3,170 (2,900–3,515)	<0.01	2,330 (1,940–2,678)	<0.01	1,310 (735–2,118)	<0.01

PTL, preterm labour; GA, gestational age; BMI, body mass index; IAI, intra-amniotic infection/inflammation; NS, not significant.

Values expressed as percentage (proportion) or median (interquartile range).

p, comparison between PTL who delivered at term and PTL without IAI.

^ap, comparison between PTL who delivered preterm without IAI and PTL with IAI.

^bp, comparison between PTL who delivered at term and PTL with IAI.

Table III.  Demographic and clinical characteristics of patients presenting with preterm prelabour rupture of membranes.

	PPROM without IAI (n = 44)	PPROM with IAI (n = 47)	p
Maternal age (years)	24.5 (20–32.8)	30 (23.8–37.5)	0.008
GA at amniocentesis (weeks)	32.6 (29.4–33.8)	30.4 (28–32.4)	0.02
GA at delivery (weeks)	33.2 (31.4–34.4)	30.7 (28.7–33)	<0.001
Birthweight (g)	2,020 (1,678–2,323)	1,660 (1,390–2,270)	0.02

Values expressed as percentage (number) or median (interquartile range).

PPROM, preterm prelabour rupture of membranes; GA, gestational age; BMI, body mass index; IAI, intra-amniotic infection/inflammation; NS, not significant.

Resistin in AF of normal pregnancies

Resistin was detected in all AF samples. Women with a normal pregnancy at term not in labour had a significantly higher median resistin concentration in AF than women in the mid-trimester [term not in labour: 23.6 ng/mL, interquartile range (IQR) 16.3–31.1 vs. mid-trimester: 10 ng/mL, IQR 6.2–15.3; p < 0.001] (Figure 1). In contrast, no significant differences were observed in the median AF resistin concentration between patients with spontaneous labour at term and those at term not in labour (term in labour: 28.7 ng/mL, IQR 19.3–39.4 vs. term not in labour: 23.6 ng/mL, IQR 16.3–31.1; p = 0.07) (Figure 1).

Figure 1. AF concentrations of resistin in normal pregnancies at mid-trimester and in those at term with and without labour. The median AF concentration of resistin was significantly higher in pregnancies at term not in labour than in those in the mid-trimester (23.6 ng/mL, IQR 16.3–31.1 vs. 10 ng/mL, IQR 6.2–15.3; p < 0.001). In contrast, no significant differences were observed in the median AF resistin concentration between women with spontaneous labour at term and those at term not in labour (28.7 ng/mL, IQR 19.3–39.4 vs. 23.6 ng/mL, IQR 16.3–31.1; p = 0.07).

AF resistin concentrations in spontaneous PTL and intact membranes

Among patients with PTL, those with IAI had a significantly higher median AF concentration of resistin than those who delivered preterm without IAI (PTL with IAI: 144.9 ng/mL, IQR 44.6–623.2 vs. PTL without IAI: 18.7 ng/mL, IQR 12.1–25.8; p < 0.001) and than those who delivered at term (PTL delivered at term: 16.3 ng/mL, IQR 12.5–22.3; p < 0.001) (Figure 2a). There were no differences in the median AF resistin concentration between patients with PTL without IAI who delivered preterm and those who delivered at term (Figure 2a). These results did not change after adjusting for storage time (ANCOVA, p < 0.001). When the analysis was restricted to patients with PTL without IAI who delivered within 48 h from the amniocentesis, this subgroup had a significantly higher median AF concentration of resistin than those with PTL and intact membranes who delivered at term (PTL without IAI delivered within 48 hours: 23.2 ng/mL, IQR 18.4–28.3 vs. PTL delivered at term: 16.3 ng/mL, IQR 12.5–22.3; p = 0.02) (Figure 2b). Similar results were found in patients with PTL and intact membranes who delivered within 7 days (p = 0.007).

Figure 2. AF concentrations of resistin among women with spontaneous PTL and intact membranes. (a) The median AF concentration of resistin was significantly higher in patients with intra-amniotic infection/inflammation (IAI) than in women without IAI (144.9 ng/mL, IQR 44.6–623.2 vs. 18.7 ng/mL, IQR 12.1–25.8; p < 0.001) as well as in those who delivered at term (144.9 ng/mL, IQR 44.6–623.2 vs. 16.3 ng/mL, IQR 12.5–22.3; p < 0.001). There was no significant difference in the median AF concentration of resistin between those who delivered preterm and those who delivered at term. (b) Patients with spontaneous PTL without IAI delivered within 48 h from the amniocentesis had a significantly higher median AF concentration of resistin than those who delivered at term (23.2 ng/mL, IQR 18.4–28.3 vs. 16.3 ng/mL, IQR 12.5–22.3; p = 0.02).

AF resistin concentrations in PPROM

Patients with PPROM and IAI had a significantly higher median AF resistin concentration than women with PPROM without IAI (PPROM with IAI: 132.6 ng/mL, IQR 32.3–869.7 vs. PPROM without IAI: 13 ng/mL, IQR 6.9–19.4; p < 0.001) (Figure 3a). These results did not change after adjusting for storage time (ANCOVA, p < 0.001). In the absence of IAI, patients with PTL who delivered preterm had a significantly higher median AF resistin concentration than those with PPROM (PTL without IAI: 18.7 ng/mL, IQR 12.1–25.8 vs. PPROM without IAI: 13 ng/mL, IQR 6.9–19.4; p = 0.003) (Figure 3b).

Figure 3. AF concentrations of resistin in women with preterm prelabour rupture of the membranes (PPROM). (a) The median AF concentration of resistin was significantly higher in patients with IAI than in those without IAI (132.6 ng/mL, IQR 32.3–869.7 vs. 13 ng/mL, IQR 6.9–19.4; p < 0.001). (b) In the absence of IAI, patients with PTL who deliver preterm had a significantly higher median AF resistin concentration than those with PPROM (18.7 ng/mL, IQR 12.1–25.8 vs. 13 ng/mL, IQR 6.9–19.4; p = 0.003).

AF resistin concentrations in clinical chorioamnionitis at term

Among patients with clinical chorioamnionitis at term, those with positive AF cultures for microorganisms had a significantly higher median AF concentration of resistin than those without (110.5 ng/mL, IQR 52.4–243.4 vs. 40.3 ng/mL, IQR 26.2–85.9, respectively; p = 0.008) (Figure 4).

Figure 4. AF concentrations of resistin in patients with clinical chorioamnionitis at term. The median AF concentration of resistin was significantly higher in patients with microbial invasion of the amniotic cavity (MIAC) than in those without MIAC (110.5 ng/mL, IQR 52.4–243.4 vs. 40.3 ng/mL, IQR 26.2–85.9; p = 0.008).

AF resistin concentrations in patients with histologic chorioamnionitis

Placental histopathologic diagnoses were available in 80% (63/79) of patients with spontaneous PTL who delivered within 72 h of amniocentesis. Of those, 51% (32/63) had evidence of placental inflammation. Patients with histologic chorioamnionitis and/or funisitis had a significantly higher median resistin concentration in AF than those without histologic inflammation (495.1 ng/mL, IQR 15.7–3554.1 vs. 29.1 ng/mL, IQR 3–846.4, respectively; p < 0.001) (Figure 5).

Figure 5. AF concentrations of resistin in patients with spontaneous PTL with and without histologic chorioamnionitis who delivered within 72 h from amniocentesis. Patients with histologic chorioamnionitis and/or funisitis had a significantly higher median resistin concentration in AF than those without histologic inflammation (495.1 ng/mL, IQR 15.7–3554.1 vs. 29.1 ng/mL, IQR 3–846.4, respectively; p < 0.001).

AF resistin concentrations and intra-amniotic inflammation

A significant correlation was observed between AF concentrations of resistin and IL-6, glucose and WBC count in patients with spontaneous PTL and those with PPROM (Spearman rho coefficient: IL-6: 0.64, p < 0.001; glucose: 0.24, p < 0.001; and WBC count: 0.4; p < 0.001).

Figure 6a,b depicts ROC curves for the identification of MIAC and intra-amniotic inflammation, respectively, in patients with spontaneous PTL and intact membranes. The cutoff values for resistin concentration in AF were derived from these ROC curves, and their diagnostic performance for the identification of MIAC and intra-amniotic inflammation are displayed in Table IV. Table V shows the diagnostic indices and the likelihood ratios of amniotic fluid resistin concentration ≥37 ng/mL for the identification of patients who delivered within 48 h, 7 days, 14 days and at <32 and <34 weeks of gestation. Using an AF resistin concentration cutoff ≥37 ng/mL, a survival analysis was conducted to determine the relationship between intra-amniotic inflammation and the duration of the amniocentesis-to-delivery interval. Patients delivered due to foetal or maternal indications were censored. The median amniocentesis-to-delivery interval was significantly shorter in patients with an AF resistin concentration ≥37 ng/mL compared to those with a concentration <37 ng/mL [median amniocentesis-to-delivery interval: 3 days (95% CI 2–4) vs. 43 days (95% CI 39–47), respectively; p < 0.001] (Figure 7). These results remained significant after adjusting for the results of the AF culture, BMI, storage time and cervical dilatation and gestational age at amniocentesis, (Cox proportional-hazards regression: 3.3, 95% CI 2.3–4.7; p < 0.001).

Figure 6. ROC curves of AF resistin concentration in patients with spontaneous PTL and intact membranes. (a) ROC curve for the identification of positive AF culture for microorganisms [area under the curve (AUC) for AF resistin: 0.932; p < 0.001]. (b) ROC curve for the identification of intra-amniotic inflammation, defined as AF IL-6 concentration ≥2.6 ng/mL (AUC for AF resistin: 0.933; p < 0.001).

Figure 7. Survival analysis of the amniocentesis-to-delivery interval (days) according to AF resistin concentration cutoff ≥37 ng/mL in patients with spontaneous PTL and intact membranes. Patients with an AF resistin concentration ≥37 ng/mL (dotted line) had a shorter median amniocentesis-to-delivery interval than those with an AF resistin concentration <37 ng/mL (solid line) (3 days, 95% CI 2–4 vs. 43 days, 95% CI 39–47, respectively; p < 0.001, log rank test).

Table IV.  Diagnostic indices and likelihood ratios of amniotic fluid resistin concentration for the detection of intra-amniotic infection and intra-amniotic inflammation in patients presenting with spontaneous preterm labour with intact membranes.

	Prevalence^a(%)	Cutoff (ng/mL)	Area under the ROC curve	Sensitivity^a(%)	Specificity^a(%)	Efficiency^a(%)	Positive likelihood ratio (95% CI)	Negative likelihood ratio (95% CI)
Intra-amniotic infection	10.1 (35/345)	≥56.1	93.2	94.3 (33/35)	91.6 (284/310)	91.9 (317/345)	11.2 (8.5–12.4)	0.06 (0.02–0.2)
Intra-amniotic inflammation	23.8 (82/345)	≥37	93.3	85.4 (70/82)	94.3 (248/263)	92.2 (318/345)	15 (10.1–21.7)	0.16 (0.1–0.2)

ROC, receiver-operating characteristic; CI, confidence interval.

Intra-amniotic infection: positive amniotic fluid culture for microorganisms.

Intra-amniotic inflammation: amniotic fluid IL-6 concentration ≥2.6 ng/mL.

^aNumbers in parentheses represent proportions.

Table V.  Diagnostic indices and likelihood ratios of amniotic fluid resistin concentration ≥37 ng/mL for the identification of patients with spontaneous preterm delivery within 48 h, 7days, 14 days and at <32 and <34 weeks of gestation

	Prevalence^a(%)	Sensitivity^a(%)	Specificity^a(%)	Efficiency^a(%)	Positive likelihood ratio (95% CI)	Negative likelihood ratio (95% CI)
Delivery within 48 h	21.2 (60/283)	66.7 (40/60)	83.4 (186/223)	79.9 (226/283)	4 (2.9–5.4)	0.4 (0.3–0.5)
Delivery within 7 days	31.4 (89/283)	61.8 (55/89)	88.7 (172/194)	80.2 (227/283)	5.4 (3.7–8.1)	0.4 (0.3–0.5)
Delivery within 14 days	38.2 (108/283)	56.5 (61/108)	90.9 (159/175)	77.7 (220/283)	6.2 (3.9–10)	0.5 (0.4–0.6)
Delivery at <32 weeks	38.4 (58/151)	72.4 (45/58)	90.3 (84/93)	83.4 (126/151)	7.5 (4.3–13.5)	0.3 (0.2–0.4)
Delivery at <34 weeks	38.7 (98/253)	64.3 (63/98)	93.5 (145/155)	82.2 (208/253)	9.9 (5.7–18.1)	0.4 (0.3–0.5)

CI, confidence interval.

^aNumbers in parenthesis represents proportions.

Discussion

Principal findings of the study

(1) Resistin is a physiologic constituent of AF; (2) its concentration in AF is significantly elevated in the presence of IAI in patients with spontaneous PTL with intact membranes, PPROM and those with clinical chorioamnionitis; (3) patients with histologic chorioamnionitis and/or funisitis have higher median AF resistin concentrations than those without; (4) AF resistin concentrations are significantly correlated with indirect AF markers for IAI (WBC count and IL-6 concentrations); and (5) AF resistin concentration increases with advancing gestation, and does not change in the presence of labour at term.

What is resistin?

Resistin is a 12.5 kDa polypeptide that belongs to the RELM family of cysteine-rich proteins, including RELM-α, RELM-β and RELM-γ, which is similar to the ‘found in inflammatory zone’ (FIZZ) family [77],[78]. In serum, resistin circulates in two different isoforms: a high molecular-weight hexamer and a low molecular-weight form of significantly increased bioactivity [79]. Resistin was originally identified as FIZZ3 in mice by Holcomb et al. [80], where FIZZ3 mRNA was expressed in white adipose tissue in association to secretion of a protein (FIZZ1) found in bronchoalveolar lavage fluid after experimental induction of an allergic inflammatory reaction in mice lungs, suggesting that FIZZ3 may play a role in inflammation. Subsequently, two independent groups [47],[81] identified resistin as a potential link between obesity and insulin resistance. Steppan et al. [47] discovered a protein that was called resistin after screening for genes induced during adipocyte differentiation but down-regulated in mature adipocytes exposed to thiazolidinediones (insulin sensitising agents). The authors found in mice that: (1) the RETN gene expression is induced during adipocyte differentiation, and is expressed and secreted by adipocytes; (2) its gene expression and protein secretion are reduced after exposure to thiazolidinediones; (3) resistin circulates in serum, and its concentration is increased in both diet-induced and genetic obesity; (4) administration of anti-resistin IgG improves blood glucose and insulin action; and (5) administration of resistin impairs glucose tolerance and insulin action in normal mice. At the same time, Kim et al. [81] reported the mRNA expression of an adipocyte-specific secretory factor (ADSF) only in murine white adipose tissue. The mRNA expression was very low or non-detectable in adipose tissue of starved or diabetic animals, but increased approximately 25-fold upon feeding or insulin administration.

These findings led to the suggestion that resistin may play an important role in adipogenesis and metabolism by impairing glucose tolerance and insulin sensitivity in mice [48],[49]. However, recent studies have shown contradictory results: (1) resistin mRNA expression was unchanged in white adipose tissue of adrenalectomised leptin-deficient ob/ob mice compared to wild-type mice, or reduced by obesity-inducing diet and fasting [82]; and (2) down-regulation of resistin mRNA was observed in adipose tissue in ob/ob and db/db murine models of type 2 diabetes when compared with wild-type controls [83]. In humans, using real-time RT-PCR, resistin gene expression was found neither in human muscle nor in most human isolated fat cells from overweight individuals or in those who had normal insulin sensitivity or were insulin-resistant or type 2 diabetic [84]. In contrast, McTernan et al. [85] found that resistin mRNA expression is similar in both omental and subcutaneous abdominal fat, but is significantly increased in abdominal depots when compared to fat from the thigh or breast. This could explain the increased risk of type 2 diabetes in patients with central obesity. Similarly, a significantly higher resistin mRNA expression in adipose tissue was found in patients with Prader-Willi syndrome (metabolically characterised by insulin resistance) than in both healthy lean controls and obese patients [86], and no association was found between serum resistin and insulin concentrations. The authors concluded that these results support a link between circulating resistin and obesity in humans but not a role for resistin in human insulin resistance, although a significant correlation has been observed between plasma concentration of resistin and insulin resistance determined by the homeostasis model assessment ratio (HOMA-R) formula [87]. Collectively, these data suggest that the role of resistin in glucose homeostasis and insulin resistance in human has not been clarified.

Resistin and inflammation

Adipocytokines are soluble mediators mainly produced by adipose tissue and include adiponectin, leptin, visfatin and resistin as well as other cytokines such as TNF-α, IL-6, IL-1 and monocyte chemotactic protein (MCP)- 1 [46]. Although adiponectin and leptin are the most abundantly expressed adipocytokines in adipose tissue, resistin is primarily expressed in circulating monocytes [50],[51] and macrophages [52]. Indeed, Anderson et al. [88] demonstrated in humans that resistin mRNA expression in adipose tissue is much lower than that of adiponectin and leptin and, while whole blood and monocyte resistin mRNA expression was high, adipose resistin mRNA expression was low or undetectable. In addition, after LPS-induced endotoxemia, the adipose resistin mRNA expression was modest compared to the marked resistin mRNA expression observed in blood.

Although resistin has been linked to insulin resistance, diabetes and obesity [47], compelling evidence of in vitro and in vivo studies support a role for resistin in inflammation. Kaser et al. [53] reported that resistin mRNA expression is increased in human PBMC exposed to IL-1β, IL-6, TNF-α and LPS. In addition, Bokarewa et al. [54] demonstrated that: (1) resistin stimulation of PBMC led to increased expression of TNF-α, IL-1β and IL-6 mRNA, as well as up-regulation of resistin mRNA itself. This was also accompanied by increased release of TNF-α, IL-1β and IL-6 into the medium in a concentration-dependent manner; (2) the concentrations of cytokines released after stimulation with the highest resistin concentration were similar to those induced by LPS; and (3) resistin induced NF-κB activity and its translocation from the cytoplasm to the nucleus, and blockage of NF-κB with parthenolide produced a marked suppression of IL-6 activity in supernatants of PBMC. Silswal et al. [89] reported that recombinant human resistin induced secretion of TNF-α and IL-12 from murine and human macrophages similar to that obtained using LPS, and also induced nuclear translocation of NF-κB. Furthermore, Lehrke et al. [90] reported: (1) a dramatic increase in RETN gene expression and protein secretion when primary human macrophages were treated with LPS and when primary human macrophages were stimulated with TNF-α; (2) blockage of TNF-α with anti-TNF-α antibodies reduced the increase in resistin gene expression; (3) rosiglitazone has a marked anti-inflammatory effect on macrophages via PPARγ, and down-regulates resistin gene expression and protein secretion in LPS-stimulated macrophages; and (4) aspirin and rosiglitazone inhibit NF-κB, which is required for LPS induction of resistin in macrophages.

Resistin has been associated with acute infection. Serum resistin concentration is significantly higher in patients with severe sepsis or septic shock than in healthy controls [55]. The authors found a positive correlation between resistin and IL-6, IL-8, IL-10 and TNF-α serum concentration at 24 h. Interestingly, the elevation of serum resistin concentration persisted at 96 h, whereas IL-6, IL-8, IL-10 and TNF-α showed a significant decrease. In addition, resistin has been associated with chronic inflammation. Resistin mRNA expression is present in the liver and is up-regulated in patients with end-stage disease, alcoholic liver disease and hepatitis C virus-induced chronic hepatitis [56]. In addition, resistin is also present in the synovial fluid and tissue [54],[57],[58]. Resistin concentration in synovial fluid is significantly higher in patients with rheumatoid arthritis than in those with osteoarthritis [54],[57],[58], its concentration in synovial fluid is higher than in serum [54],[58] and it correlates significantly with synovial WBC count and IL-6 concentration in patients with rheumatoid arthritis [54]. Moreover, intra-articular injection of resistin in the knee joints of healthy mice caused arthritis, and histological examination revealed leukocyte infiltration of synovial tissue [54].

Collectively, there is compelling evidence suggesting that resistin has features of a pro-inflammatory cytokine and plays an important role in inflammation by eliciting cytokine production and NF-κB activation.

Resistin in human pregnancy

A limited number of studies have investigated resistin in maternal circulation. Recently, Nien et al. [61] found higher plasma resistin concentrations in pregnant women than in non-pregnant subjects, which is consistent with previous reports by Yura et al. [60], Cortelazzi et al. [62] and Palik et al. [63] In contrast, Chen et al. [64] reported that differences in serum resistin concentrations between non-pregnant and pregnant women are significant only in the third trimester. In addition, plasma resistin concentrations in normal pregnancy increase with gestational age [61]. This is in contrast with a previous report [62]; however, differences in study design and sample size may contribute to the differences between studies. Cortelazzi et al. [62] determined plasma resistin concentrations in patients with pregnancy complications and found no significant differences between normal pregnant women and patients with gestational diabetes, chronic hypertension and gestational hypertension. In contrast, patients with preeclampsia had a significantly lower plasma resistin concentration than women with normal pregnancies [62]. Although this is consistent with a previous study [64], it is not with others reporting no differences between patients with mild and severe preeclampsia and those with a normal pregnancy [65], or higher plasma resistin concentrations in patients with preeclampsia compared to controls [66]. In addition, other studies have reported that the maternal serum concentration of resistin is significantly higher [67] or lower [68] in patients with gestational diabetes compared to those without.

Resistin has also been measured in umbilical cord blood. Cho et al. [91] found that the mean umbilical cord serum resistin concentration is significantly higher than that of the maternal circulation, and that no differences were observed between male and female neonates. Similar findings were reported in another study [62] which also found no differences in umbilical cord blood resistin concentration between either the umbilical artery and vein or between neonates from mothers with gestational diabetes and those from women with normal pregnancies. Moreover, Ng et al. [92] found that neonates born from mothers with pre-existing or gestational diabetes who required exogenous insulin for glycaemic control during pregnancy had an umbilical cord plasma resistin concentration significantly lower than neonates born from mothers with normal pregnancies or neonates born to patients with gestational diabetes that required only dietary treatment. In a different study, the same authors reported that the umbilical cord plasma resistin concentration is significantly higher in term than preterm neonates, as well as in neonates delivered vaginally compared to those by caesarean delivery [93]. Finally, in a study including monochorionic twins with twin-to-twin transfusion syndrome, umbilical cord serum resistin concentration correlated positively with birthweight. Furthermore, only approximately 40% of the small-for-gestational age twins had a resistin concentration lower than that of the normal co-twin [94].

Resistin in AF

This study reports, for the first time, the identification of resistin in AF. Resistin was detected in all AF samples included in this study, suggesting that it is a physiologic constituent of the AF. In addition, AF resistin concentration increases with gestational age, which is in parallel to the results of previous studies reporting that the maternal plasma/serum resistin concentrations increase as gestation progresses. Moreover, previous studies [60],[95],[96] have found that resistin is expressed in the placenta and chorioamniotic membranes, and its mRNA expression is significantly higher in term placentas than in the chorionic villi in the first trimester, suggesting that the placenta may be the source of resistin in AF. Sagawa et al. [95] and Yura et al. [60] identified resistin mRNA expression in the syncytiotrophoblast as well as in the decidua. In patients with elective caesarean section at term, resistin mRNA expression was higher in the placenta than in the subcutaneous fat from the abdominal wall [60]. Lappas et al. [96] determined resistin release from human placenta and foetal membranes, maternal subcutaneous adipose tissue and skeletal muscle from patients with gestational diabetes and those with normal pregnancies at the time of elective caesarean section at term. Tissue explants were incubated with and without LPS, TNF-α, IL-6, IL-8, phorbol myristate acetate (PMA, a potent activator of protein kinase C), glucose, insulin, dexamethasone, progesterone and oestrogen. The authors demonstrated that, in basal conditions, all tissues secreted resistin and that placenta and choriodecidua secreted significantly more resistin than the amnion. In addition, PMA significantly increased the release of resistin from placenta and adipose tissue; insulin increased placental resistin release, whereas dexamethasone, progesterone and estrogen significantly decreased placental resistin release. The observation that resistin is expressed in the chorioamniotic membranes suggests that these tissues may contribute to the increased AF concentration of resistin with advancing gestation and in patients with IAI.

Resistin in intra-amniotic infection and inflammation

A novel observation of this study is that the median AF resistin concentration was increased in patients with IAI regardless of the membrane status, as well as in patients with clinical and histologic chorioamnionitis. Indeed, in patients with spontaneous PTL and intact membranes and those with PPROM who had IAI, the median AF resistin concentration was approximately 10-fold higher than in those without IAI. In addition, among patients with spontaneous PTL who delivered within 72 h from the amniocentesis, the median AF resistin concentration was 17-fold higher in patients with histologic chorioamnionitis than in those without. Among patients with spontaneous PTL with intact membranes, those with IAI had a significantly higher median AF concentration of resistin than those without IAI who delivered either preterm or at term, and no differences were found between the last two groups. Interestingly, the analysis including patients with spontaneous PTL without IAI who delivered within 48 h and 7 days from the amniocentesis demonstrated that these subgroups had a significantly higher median AF concentration of resistin than those who delivered at term, suggesting that resistin may play a role in the process of preterm parturition in the absence of intra-amniotic infection and inflammation. Alternatively, some patients with spontaneous PTL classified as without IAI could have subclinical infection and/or inflammation that may be responsible for initiating the parturition pathway.

In patients with spontaneous PTL as well as in those with PPROM, AF resistin concentration significantly correlated with IL-6 concentration and WBC count, which are indices of intra-amniotic infection and inflammation [38],[97-99]. It is possible that AF WBC may be the source of resistin in the amniotic cavity. Because it has been proposed that AF WBC are of foetal origin [100], and because the chorioamniotic membranes are considered to be foetal tissue, it is likely that the foetus may contribute to the increased concentration of resistin in the AF. Other possible sources of resistin in AF may be maternal, placenta and decidua.

We tested the diagnostic performance of resistin concentration in AF to identify a positive AF culture for micro-organisms and intra-amniotic inflammation (defined as an IL-6 AF concentration >2.6 ng/mL [14]) in patients with PTL and intact membranes. There was no difference in the identification of microbial invasion of the amniotic cavity and intra-amniotic inflammation, because the area under the ROC curve was almost identical for both (93.2% and 93.3%, respectively; see Table IV). We focus on the diagnostic performance of resistin for identification of intra-amniotic inflammation because the prevalence of intra-amniotic inflammation is approximately twice of that of intra-amniotic infection in patients with spontaneous PTL and intact membranes. In addition, the outcome of patients with microbiologically proven intra-amniotic infection is similar to that of patients with intra-amniotic inflammation and a negative AF culture [14]. AF resistin concentration had a sensitivity of 85.4% and a specificity of 94.3% for the diagnosis of intra-amniotic inflammation using a cutoff of ≥37 ng/mL derived from a ROC curve. Using this cutoff in patients with spontaneous PTL, the median amniocentesis-to-delivery interval was significantly shorter in patients with an AF resistin concentration ≥37 ng/mL than in those with <37 ng/mL (3 days vs. 43 days; p < 0.001). Similarly, AF resistin concentration was related to the detection of histological evidence of chorioamnionitis in patients with spontaneous PTL who delivered within 72 h of amniocentesis. Among patients who had elevated resistin in AF (≥37 ng/mL) and delivered within 72 h, 71% of them (30/42) had histologic chorioamnionitis and/or funisitis.

The findings reported herein support the view that resistin is a pro-inflammatory cytokine [46],[49],[54],[89], and suggest that resistin participates in the host response to intrauterine infection or is involved in the inflammatory response observed in patients with clinical and histologic chorioamnionitis. These results are in agreement with those reported for visfatin in AF, another adipocytokine with pro-inflammatory characteristics [101].

Conclusions

This is the first study describing the presence of resistin in AF and its association with intra-amniotic infection and inflammation. Collectively, these results suggest that resistin plays a role in normal gestation, preterm parturition and in the inflammatory process that occurs during IAI. Moreover, the concentration of resistin in AF relates to the risk for impending preterm delivery and histological chorioamnionitis.

Acknowledgement

This research was supported in part by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.