Journal of Reproductive Immunology
Volume 84, Issue 1 , Pages 111-116, January 2010

Cytokine production by peripheral blood mononuclear cells of women with a history of preterm birth

  • Morgan R. Peltier

      Affiliations

    • Women's and Children's Research Institute, Winthrop University Hospital, 222 Station Plaza N, Suite 505, Mineola, NY 11501, USA
    • Corresponding Author InformationCorresponding author. Tel.: +1 516 663 2035; fax: +1 516 663 8871.
    • ,
  • David S. Faux

      Affiliations

    • Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Utah-School of Medicine, Salt Lake City, UT, USA
    • ,
  • Steven D. Hamblin

      Affiliations

    • Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Utah-School of Medicine, Salt Lake City, UT, USA
    • ,
  • Robert M. Silver

      Affiliations

    • Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Utah-School of Medicine, Salt Lake City, UT, USA
    • ,
  • M. Sean Esplin

      Affiliations

    • Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Utah-School of Medicine, Salt Lake City, UT, USA

Received 12 December 2008; received in revised form 28 September 2009; accepted 12 October 2009. published online 18 November 2009.

Article Outline

Abstract 

Preterm birth is associated with elevated production of pro-inflammatory cytokines such as TNFα at the maternal–fetal interface. Previous studies have suggested that women with a history of preterm birth produce aberrantly strong inflammatory responses to bacterial lipopolysaccharide (LPS). However many intrauterine infections in women are associated with pathogens including Ureaplasma urealyticum, Mycoplasma hominis and Streptococcus agalactiae (group B streptococcus) that contain pro-inflammatory factors other than LPS. We evaluated whether peripheral blood leukocytes from women with a history of preterm birth produce elevated amounts of TNFα upon stimulation with pathogens associated with preterm birth and if pre-treatment with aspirin, an anti-inflammatory medication, decreases the ex vivo production of this cytokine. Heat-killed bacteria elicited increased TNFα production from leukocytes in a dose-dependent manner, but no differences in TNFα production between leukocytes from women with preterm birth and control women with term birth were detected. In women who consumed aspirin each day for one week, TNFα production was increased in leukocytes from control women stimulated with Escherichia coli and U. urealyticum, but was reduced or unchanged in leukocytes from women with preterm birth. Similar trends were observed for a subset of samples stimulated with U. urealyticum and assayed for IL-6, IL-10, IL-1β and TNFα by bead array. We conclude that leukocytes from women with a history of preterm birth do not have elevated pro-inflammatory responses to pathogens, and that reproductive history is associated with different effects of aspirin on pro-inflammatory cytokine production.

Keywords: Preterm birth, Infection, Reproductive history, TNFα

 

Back to Article Outline

1. Introduction 

Reproductive history is one of the most significant risk factors for preterm birth and appears to be consistent with regard to preterm birth subtype. Women who have had a previous spontaneous preterm birth are at greater risk for having another spontaneous preterm birth than for having a medically indicated preterm birth (Ananth et al., 2006). Preterm birth also appears to recur with similar severity, with the recurrent preterm birth often occurring at approximately the same gestational age as the initial preterm birth (Ananth et al., 2006). These characteristics, as well as the observation that women who were born preterm are at increased risk for giving birth preterm (Porter et al., 1997), suggest that genetic factors may play an important role in the condition. Other factors however, also influence the condition because most women with a prior preterm birth deliver at term in their next pregnancy.

Intrauterine infection has been strongly associated with spontaneous preterm birth (Romero et al., 2002). Bacteria are thought to stimulate maternal immune cells to produce pro-inflammatory cytokines as a part of the host response to infection (Peltier, 2003). Although many pro-inflammatory cytokines are associated with preterm birth in clinical cases, TNFα appears to be especially important. Administration of this cytokine causes preterm birth in mice (Silver et al., 1994), and administration of antibodies to TNFα blocks LPS-induced preterm birth in mice (Holmgren et al., 2008).

Most of the infections associated with preterm birth are subclinical in nature, and it is possible that women who succumb to preterm birth do so because they have aberrantly high immune responses to bacteria that colonize their reproductive tract during pregnancy. A number of studies have examined if certain TNFα gene promoter polymorphisms increase the risk for preterm birth (Engel et al., 2005, Moore et al., 2004, Roberts et al., 1999) but the results have largely been conflicting (Menon et al., 2006). It is possible that these polymorphisms alone are insufficient to cause preterm birth in the absence of infection (Macones et al., 2004) or that these polymorphisms do not result in increased TNFα secretion (Menon et al., 2006). Previous studies have demonstrated that women with a history of preterm birth have enhanced TNFα production in response to LPS relative to controls (Amory et al., 2001). It is unclear, however, whether or not there is also enhanced production of TNFα in response to bacteria that are more frequently associated with preterm birth such as genital mycoplasmas that lack LPS but contain other macrophage stimulating factors (Peltier et al., 2005, Peltier et al., 2007). Therefore, we sought to explore whether or not peripheral blood mononuclear leukocytes (PBML) from women with a history of preterm birth produce more TNFα in response to Escherichia coli, Streptococcus agalactiae (group B streptococcus), and Ureaplasma urealyticum compared to women with prior uncomplicated term deliveries. These bacterial species cause preterm birth in animal models (Gravett et al., 1994, Hirsch et al., 1995, Novy et al., 2009) and are representative of Gram negative, Gram positive and genital mycoplasmas, respectively.

Back to Article Outline

2. Materials and methods 

2.1. Preparation of bacterial pathogens 

Low-passage strains of E. coli, S. agalactiae, and U. urealyticum were purchased from the American Type Culture Collection and cultivated as directed by the supplier to late log phase. Bacteria were then concentrated by centrifugation and resuspended in culture medium (RPMI 1640+10% fetal bovine serum). Quantitative cultures of the resuspended bacteria were then established for determining the number of colony forming units of this sample. Bacteria were then heat-killed by incubation at 70°C for 1h and stored in aliquots at −70°C until use.

2.2. Subjects and blood sampling 

Subjects for the case group consisted of women with a history of at least one preterm birth between 24 and 32 weeks gestation and no previous term deliveries. Each woman had histological or culture-proven chorioamnionitis, or both. The control group consisted of women who were identified from delivery records from the last five years matched with cases for age, race and calendar date of last delivery. All control patients had at least one uncomplicated delivery at >37 weeks gestation, no history of preterm birth, and no history of chorioamnionitis. Exclusion criteria for the study included autoimmune disease, active inflammatory disease, peptic ulcer disease, use of anti-inflammatory medications, medication allergies or prior adverse reactions to aspirin, pregnancy within the previous six weeks and a positive pregnancy test at the start of the study. All women were using a reliable form of contraception during the study interval. Published data on TNFα responsiveness by whole blood cultures to LPS were used to estimate the number of patients needed for this study (Amory et al., 2001). We expected that average bacteria-stimulated TNFα concentrations would be approximately 11,243pg/ml and 3649pg/ml for women with a history of preterm and term birth, respectively and that the standard deviation would be 3568pg/ml. Under these assumptions, we found that 10 patients total (5 per group) would give the experiment 83.7% power. Institutional Review Board approval for this study was obtained at the University of Utah and patients gave written informed consent.

2.3. Culture of PBML 

PBML contain the monocyte/macrophage cell population which coordinates host defenses against bacterial pathogens through the production of pro-inflammatory cytokines such as TNFα. Therefore, quantifying TNFα production by maternal PBML provides a rapid and convenient method for evaluating an individual's innate immune responses to bacterial pathogens such as those that cause preterm birth. Peripheral blood was collected into endotoxin-free heparin tubes and transported to the research laboratory for processing within 2h. PBML were isolated by centrifugation as previously described (Peltier et al., 2000a, Peltier et al., 2000b), except RPMI was used as the culture medium. Cells were plated at 106cells/ml in RPMI 1640+10% FBS in 24-well plates and stimulated with 103, 105, or 107CFU/ml E. coli, 103, 105 or 107CFU/ml S. agalactiae, or 104, 105 or 106CFU/ml U. urealyticum for 24h in a final volume of 500μl. These levels of colonization are in the range of those in the lower genital tract of women at risk for preterm delivery (Abele-Horn et al., 2000, Krohn et al., 1991, Regan et al., 1996). Conditioned medium was then harvested and stored at −20°C until TNFα assay using ELISA reagents purchased from eBiosciences (San Diego, CA).

A subset of the samples (cultures stimulated with 106CFU/ml U. urealyticum for 24h), were further evaluated for concentrations of TNFα, IL-6, IL-1β, and IL-10 using a cytometric bead array system (Becton Dickerson, San Diego, CA) to simultaneously quantify these cytokines in the limited volume of sample remaining. Bead array assays were performed as directed by the manufacturer and evaluated on a FACScan flow cytometer (Becton Dickinson, San Diego, CA).

2.4. Anti-inflammatory treatment 

After collecting blood for the cell culture studies described above, full strength aspirin (325mg/day orally for one week) was provided to participants. After one week of treatment, a second sample of blood was collected and cultures of PBML were established to re-test responsiveness to bacteria as described above. We chose aspirin for our study because it inhibits cyclooxygenase in a fashion similar to many non-steroidal anti-inflammatory drugs such as indomethacin that are used clinically and it is safe for studies such as this one. Furthermore, low-dose aspirin is frequently used clinically during pregnancy in an attempt to improve pregnancy outcome.

2.5. Statistical analysis 

Preliminary analyses and F-tests suggested that heteroscedasticity (heterogeneity of variance) may be present; therefore, data were log-transformed. Unstimulated samples were omitted from the study because they were all below the sensitivity of the assay (15.7pg/ml) and showed no variance before and after aspirin treatment. Potential differences in TNFα concentrations were compared using the generalized linear models procedure of SAS. The mathematical model evaluated the effects of group (case vs. control), aspirin treatment (before and after aspirin treatment), concentration of bacteria, and all interactions. Clustering of observations was corrected for by specifying patient as the unit upon which repeated measures were taken. Least-squares means±S.E.M. were estimated from the model and potential differences between groups and aspirin treatments were compared using the diff option for pre-planned comparisons. For clarity of presentation, least-squares means and standard errors from untransformed data are shown; however, all tests of hypothesis (P-values) are from analyses conducted on the transformed datasets. P-values less than 0.05 were considered to be significant.

Back to Article Outline

3. Results 

All subjects recruited for this study were Caucasian women who did not use tobacco presently or during past pregnancies. There were no differences between groups with regard to maternal age (29.8±8.0 years vs. 30.2±8.6 years, case vs. controls), but delivery of the index pregnancy was significantly earlier for cases than controls (29.4±3.4 weeks vs. 39±1.2 weeks). The concentration of TNFα in unstimulated samples was at the sensitivity of the assay for all samples (15.7±0.0; mean±standard error). Bacteria treatments, in general, increased TNFα production in a dose-dependent manner. E. coli-stimulated TNFα production did not differ between women with a history of term birth or preterm birth prior to aspirin treatment. Aspirin treatment tended to reduce or have no effect on E. coli-stimulated TNFα production in cases but significantly increased the production of this cytokine in controls (Fig. 1). Similar trends in responses to aspirin between groups were observed for cultures stimulated with S. agalactiae (Fig. 2) and U. urealyticum (Fig. 3) with the exception that aspirin slightly increased TNFα production by leukocytes from cases but not controls for cultures stimulated with 103CFU/ml S. agalactiae.

  • View full-size image.
  • Fig. 1. 

    Effect of reproductive history and aspirin treatment (325mg/week) on TNFα production by PBML stimulated with E. coli. Shown are least-squares means+standard errors for 5 patients with a history of preterm birth (cases) and 5 women with a history of term birth (controls) before and after aspirin treatment for cultures stimulated with up to 107CFU/ml heat-killed bacteria (CFU/ml is indicated) for 24h. Data points marked with an asterisk indicate significant (P<0.05) effects of aspirin for that group at that concentration of bacteria. No effects of reproductive history were detected prior to aspirin treatment for any pathogen tested. Concentrations of TNFα in unstimulated samples were below the sensitivity of the assay. Note that data were log-transformed prior to analysis.

  • View full-size image.
  • Fig. 2. 

    Effect of aspirin on group B streptococci-stimulated TNFα production by women with a history of preterm (cases) or term (control) birth. Bars marked with an asterisk indicate significant effects of aspirin for that group at the indicated concentration of S. agalactiae.

  • View full-size image.
  • Fig. 3. 

    Effect of aspirin on U. urealyticum-stimulated TNFα production by women with a history of preterm or term birth. Bars marked with an asterisk indicate statistically significant effects of aspirin on TNFα production for the group (preterm birth or term birth) at the indicated concentration of U. urealyticum.

Bead array analysis of samples stimulated with 106CFU/ml U. urealyticum at 24h provided similar results for TNFα production as that observed by ELISA (Fig. 4A). Aspirin increased TNFα production from leukocytes taken from controls (P=0.001) but not cases (P=0.147). No significant differences between cases and controls were detected for levels prior to aspirin treatment for bacteria-stimulated IL-10, IL-1β, TNFα or IL-6 production. Aspirin decreased IL-β production (Fig. 4B) in cases (P=0.0259) but not controls. IL-10 levels (Fig. 4C) were not significantly affected by aspirin consumption for cases but tended to be higher (P=0.083) for controls. Control patients had increased IL-6 production in response to U. urealyticum after treatment with aspirin (P=0.029) but no effects of aspirin were detected for cases (Fig. 4D).

  • View full-size image.
  • Fig. 4. 

    Bead array analysis of conditioned medium from PBML cultures established from women with a history of term (n=5) or preterm (n=5) labor after stimulation with 106CFU U. urealyticum for 24h. Shown are least-squares means±S.E.M. for TNFα (A), IL-1β (B), IL-10 (C) and IL-6 (D). Points marked with an asterisk indicate significant effects of aspirin for that group. No effects of reproductive history were detected prior to aspirin treatment for any pathogen tested.

Back to Article Outline

4. Discussion 

We hypothesized that peripheral blood mononuclear leukocytes isolated from women with a history of preterm birth would produce greater amounts of TNFα even when stimulated with low amounts of bacteria associated with preterm birth such as E. coli, S. agalactiae and U. urealyticum compared with women without a history of preterm labor. However, we detected no differences between women with a history of preterm and term birth. Our findings differ from those of Amory et al. (2001) who found that women with a history of preterm birth produce elevated amounts of TNFα ex vivo. This discrepancy may be due to several factors such as criteria for selection of cases, types of cell cultures used and the method in which the cells were stimulated. Cases for our study were randomly selected from a patient population of women with preterm birth and histological evidence of intrauterine inflammation. Amory et al. (2001) included the additional criteria of having elevated serum levels of TNFα or IL-6. In doing so they may have introduced a selection bias for women who naturally produce high levels of pro-inflammatory cytokines that, in the absence of infectious stimuli, may not be risk factors for preterm birth (Dizon-Townson et al., 1997). Another significant difference is that our study utilized cultures of PBML stimulated with heat-killed bacteria associated with preterm birth. In contrast, Amory used purified LPS from E. coli to stimulate cultures of whole blood (Amory et al., 2001). Whole bacteria may have stimulated the immune system through Toll-like receptor (TLR)-2, TLR-9 as well as TLR-4, whereas purified LPS would only be expected to interact with TLR-4.

For the most part, aspirin increased TNFα production by leukocytes in response to bacteria in controls, but had no effect on the production of this cytokine by leukocytes obtained from patients with a history of preterm birth. Our finding that aspirin treatment increases the production of TNFα in women in the control group is consistent with previous reports (Endres et al., 1996). Aspirin may enhance the stimulation of TLR-4 in vivo to create a pro-inflammatory environment. In vitro experiments by Moratz et al., found that aspirin or heat shock induced the production of HSP-70, an endogenous TLR-4 agonist, by bone marrow-derived mast cells that was followed by increased production of TNFα and IL-6 but not in TLR-4 null animals (Mortaz et al., 2006). Alternatively aspirin may increase intestinal permeability and result in absorption of LPS and whole bacteria into the blood stream as has been shown for other non-steroidal anti-inflammatory drugs. This may lead to a pro-inflammatory phenotype (Tugendreich et al., 2006). Exposure to more pathogens may then cause an elevated immune response. The mechanism by which aspirin failed to increase TNFα production in cases (and may have decreased it) is unclear and further studies on prostaglandin production by women with a history of preterm birth are needed.

Strengths of our study include well-characterized cases and control subjects, use of multiple classes of pathogens associated with preterm birth, and multiple concentrations of bacteria tested. Limitations include small numbers of patients enrolled and the unexpectedly large patient-to-patient variation in TNFα production in response to bacteria that may have made the study underpowered. Our findings are also limited by the absence of a placebo group for studying the effects of aspirin and a lack of biochemical confirmation of aspirin consumption. Therefore, the effects of aspirin need to be considered on an “intention to treat” basis (as would be done in a randomized clinical trial). An additional limitation of this study is that we did not have the means to sequence or perform SNP analyses for the TNFα gene and its regulatory regions in these women at the time the study was performed. A recent meta-analyses, however, reported no significant effect of the more common TNFα polymorphisms (−308 G→A) on risk of preterm birth for women (Menon et al., 2006). It is noteworthy that these limitations are non-differential in nature and would tend to bias results towards the null.

In summary, patients with a history of preterm birth had similar production of TNFα in leukocytes cultured in vitro in response to bacteria as those who only had previous term births. Although there was no evidence of an exaggerated inflammatory response in PBML of women with prior preterm birth (sampled outside of pregnancy) as judged by the cytokines evaluated; however, there may be other causes of an exaggerated inflammatory response such as increased recruitment of leukocytes to uterine/placental tissues and increased responsiveness of other immune cells that were not examined in this study. Patients with and without a history of preterm birth had different responses to aspirin with regard to TNFα production after stimulation with heat-killed bacteria. Thus, inflammatory cytokines may be differentially regulated by prostaglandins in women with a history of preterm birth.

Back to Article Outline

Acknowledgements 

The authors wish to thank Ms. Michelle Lewandowski, Mr. Jess Dalton, and Drs. Antonio Frias and Torry Petersen for assistance with patient recruitment and collection of blood samples. Further thanks are extended to Dr. Cande Ananth, Division of Biostatistics and Epidemiology, Department of Obstetrics, Gynecology and Reproductive Sciences, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, New Brunswick, NJ for statistical assistance.

Back to Article Outline

References 

  1. Abele-Horn M, Scholz M, Wolff C, Kolben M. High-density vaginal ureaplasma urealyticum colonization as a risk factor for chorioamnionitis and preterm delivery. Acta Obstet. Gynecol. Scand. 2000;79:973–978
  2. Amory JH, Hitti J, Lawler R, Eschenbach DA. Increased tumor necrosis factor-alpha production after lipopolysaccharide stimulation of whole blood in patients with previous preterm delivery complicated by intra-amniotic infection or inflammation. Am. J. Obstet. Gynecol. 2001;185:1064–1067
  3. Ananth CV, Getahun D, Peltier MR, Salihu HM, Vintzileos AM. Recurrence of spontaneous versus medically indicated preterm birth. Am. J. Obstet. Gynecol. 2006;195:643–650
  4. Dizon-Townson DS, Major H, Varner M, Ward K. A promoter mutation that increases transcription of the tumor necrosis factor-alpha gene is not associated with preterm delivery. Am. J. Obstet. Gynecol. 1997;177:810–813
  5. Endres S, Whitaker RE, Ghorbani R, Meydani SN, Dinarello CA. Oral aspirin and ibuprofen increase cytokine-induced synthesis of il-1 beta and of tumour necrosis factor-alpha ex vivo. Immunology. 1996;87:264–270
  6. Engel SA, et al. Risk of spontaneous preterm birth is associated with common pro-inflammatory cytokine polymorphisms. Epidemiology. 2005;16:469–477
  7. Gravett MG, et al. An experimental model for intraamniotic infection and preterm labor in rhesus monkeys. Am. J. Obstet. Gynecol. 1994;171:1660–1667
  8. Hirsch E, Saotome I, Hirsh D. A model of intrauterine infection and preterm delivery in mice. Am. J. Obstet. Gynecol. 1995;172:1598–1603
  9. Holmgren C, et al. Evaluation of the use of anti-TNFalpha in an lps-induced murine model. J. Reprod. Immunol. 2008;78:134–139
  10. Krohn MA, Hillier SL, Lee ML, Rabe LK, Eschenbach DA. Vaginal bacteroides species are associated with an increased rate of preterm delivery among women in preterm labor. J. Infect. Dis. 1991;164:88–93
  11. Macones GA, et al. A polymorphism in the promoter region of tnf and bacterial vaginosis: preliminary evidence of gene–environment interaction in the etiology of spontaneous preterm birth. Am. J. Obstet. Gynecol. 2004;190:1504–1508
  12. Menon R, et al. Analysis of association between maternal tumor necrosis factor-alpha promoter polymorphism (−308), tumor necrosis factor concentration, and preterm birth. Am. J. Obstet. Gynecol. 2006;195:1240–1248
  13. Moore S, et al. An investigation into the association among preterm birth, cytokine gene polymorphisms and periodontal disease. BJOG. 2004;111:125–132
  14. Mortaz E, Redegeld FA, Nijkamp FP, Wong HR, Engels F. Acetylsalicylic acid-induced release of hsp70 from mast cells results in cell activation through tlr pathway. Exp. Hematol. 2006;34:8–18
  15. Novy MJ, et al. Ureaplasma parvum or mycoplasma hominis as sole pathogens cause chorioamnionitis, preterm delivery and fetal pneumonia in rhesus macaques. Reprod. Sci. 2009;16:56–70
  16. Peltier MR. Immunology of term and preterm labor. Reprod. Biol. Endocrinol. 2003;1:122
  17. Peltier MR, Freeman AJ, Mu HH, Cole BC. Characterization and partial purification of a macrophage-stimulating factor from mycoplasma hominis. Am. J. Reprod. Immunol. 2005;54:342–351
  18. Peltier MR, Freeman AJ, Mu HH, Cole BC. Characterization of the macrophage-stimulating activity from ureaplasma urealyticum. Am. J. Reprod. Immunol. 2007;57:186–192
  19. Peltier MR, Grant TR, Hansen PJ. Distinct physical and structural properties of the ovine uterine serpin. Biochim. Biophys. Acta. 2000;1479:37–51
  20. Peltier MR, Liu WJ, Hansen PJ. Regulation of lymphocyte proliferation by uterine serpin: interleukin-2 mrna production, cd25 expression and responsiveness to interleukin-2. Proc. Soc. Exp. Biol. Med. 2000;223:75–81
  21. Porter TF, Fraser AM, Hunter CY, Ward RH, Varner MW. The risk of preterm birth across generations. Obstet. Gynecol. 1997;90:63–67
  22. Regan JA, et al. Colonization with group b streptococci in pregnancy and adverse outcome. Vip study group. Am. J. Obstet. Gynecol. 1996;174:1354–1360
  23. Roberts AK, et al. Association of polymorphism within the promoter of the tumor necrosis factor alpha gene with increased risk of preterm premature rupture of the fetal membranes. Am. J. Obstet. Gynecol. 1999;180:1297–1302
  24. Romero R, Espinoza J, Chaiworapongsa T, Kalache K. Infection and prematurity and the role of preventive strategies. Semin. Neonatol. 2002;7:259–274
  25. Silver RM, Lohner WS, Daynes RA, Mitchell MD, Branch DW. Lipopolysaccharide-induced fetal death: the role of tumor-necrosis factor alpha. Biol. Reprod. 1994;50:1108–1112
  26. Tugendreich S, Pearson CI, Sagartz J, Jarnagin K, Kolaja K. Nsaid-induced acute phase response is due to increased intestinal permeability and characterized by early and consistent alterations in hepatic gene expression. Toxicol. Pathol. 2006;34:168–179

PII: S0165-0378(09)00493-8

doi:10.1016/j.jri.2009.10.002

Journal of Reproductive Immunology
Volume 84, Issue 1 , Pages 111-116, January 2010

Article Tools

* Image Usage & Resolution