WWIHNIHIIHHHHWWNWW!HWNWH'HWII oo IMO-A TH it'll LIBRARY Michigzr. State University This is to certify that the dissertation entitled lNFLAMMATlON-RELATED GENE POLYMORPHISMS AND PRETERM DELIVERY SUBTYPE IN THE PREGNANCY OUTCOMES AND COMMUNITY HEALTH STUDY presented by NICOLE MARIE JONES has been accepted towards fulfillment of the requirements for the Doctoral degree in Epidemiologx (Lain/ii; WW Major PI‘oféssoi’s Signature AM £¥fi2007 Date MSU is an Affirmative Action/Equal Opportunity Employer .-J-I-_IJAL-A-l‘h-l-I-I-I-l-I-‘-'-'-I-I-I-I-I-I-I-C_I-o-a. _._._._._._.-. .. PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5/08 K:IProlecc&Pres/CIRC/DateDue Indd INFLAMMATION-RELATED GENE POLYMORPHISMS AND PRETERM DELIVERY SUBTYPE IN THE PREGNANCY OUTCOMES AND COMMUNITY HEALTH STUDY By Nicole Marie Jones A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Epidemiology 2009 ABSTRACT INFLAMMATION-RELATED GENE POLYMORPHISMS AND PRETERM DELIVERY SUBTYPE IN THE PREGNANCY OUTCOMES AND COMMUNITY HEALTH STUDY By Nicole Marie Jones Genetic variability in pregnant women and/or their fetuses may influence the risk of preterm birth. The Pregnancy Outcomes and Community Health Study assessed eleven functional polymorphisms in nine genes involved in innate immune function. Within a subcohort, polymorphism results were tested for maternal and child interactions, gene environment interactions, and gene- gene interactions. Maternal and Child Genotype Interactions (tested in ten polymorphisms) Among non-Hispanic white women, the risk of spontaneous preterm birth was highest in the group where both mother and child carried allele 2 of interleukin 1 receptor antagonist (IL-IRN) intron 2 repeat (odds ratio= 2.4; 95 % confidence interval 1.3, 4.2). Among African American women, the risk of spontaneous preterm birth was highest when both mother and child carried the tumor necrosis factor receptor 2 198 G allele (odds ratio= 2.3; 95% confidence interval 1.0, 5.3). These finding suggest that carriage of a specific gene polymorphism by both mother and fetus may result in a synergy of susceptibility to preterm birth. Vaginal Flora and Genotype Interactions (tested in three polymorphisms) Among non-Hispanic white and African-American women with both vaginal smear data and DNA, a Nugent Score 2 4 and tumor necrosis factor-alpha genotype -238 A/G or A/A was associated with an increased risk for preterm birth (race adjusted odds ratio= 2.6; 95% confidence interval 1.2, 5.8; p-value for genotype and vaginal flora interaction=0.02). There was no increased risk with just TNF-alpha genotype or Nugent Score. These findings suggest that the relationship between genotype and preterm birth may be strengthened in the presence of specific vaginal environments. Gene-Gene Interactions (tested in eleven polymorphisms) The impact of gene—gene interactions on the risk of preterm birth was assessed with the use of multifactor dimensionality reduction. T wo significant interactions were identified in non-Hispanic white women; a redundant interaction between IL-lRN and interleukin-1 beta and a synergistic interaction between CD-l4 and tumor necrosis factor receptor type II. These findings suggest that the clinical circumstances leading to preterm delivery may be indicative of biologically alternate pathways to preterm. Future studies should continue to test the associations between inflammation-related gene and preterm birth diverse populations. As investigators develop models, careful consideration of phenotypic definition, interactions between genes and environment and interactions between maternal and fetal genotype along specific pathways to preterm birth will all be important in the search to understand causal mechanisms. Dedicated to Claudia and to my family. Thank you for believing in me. iv ACKNOWLEDGEMENTS Thank you to my dissertation guidance committee and the CDC analytic team members for valuable input and oversight: Dr. Claudia Holzman, Dr. Rachel Fisher, Dr. Karen Friderici, Dr. Mehmet Gene, Ms. Katherine Jemigan, Ms. Yan Tian, and Dr. Steven Witkin. Thank you to the members of the POUCH team for their dedication to doing the best work possible including our project directors, community research assistants, recruiters, data entry staff, and support staff. Thank you to the Epidemiology Department faculty, staff, and students for their incredible support. I have made many friends along the way and for this I am truly grateful. Finally, I would like to thank the POUCH participants for their willingness to share their stories and time with us. This work was supported by funding from the National Institute of Child Health and Human Development and the National Institute for Nursing Research, the March of Dimes Perinatal Epidemiological Research Initiative Program, the Thrasher Research Foundation, and a cooperative agreement from the Centers for Disease Control and Prevention. TABLE OF CONTENTS LIST OF TABLESvu LIST OF FIGURES ............................................................................... ix KEY TO ABBREVIATIONS .................................................................... x PART I INTRODUCTION TO PRETERM BIRTH ............................................ 1 PART II EPIDEMIOLOGIC EVIDENCE SUPPORTING A GENETIC . CONTRIBUTION TO PRETERM BIRTH ...................................................... 5 PART III THE POUCH STUDY 15 PART IV THE FOLLOW UP CHILD DNA COLLECTION..................................... 21 PART V POUCH STUDY INFLAMMATION RELATED GENE POLYMORPHISMS ................................................................................ 24 PART VI PROJECT 1- MATERNAL AND CHILD GENOTYPE COMBINATION AND PRETERM BIRTH RISK ............................................. 40 PART VII PROJECT 2- INTERPLAY OFCYTOKINE POLYMORPHISMS AND BACTERIAL VAGINOSIS IN THE ETIOLOGY OF PRETERM BIRTH 51 PART VIII PROJECT 3- GENE-GENE INTERACTIONS IN CYTOKINE 60 POLYMORPHISMS AND PRETERM DELIVERY SUBTYPE ........................... PARTIXCONCLUSION 7O 73 LIST OF TABLES TABLE 1. SUMMARY OF EPIDEMIOLOGIC STUDIES ASSESSING PRETERM BIRTH RECURRENCE RISK ..................................................... 12 TABLE 2. MATERNAL CHARACTERISTICS OF POUCH STUDY PARTICIPANTS BY RACE/ETHNICITY, MICHIGAN, 1998-2001 ..................... 20 TABLE 3. PREVIOUS STUDIES OF CANDIDATE POLYMORPHISMS ............... 33 TABLE 4. MATERNAL CHARACTERISTICS IN NON-HISPANIC WHITE AND AFRICAN-AMERICAN WOMEN WITH BOTH MATERNAL AND CHILD DNA COLLECTED ....................................................................................... 47 TABLE 5. MATERNAL MINOR ALLELE FREQUENCIES FOR TERM DELIVERIES WITH NORMAL MID-PREGNANCY MATERNAL SERUM ALPHA-FETOPROTEIN ........................................................................... 48 TABLE 6. MATERNAL OR CHILD GENE POLYMORPHISMS AND PRETERM DELIVERY .......................................................................................... 49 TABLE 7. MATERNAL AND CHILD GENOTYPE INTERACTIONS .................. 50 TABLE 8. RACE-SPECIFIC MATERNAL CHARACTERISTICS OF POUCH SUBCOHORT WOMEN WITH MATERNAL DNA AND VAGINAL SMEARS COLLECTED ....................................................................................... 58 TABLE 9. UNWEIGHTED CYTOKINE MINOR ALLELE FREQUENCY .............. 59 TABLE 10. WEIGHTED MINOR ALLELE PERCENTAGES AND VAGINAL FLORA IN NON-HISPANIC WHITE AND AFRICAN AMERICAN SUBCOHORT PARTICIPANTS ................................................................ 59 TABLE 11. TNF-ALPHA -G238A, VAGINAL FLORA, AND PRETERM DELIVERY SUBTYPES .......................................................................... 59 TABLE 12. PREVIOUS STUDIES USING MULTI-FACTOR DIMENSIONALITY TO ASSESS GENE-GENE INTERACTIONS AND RISK OF PRETERM BIRTH ............................................................. 66 vii LIST OF TABLES CONTINUED TABLE 13. CHARACTERISTICS OF NON-HISPANIC WHITE AND AFRICAN AMERICAN POUCH STUDY PARTICIPANTS WITH MATERNAL DNA ................................................................................. 67 TABLE 14. RESULTS FROM MULTI-FACTOR DIMENSIONALITY REDUCTION ANALYSIS FOR MATERNAL GENOTYPE .............................. 68 viii LIST OF FIGURES FIGURE 1. ENROLLMENT IN THE CHILDREN’S DNA FOLLOW-UP COLLECTION ...................................................................................... 23 FIGURE 2. PROPOSED BACTERIAL MEDIATED INFLAMMATION PATHWAY TO PRETERM DELIVERY WITH SELECTED CANDIDATE GENES ................................................................................................. 30 ix KEY TO ABBREVIATIONS A: Adenosine BV: Bacterial vaginosis CI: Confidence interval C: Cytosine G: Guanine IL-IB: Interleukin-1 beta IL-lra: Interleukin-l receptor antagonist MBL: Mannose binding Iectin MDR: Multi-factor dimensionality reduction MI: Medically indicated preterm birth MSAFP: Maternal serum alpha-fetoprotein MMP-9: Matrix metalloproteinase 9 OR: Odds ratio PCR: Polymerase chain reaction POUCH Study: Pregnancy Outcomes and Community Health Study PROM: Premature rupture of membranes SNP: Single nucleotide polymorphism PTD: Preterm delivery SPT: Spontaneous preterm birth TLR-4: Toll-like receptor-4 I'N F-a: Tumor necrosis factor alpha PART I INTRODUCTION TO PRETERM BIRTH In 2006 preterm delivery (the birth of an infant before 37 weeks gestation) occurred in 12.8 % of all US. pregnancies. [l] The US. preterm birth rate is much higher than that reported in other industrialized nations, for example: 6.2% for France in 1998, and 6.4 % for Sweden in 2000 and has increased by 30% since 1981.[1, 2] Most of this increase is in the number of late preterm births (those that occur at 34-36 weeks of gestation). [3] Included in the preterm birth rate are deliveries from multifetal pregnancies, malformations, medically indicated births (either induced or surgically delivered) and spontaneous live births. Several factors may have contributed to the rise in the US. preterm birth rate including an increase in the rate of multiple births, [4] an increase in rates of maternal complications such as diabetes, chronic hypertension, and anemia [5] and changes in the medical management of pregnancy (an increase in preterm births by Cesarean delivery and induction). [3] Preterm birth is a leading cause of infant mortality in the US accounting for up to one-third of all infant deaths in 2002. [6] Premature birth is a major contributor to childhood neurodevelopmental problems with neonatal complications that may include respiratory, gastrointestinal, immune system, central nervous system, hearing, and vision problems. Longer term sequelae can include cerebral palsy, mental retardation, visual and hearing impairments, behavior and social-emotional concerns, learning difficulties, and poor health and growth. [7] The Institute of Medicine estimated the 2005 societal economic burden of preterm birth in the US. to be at least $26.2 billion.[8] This estimate included the cost of medical care services, maternal delivery costs, early intervention services, special education services, and lost household and labor market productivity. There are significant racial/ethnic differences in the incidence of preterm birth in the US. African Americans are about twice as likely to deliver preterm as white women while the incidence among Asian women is much lower. These racial/ethnic differences are true for the overall preterm birth rate (less than 37 weeks) and are even stronger between whites and Afiican Americans who delivery extremely preterm (less than 28 weeks). The difference appears to persist even after adjustment for socioeconomic factors and maternal behaviors such as smoking and drug use. [9-11] In Michigan, the same racial/ethnic differences are found in preterm birth rates as at those in the US. overall. Preterm delivery can be categorized based upon clinical circumstances at delivery (hereafter referred to as preterm subtype) into medically indicated or spontaneous with or without premature rupture of membranes. Medically indicated preterm births comprise approximately 25% of all preterm births (range 8.7 % to 35.2%) [12] and include cases of poor fetal growth, fetal distress and maternal complications such as severe hypertension, placental abruption, preeclampsia or antepartum bleeding. Spontaneous preterm births without rupture of membranes prior to the onset of delivery account for 50% of all preterm births (range 23.3% to 64.1%).[12] Premature birth with rupture of membranes before the onset of labor (PROM) comprises another 25% of all preterm births (range 7.1% to 51.2%) [12] Even though the rate of preterm birth varies up to three fold across populations, the proportions of spontaneous and medically indicated are consistent estimates. [12] Established maternal risk factors for preterm birth include maternal demographic characteristics, (i.e. extremes of maternal age, African-American race, low socioeconomic status) maternal obstetric factors, (i.e. multifetal pregnancy, abnormal vaginal flora/infection, uterine and cervical anomalies, previous preterm birth) maternal behaviors, (i.e. smoking, strenuous physical work load) and maternal nutritional status (low pre-pregnancy BMI, poor weight gain during pregnancy). Other potential risk factors relate to maternal psychosocial status including maternal anxiety and maternal stress. Risk factors may be both distinct and overlapping for preterm delivery subtype. [13] For example, a study by Berkowitz found that Afiican—American women or women who initiate prenatal care after 13 weeks are at an increased risk for all subtypes of preterm birth, while high weight gain (greater than 1 pound per week) was associated only with medically indicated preterm birth and smoking was associated only with PROM.[14] There is growing evidence that both infection and/or inflammation may play a role in some pathways to preterm birth. Spontaneous preterm births have higher levels of pro- inflammatory cytokines in mid-pregnancy maternal serum, mid-pregrancy amniotic fluid, and mid-pregnancy cervical samples as compared to term births.[15-17] Higher levels of pro-inflammatory cytokines were also reported in the placental tissues of spontaneous preterm births and evidence of histologic chorioarnnionitis is more common in placentas of spontaneous preterm births.[18, 19] In non-human primate models, intramniotic infusion of pro-inflammatory cytokines resulted in the induction of preterm labor and birth.[20] One theory is that activation of inflammatory pathways and cytokine production leads to prostaglandin synthesis which initiates uterine contractions and remodeling of the collagen in the cervix.[21] The strongest and most consistent risk factor for a woman to deliver prematurely is having a history of a prior preterm birth. [22, 23] This finding is robust even after adjustment for demographic and socioeconomic risk factors.[24, 25] This observation lead to the hypothesis that individual predisposition to preterm birth may have a genetic component. Part II will present a review of epidemiologic evidence that supports a genetic contribution to preterm birth risk. PART II EPIDEMIOLOGIC EVIDENCE SUPPORTING A GENETIC CONTRIBUTION TO PRETERM BIRTH There are several types of epidemiological evidence that suggest there may be a genetic contribution to preterm birth including studies assessing recurrence risk within a woman, studies comparing maternal twin preterm birth rates, studies of preterm birth risk across generations, and studies assessing clustering of preterm birth within families. Preterm Birth Recurrence Risk (Table one) Table 1 presents a brief review of studies measuring recurrence risk for preterm delivery among singleton pregnancies. Data sources include large national registries, U.S. vital statistics, hospital records, and large birth cohort studies. National Birth Registries Data from national birth registries in three Scandinavian countries (Norway, Denmark, and Sweden) have assessed a woman’s risk of recurrence. [24, 26-28] In data from Denmark, the odds ratio for having a preterm birth following a prior preterm birth compared to having a preterm birth following a prior term birth was 5.28. [26] This analysis was adjusted for multiple risk factors for preterm birth including maternal disease, cervical colonization, spontaneous and induced abortion, pregnancy complications, and elective delivery. The risk for a recurrent preterm birth was stronger with earlier gestational age at delivery of the first pregnancy. Women who had a medically indicated preterm birth were more likely to have a second medically indicated preterm birth than a spontaneous preterm birth. Separate analyses in Danish data revealed that a change in mother for the second pregnancy significantly reduced the risk of a repeat preterm birth (adjusted odds ratio = 0.4; 95% confidence intervals=0.3, 0.6 adjusted for change of municipality, change of occupation, SES change, age, parity, social status) while change in father had very little impact on the risk of repeat preterm. [28] Norwegian data collected from 1967-1995 found an adjusted odds ratio for recurrence of 5.3 (adjusted for age, marital status, study time period, and inter-pregnancy interval). The strength of the odds ratio increased from 4.9 to 6.0 from the beginning to the end of the thirty year study. [27] Swedish data considered both the impact of preterm birth severity and changes in smoking habits between pregnancy.[24] Women who delivered very preterm (<32 weeks) had significantly increased odds of delivering very preterm in their second pregnancy (odds ratio for recurrent very preterm = 12.1; odds ratio for moderate preterm following very preterm =7.1). Women who delivered moderately preterm had a stronger recurrence risk for moderate preterm birth although their risk for very preterm was elevated (odds ratio for recurrent moderate preterm = 5.9; odds ratio for moderate following very preterm: 2.3). These analyses were adjusted for smoking, age, education, living arrangements with father, country of birth, inter-pregnancy interval, and year of delivery. Changes in smoking habits between pregnancies had very little impact on risk of recurrence. In general the findings from the Scandinavian studies suggest that a woman with a preterm birth is at least 5-fold more likely to have a subsequent preterm birth. The strengths of data collected by Scandinavian Registries include the large sample sizes and consideration of demographic, behaviOral, and medical risk factors as potential explanatory variables for the increased recurrence risk. One major limitation to these studies is that the findings are from Scandinavian data and may not be generalizable to more socioeconomic and ethnically diverse populations such as the US. US. Vital Data Five studies have used U.S. vital data to calculate a woman’s risk for preterm birth following a preterm birth.[22, 25, 29, 30] Four of these studies used linked maternal and child birth records from Missouri. In this data, women whose first pregnancy was delivered preterm (< 37 weeks) were an increased risk for a preterm birth in their next delivery. [25, 29, 30] This association persisted after adjustment for maternal age, education, marital status, smoking and alcohol use, prenatal care, prepregnancy BMI, and ”interpregnancy interval. The recurrence risk was strongest for African Americans and for women with shorter interpregnancy interval (less than six months). Further analyses found that women with more than one preterm birth had an even stronger risk for recurrence. In addition grouping women by preterm subtype showed that women were more likely to repeat with the specific subtype (i.e. medically indicated or spontaneous). There was effect modification by timing of preterm birth such that women who delivered very preterm (less than 32 weeks) had a stronger risk of delivering very preterm in a subsequent pregnancy. Data from Georgia vital records found similar results with adjustment for infant gender, year of delivery, interpregnancy interval, father’s name on the birth certificate, maternal education, maternal Smoking, and pregnancy outcome.[22] The risk of recurrence for white women was 19.9% and 27.6% for Afi'ican-American women while only 6.3% of white women with a prior term delivery had a subsequent preterm birth and 13.7% of Afi'ican-American women had a term delivery following a preterm birth. The risk of recurrence was strongest among Afiican-American teenagers. The data from US. vital statistics supports the findings from the Scandinavian registries, although both analyses lacked quality detailed information about maternal environment and could only group preterm birth subtypes based on administrative data collected after birth. Hospital Data Bases Two studies used hospital record databases to estimate the risk of preterm birth recurrence. The recurrence risk among women in a sample of low income women of all races found an unadjusted odds ratio for recurrence of 5.6; 95 % confidence interval=4.5, 7.0.[31] Once again the risk was higher for women with multiple preterm births and for the specific subtype of preterm birth. The authors of that study noted that in their population, although women were at high risk to repeat preterm delivery these women only accounted for 10 % of all preterm births. In a tertiary care inner-city hospital sample, women with a prior preterm had a preterm birth in their second pregnancy 33.6% of the time as compared to women who had a prior term delivery who delivered preterm in their second pregnancy 8.0% of the time.[32] No adjusted analyses were performed in either of the hospital record studies. These studies were designed to assess recurrence risk for the benefit of a clinician and did not attempt to consider the impact of specific maternal characteristics on the risk of repeat preterm. Birth Cohorts Two preterm birth cohort studies examined the risk of recurrent preterm birth. Analysis of data from the Preterm Prediction Study included only spontaneous births and found an unadjusted relative risk of 2.5 which was much stronger for women with early preterm (< 28 weeks odds ratio = 10.6).[33] In this cohort, there was a stronger risk of recurrence among women with a positive fetal fibronectin and shorter cervix.[34] In the Alabama Preterm Birth Project, women with consecutive singleton preterm (22-32 weeks) births women were more likely to have a spontaneous birth if they had a previous spontaneous birth than women who had a previous term, or medically indicated preterm birth. Women with previous medically indicated preterm births were more likely to have a medically indicated preterm birth as compared to women with a history of a term or spontaneous preterm birth.[35] Twin Studies Findings in two twin studies support the idea that genes play an important role in determining risk of preterm birth. The Swedish Twin and Birth Registry calculated the concordance among monzygotic and dizygotic birth mothers. The concordance was 0.22 among monozygotic and 0.11 among dizygotic twins.[36] This study estimated the heritability of preterm birth to be 37%. In the Australian National Health and Medical Council Twin Registry, the monozygotic concordance was 0.30 and the dizygotic twin concordance was .03 with an estimated heritability of 17% for the first pregnancy and 27% for any other pregnancy.[3 7] In twin studies, comparison of concordance rates among monozygotic and dizygotic twins are less likely to be due to unmeasured environmental factors and are more likely to be due to shared genetic factors than in studies of preterm birth recurrence rates. Other Epidemiologic Evidence Other sources of evidence include generation studies, genealogy databases, and preterm birth rates among relatives. One study assessed gestational age across generations and found a very weak correlation (0.086).[3 8] The adjusted odds ratio (adjusted for maternal birthweight, and birth order) for a preterm birth given that the grandmother was preterm was 1.46; 95% confidence interval=0.96-2.21. In another study, the risk of delivering preterm was increased if the mother was born preterm (odds ratio for mothers born at S 30 weeks = 2.38; 95% confidence interval=1.4, 4.2).[39] Within a Utah genealogy database, 28 families with clusters of preterm births were much more closely related to each other than randomly selected individuals fi'om the population (the coefficient of kinship was more than 50 standard deviations hi gher).[40] Another genealogic data base fi'om Lancaster County, Pennsylvania found that a lower mean gestational age and higher risk of preterm birth was associated with maternal but not paternal inbreeding.[4l] A Scottish study compared self-reported preterm birth rates among relatives of women with low birthweight babies. The preterm birth rate of sisters was 16% and among sister-in-laws was 9%.[42] In summary, epidemiologic data suggest that there is a genetic component to a woman’s risk of preterm birth. Potential effect modifiers include preterm subtype, timing of preterm birth, and race/ethnicity. Maternal factors appear to have a stronger influence than paternal factors. The major problem with epidemiologic data is that it fails to indicate whether risk truly has a genetic component or merely is a reflection of shared adverse environment. Multiple gene association studies have attempted to measure the impact of specific genes on the risk of delivering preterm. Over 70 studies have investigated possible associations between preterm birth and genetic variation in inflammatory, metabolic, and vascular related genes. The most commonly investigated pathway included genes involved in the immune system. The approaches have included studies examining a 10 single polymorphism in a single gene, multiple polymorphisms in a single gene, a single gene haplotype, single polymorphisms in multiple genes, and haplotypes in multiple genes. The number of polymorphisms examined in a single publication ranges from one to over 1500. Studies with larger numbers of polymorphisms have adjusted for multiple comparisons but most studies do not adjust. All but six studies reported at least one significant finding. Studies have included as few as 14 cases of preterm birth and as many as 593. Very few have included gene-gene or gene-environment interactions (diet, smoking, vaginal flora). No studies have separated out medically indicated preterm birth as a distinct subtype. I The Pregnancy Outcomes and Community Health (POUCH) Study assessed eleven polymorphisms in nine genes involved in innate immune function. Part III describes the methods used to assemble the POUCH cohort. 11 03E memo _c8m:.€unmo< :wESE concemnoonmo ”.0328 8535 n PE Sufi «.26 ”MO dqnmS Vnmdmvnz 3 e 25$: :3. 82%.: Bow @8853 one.— mEcO $83 3.? £5; gem n PE 9808 8‘83 VTNm 95 3&2 E 55 5:83 Maui—ow me it 838.2 3323 5389595 $603 2.3 @3qu 3&3. «Hume .8an 85 seem was” 3.58 a? asaaefiom £83 58 co hang in ages: 32 .0202 Go .3 6 $3 msoumov .mueo 3328c mafia 308 $536 EEEHZ Hogan DEED.“ E owgo 55m .oma .omgo mmw docuasooo $3-33 $53 xmt oonobzoom mo omega 5:36:38 Co 0330 §$afie§om £83 hm v 3&on .2832 swimm— ooa .ommmm denounce no comes: 2:: b? can macaw E 893:0 www.mvmuz mNuMO vauem-mm fictive moamefl qmnmo om-mmncm-mm mo So» .2352: 55505;an .53 95m 28 a: £59 :mmo wm-~mnmmv mo B58 .566 55 $880qu 383 mmv o>= ”.9295 gasoowdou 32 edumo vanmmv was: £883... awe .mfioem Ssgesom 383 3-2 Emcee Em 283m .wamafiao SQSHZ .3329: 237m:— >=mo€oa a “non“: 9 .385 $8.3 can 98 6: 808 @0339: 38:52 .0033“. 03820 was 89.385 BED Ea a 3 @8385 .3238388 heaven doEona EEmoE mo flawed SEE “EN Ohm («0 {cam 33?: e5 msoogqonm demands—co .3282 use crammed 39 56 u MO< .w~.nnm0 3228 633% 38032 nfi>§v§om $83 RV gm 73°st #3889 . xw: «013.58.. .3 non—Ewen v0.29. «names—E weir—«EU 8am Dot—on 3.83...”— J—mt 35:58.. 5:: .539. 5333 355m 0 12 moisten»: @0263 xoquuwoi n Em 369: m3: boom” film 638 3:6 vogwagnm0< swig: ooaowmaooufio ”Eel—on Shana ”PE 6:8 mEvo HMO 3:88 ommdflnz ev Eta... $5.83 $2.32 -883 5m “magma 850 "£83 onA 83503— oowumo M: v £8851”. mayo—mam 86:03qu boom $03 88 ooaobsoom own 88:82 .88 .88 3805 oz 505.... £83 mmv 8502 in «5882 .oofiumun 82.. 88:83 88.85 5 ca: 5an Swag ofim nEmem 83828 98 Emotofixx :85 . .EE .3395 $8235 35888 . mmoiomuz mm.~ HMO 833 £232.? :38an .omv 95 5085 32-33 wc.m HMO .88 Egg on .0; .8586 coo.“ G853. 3:5 v.58 ho 03H going :85 .2860: .mV 5:838 .omv om< .833 £83 nmv 8:88 5:9 €59.82 Son .839 .650 mowém _ HZ .988: 28 2:58 3288 zonaoafixfi .8885 32-33— ommoomm 8m “mowaobm 42m 8550.505 .88 78895 .5055 $303 ONA 3:53: SHE 8808 .8: 3:8? 23 macaw .383 98.5.4... .8898. 033838 9.3. ooom .8 an: 8382.: 2o.“ 5 -m 3.58 £23260 .owm 3:884 .833 £83 SV $502 582 :38de .555. .mmwwflwog 50.55 80.32 Em xmt «mowaobm .918 on 8 .385 208 EL 38% .od.wEEu How «mowuobm $5.2 u 888:3. 505 80:85.. 885 vfidmnz ciao n 8.33 08088 8:885 823 www.mmflnz E8QEAEoH .mfixofim “8:888 483838 33-32 $0.5m H S855 5851‘ E888: .88Etoo in no 380.5 8055 .339 3: $0.2 u 833 ea“: $8.88 .3385 85:85 885 8803:: 333838 9:. ooom 8805-880E -883 $8338 mo H8» .895» E83 98 8.33 383 nmv £888 :33 8980 .§< . , , 355m 33> .m.D 8.2:qu it 35:33.. b.9509 aim 2..an use? 3 6820.. “9555 8.8.3.30 3am 5.3.95 .958» 3.5m Jam 82— £053.. .3358:er — 935,—. 13 355802 we 2393 253895 n SOME 6E: «35 voggwénanw €383 oouocmqooua ”huge—u .5205 up: 6:2 £5 "mo .29: 550 a :Snz 55 23 PE 3.80 55 .833 wmvnomv £88". mb 9 it 8558“ hownobm 50:05 . $603 vm-mm am 3:25 33 Wm H MO 252 5o£< mdevamv 33m 5:335 8.8.0:— .5202 3865 =8§a .33 Essa up 8 box: 805 “.8863 550 nee—mam 3328200 33232 28.58% B .oeami “86$ 88 2 box: 205 385.5% 282 .833 Jon—m £83 QWNN in 8.5on «Esau? $588200 4 a , _ _ , $2.5m 95:90 8:55 $28:on .3855 Eamon 93039 50322 mo $.on u 885.8% 282 58%? £83 nmv $2-on 35%: nos—mam mg“ .onm PE =a mo $2 how E268 Ego magnum 958% 2593 can 8:80 Sm Swaobm $3 n GE 8.8 83H $3. n DE SE 95H 5&0 mcwfi T2 83 n 9.5 8E 0:0 Gasman 2:85 32 8% WmuEOyE 505:3. 3982 85: 938B mmzmn— Sumo 308.38% 58:? Eamom 82-32 {N Ba 88 ed HMO ooaobsoom 252 .853 8.83 2V 5: 885:qu @338800 .8005 , woman Sun 3: 3H 9:.de .th Sta mo 3.558 380 32-32 can .590 £825on 28 5855 .5055 "£83 omA 35903— 3 homaobm nouumoommzw .8: .38? Ed goaws 688.: 585 aofiwim o>zsoom=8 8E boom .m.w-¢.~ n MO< .383 358 “32:50 .38 dw< .233 8.003 nmv 3:88 in 58$: $805202 , , 1055230 35.5w ~85 .mfi.’ =33:qu an: 35:32 33:0: 8 Baa—9. nus—=5— nSatgoo 35— 8.335 14 PART III THE POUCH STUDY The POUCH Study was a prospective cohort of women enrolled mid-pregnancy (15- 27 weeks gestation) from 52 participating clinics in five Michigan communities (Lansing, Flint, Saginaw, Grand Rapids, Kalamazoo) from 1998 through 2004. The study was designed to assess biologic and psychosocial risk factors for preterm birth. The study protocol was approved by the Michigan State University Committee for Research on Human Subjects, institutional review boards at nine participating hospitals, and the Michigan Department of Community Health human subjects committee. Data Collection and Variable Definition Enrollment in the POUCH cohort Women were sampled at the time of prenatal maternal serum alpha-fetoprotein (MSAF P) screening (1 5-22 weeks) from the Michigan State University Prenatal Screening Laboratory records and invited to participate in research study on ‘women’s health in pregnancy’. All women with unexplained high (two or more multiples of the median) MSAF P, and a stratified random sample of normal MSAFP women (stratified by race) were sent letters describing the study protocol. Participants were recruited through the use of study invitation letters, scripted recruitment phone calls, and post card reminders. The eligibility criteria included maternal serum alpha-fetoprotein (MSAF P) screening at 15-22 weeks of pregnancy, maternal age of at least .1 5, English-speaking, no pre-existing diabetes, no known fetal anomalies or chromosomal defects, and singleton pregnancy. At the time of enrollment, women met with a nurse researcher at a clinic in their community. The nurse researcher obtained consent, administered an in-person structured interview and collected samples of blood, urine, vaginal fluid, and hair. In 15 addition, women completed a self-recorded questionnaire. Participants were paid $50 in appreciation for their time. Interview Self-Reported Data/Assessment of Maternal Race and Ethnicity During the interview, participants were first allowed to choose multiple racial/ethnic groups. In a separate question, women selected a single racial/ethnic group. The interview and questionnaire recorded other demographic information, medical history, obstetric history, maternal behaviors and contained several validated psychosocial instruments. [43] At—home data collection at enrollment (mid-pregnancy) After the initiation of the POUCH Study, additional funding was awarded which allowed the study to include an at-home collection portion. This add-on to the original protocol was designed to assess maternal stress and included daily collection of biologic samples (urine and saliva) and completion of a daily diary over the period of three days.[44] Optional participation in the at-home collection was offered to women at the time women enrolled in the POUCH cohort. Women received up to $40 for participating in the at-home collection portion. Cohort follow-up prior to delivery Afier enrollment women were contacted by telephone at 30 and 37 weeks gestation. The purpose of the phone calls were to remind women to identify themselves as a study participant to labor and delivery nurses upon presentation to the hospital. In addition, follow up phone calls presented study personnel an opportunity to identify women who moved after enrollment and prior to delivery. 16 Subcohort sampling To maximize statistical power while conserving financial resources a case-cohort sampling scheme was used. The subcohort included all women with preterm births, all women with unexplained high MSAF P, and a random sample of women with normal MSAF P and a term delivery with oversampling of Afiican-American women. Collection of data in the subcohort after delivery Placental tissues along with the umbilical cord were saved by labor and delivery staff. For subcohort women these were later subjected to gross and microscopic examination by a placental pathologist. Prenatal care records and hospital labor & delivery records were collected and abstracted for women in the subcohort. Definition of preterm birth Gestational age at delivery was calculated fiom the date of the last menstrual period (LMP). Gestational age estimates from early ultrasound (before 25 weeks) were substituted if the ultrasound gestational age differed from the LMP gestational age by at least two weeks or if the LMP was missing (18% of the cohort). Information about timing of the onset of labor and the timing of rupture of membranes was abstracted (independently by two study members) from post-partum medical records. The onset of labor was defined as the point when regular contractions occurred along with cervical dilatation of at least two centimeters. Preterm births (delivery before the 37‘h week of gestation) were grouped into two subtypes based on clinical circumstances at delivery. Births that occurred by Cesarean section or induction for the health of the mother or neonate without the onset of spontaneous labor or rupture of membranes were identified 17 as medically indicated preterm births. Preterm births in which the onset of labor or rupture of membranes occurred as the initiating event were considered spontaneous. Assessment of Bacterial Vaginosis (BV) Dacron swabs were used to obtain vaginal fluid samples at enrollment in POUCH study women. Heat-fixed gram-stained vaginal smears were prepared and bacterial morphotypes were scored according to Nugent’s criteria.[45] A score of 0-3 is considered normal vaginal flora, 4-6 is an intermediate score, and a score of 7 or greater is defined as BV positive. DNA extraction and genotyping (IL- 1,6 +3954, TNF-a -238, INF-a ~308) Maternal blood was obtained by venipuncture at enrollment. Genomic DNA was prepared from venous samples using the Gentra Systems (Minneapolis, MN) PuregeneTM kit. SNP genotyping was performed using TaqMan® Assays-on-Demandm, which consists of PCR primers and a fluorescently-labeled probe (Applied Biosystems, Foster City, CA). Three primer/probe sets were used that were specific to SNPs in IL- 10 (rsl 143634) or TNF-or (13361525 and r81800629). Assays were performed in 384-well plates in 5 uL reactions containing 25 ng of template DNA or no-template control (NTC) and 0.25 “L of primer/probe in 1X TaqMan® PCR Universal Master Mix. Thermal cycling conditions were according to manufacturer guidelines, as follows: 2 min at 50°C, 10 min at 95°C, and 40 cycles of 95°C for 15 sec and 60°C for 1 min. End-point fluorescent detection was performed on an ABI Prism® 7900HT Sequence Detection System. Genotypes were evaluated using SDS software (v.2.1). Duplicate genotypes were generated on a random sampling representing 2 15% of the individuals with a concordance rate of 99.9%. 18 Final Study Sample POUCH Cohort (Table Two) The POUCH cohort enrolled 3,038 women with 19 women lost to follow up. The preterm birth rate was 10% in non-Hispanic white women and 15% in African-American women. The spontaneous preterm birth rate was 7% in non-Hispanic white women and 10% in Afiican-American women. The rate of medically indicated preterm births was 3% in non-Hispanic white women and 5% in African-American women. In non-Hispanic white women, 89% had at least a high school education, 35% had Medicaid Insurance and 44% were primiparous. In African-American women, 64% had at least a high school education, 80% had Medicaid Insurance, and 40 % were primiparous. Characteristics of study participants were grouped by race and compared to birth certificate data fi'om the five Michigan communities for the year 2000. There were no significant differences between POUCH participants and birth certificate data in parity, education level, proportion with Medicaid insurance, preterm birth rate, rate of previous stillbirth, rate of previous preterm infant, and rate previous low birthweight infant. The percentage of African Americans over 30 years of age was lower in the POUCH Study than in the birth certificate data (14% versus 21%). POUCH Subcohort The final subcohort was comprised of 1,371 women. Funding for a follow-up study to collect child DNA within the subcohort was received in 2005. Part IV describes the methods used to collect the children’s DNA. 19 Table 2. Maternal characteristics of POUCH Study participants by race/ethnicity, Michigan, 1998-2001 Maternal characteristics Non-Hispanic African Other White American Ethnicity N (%) N (%) N (%) Maternal Age <20 209 (10) 206 (28) 45 (17) 20-30 1125 (56) 432 (58) 154 (57) >30 684 (34) 105 (14) 59 (23) Education Level and Age (years) < 12 and <20 years old 119 (6) 140 (19) 32 (12) < 12 and >20 years old 102 (5) 126 (17) 36 (14) 212 1797 (89) 477 (64) 190 (74) Marital Status Married or living with husband/partner 1695 (84) 298 (40) 191 (74) Not married or living with husband/partner 319 (16) 442 (60) 67 (26) Medicaid Insured“ No 1307 (65) 145 (20) 1 10 (42) Yes 710 (35) 596 (80) 148 (57) Primiparous’” No 1132 (60) 443 (60) 148 (58) Yes 885 (44) 300 (40) 108 (42) Pregnancy Outcome Term delivery 1821 (90) 633 (85) 230 (89) Spontaneous Preterm Birth 134 (7) 76 (10) 18 (7) Medically Indicated Preterm Birth 63 (3) 34 (5) 10 (4) *Missing data on 3 women ** Missing data on 1 woman 20 PART IV THE FOLLow UP CHILDREN’S DNA COLLECTION Recruitment Subcohort women who had not declined future contact for follow-up studies were mailed a newsletter describing the children’s DNA collection study. Women who had moved since their last communication with the POUCH Study were located with the assistance of participant identified personal contacts or through the use of intemet searches of public data bases. Following the newsletter, telephone recruiters contacted women to describe the study protocol. Women were asked to collect DNA from the child born at the end of the POUCH Study. Telephone recruiters confirmed the child’s date of birth to prevent collections from the wrong child. Interested women were mailed a DNA collection kit and consent form to return through the mail. DNA collection, extraction, and genotyping The collection kit contained five cotton tipped swabs and solution. The kit instructed women to rub five swabs on the inside of each cheek. Women who failed to return the collection kit afier two weeks were followed up with reminder phone calls and post cards. Upon completion of the children’s DNA collection study, women received a check for $40. Genomic DNA was extracted from the childen’s buccal swabs (Puritan Cotton Tipped Applicators REF 803-PC) and was prepared using a phenol/chloroform extraction method.[46] Eight maternal and all child polymorphisms were genotyped using PCR and primer pairs that spanned the polymorphic regions followed by restriction enzyme digestion and agarose gel electrophoresis. PCR conditions and primers sequences have been described in previously published protocols: CD14, [47] TLR4, [48] Il-lRN (IL- lra), [49] Il-lB, [49] TNF-a, [50] TNFRSFIB (TNFRZ), [51] TNF SR6 (Fas), [52] MMP- 9, [53] MBL54, [54] MBL57 [55]. The remaining three maternal polymorphisms (IL- 10 21 +3954, TNF-or -238, and TNF-a -308) were genotyped as part of the original POUCH Study protocol using previously described methods (p.20 of dissertation). Participation (Figurel) Funding for a follow-up collection of children’s DNA within the subcohort was received in 2005. Women whose fetus/child had died (n= 19), who had declined firture contact (n=69), or who were themselves deceased (n=3) were not included in this follow up study. Attempts to contact the remaining subcohort women (n=1280) were made via mail and telephone; 1024 subcohort women were located and 1006 were successfully contacted during the study follow up period. Ten women were no longer living with their child and therefore were not able to give consent for their child to participate in the study. Forty-one women declined to participate. Children’s DNA collection kits were mailed to 955 women and 865 kits were returned. 22 POUCH Subcohort N=1 371 -19 Fetal/Child Death -3 Maternal Deaths -69 Declined future contact, Eligible Subjects N=1,280 -256 Lost to follow up Located Subjects N=1,024 -18 Not contacted -10 Without child custody -131 Declined (90 did not return kit) Child DNA Collected N=865 Figure 1. Enrollment in the children’s DNA follow-up collection 23 PART V POUCH STUDY INFLAMMATION RELATED GENE POLYMORPHISMS Maternal DNA collected during pregnancy and children’s DNA collected during the follow-up study was assayed for ten innate immune system gene polymorphisms in nine genes (CD14, toll-like receptor-4, tumor necrosis factor or, interleukin-1 B, interleukin—1 receptor antagonist, matrix metalloproteinase 9, Pas, mannose binding Iectin, and tumor necrosis factor receptor 2). Maternal DNA was assessed for an additional eleventh polymorphism in the tumor necrosis factor or gene (-23 8). The polymorphisms were selected based on evidence of functionality. We hypothesized that the polymorphisms impacted the risk of PTD by causing variation in the bacterial-mediated inflammation processes (see Figure 2 adapted from Campos[56]). Previous publications that used candidate gene approaches with the selected polymorphisms and preterm birth are summarized in Table 3 while three studies that used a gene-wide search are discussed later in part VIII. CD14, an innate immune system component, plays at least two roles that are counterbalancing. As a cell surface receptor it forms a complex with lipopolysaccharide (LPS- bacterial antigen) and LPS binding protein to activate pro-inflammatory immunity and TNF-a release from monocytes/macrophages and neutrophils. [57] In its soluble form, CD14 binds to the LPS-LPS binding protein complex and inhibits LP S-induced pro-inflammatory cytokine responses.[58] A cytosine (C) to thymine (T) polymorphism at position -159 in the gene promoter was found in some studies to be associated with elevated serum CD14 levels and pro-inflammatory immune system activation in response to bacteria. [59, 60] This polymorphism has been investigated previously in two studies of preterm birth (Table 3). A German study of preterm very low birthweight infants 24 found no-significant associations with preterm birth in either maternal or infant genotype. [61] A small study (n=28 PROM births and n=72 term births) of multifetal PROM births found no significant main effect for child genotype but did find that maternal genotype was more likely to be TT among preterm (39%) than term (18%) births.[47] Although this study had limited power, it found a significant gene-gene interaction for CD14, IL- lra, and heat shock protein-70. Toll-like receptor-4 (TLR4) is a membrane-bound pattern recognition receptor which recognizes the LPS binding protein-LPS-CD14 complex.[57] Binding triggers a cascade of intracellular events leading to the transcription of genes coding for pro-inflammatory cytokines.[62] An adenosine (A) to guanine (G) polymorphism at position 896 in the TLR4 gene is associated with a markedly reduced responsiveness to LPS in vivo and in vitro, although this is not a consistent finding across populations.[63 ~66] Five previous studies have published data on the TLR4 A8966 polymorphism and preterm birth (Table 3). Three reported no significant associations with spontaneous or PROM preterm births in white or African-American maternal or fetal genotypes. [61, 66, 67] A Finnish study of all preterm birth in both maternal and fetal DNA found a higher allele frequency in fetal DNA from preterm births as compared to term births.[48] A Uruguayan study found a higher allele frequency in neonates who were born before 33 weeks with PROM as compared to neonates who were born before 33 weeks without PROM. [68] Tumor necrosis factor or (TNF-a) is a pleuripotent cytokine active in the induction of both pro-inflammatory immunity and programmed cell death (apoptosis). A single G to A transition at position -308 has been associated with an increased transcription rate and higher TNF-or secretion. [69-71] The association between TNF-a -308A and 25 spontaneous preterm birth has investigated in thirteen studies. (Table 3) Eleven reported on the association in maternal DNA. Three of four studies in white women found no significant association.[72-74] The single study with a positive finding in white women was an Australian study with a haplotype analysis.[75] Four studies included data for both African- American and white women.[76-79] Moore found a stronger association for smokers but not for women with periodontal disease while Engel found a stronger association for white women and women with BV.[78, 79] A study in Mexican women found a protective effect of the A allele. [80] Two studies reported on maternal DNA in just Afi'ican-American women.[81, 82] Both had positive aSsociations with one study reporting a stronger effect among Afiican-Arnerican women with BV. A meta-analysis in 2006 reported a non-significant overall summary odds ratio for maternal genotype (odds ratio: 1.41; 95% confidence interval 0.90-2.19) although this single summary included studies with different racial/ethnic compositions. Four studies provided data on child genotype. One study in white women found no significant association. [83] Two studies included multiple races and found no association for fetal genotype within racial group. [76, 77] A single study in just Afi'ican Americans found a significant association.[84] A second polymorphism at position -238 results in a G to an A substitution. While no significant associations have been reported for this polymorphism and preterm birth, in a sample of preterm births, the heterozygote genotype was more common in fetuses with placental evidence of severe chorioamnionitis.[85] The biologic functions of TNF-or are regulated by its binding to two receptors, TNFRl and TNFR2. TNFRl is present on most cell types and is involved in the apoptosis pathway. TNFR2 is found predominantly on immune system and endothelial cells and 26 TNF-a binding is associated with activation of pro-inflammatory immunity.[86] There is a T to G polymorphism at position 422 in exon 6 of the TNFR2 gene (gene symbol TNFRSFl B). [87] Possession of the G allele is associated with increased susceptibility to chronic inflammatory disorders.[87, 88] A single study in white women found no significant association with spontaneous preterm and maternal genotype. (Table 3) [72] Interleukin-113 (IL-10) is a multi-functional pro-inflammatory cytokine that can stimulate the production of prostaglandins [89, 90]. The C to T polymorphism in the IL- 113 gene at +3953 results in increased IL-lB secretion. [71] Four previous studies have reported on a possible association between preterm birth and IL—lB +3953 (Table 3). Two found no significant associations in white or Afiican-American mothers or fetuses. [78, 91] One study found an increased risk of spontaneous preterm birth (less than 34 weeks) among Afiican-American homozygous fetuses.[49] Another study reported a significant interaction with maternal genotype and BV for risk of spontaneous preterm birth. [49, 79] Interleukin—1 receptor antagonist (IL-Ira) is a competitive inhibitor of IL-1. IL-lra binds to IL-1 receptors without signal transduction. A tandem repeat in the second intron of the gene (IL-lRN) is present in five allelic versions. Allele 2 is associated with an increased risk of chronic inflammatory disorders. [92] The genes coding for IL-lra and IL-1[3 are linked and possession of IL-lra allele 2 results in a net increase in IL-lB-related pro- inflammatory activity.[93, 94] The possible association between and IL-1 ra and preterm birth has been reported in seven previous studies (Table 3). The two largest studies found no significant association.[75] [95] Other studies found significant associations in maternal or fetal genotype in Afi'ican-American, white, and Hispanic women. 27 Matrix metalloproteinases are a family of zinc dependent enzymes capable of degrading the extracellular matrix. Matrix metalloproteinase 9 (MMP-9) is an enzyme that degrades collagen type IV, elastin, and fibronectin. High amniotic fluid levels of MMP-9 are associated with fetal membrane rupture, parturition, and placental detachment.[96, 97] In vitro studies find an increased production of MMP-9 in the presence of other pro-inflammatory cytokines or LPS. [98, 99] The C to T transition at position -1562 results in an up-regulation of promoter activity. The T allele is associated with increase in gene expression and serum MMP-9 levels.[100] A single study found no significant association between PROM births and fetal genotype in an Afiican—American population (Table 3). [101] Fas (gene symbol TNFRSF6) is a transmembrane cell surface receptor for F as ligand. Fas-Fas ligand binding results in activation of apoptotic signal transmission. Pro- inflammatory cytokines can induce Fas expression in trophoblasts and increase likelihood that these cells will undergo apoptosis.[102] LPS can also induce expression of F as in arnniochorionic membranes.[103] A G to A substitution at position -670 is associated with reduced promoter activity and gene expression.[104] Three studies have examined the association between the Fas -670 polymorphism and preterm birth (Table 3). In a small case control study (n=18 PROM preterm and n=18 term) with fetal DNA, the AA genotype was not present in the PROM births.[105] In a study of multifetal preterm births, the GG genotype was more common in both maternal and child DNA.[52] Mannose binding lectin (MBL) is an acute phase serum protein involved in antimicrobial innate immunity and activation of the lectin complement pathway. Polymorphisms in codons 54 (MBL 54) and 57 (MBL 57) are associated with reduced 28 serum MBL levels. [106] Maternal but not child genotype was associated with preterm birth in Australian studies of MBL 54. [75, 107] A separate study found a an increased risk for preterm birth among maternal genotype in white women using a haplotype analysis. [108] Parts VI, VII, & VIII describe three projects that examine the relationship between preterm birth and the polymorphisms measured in the POUCH Study. 29 Macrophage Regulation Immune Cell e, ,3, Uterine ‘ :3 /7Contractions in lL-1ra Prostaglandins . lL-1B __ LPS Preterm Birth MMP-9 (-—"| TNF-a \ Membrane / o Fas .————9 Degradation TNF-R2 Immune Cell Figure 2. Proposed bacterial mediated inflammation pathway to preterm delivery with selected candidate genes. Lipopolysaccharide or LPS is a component of gram negative bacterial cell wall that binds to LPS binding protein (LBP). LBP transfers bacteria to membrane bound CD14. CD14 is associated with Toll like receptor four (TLR4). This complex induces a cascade of pro-inflammatory cytokines including increased production of tumor necrosis factor alpha (TNF-a) and interleukin one beta (IL-1B).[57, 62] The action of IL-lB is competitively inhibited by interleukin one receptor antagonist (IL-1ra). Binding of IL-lra to an IL-1 receptor results in a dampening of the inflammatory response, while binding of Il-l B firrther activates the inflammatory response.[92] Binding of TNF-o. to tumor necrosis factor receptor type two (TNF-R2) also activates the pro-inflammatory response.[86] Both LPS and TNF-or may stimulate production of matrix metalloproteinases (including MMP-9) and the Fas protein which participate in membrane degradation and potential rupture of membranes leading to preterm birth.[98, 99, 102] TNF-alpha, IL-l B, and MMP-9 are capable of stimulating production of prostaglandins which can cause uterine contractions leading to preterm birth.[90] In addition to the eight genes in this figure, we also selected two polymorphisms in the mannose binding lectin gene. Mannose binding lectin stimulates the complement pathway in response to carbohydrates on the surface of bacteria, fimgi, and viruses. 30 RED ~88 mom 8.333 8.3% 5 8....8053 38 28:53 88 womanbm-3§ mwfiwam 389.36 02 20am :35 253280 .8353 SSE< canton 3:882 480m 3.08% 3:5 :5 an reggae .PE <29 HE E a: e E33 SN 88 330 .8260 kiwi homo—50¢ 22?. mega—am =< mod 5:85 .58qu 68%? 82 6:5 Each $3 .E. Each Q. E $2” PH 20am mm ”3:532 SOME 230 ES ~§~2 woom SOME Q 38:88am iHmEHHDE dam—«M SWEEP wooinwmokfifin 32 E3 @305 ”EEO Swans—8 gauge... 5va :83 ea vows—8m— mecvufiwwoB 5.5 32 392, €8on .Beoveoomoa Emmoagtfi 32 8.353 ”35032 388va $595 ugomiwmm oz bo> 9:8on 2 .uEBuoE 58.80 230 .8 13.—8:2 .82.:qu 33:230.:— Jfiw 2.5—am Sauna—um nouamugeu $525— 53: 5.33.5 53.9.3 .9550 acts—sack 33w «Ema—32:22— SaEE—uo be 8:25». 2.35...— .n 93:. 31 SOME Signage 223 Eve use 250 293 8 GE ram 29E .238 ommueae m H .n zoaozwoa 20:“ 6:5on ommnfiuououm $3.538 02 .29: ea em 3am 38:38am 3o heme: 88 sex fiasco—a .9595 £83 2V 3% 358m: £33m 8:333 own 35:88» 30.3933 8ng 56:55 5:3 so: 3 remotes“... 550 Gun 3:532 3596 «5095a 02 maoouSuomm §o£< Emotes anew—.3: 65.35 .833 voom .050 Svnfiuoh oooHEwEB 53 32 b3 8.82m NEEO OE Sta £3828 SNHHEQH .525» :53 .05 vowzdoxm covnfiwmuk, :39 32 bo> E205 .ooqouuoomov E30353 32 _ “Wanna: imam: museum “some? 02 9> E35 grease: flog—Eon mono—.3...“— uouatagu 3:35pm 552 E3395 3E3 3.5.50 32 .52 mm: mfioam tmnmo< £83 8565 8m 3.: 3:835 "£83 omv mmv 20mm 38032 woo” 3:82 099293 E a. msoogaonm o~§-§=abm=< £353 an. .mouofiufi 485:8:0 an own 38383» .>m .92 8% magenta an 38:58» E v 4E a 230 a. secs 2 F2 38 5E9 .008 dw< “58¢?me oz 582 2.0 850 ageing 505 “233 505 BED E8 Sm fionmo wouowhneoaom Shana VM SON $3338 02 ”finnmo moaowxuofio: 38.38% S .o 505 .082 E2 o: a: 6.: 3 ”mo 29: E205 2 DE SE 5.5a $.53 :8 E5 .29: been 82 563an .ow< E358 $3 29: + 38:88am omd 585:2 50E .mtonom 30 E8 3 8365 SN 32 :«EBaZ damnaoh 83530 oz 3826 ugomawfi oz waooqfiaoaw 26 525qu 5:25on «93:52.0 3.5.5,.— EaE .5895 not?» 35.80 33 .>m 5:5 5:03:25 02 8:332: .0885 8:2 Hmm £23033 5:03:28 855:: 2: ooom .omn 3:532 232.2:me 02 385.55% 2. 3:532 22>? ”8:52 3:532 23950 05 .33 5:2 3% 52,5 05 .5803 SE? S. n 2.33 555:: on“ .>m E .>m :35 885:2. mm. H :358< ”305:3 §o£< 55, 85825 :85 E 5320 38:8:onm 5.6% .22? noon ._ow:m HE2 N2 God 2555 855:: on 2.8 302:8 :_ 5:33 3:532 30m .8525 35:32 223 552 59025 =< S. 53:52 .Now3> 2333 332.25 :23 3:532 .3225 .5803 E 223583 8.2 mm 55 830825 o: 283:; :8 5w:obm E555 3 38 .mioam Sumo 83:3 :< 3.. 333 was: "282 .>m 3E2~2 65 85225 odumo< w _ .5233 8:2 0mm .>m 2530: >m :2 5w:obm 20M: 2 _.n:uo58< 522: mi 38 £382 .023 QTMO< a. 38:88am :85 223 dautofi< :85 £2582 , flfiri. ,V ,, : g , , ,Ezthzou Vmcmmxg. E. EEO .8 53qu 5:33: 352:2 .35 2:55 .5255— 533950 “ME—95m :32 855.5 53.5... 3.550 34 088de .0885 E8 Sm £23260 E23a 8. 88 Gun 35032 .mwfivna “50¢?me 02 302553 3:332 833 dog , .. . . ,, , . . , . . , . 08.: QE . , . E2 Rm ~58on S 33538 02 .mwfiuww agowawmm 02 89085 =< Z. Mom—m gamma 833 Sam 45:0 ,..,.., .....,,. ,. ,;...,.,,, ,. ,,,,,, L3,... ._ .5.,.: . ...,,. j .. . . _. ,H.§%§U,§ab-ueserg 230 .8 5.25%: ago—.3; 13.532 .35 29:5 Saginaw $31560 3565."— Eaz 5.3.9.5 .3226 15.30 £53353.— 35 1:538 .230 EB 52 E803 mm SOME 5? 3368mm 585 33 p.833 gamma 383 va m a .HEEO gmoSm 36:35 macaw uwmwnao m 20: 303.38% SWEESZ 259mg «com .080 .L... . ..:.,... , . .. . , . :..,..,, ,.,.,...§$ fiwdohEuNNkfi MNWBEB Show Em .>m >m 53» m _ .nanotona 8.5on 02 5:5 8:08.35 530885 23336 Ecogaomm 50% going 535 .833 noom 4095 .3656 203 Bowmbnoouoo . End 30.5.60 EEO 48532 £9.85? 0930qu EB mu £5 553083 SOME 5055 85.05 mm ooom “Sommqwfi 02 38:88am 3.0 6&0 anew—08¢. §o£< .833 .3633 388.22 5.8 mm E895 xv ”305 $56 Hausa... oz 8.5%; =< w _ .o 833 £85 38 .282 €532 .230 EB 53 P: .«o ME 33803 5.805 mm 5 was 8893080: :85 .mowaom 505:2 8603 3” v 34220 goqoBm 56:55 565 unofi< maoogdoaw 3.;ME032 unsung meow .050 320 .3 .85:qu zones—5.....— .aESa—z £wa 2955 .bfiififl 32.1560 nus—vim 5a: 5.53.:— Lotuau .3300 35am 36 3532 big ESP m2 .333 3ng .x. 2 n £8300 2 .HBEB 833$ mm 5.25% :8 u 8.30 £83 2v 2 .ufiocoa... 25%.: 88 Jun 3:332 ”36532.2 22?. 38:85am 30% .833 iguana .305 £5.52 3:332 Ehx wmv “EBB.“ oom 83>» .mwaué 68%»... oz 8.533 5. .25.ng .5055. 525. 38 .8m .53 833E 385323 38032 mo 8S 3030th vaBEB Ebb NM: 93 6983.8 Son: :.u§otofi< ~52on on B 30:32 in 333$ 38:83am 30% 30.5% 505 93 8.33 33 .050 4..ng :35 E2 2 age... a: a 230 E303 3 38 bUEfim MERE zonoscob 8.530 39393 =< VN. =25; uofimmom _Scu32 MSU—08m £83 8.5“ mm: um: 08835 wmv 20~Eum 833E mom 38 3:82 Enema—ME “oz geognoam 833-32gm:< “£052 33m .223 8.8% new 8533 .«o xmt 39835 583$ E «3.75:1: 33m 383303 on. $me meow 4523 . . ,u§§§ur..wu uxmfigémfi, 330 S .52.:th mung—.2...“ 3:332 .35 can—am 53:5”.— uo3_.3>90 355:..— 532 5.53.5 .5239 3.5.50 iota—sack 33m 32:56 m 2...; 37 EEO 20~E .x. o << Ebb M: 3.53 awn E2 $2 3 583 M: 35038 .ooqm 2OMm gmw O< 20~E 36 233-82 .833 moon .815 .38 2833 was unboaow ESP 2. £5 momaom 50m 2OMm mm 45895 mo 88 Gamma 2OyE .Som Ea 35062 220 was 3:832 a. Soognoam A EEO ES NE E 4m: 5 8.205 NS 88 oouohotmc Enemawa 02 8.83 cmv 5595. .bgom «Eu... mm“ wfioam 383 on v 383 QE SN 8: 8:835 Wm H MO< wmv Sanka. 38032 voom .2802 E .52 «noon—35mm mud 233-5%33‘. 5:053 H 2 a3 .2 an: 230 .8 52.509 mung—:5...— EESa—z Jim 9355 .bEE-fi 8.2.3.30 magi—E 53% 5.53.5 .523 .3250 39 PART VI PROJECT 1- MATERNAL AND CHILD GENOTYPE COMBINATION AND PRETERM BIRTH RISK Purpose A primary goal of this study was to evaluate maternal and child genotype combinations of ten gene polymorphisms just described as contributors to PTD risk. Of particular interest was the possibility of interactions between maternal and child genotypes. Rationale Genetic variability in pregnant women and/or their fetuses likely influence the degree of susceptibility to exogenous and endogenous factors that contribute to preterm birth. Women who deliver preterm are at increased risk of delivering preterm in subsequent pregnancies and findings from twins studies estimate the heritability of PTD to be about one-third. [22, 37] Over 70 studies have investigated the association between preterm birth and genetic variation in inflammatory, metabolic, and vascular related genes but few have considered the impact of combinations of maternal and fetal genotype. Specific Aims The aims of this study were to examine: 1) relations between the ten candidate polymorphisms and risk of PTD within the maternal genotypes; 2) relations between the ten candidate polymorphisms and risk of PTD within the child genotypes; and 3) the impact of maternal and child genotypes together on the risk of PTD. Study Sample Afier excluding other races, 489 non-Hispanic white (n=132 PTD) and 290 African- American (n=50 PTD) maternal child pairs were included (Table 4). 4O Analytic Approach Descriptive Statistics Chi-square statistics were used to compare the distribution of maternal characteristics in the POUCH study sub-cohort to the distribution of maternal characteristics in the sample that participated in the child DNA collection. Hardy Weinberg equilibrium for each polymorphism was tested in the group of women who delivered at term with normal MSAF P. Chi-square was used to compare the race-specific allele frequencies for women who participated in the children’s DNA collection and women who did not provide a DNA sample from the child. Modeling Strategy All logistic regression analyses incorporated weights that were based on the case- cohort sampling scheme (oversampling of high MSAFP, Afiican-American women, and PTD into the subcohort). In weighted analyses, model estimates are based on prevalence in the original cohort; therefore the effect of oversampling high-risk subgroups into the subcohort is statistically accounted for. All models were stratified by race (defined for both mother and child by maternal self-report) to consider potential effect modification. Genotype Classification For logistic regression models, genotypes were divided into two classes with the homozygous common allele genotype as the reference group and the heterozygous and the homozygous rare-allele genotypes grouped as the ‘risk’ group. For IL-lra, genotypes containing alleles three, four, or five were very rare and were excluded (n=18). Allele two was treated as the ‘risk’ allele and the allele one homozygotes were used as the reference group. 41 Assessment of the Main Effect of Maternal or Child Genotype on PT D Two sets of models were used to test for significant main effects of each polymorphism on PTD risk. The first set of models used weighted logistic regression (SAS Version 9.1 .3- Proc SurveyLogistic) to calculate the odds ratio for PTD and the rare allele for each maternal and child polymorphism. The second set of models used weighted polytomous logistic regression to calculate the odds ratio for PTD subtype and the rare allele for each maternal and child polymorphism. The outcome in the polytomous model is a three-level variable: term (referent), and two PTD subtypes, i.e. M1, or SPT. The main effects of each maternal polymorphism were tested in the sample of women with both child and maternal DNA and compared to those that included all African-American and non-Hispanic white women with maternal DNA (n=] ,327 data not shown) to see if loss to follow-up and child deaths could affect results. Assessment of Maternal and Child Genotype Interactions The effect of the combination of maternal and child genotype on PTD risk was tested for three polymorphisms with significant main effects on preterm birth: IL-lra, MMP-9 and TNFR2. Two sets of models were used to test for significant maternal and child genotype interactions on PTD risk. The first set of models used weighted logistic regression to calculate the odds ratio for PTD and the rare allele for maternal and child genotype combinations. The second set of models used weighted polytomous logistic regression to calculate the odds ratio for PTD subtype (M1 or SPT) and the rare allele for maternal and child genotype combinations. Analyses were also repeated with the removal of sampling weights to make sure there were no biases introduced using the 42 original weights. The findings did not differ and the results of the weighted analyses are presented below. Results Descriptive Statistics Race stratified maternal characteristics (Medicaid Insurance use, education level, maternal age, parity, and preterm birth rates) in women who participated in the child DNA follow up collection did not differ significantly fiom those in the entire POUCH Study sub-cohort. The MMP-9 gene in the non-Hispanic white women was the only genotype to deviate from Hardy-Weinberg equilibrium. Race—specific gene frequencies among women with normal MSAFP who delivered at term were similar to previously published findings (Table 5). The maternal allele fi'equencies were not different from the children’s allele frequencies (participants in the follow up collection) or the frequencies in women for whom we had no children’s DNA (non-participants in the follow-up collection). Main Eflects of Maternal and Child Genotype on risk of PT D The significant findings for the main effects of maternal and child genotypes are presented in Table 6. For maternal genotypes, the IL-lra allele 2 was significantly related to an increased risk of PTD in non-Hispanic white women but not in African- American women (p-value for race interaction in combined model =.O3) (Table 6). The effect remained significant in relation to SPT. The MMP-9 T allele and the TNF R2 G allele were associated with an increased risk of ASPT in the maternal genotype for non- Hispanic white women and Afiican-American women respectively. 43 For child genotypes, the IL-lra was significantly related to PTD overall and SPT in Non-Hispanic white women. No other polymorphisms were significantly (p>.05) related to PTD overall. No alleles were significantly related to medically indicated preterm birth. Maternal and Child Genotype Interactions Among non-Hispanic white mother-child pairs, the odds of SPT were increased when both mother and child had the IL-lra 2 allele (odds ratio = 2.4; 95% confidence intervals 1.3, 4.2) as compared to when only mother or child had allele 2 (p-value for multiplicative interaction < 0.05) (Table 7). Among African-American mother-child pairs, the odds of SPT were increased when both mother and child had the TNFR2 G allele (odds ratio 2.3; 95% confidence interval 1.0, 5.3) as compared to when only mother or child had the G allele (p-value for additive interaction<.01). Discussion The POUCH Study considered 10 polymorphisms in 9 genes associated with innate immunity in relation to risk of PTD. Eight of these polymorphisms had been associated with PTD in some, but not all, previous studies. We replicated findings related to IL-lra and report new associations for MMP-9 1562 and TNFR2 198. For IL-lra and TNF R2, the presence of the variant allele in both the maternal and child genotype significantly increased the risk of spontaneous PTD in whites and African Americans, respectively. Since IL-lra allele 2 is associated with the highest unopposed pro-inflammatory activity as compared to the other IL-lra alleles, our findings suggest that intense and/or prolonged inflammation in both the maternal and fetal compartments contribute to the induction of SPT. A similar pro-inflammatory mechanism might be responsible for the observed interaction between the maternal and child TNFR2 gene 44 polymorphism and risk of PTD. The race differences in our findings may be explained in several ways. The IL-lra allele 2 was rare in African-American study participants, and this limited our ability to accurately measure the association in this sub-group. The TNFR2 genotype distributions were similar in Whites and African Americans, but differences between these two racial groups in environmental exposures and/or genotype fiequencies in other relevant genes not evaluated here likely influence the net effect of TNFR2 on pregnancy outcome. For MMP-9, there was a significant main effect for maternal genotype but no significant interaction for combinations of maternal and child genotypes. In the child DNA follow-up study, the final sample was comprised of maternal and child pairs who were willing to be re-contacted (a question posed as part of the original pregnancy cohort study) and could be located several years after delivery. Participants in the children’s DNA follow-up collection included 865 (63%) of the 1371 women in the POUCH Study sub-cohort. We compared women in the entire sub-cohort and those who participated in the children’s DNA collection study and found no significant differences in maternal demographic characteristics or allelic frequencies. We also compared the main maternal genotype effects on risk of PTD. The only clear difference was that MBL 57 was significantly related to PTD among non-Hispanic white women in the entire sub- cohort but not in the subset of women with child DNA. This dyad subset was missing preterm births that did not survive to the time of follow-up, and this might explain the differences in the MBL 57 result. It is also possible that some of the polymorphisms in children are related to both survivorship and pregnancy outcome and were missed or underestimated in the analyses of the follow-up mother-child sample. Our comparisons of 45 maternal variables suggest that overall the loss to follow-up resulted in minimal bias but did affect our statistical power and precision. We can not rule out the possibility of chance findings given the number of comparisons that were necessary to estimate the main effects of each polymorphism. However, we focused on functional polymorphisms with plausible evidence that gene products are involved in inflammation-related PTD. The findings related to MMP-9 1562 and TNFR2 198 are new and will need to be replicated in future studies. We can not rule out the possibility that the negative polymorphisms are important only relevant in combination with other unmeasured genes or environmental factors. Potential genetic contributions along pathways to PTD most likely involve complex gene-gene and gene-environment interactions; both will be the subjects of future analyses. Our findings here suggest that the development of genetic models should carefully consider the contribution of combinations of maternal and fetal genotype on PTD risk. 46 I Table 4. Maternal characteristics in Non—Hispanic white and African-American women with both maternal and child DNA collected. (Weighted percentages reflect sub-cohort sampling scheme) Non-Hispanic White African-American (N=489) (N=290) thg th thg th Maternal Characteristics N % % N % % Maternal Age (years) <20 45 9 9 68 23 24 20-29 259 53 53 179 62 62 230 185 38 38 43 15 14 Maternal Education (years) <12 49 10 9 94 32 33 12 118 24 23 86 3O 30 >12 322 66 68 110 38 37 Medicaid Insured* No 335 69 7O 61 21 20 Yes 153 31 30 229 79 8O Parity* None 210 43 4O 1 16 4O 41 At least one 278 57 60 174 60 59 Pregnancy Outcome Term 357 73 91 240 83 86 Preterm Medically Indicated 38 8 3 11 4 3 Preterm Spontaneous 94 19 6 39 13 11 Percentages are presented both unweighted (unwgt) and weighted (wgt) for subcohort sampling scheme. "' Data on Medicaid status and parity is missing for l non-Hispanic White woman 47 Table 5. Maternal minor allele frequencies for term deliveries with normal mid—pregnancy maternal serum alpha-fetoprotein. (n=503) Non-Hispanic white African American Gene/Polymorphism Minor Allele Minor Allele Frequency Frequency CD14 -159 C>T .50 .37 IL-IB +3953 C>T .23 .14 IL-lRN allele 2 .25 , .lO MBL 54 .19 .04 MBL 57 .03 .35 MMP-9 -1562 C>T .13 .11 TLR4 896 A>G .05 .06 TNF-a ~308 G>A .18 .15 TNFR2 198 ”PG .26 .20 TNFRSF6 -670 A>G .47 .33 48 Table 6. Maternal or child gene polymorphisms and preterm delivery Preterm Subtype Medically Gene Ethnicity Preterm Overall Indicated Spontaneous OR [95% C 1] OR [95% C 1] OR [95% C I] MATERNAL MMP-9 Non—Hispanic White 1.5 [0.9, 2.3] 1.1 [0.5, 2.3] 1.6 [1.0, 3.3]* African American 1.2 [0.6, 2.4] 1.1 [0.3, 4.6] 1.2 [0.5, 3.2] IL-lRN Non-Hispanic White 1.8 [1.2, 2.8]* 1.5 [0.8, 3.1] 1.9 [1.2, 3.1]* African American 0.5 [0.2, 1.5] 0.8 [0.1, 4.0] 0.5 [0.1, 1.6] TNFR2 Non-Hispanic White 1.1 [0.7, 1.7] 0.7 [0.3, 1.4] 1.3 [0.8, 2.1] African American 1.5 [0.8, 2.8] 0.4 [0.1, 2.0] 2.0 [1.0, 4.0]* CHILD MMP-9 Non-Hispanic White 0.9 [0.6, 1.5] 0.8 [0.3, 1.8] 1.0 [0.6, 1.7] African American 1.0 [0.5, 2.1] 3.0 [0.9, 11] 0.7 [0.3, 1.6] IL—lRN Non-Hispanic White 1.6 [1.0, 2.4]* 1.5 [0.7, 3.0] 1.6 [1.0, 2.7]* African American 1.2 [0.5, 2.5] 1.8 [0.4, 7.5] 1.0 [0.4, 2.5] TNFR2 Non-Hispanic White 1.0 [0.7, 1.6] 0.3 [0.1, 0.7] 1.5 [0.9, 2.4] African American 1.1 [0.5, 2.0] 0.2 [<0.1, 1.8] 1.4 [0.7, 2.9] *p<.05; OR= Odds Ratio; CI= Confidence Interval 49 Table 7. Maternal and chiltLgenotype interactions Preterm Subtype Preterm Medically Maternal & Child Overall Indicated Spontaneous Minor Allele Gene/Ethnicity Carriage OR [95% CI] OR [95% CI] OR [95% CIL IL-lRN/ Non-Hispanic white Morn No- Child No Reference Reference Reference Mom Yes-Child No 0.9 [0.5, 1.9] 0.9 [0.3, 2.9] 0.9 [0.4, 2.1] Mom No - Child Yes 1.1 [0.6, 2.1] 0.9 [0.3, 2.9] 1.2 [0.5, 2.4] Mom Yes- Child Yes 2.2 [1.3, 3.7]* 1.9 [0.8, 4.4] 2.4 [1.3, 4.2]* TNFR2/ African American Mom No- Child No Reference Reference Reference Mom Yes—Child No 0.2 [<0.1, 1.1] Not Estimable 0.4 [0.1, 1.8] Mom No - Child Yes 0.8 [0.3, 2.1] 0.3 [<0.1, 2.9] 1.1 [0.4, 3.1] Mom Yes- Child Yes 1.6 [0.8, 3.4] 0.3 [<0.1, 2.7] 2.3 [1.0, 5.3]* ‘P<.05 OR= Odds Ratio, CI= Confidence Interval 50 PART VII PROJECT 2- INTERPLAY 0F CYTOKINE POLYMORPHISMS AND BACTERIAL VAGINOSIS IN THE ETIOLOGY OF PRETERM DELIVERY Purpose The purpose of this project is to consider the impact both vaginal flora and immune system polymorphisms on the probability of preterm delivery. Rationale Genetic variants in both the tumor necrosis factor-alpha (TNF-a) and interleukin 1- beta (IL-18) genes have been inconsistently associated with an increased risk for preterm delivery. [72, 91, 109] TNF-a and IL-1 [5 are both pro-inflammatory cytokines involved in the innate immune response to microbial products and intramniotic infusions of either can stimulate increased synthesis of prostaglandins and subsequent preterm labor in non- human primates.[20] A gene-environment interaction was reported in a study of African-American women where the strength of the association between a single nucleotide polymorphism within the promoter of TNF—a (-G308A) and spontaneous preterm delivery was increased in the presence of bacterial vaginosis (BV). [81] The same interaction was reported in Afiican Americans delivering spontaneously preterm with a polymorphism (+3954) in the IL-1 [3 gene and BV. [79] BV is a condition that occurs when the flora in the vaginal tract is altered from one typically predominated by Lactobacillus sp. to a more polyrnicrobial community including Gardnerella vaginalis, Mobiluncus, Bacteroides, and Mycoplasma. BV has been associated with an increased risk for preterm delivery and chorioarnnionitis 51 although these findings have not been consistent and antibiotic trials to treat BV during pregnancy have not universally resulted in a lowering of the risk for preterm delivery. [1 10, 111] Further investigation into genetic variability in immune response to changes in vaginal flora may assist in elucidating mechanisms through which BV and preterm birth may be linked. The purpose of this project is to test the associations of three pro-inflammatory cytokine gene polymorphisms (two TNF-a promoter polymorphisms -G308A and - G238A and a single lL-l [3 polymorphism +C3954T) with preterm delivery subtypes (medically indicated or spontaneous), along with their potential interaction with vaginal flora in the Pregnancy Outcomes and Community Health (POUCH) Study. Specific Aim The aim of this project is to assess the interaction between maternal genotype (two TNF-or promoter polymorphisms -G308A and -G23 8A and a single IL—lB polymorphism +C3954T) and vaginal flora with regards to preterm birth risk and preterm birth subtype risk within African-American and non-Hispanic white subcohort women with both maternal DNA and vaginal smear findings. Study Sample Maternal polymorphisms were determined on subcohort women with available stored DNA samples (N =1 ,317). Vaginal fluid smears were collected for 82% of subcohort women with DNA (N =1 084). Racial groups that contained small numbers of women were excluded (n=18 Asian women, n= 44 Hispanic women, and n = 15 Native American women) in order to test for potential effect modification within racial group. The final study sample of 1,007 women (777 term and 230 preterm deliveries) included African- 52 American and non-Hispanic white subcohort women with both DNA results and vaginal smears. Analytic Approach Women with intermediate flora were grouped with women who were BV positive. The goals of this analysis were to: 1) test for significant associations between genotype and vaginal flora, 2) test for significant associations between genotype and preterm birth, 3) test for significant associations between vaginal flora and preterm birth, 4) test for an interaction between vaginal flora and genotype in relation to risk of preterm birth. The three polymorphisms were considered separately in all analyses. Individuals who were heterozygous or homozygous for the rare allele were grouped together. In all regression models, inverse probability weighting was used to incorporate the sampling scheme into analyses so that model estimates reflect distributions in the original cohort. The associations between genotypes and vaginal flora were tested using Rao-Scott chi-square to compare the race-specific genotype distributions in women with intermediate and positive BV scores versus women with negative BV scores. The association between genotype and preterm birth was tested using weighted polytomous logistic regression to estimate the odds ratios for carrying the rare allele and delivering preterm. The association between vaginal flora and preterm birth was tested using weighted polytomous logistic regression to estimate the odds ratio for having abnormal vaginal flora and delivering preterm. The risk of preterm birth was estimated in weighted polytyomous logistic models that included vaginal flora, genotype, and race, and three interaction terms: race with genotype, race with vaginal flora, and vaginal flora with genotype. The Wald test (p<0.05) was used to test for significance of interaction. The 53 interaction between genotype and vaginal flora was modeled by using the EV negative common allele homozygotes as the reference group. In the absence of significant interactions with race, all logistic regression models were adjusted for race/ethnicity. Results: Descriptive Statistics The weighted and unweighted maternal characteristics of the subcohort women with vaginal smears and DNA are presented in Table 8. The weighted percentages reflect the distributions in the overall cohort while the unweighted reflect the distributions in the subcohort. Afler exclusion of women without DNA or BV scores, 60% of non-Hispanic white POUCH participants and 33% of African-American POUCH participants had greater than a high school education. The weighted prevalence of BV in Non-Hispanic white women was 7% and in African-American women was 22%. The weighted prevalence of preterm birth was 10% in Non-Hispanic white women and 14% in African- American women. Unweighted race/ethnic—specific allele frequenCies for term women with normal MSAF P were calculated in both race/ethnicities and did not differ from Hardy-Weinberg equilibrium (Table 9). The allele frequencies in our Afiican-American and non-Hispanic white women were similar to previous reports. [78, 79, 82] Genotype and Vaginal Flora In both racial groups, the TNF-a —238 A allele was more common in women with abnormal flora than in BV negative women with borderline statistical significance (Table 10 non—Hispanic white weighted distribution 15% versus 9%, p=0.16;African American weighted distribution 10% versus 6%, p=0.10). There was no association between IL-1 [3 54 +3954 or TNF-a -308 and vaginal flora. There was no significant interaction between race and genotype on risk of abnormal vaginal flora. Genotype and Preterm Delivery In race adjusted models, women who had TNF-a -308 A allele had decreased risk of spontaneous preterm birth (adjusted odds ratio = 0.6; 95% CI 0.4, 0.9) . There was no significant main effect of the TNF-a -238 A allele. Women with the Il-1 B+3954 A allele were less likely to have a medically indicated preterm birth than deliver at term (adjusted odds ratio = 0.5; 95% confidence interval 0.3, 0.9). There was no significant interaction between race and genotype on preterm birth risk. Vaginal Flora and Preterm Birth In race adjusted models, women who had abnormal flora were not at significantly increased risk for delivering preterm (preterm birth overall OR=1.2; 95% CI 0.8, 1.7; spontaneous preterm birth OR=] .3; 95% CI 0.8, 1.9; medically indicated preterm birth OR=1.0; 95% CI 0.5, 1.9). There was no significant interaction between race and BV on preterm birth risk. Genotype, Vaginal Flora, and Preterm Delivery Compared to women who were BV negative and TNF-a -238 G/G, women who had abnormal flora and were TNF-a -238 A/G or -238 NA were at a significantly increased risk for PTD (Table 11 OR= 2.6 95% CI 1.2, 5.8; p-value for interaction=0.02). When grouped by circumstances at delivery the magnitude of the association did not change although the result for medically indicated no longer reached statistical significance. There was not a significant interaction for either the TNF-a -308 and IL1-B +3954 polymorphism and vaginal flora. 55 Discussion We report here a significant interaction between the TNF-a —238 polymorphism and bacterial vaginosis in relation to risk of preterm birth. Women with both the at-risk allele and who had abnormal vaginal flora were at greatest risk for preterm birth while women with just the at-risk allele or just abnormal flora were not at increased risk to deliver preterm. The strength of the association did not vary by preterm birth subtype (i.e. medically indicated or spontaneous). In our attempt to replicate the previously reported interaction between BV and two polymorphisms, TNF-or-308 and IL-1 [3, our sample did not show significant interactions for these two polymorphisms. There were several strengths to our findings. Our study was exploratory but we focused on three pre-hypothesized candidate polymorphisms with potential functionality therefore we did not have to adjust for multiple comparisons. Our population included a diverse sample of women from multiple communities. Unlike one previous study, we assessed BV as part of the study protocol instead of relying on symptoms reported in medical records. We extracted data fiom medical records on circumstances at delivery and we considered the impact of the polymorphism on preterm birth subtype including medically indicated births. We were able to test for potential effect modification by race. There were some important limitations. We did not have measures of cytokine expression to test for functionality of the polymorphisms in our population. Although we had a large cohort, the frequency of the rare alleles is low, so there was insufficient power to consider gene-gene interactions or to consider subgroups of early preterm birth where infection is more strongly implicated as a mechanism leading to preterm birth. We combined both heterozygotes and homozygyotes into a single group and therefore we 56 could not test for a gene-dose effect. Finally our measurement of vaginal flora was limited to a single sample in mid-pregnancy. This study adds to previously reported findings related to gene-environment interactions because it takes place in a population with different races and therefore tests the importance of shared genetic background in relation to shared environment exposures. In our study, the association between vaginal flora, genotype, and preterm birth did not vary by race. The relationship between BV and preterm birth is not fully understood. The TNF-or - 238 polymorphism increases the production of TNF-a in response to infection. TNF- alpha is a pro-inflammatory cytokine with multiple functions which include mediating the secretion of vasoactive substances that facilitate coagulation and stimulation of the release of prostaglandins and matrix metalloproteinases.[89, 112-115]. Elevated levels of TNF-alpha in connection with the development of BV could stimulate degradation of membranes or myometrial contractions as part of initiation of the process of early labor. Models that incorporate both specific functional polymorphisms and vaginal flora may help elucidate mechanisms to preterm birth through which successfirl interventions can be developed. Future studies should continue to test additional inflammation—related polymorphisms in diverse populations. It is unclear whether combinations of polymorphisms would more strongly impact the risk of preterm birth. Replication of findings from previous genetic studies of preterm birth have been limited and may need to be tested in larger studies. 57 Table 8. Race-Specific maternal characteristics of POUCH subcohort women with maternal DNA and vag'mal smears collected Non-Hispanic White African American Maternal characteristics N (%) (%) N (%) (%) Unwgt Wgt Unwgt Wgt Maternal Age <20 53 (10) (9) 123 (28) (28) 20-30 309 (55) (57) 265 (59) (60) >30 198 (35) (34) 59 (13) (13) Education Level (years) 5 12 223 (40) (38) 306 (68) (69) > 12 337 (60) (62) 141 (32) (31) Medicaid Insured" No 351 (63) (64) 72 (16) (16) Yes 208 (37) (36) 375 (84) (84) Primiparous" No 333 (60) (61) 261 (58) (58) Yes 226 (40) (39) 186 (42) (42) Gram Stain Results BV — (Score 0-3) 474 (85) (85) 305 (68) (68) BV Intermediate (Score 4-6) 43 (8) (8) 53 (12) (12) BV + (Score 2 7) 43 (8) (7) 89 (20) (20) Pregnancy Outcome Term delivery 411 (73) (91) 366 (82) (86) Spontaneous Preterm Birth 101 (18) (6) 57 (13) (10) Medically Indicated Preterm Birth 48 (9) (3) 24 (5) (4) Percentages are presented both unweighted (unwgt) and weighted (wgt) for subcohort sampling scheme. *Missing data for 1 non-Hispanic white women. 58 Table 9. Unweighted cytokine minor allele frequency Allele Non_Hispanic White African American N=56O N=447 TNF-a -308A .19 .14“ TNF-a -238A .05 .04 IL-IB +3954" .23 .16* *Significantly different by race p<.05 "Two non-Hispanic white woman are missing IL-1 [3 genotype Table 10. Weighted minor allele percentages and vaginal flora in non-Hispanic white and African- American subcohort participants African-American subcohort participants Non-Hispanic White African American Allele Nugent Score Nugent Score < 4 2 4 < 4 2 4 (N=474) (N=86) (N=305) (N=142) N (Eighted %) N (Wflghted %) N (Weighted %) N (Weighted %) TNF-a -308 Minor Allele 151 (32%) 27 (35%) 74 (25%) 35 (24%) TNF-a -238 Minor Allele 38 (9%) 16 (15%) 19 (6%) 14 (10%)* IL-lB +3954" Minor Allele 189 (40%) 32 (39%) 76 (26%) 37 (26%) ‘p=. 10 for African American genotype Nugent 24 versus Nugent <4 “Two non-Hispanic White woman are missing IL-IB genotype Table 11. TNF-alpha -G238A, vaginal flora, and preterm deliverv subtypes All Spontaneous Medically Preterm Preterm Indicated flgent Score and TNF-a genotype AOR (95% Cl) AOR (95% CI) AOR (95% Cl) Nugent Score < 4 and -238 G/G N=722 Reference Reference Reference Nugent Score 2 4 and -238 G/G N=198 1.0 (0.6, 1.5) 1.1 (0.7, 1.7) 0.8 (0.4, 1.6) Nugent Score < 4 and -238 NO or A/A N=57 0.7 (0.3, 1.4) 0.8 (1.0, 2.4) 0.5 (0.1, 1.9) Nugent Score 2 4 and -238 A/G or A/A N=30 2.6 (1.2, 5.8) 2.6 (1.1, 6.5) 2.6 (0.8, 8.4) AOR- race adjusted odds ratio; CI — confidence interval 59 PART VIII PROJECT 3- GENE-GENE INTERACTIONS IN CYTOKINE POLYMORPHISMS AND PRETERM DELIVERY SUBTYPE Purpose The purpose of this project is to consider the impact of gene-gene interactions on the risk of preterm delivery. Rationale Assessment of gene-gene interactions is an important step in developing etiologic models of preterm birth, however, few studies have modeled interactions between genes. In the POUCH Study, candidate genes were selected along a common pathway to preterm birth (Figure 2). Variability in immune system response from polymorphisms within multiple genes could lead to effects that are synergistic for the risk of preterm birth. Multi-factor (MDR) is a technique which measures interactions by classifying combinations of genes into high risk and low risk groups in relation to the outcome.[116] The benefits of MDR include the fact that no mode of genetic inheritance is assumed and problems with type I error due to multiple comparisons of possible gene-gene combinations are reduced. First the data is randomly divided into ten parts. Nine parts are used as a training set in which the most high risk gene-gene combinations are selected based on the ratio of cases to controls in each genotype by genotype group. In the tenth portion or testing set of the data, the ability of those genes to predict the outcome is calculated as accuracy (the ratio of correct classifications to the total number of instances classified). The process is repeated ten times and the best model is chosen based on the number of times the same polymorphisms were selected in the training data sets (cross validation) and the ability of those polymorphisms to classify the outcome in the testing data (accuracy). A nonparametric sign test is used to calculate a p-value based on the 60 number of testing accuracies greater the 0.5 (proportion by chance). Models are classified as synergistic (positive information gain) or redundant (negative information gain). There are four previous publications with data using MDR to assess gene-gene interactions and risk of preterm birth (Table 12). A study in Korean women considered four polymorphisms in three genes related to metabolic detoxification. [117] In that study a two-locus model (glutathione S-transferase u 1 and cytochrome P45011a 1 — T6235C) had a testing accuracy of 60.9% with a cross validation consistency of 10/10. The three other publications all use data from the same Tennessee study that is focused specifically on the genetics of spontaneous preterm birth. Two of the publications[72, 76] are limited to polymorphisms within the inflammatory pathway while the third publication includes over 1,500 SNPs in four different pathways (activation of the hypothalamic-pituitary-adrenal axis, inflammatory response, decidual hemorrhage, and uterine distention). [76] In the Menon publication, a significant three locus model was found among white women with spontaneous preterm births. (TNF-a, -7227, IL-6, 33314, and IL-6R 33314). [72] In the Fortunato publication, this same model was replicated in a larger group of white women with spontaneous preterm births but no significant gene-gene interactions were found in the white fetuses. (The authors report the significant SNP as IL—6R 2215 in this publication but the rs number remains rs4845622. We believe there to be a typographical error somewhere in the labeling of the SNPs between the two studies). Among Afiican-American women, no significant multilocus model was found for the maternal DNA but a significant fetal model includes two polymorphisms TNF-a -3448 and TNF R1 17691. The Velez study reported a significant 61 model in the coagulation pathway with an interaction between polymorphisms in Factor V and Factor VII. In contrast to the POUCH Study, these three studies did not focus specifically on functional polymorphisms and did not include medically indicated preterm births. Specific Aim The aim of this project was to assess gene-gene interactions within maternal genotype (TNF-a -238, TNF-a -308, IL-IB +3953, IL-lRN intron 2 repeat, TLR—4 896, MBL 54, MBL 57, Fas -670, CD 14 -159, MMP-9 -1562, and TNFR2 198) among African- American and non-Hispanic white subcohort women with bOth preterm birth and preterm birth subtype as outcomes. Study Sample (Table 13) Maternal polymorphisms were assayed on 1,317 (96% of subcohort) women. The final sample included Afiican-American (N=553) and non-Hispanic white women (N=675). Initial analyses included both mothers and children. However, the significant findings in the maternal subcohort (n=1,317) and the group with both maternal and children’s DNA (n=779) were not consistent. Due to concerns that the missing children’s DNA may not be missing at random and that it represented a significant loss to follow up (greater than 20% of the original subcohort), the results are presented here for the maternal subcohort only. This group was chosen in order to minimize the potential for bias in the reported results. Analytic Approach The impact of gene-gene interactions on the risk of preterm birth was assessed with the use of MDR. (MDRware version 1.1.0). Two sets of MDR models were used to test 62 for gene-gene interactions for all eleven maternal polymorphisms. The first set of models assessed gene— gene interactions in race—stratified models with preterm delivery as the outcome. The second set of models assessed gene-gene interactions with preterm subtype (medically indicated or spontaneous) as the outcome. Genotypes for IL-1 RN were grouped into three levels: no copies of allele two, one copy of allele two, or three copies of allele two. Genotypes for the remaining nine polymorphisms were entered as three levels: homozygous rare allele, heterozygous, homozygous common allele. The best multi-gene model had the highest level of cross validation in the testing data set and had to be selected in the training data set at least 8 out of 10 times. Following the selection of the best maternal gene-gene interaction models, weighted logistic regression was used to calculated odds ratios for the rare and common allele combinations. Results (Table 14) In non-Hispanic whites, MDR identified two significant gene-gene interactions with at least 80% consistency. A redundant interaction between IL-IRN and IL-1 [3 had an accuracy of 58% and a cross validation of 9/10 in the model with spontaneous preterm delivery. A synergistic interaction between CD14 and TNFR2 had an accuracy of 58% and a cross validation of 10/10. In Afiican—American women, no gene-gene interactions were identified however, a significant single gene model was found for IL-1 [3 and preterm birth. Discussion A significant redundant interaction was identified for IL—lRN and IL—lB in non- Hispanic whites. These genes are physically linked and their products biologically compete for opposing roles in the inflammatory process. In the MDR results, women 63 with the rare IL-l RN allele (decreased production and increase in inflammation) and common IL-l B (no change regular inflammation) were grouped into the high risk category although this was not clearly demonstrated in the odds ratios estimated from logistic models (Table 15). This was the first report using MDR with medically indicated preterm birth. The final model included two receptors’ polymorphisms. The CD14 is associated with a increased inflammatory response while the effect of the TNFR2 polymorphism is less clear. In the MDR results, the low risk group did not follow a clear pattern. In the logistic models, the group with the rare CD14 allele and common TNFR2 alleles had an increased risk of medically indicated preterm birth (odds ratio=5.8, 95% confidence interval 1.5, 22.5) and the group that had the common CD14 alleles and rare TNFR2 allele had an increased risk of medically indicated preterm birth (odds ratio=4.8, 95% confidence interval 1.4, 16.4) as compared to the group with all common alleles. A new single polymorphism association was found in our data for IL-1 B and preterm birth in Afiican Americans. Both the homozygous rare allele and homozygous common allele groups were at increased risk for preterm birth (Table 15 odds ratio=3.2, 95% confidence interval 1.0, 10.9; odds ratio=1.9 95% confidence interval 1.0, 3.4). This association was not identified in previous logistic models because a dominant effect model had been previously used. The preterm birth models developed with MDR were similar to previous work in two ways. First we identified a significant gene-gene interaction for a polymorphism that had no significant main effect on PTD (CD14). Second, our models were not the same across 64 racial groups and much like the Fortunato study we found no significant gene- gene interactions using MDR with the maternal African-American data. If all possible two-way and three-way interactions had been assessed, over 800 logistic regression models would need to be run for each genetic model (dominant and recessive). With a significance level of .05, up to eighty significant interactions would be found by chance alone. Although methods are available to adjust for multiple comparisons, the small numbers of POUCH women with each subtype of preterm birth and the likely small effect of each interaction would make it possible that many findings would border on statistical significance. Adjustment may not help separate the true positive results from the false positives. MDR provided a useful alternative method for dealing with multiple comparisons. In our study, the selected best models had high cross validation consistency (10/10) but low accuracy (58%). Previous authors have reported accuracies ranging from 59%-68%. The findings in the POUCH study were based on fewer cases of preterm birth perhaps this contributed to a weaker overall fit of our final model. The use of MDR in our data set provides preliminary evidence of gene— gene interactions within the inflammatory pathway. This is the first comparison of MDR- based models across preterm birth subtypes. The best models for spontaneous preterm birth and medically indicated preterm birth did not include the same polymorphisms. Our findings suggest that the clinical circumstances leading to preterm delivery (i.e. preterm birth subtype) are indicative of biologically alternate pathways to preterm and that genetic models should be developed separately for medically indicated and spontaneous preterm birth. 65 mSnanH mono» 83650 sing—82m <29 gum E Emu—mama N: 02 5 £75 own a Eon“ . 58am zap—Bum fisEoovdoquBéE .2532 woom 0:2 «3558 > 88$ can 9295 com 333.44 .38 55 mo noua>uo< 588an .No_o> £53885 13$ 95 358$: 02 39: a? $26 80: SEZH $4M- ”mm? 5.2 o:o_ wvvm- NYE KNNF- men: “505:2 =85 Eon ”neurone gov—fix .Som fleece HuaomEmmw 026m 58.3 2% googuomm oz 63.58" 50.554. 3:032 I m.mm ”N505.“ Mod SWEEP 358 ”505$ §o£< EEBcE obnfinofinm Enema? 02 ”BE? .80..“ 326683 02 :85 mos: m fl mm "233 Eva mean can ca: .9529 vognquok can as. coco SEE 4me 6.2 87458on gmom 549:. wvvm- vamp- 3E7? macaw m 5 $78 mm BE? moon 3 \2 ”833 388va “2.53 38032 ”noumEEmmE owammiaoz .Swfifiom men: 63 3852on .9572. Hmmnnuoh Mon: 3 mmm .5373. 8.? STEBDE £5.60 0%: SN». ocovm «ME muaow m 5 £75 E 3E3 coca 53m 82 PE 34:- womb- 557E. Bowman—m5 unamflméoz 6252 H $3 -w 25.332» 23:- .888 8835.. >327 3:4;0 I M «Ema oEoEooSo vows—oxm 3me “23% 38:83» 5:8 .Em fizm v E“ as» m ofluase £660 flobaoo ._o>o_ noumozco .owa 3503.: fig “coasoumxouofl wing—535 2:: 98 830 Enema 5w waofi< 8&5 com 30835.5... 55 32.50 0209802 8:83 523— woom dam 3d. 35 a. 35 a Danni—.8 acting—“H nous—9:9» $qu 352 mag in; 398% 5:2 «Emu—505293050 dean—snea— 3.5m 5:.— EHSP. me it E: 388335 25 .28 $8: 8 5:82.955 cough—-31.: 5m: 8.63m 2.3.5 .5 .N— 033—. 66 Table 13. Characteristics of Non-Hispanic White and African-American POUCH Study participants with maternal DNA. Non-Hispanic White African-American (N=675) (N=553) thg th thg th Maternal Characteristics N % % N % % Maternal Age (years) <20 66 10 9 151 27 28 20-29 369 55 56 325 59 59 230 240 35 34 77 14 13 Maternal Education (years) <12 80 12 11 205 37 38 12 182 27 26 165 30 30 >12 413 61 63 183 33 32 Medicaid lnsured* . No 428 64 65 99 18 18 Yes 246 36 35 453 82 82 Parity* None 286 42 41 227 41 42 At least one 388 58 59 326 59 58 Pregnancy Outcome Term 483 72 90 450 81 86 Preterm Medically Indicated 62 9 3 29 5 4 Preterm Spontaneous 130 19 7 74 14 10 Percentages are presented both unweighted (unwgt) and weighted (wgt) for subcohort sampling scheme. *Missing data on Medicaid for one African-American and one non-Hispanic white woman. Missing parity data for one non-Hispanic white woman. 67 Table 14. Results from Multi—Factor Dimensionality reduction analysis for maternal genotype Model Testing Balance Cross Validation Sign test (p) Selected as Accuracy Consistency final best model? Preterm birth in non-Hispanic whites IL-lRN 53% 8/10 7 (.17) No IL-lRNzlL-lfi 54% 6/10 9 (.01) No TLR4:CD14:IL—IB 53% 5/10 7 (.17) No Preterm birth in African Americans Il-IB 55% 10/10 8 (.05) Yes CD14:TNF-a -308 58% 3/10 2 (.99) No MBL57:Il-1 BzTNF-a—308 63% 4/10 3 (.95) No Medicallyindicated preterm birth in non-Hispanic whites CD14 49% 9/10 5 (.62) No CD14:TNFR2 58% 10/10 8 (.05) Yes CD14zTNFRzFas 53% 7/10 7 (.17) No MedicalbLindicated preterm birth in African Americans IL-lB 52% 8/10 6 (.38) No TNFR2:TNF-a -308 44% 4/10 4 (.83) No CD14:IL-1b:TNF—a -308 53% 5/10 4 (.83) No Spontaneous preterm birth in non-Hispanic whites IL-lRN 55% 9/10 7 (.17) No IL-lRN:IL-l[3 58% 9/10 8 (.05) Yes CD14:Fas:TNF-a -308 58% 8/10 8 (.05) No Spontaneous preterm birth in African Americans TNFR2 42% 4/10 1 (.99) No TNFR2:IL-1b 44% 3/10 2 (.99) No CD14:TNFR2:Fas 51% 6/10 6 (.38) No Testing balance accuracy — Percentage of outcomes corrected classified by using the model developed in the training data sets in the testing data. Cross validation consistency- The number of times the same polymorphisms were selected in the training data sets. Sign test-A nonparametric test used to calculate a p-value based on the number of testing accuracies greater the 0.5 (proportion by chance). 68 Table 15. Odds Ratios from weighted logistic regression models for significant models identified with multi-factor dimensionality reduction. Model OR (95% CI) Non-Hispanic white Spontaneous IL-lRNzlL-IB Commoanare Reference Common:Common 0.9 (0.4, 2.2) Rare: Common 1.4 (0.7, 3.1) RarezRare 1.2 (0.4, 3.1) Medically Indicated CDl4zTNFR2 CommonzCommon Reference Rare:Common 5.8 (1.5, 22.5) Commoanare 4.8 (1.4, 16.4) RarezRare 2.1 (0.6, 7.8) African American Preterm IL-lB CT Reference Tl‘ 3.2 (1.0, 10.9) CC 1.9 (1.0, 3.4) Common — Common allele homozygous women Rare — Heterozygous and rare allele homozygous women 69 PART IX CONCLUSION According to Menon, “preterm birth is a complex disease with multiple pathways that culminate in a common terminal pathway.” Most likely these pathways involve both genetic and environmental factors. In the POUCH Study our focus on specific functional polymorphisms within inflammation-related genes allowed us to identify new significant interactions between maternal and fetal genotype, between vaginal flora and maternal genotype, and between genes within maternal genotype. These findings lend support to the hypothesis that alterations in immune system response can cause preterm birth. Our findings also further illustrate the inherent complexity of developing models that include individual level genetic variability. The strength of the POUCH study design is its ability to collect detailed individual level factors (for example, vaginal flora, perceived stress) in mid-pregnancy prior to delivery. Other gene association studies of preterm birth have primarily utilized the case- control design and therefore had to rely upon information available at birth such as medical records, administrative data bases, or biologic samples collected at birth. One limitation of identifying cases using the cohort design is the difficulty in accruing large numbers of cases when the outcome is rare. For example, the majority of cases in the POUCH Study occur within the 35-36 week range and therefore we lack sufficient power to assess the impact of stratifying by early preterm birth. The POUCH cohort is made up of an ethnically diverse group of women with a variety of socioeconomic backgrounds. While these Characteristics maximum the potential generalizibility of findings from the POUCH Study, they may also make it difficult to separate out the effects specific genes when genetic background may play a role. In the 70 POUCH Study, women were allowed to self-select a single race. Other studies have focused on less diverse populations or have excluded women with multi-racial backgrounds. Gene association studies within the inflammatory pathway suffer fiom a lack of replicability. Some potential reasons include that within an individual different genetic etiologies can lead to the same trait. A single gene mutation is unlikely to have a large effect for such a complex phenotype and multi-locus gene-gene interactions may have an important role. The measured polymorphisms may not be functional but may be markers of nearby functional changes within a gene. Within different populations, different patterns of correlations between polymorphisms may affect the strength of the association. The definition of preterm birth varies across studies. Many published studies have limited sample sizes that are underpowered to detect effects. Finally, the genotype of both mother and fetus may be important. In 2007, The PREBIC study group published guidelines for genetic epidemiologic studies of preterm birth.[12] They discussed issues related to phenotypic criteria, study design, selection of controls, and candidate gene selection. In their review, the authors argue for both a minimal and an optimal set of data that should be available in genetic studies of preterm birth. The authors hope that by following specific criteria that future genetic studies will include greater opportunities for verification in independent populations, comparisons of results between multiple studies, and combining of data sets from different studies. As technology advances, the ability to measure thousands of polymorphisms within a single study is becoming a reality. The future of genetics and preterm birth may include 71 larger data sets with more complex analyses. Preterm birth cohorts must continue to play a role so that exposures during the prenatal period can be accurately measured. As investigators develop models careful consideration of phenotypic definition, interactions between genes and environment and interactions between maternal and fetal genotype along specific pathways to preterm birth will all be important in the search to understand causal mechanisms. 72 REFERENCES 10. 11. Hamilton, B.E., J. Martin, and S. Ventura, Births: Preliminary Data for 2006, in National Vital Statistics Reports. 2007, CDC. p. 1-18. Buitendijk, S., J. Zeitlin, M. Cuttini, J. Langhoff-Roos, and J. Bottu, Indicators of fetal and infant health outcomes. Eur J Obstet Gynecol Reprod Biol, 2003. 111 Suppl 1: p. S66-77. Hamilton, B.E., J .A. Martin, and SJ. Ventura. Births: preliminary data for 2005. Natl Vital Stat Rep 2006 [cited 1/25/07]; 1-17]. Available from: http://www.cdc.gov/nchs/products/pubs/pubd/hestats/prelimbirthsOS/prelimbirths 05.htm. Blondel, B. and M. Kaminski, Trends in the occurrence, determinants, and consequences of multiple births. Semin Perinatol, 2002. 26(4): p. 239-49. MacDorman, M.F., J .A. Martin, T.J. Mathews, D.L. Hoyert, and SJ. Ventura, Explaining the 2001-2002 infant mortality increase in the United States: data from the linked birth/infant death data set. Int J Health Serv, 2005. 35(3): p. 415- 42. Callaghan, W.M., M.F. MacDorman, S.A. Rasmussen, C. Qin, and EM. Lackritz, The contribution of preterm birth to infant mortality rates in the United States. Pediatrics, 2006. 118(4): p. 1566-73. Ward, R.M. and J .C. Beachy, Neonatal complications following preterm birth. BJOG, 2003. 110 Suppl 20: p. 8-16. Behrman, RE. and AS. Butler, Preterm birth: Causes, consequenses, and prevention. 2006, Institute of Medicine: Washington, DC: National Academies Press. Zhang, J. and DA. Savitz, Preterm birth subtypes among blacks and whites. Epidemiology, 1992. 3(5): p. 428-33. Goldenberg, R.L., S.P. Cliver, F.X. Mulvihill, C.A. Hickey, H.J. Hoffman, L.V. Klerman, and M.J. Johnson, Medical, psychosocial, and behavioral risk factors do not explain the increased risk for low birth weight among black women. Am J Obstet Gynecol, 1996. 175(5): p. 1317-24. Mercer, B.M., R.L. Goldenberg, A. Das, A.H. Moawad, J .D. Iams, P.J. Meis, R.L. Copper, F. Johnson, E. Thom, D. McNellis, M. Miodovnik, M.K. Menard, S.N. Caritis, G.R. Thumau, S.F. Bottoms, and J. Roberts, The preterm prediction 73 12. l3. 14. 15. l6. 17. 18. 19. 20. study: A clinical risk assessment system. Am J Obstet Gynecol, 1996. 174(6): p. 1885-95. Pennell, C.E., B. Jacobsson, S.M. Williams, R.M. Buus, L.J. Muglia, S.M. Dolan, N.H. Morken, H. Ozcelik, S.J. Lye, and C. Relton, Genetic epidemiologic studies of preterm birth: guidelines for research. Am J Obstet Gynecol, 2007. 196(2): p. 107-18. Moutquin, J .M., Classification and heterogeneity of preterm birth. BJOG, 2003. 110 Suppl 20: p. 30—3. Berkowitz, G.S., C. Blackmore-Prince, R.H. Lapinski, and DA. Savitz, Risk factors for preterm birth subtypes. Epidemiology, 1998. 9(3): p. 279-85. Goepfert, A.R., R.L. Goldenberg, W.W. Andrews, J .C. Hauth, B. Mercer, J. Iams, P. Meis, A. Moawad, E. Thom, J .P. VanDorsten, S.N. Caritis, G. Thumau, M. Miodovnik, M. Dombrowski, J. Roberts, and D. McNellis, The Preterm Prediction Study: association between cervical interleukin 6 concentration and spontaneous preterm birth. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol, 2001. 184(3): p. 483-8. Genc, M.R., S.S. Witkin, M.L. Delaney, L.R. Paraskevas, R.E. Tuomala, E.R. Norwitz, and AB. Onderdonk, A disproportionate increase in IL-1 beta over IL- Ira in the cervicovaginal secretions of pregnant women with altered vaginal microflora correlates with preterm birth. Am J Obstet Gynecol, 2004. 190(5): p. 1 191-7. Wenstrom, K.D., W.W. Andrews, J .C. Hauth, R.L. Goldenberg, M.B. DuBard, and SP. Cliver, Elevated second-trimester amniotic fluid interleukin-6 levels predict preterm delivery. Am J Obstet Gynecol, 1998. 178(3): p. 546-50. El-Shazly, S., M. Makhseed, F. Azizieh, and R. Raghupathy, Increased expression of pro-inflammatory cytokines in placentas of women undergoing spontaneous preterm delivery or premature rupture of membranes. Am J Reprod Immun, 2004. 52(1): p. 45-52. Holzman, C., X. Lin, P. Senagore, and H. Chung, Histologic chorioarnnionitis and preterm delivery. Am J Epidemiol, 2007. 166(7): p. 786-94. Sadowsky, D.W., K.M. Adams, M.G. Gravett, S.S. Witkin, and M.J. Novy, Preterm labor is induced by intraamniotic infusions of interleukin-1 beta and tumor necrosis factor-alpha but not by interleukin-6 or interleukin-8 in a nonhuman primate model. Am J Obstet Gynecol, 2006. 195(6): p. 1578-89. 74 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. Goldenberg, R.L., J .C. Hauth, and W.W. Andrews, Intrauterine infection and preterm delivery. New Eng J Med, 2000. 342(20): p. 1500-7. Adams, M.M., L.D. Elam-Evans, H.G. Wilson, and DA. Gilbertz, Rates of and factors associated with recurrence of preterm delivery. J AMA, 2000. 283(12): p. 1 591-6. Carlini, L., E. Somigliana, G. Rossi, F. Veglia, M. Busacca, and M. Vignali, Risk factors for spontaneous preterm birth: a Northern Italian multicenter case- control study. Gynecol Obstet Invest, 2002. 53(3): p. 174-80. Cnattingius, S., F. Granath, G. Petersson, and BL. Harlow, The influence of gestational age and smoking habits on the risk of subsequent preterm deliveries. N Engl J Med, 1999. 341(13): p. 943-8. Kistka, Z.A., L. Palomar, K.A. Lee, SH. Boslaugh, M.F. Wangler, F.S. Cole, M.R. DeBaun, and L.J. Muglia, Racial disparity in the frequency of recurrence of preterm birth. Am J Obstet Gynecol, 2007. 196(2): p. 131 e1-6. Kristensen, J ., J. Langhoff-Roos, and RB. Kristensen, Implications of idiopathic preterm delivery for previous and subsequent pregnancies. Obstet Gynecol, 1995. 86(5): p. 800-4. Melve, K.K., R. Skjaerven, H.K. Gjessing, and N. Oyen, Recurrence of gestational age in sibships: implications for perinatal mortality. Am J Epidemiol, 1999. 150(7): p. 756-62. Basso, 0., J. Olsen, and K. Christensen, Study of environmental, social, and paternal factors in preterm delivery using sibs and half sibs. A population—based study in Denmark J Epidemiol Community Health, 1999. 53(1): p. 20-3 Ananth, C.V., D. Getahun, M.R. Peltier, H.M. Salihu, and AM. Vintzileos, Recurrence of spontaneous versus medically indicated preterm birth. Am J Obstet Gynecol, 2006. 195(3): p. 643-50. McManemy, J ., E. Cooke, E. Amon, and T. Leet, Recurrence risk for preterm delivery. Am J Obstet Gynecol, 2007. 196(6): p. 576 el-6. Bloom, S.L., N.P. Yost, D.D. McIntire, and K.J. Leveno, Recurrence of preterm birth in singleton and twin pregnancies. Obstet Gynecol, 2001. 98(3): p. 379-85. Ekwo, E. and A. Moawad, The risk for recurrence of premature births to African- American and white women. J Assoc Acad Minor Phys, 1998. 9(1): p. 16-21. 75 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. Mercer, B.M., R.L. Goldenberg, A.H. Moawad, P.J. Meis, J .D. Iams, A.F. Das, S.N. Caritis, M. Miodovnik, M.K. Menard, G.R. Thumau, M.P. Dombrowski, J .M. Roberts, and D. McNellis, The preterm prediction study: effect of gestational age and cause of preterm birth on subsequent obstetric outcome. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol, 1999. 181(5 Pt 1): p. 1216-21. Iams, J .D., R.L. Goldenberg, B.M. Mercer, A. Moawad, E. Thom, P.J. Meis, D. McNellis, S.N. Caritis, M. Miodovnik, M.K. Menard, G.R. Thumau, S.E. Bottoms, and J .M. Roberts, The Preterm Prediction Study: recurrence risk of spontaneous preterm birth. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol, 1998. 178(5): p. 1035-40. Mercer, B.M., C.A. Macpherson, R.L. Goldenberg, A.R. Goepfert, S. Haugel-De Mouzon, M.W. Vamer, J .D. Iams, P.J. Meis, A.H. Moawad, M. Miodovnik, S.N. Caritis, J .P. Van Dorsten, Y. Sorokin, G.R. Thumau, and CY. Spong, Are women with recurrent spontaneous preterm births diflerent from those without such history? Am J Obstet Gynecol, 2006. 194(4): p. 1176-84. Clausson, B., P. Lichtenstein, and S. Cnattingius, Genetic influence on birthweight and gestational length determined by studies in oflfspring of twins. BJOG, 2000. 107(3): p. 375-81. Treloar, S.A., G.A. Macones, L.E. Mitchell, and NC. Martin, Genetic influences on premature parturition in an Australian twin sample. Twin Res, 2000. 3(2): p. 80—2. ‘ Magius, P., L.S. Bakketeig, and R. Skjaerven, Correlations of birth weight and gestational age across generations. Ann Hum Biol, 1993. 20(3): p. 231-8. Porter, T.F., A.M. Fraser, C.Y. Hunter, R.H. Ward, and M.W. Vamer, The risk of preterm birth across generations. Obstet Gynecol, 1997. 90(1): p. 63-7. Ward, K., V. Argyle, M. Meade, and L. Nelson, The heritability of preterm delivery. Obstet Gynecol, 2005. 106(6): p. 1235-9. Khoury, M.J. and EH. Cohen, Genetic heterogeneity of prematurity and intrauterine growth retardation: clues from the Old Order Amish. Am J Obstet Gynecol, 1987. 157(2): p. 400-10. Johnstone, F. and L. Inglis, Familial trends in low birth weight. Br Med J, 1974. 3(5932): p. 659-61. 76 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. Holzman, C., B. Bullen, R. Fisher, N. Paneth, and L. Reuss, Pregnancy outcomes and community health: the POUCH study of preterm delivery. Paediatr Perinat] Epidemiol, 2001. 15 Suppl 2: p. 136-58. Jones, N.M., C.B. Holzman, A.J. Zanella, C.M. Leece, and M.H. Rahbar, Assessing mid-trimester salivary cortisol levels across three consecutive days in pregnant women using an at-home collection protocol. Paediatr Perinat Epidemiol, 2006. 20(5): p. 425-37. Nugent, R.P., M.A. Krohn, and S.L. Hillier, Reliability ofdiagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol, 1991. 29(2): p. 297-301. Meulenbelt, I., S. Droog, G.J. Trommelen, D.I. Boomsma, and RE. Slagboom, High-yield noninvasive human genomic DNA isolation method for genetic studies in geographically dispersed families and populations. Am J Hum Genet, 1995. 57(5): p. 1252-4. Kalish, R.B., S. Vardhana, N.J. Normand, M. Gupta, and SS. Witkin, Association of a maternal CD14 -159 gene polymorphism with preterm premature rupture of membranes and spontaneous preterm birth in multi-fetal pregnancies. J Reprod Immunol, 2006. 70(1-2): p. 109-17. Lorenz, E., M. Hallman, R. Marttila, R. Haataja, and D.A. Schwartz, Association between the Asp299Gly polymorphisms in the T all-like receptor 4 and premature births in the Finnish population. Pediatr Res, 2002. 52(3): p. 373-6. Genc, M.R., S. Gerber, M. Nesin, and SS. Witkin, Polymorphism in the interleukin-I gene complex and spontaneous preterm delivery. Am J Obstet Gynecol, 2002. 187(1): p. 157-63. Wilson, A.G., F.S. di Giovine, A.I. Blakemore, and G.W. Duff, Single base polymorphism in the human tumour necrosis factor alpha (TNF alpha) gene detectable by NcoI restriction of PCR product. Hum Mol Genet, 1992. 1(5): p. 353. - Al-Ansari, AS, WE. Ollier, J. Villarreal, J. Ordi, L.S. Teh, and AH. Hajeer, Tumor necrosis factor receptor II (TNFRII) exon 6 polymorphism in systemic lupus erythematosus. Tissue Antigens, 2000. 55(1): p. 97-9. Kalish, R.B., D.P. Nguyen, S. Vardhana, M. Gupta, S.C. Perni, and SS. Witkin, A single nucleotide A>G polymorphism at position -6 70 in the Fas gene promoter: relationship to preterm premature rupture of fetal membranes in multifetal pregnancies. Am J Obstet Gynecol, 2005. 192(1): p. 208-12. 77 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. Gremlich, S., D. Nguyen, D. Reymondin, P. Hohlfeld, Y. Vial, S.S. Witkin, and S. Gerber, Fetal MMPZ/MMP9 polymorphisms and intrauterine growth restriction risk. J Reprod Irnmunol, 2007. 74(1-2): p. 143-51. Babula, 0., I. Danielsson, I. Sjoberg, W.J. Ledger, and SS. Witkin, Altered distribution of mannose-binding lectin alleles at exon I codon 54 in women with vulvar vestibulitis syndrome. Am J Obstet Gynecol, 2004. 191(3): p. 762-6. Giraldo, P.C., O. Babula, A.K. Goncalves, I.M. Linhares, R.L. Amara], W.J. Ledger, and SS. Witkin, Mannose-binding lectin gene polymorphism, vulvovaginal candidiasis, and bacterial vaginosis. Obstet & Gynecol, 2007. 109(5): p. 1123-8. Campos, J ., F. Gude, C. Quinteiro, C. Vidal, and A. Gonzalez-Quintela, Gene by environment interaction: the -1 59C/T polymorphism in the promoter region of the CD14 gene modifies the effect of alcohol consumption on serum IgE levels. Alcohol Clin Exp Res, 2006. 30(1): p. 7-14. Wright, S.D., R.A. Ramos, P.S. Tobias, R.J. Ulevitch, and J .C. Mathison, CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science, 1990. 249(4975): p. 1431-3. Haziot, A., G.W. Rong, V. Bazil, J. Silver, and SM. Goyert, Recombinant soluble CD14 inhibits LPS-induced tumor necrosis factor-alpha production by cells in whole blood. J Irnmunol, 1994. 152(12): p. 5868-76. Baldini, M., I.C. Lohman, M. Halonen, R.P. Erickson, P.G. Holt, and FD. Martinez, A Polymorphism* in the 5 ' flanking region of the CD14 gene is associated with circulating soluble CD14 levels and with total serum immunoglobulin E. Am J Respir Cell Mol Biol, 1999. 20(5): p. 976-83. Temple, S.E., KY. Cheong, C.M. Almeida, P. Price, and G.W. Waterer, Polymorphisms in lymphotoxin alpha and CD14 genes influence INF alpha production induced by Gram-positive and Gram—negative bacteria. Genes Irnmun, 2003. 4(4): p. 283-8. Harte], C., D. Finas, P. Ahrens, E. Kattner, T. Schaible, D. Muller, H. Segerer, K. Albrecht, J. Moller, K. Diedn'ch, and W. Gope], Polymorphisms of genes involved in innate immunity: association with preterm delivery. Mol Hum Reprod, 2004. 10(12): p. 911-5. Baumgarten, G., P. Knuefermann, H. Wrigge, C. Putensen, H. Stapel, K. Fink, R. Meyer, A. Hoefi, and C. Grohe, Role of T all-like receptor 4 for the pathogenesis of acute lung injury in Gram-negative sepsis. Eur J Anaesthesiol, 2006. 23(12): p. 1041-8. 78 63. 64. 65. 66. 67. 68. 69. 70. 7]. Schwartz, D.A., The genetics of innate immunity. Chest, 2002. 121(3 Suppl): p. 62S-68S. Erridge, C., J. Stewart, and LR. Poxton, Monocytes heterozygous for the Asp299Gly and Thr39911e mutations in the T oil-like receptor 4 gene show no deficit in lipopolysaccharide signalling. J Exp Med, 2003. 197(12): p. 1787-91. Michel, 0., TD. LeVan, D. Stern, M. Dentener, J. Thorn, D. Gnat, M.L. Beijer, P. Cochaux, P.G. Holt, F .D. Martinez, and R. Rylander, Systemic responsiveness to lipopolysaccharide and polymorphisms in the toll-like receptor 4 gene in human beings. J Allergy Clin Irnmunol, 2003. 112(5): p. 923-9. Genc, M.R., S. Vardhana, M.L. Delaney, A. Onderdonk, R. Tuomala, E. Norwitz, and SS. Witkin, Relationship between a toll-like receptor-4 gene polymorphism, bacterial vaginosis-related flora and vaginal cytokine responses in pregnant women. Eur J Obstet Gynecol Reprod Biol, 2004. 116(2): p. 152-6. Ferrand, P.E., T. Fujimoto, V. Chennathukuzhi, S. Parry, G.A. Macones, M. Samme], H. Kuivaniemi, R. Romero, and J .F . Strauss, 3rd, The CARDI5 2936insC mutation and TLR4 896 A>G polymorphism in African Americans and risk of preterm premature rupture of membranes (PPROM). Mol Hum Reprod, 2002. 8(11): p. 1031-4. Rey, G., F. Skowronek, J. Alciaturi, J. Alonso, B. Bertoni, and R. Sapiro, Toll receptor 4 Asp299Gly polymorphism and its association with preterm birth and premature rupture of membranes in a South American population. Mol Hum Reprod, 2008. 14(9): p. 555-9. Wilson, A.G., J .A. Symons, T.L. McDowell, H.O. McDevitt, and G.W. Duff, Effects of a polymorphism in the human tumor necrosis factor alpha promoter on transcriptional activation. Proc Natl Acad Sci U S A, 1997. 94(7): p. 3195-9. Pociot, F., L. Briant, C.V. Jongeneel, J. Molvig, H. Worsaae, M. Abba], M. Thomsen, J. Nerup, and A. Cambon-Thomsen, Association of tumor necrosis factor (TNF) and class 11 major histocompatibility complex alleles with the secretion of INF-alpha and TNF-beta by human mononuclear cells: a possible link to insulin-dependent diabetes mellitus. Eur J Irnmunol, 1993. 23(1): p. 224- 31. Hernandez-Guerrero, C., F. Monzon-Bordonaba, L. Jimenez-Zamudio, R. Ahued- Ahued, F. Arechavaleta—Velasco, J .F. Strauss, 3rd, and F. Vadillo-Ortega, In-vitro secretion of proinflammatory cytokines by human amniochorion carrying hyper- responsive gene polymorphisms of tumour necrosis factor-alpha and interleukin- 1 beta. Mo] Hum Reprod, 2003. 9(10): p. 625-9. 79 72. 73. 74. 75. 76. 77. 78. 79. 80. Menon, R., D.R. Velez, H. Simhan, K. Ryckman, L. Jiang, P. Thorsen, I. Vogel, B. Jacobsson, M. Merialdi, S.M. Williams, and SJ. Fortunato, Multilocus interactions at maternal tumor necrosis factor-alpha, tumor necrosis factor receptors, interleukin-6 and interleukin-6 receptor genes predict spontaneous preterm labor in European-American women. Am J Obstet Gynecol, 2006. 194(6): p. 1616-24. Chen, D., Y. Hu, B. Wu, L. Chen, Z. Fang, F. Yang, and L. Wang, Tumor necrosis factor-alpha gene G3 08A polymorphism is associated with the risk of preterm delivery. Beijing Da Xue Xue Bao, 2003. 35(4): p. 377-8]. Dizon-Townson, D.S., H. Major, M. Vamer, and K. Ward, 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(4): p. 810-3. Annells, M.F., P.H. Hart, C.G. Mullighan, S.L. Heatley, J .3. Robinson, P. Bardy, and HM. McDonald, Interleukins-I, -4, -6, -10, tumor necrosis factor, transforming growth factor-beta, FAS, and mannose-binding protein C gene polymorphisms in Australian women: Risk of preterm birth. Am J Obstet Gynecol, 2004. 191(6): p. 2056-67. Fortunato, S.J., R. Menon, D.R. Velez, P. Thorsen, and SM. Williams, Racial disparity in maternal-fetal genetic epistasis in spontaneous preterm birth. Am J Obstet Gynecol, 2008. 198(6): p. 666 el-9; discussion 666 e9-10. Amory, J .H., K.M. Adams, M.T. Lin, J .A. Hansen, D.A. Eschenbach, and J. Hitti, Adverse outcomes after preterm labor are associated with tumor necrosis factor- alpha polymorphism -863, but not -3 08, in mother-infant pairs. Am J Obstet Gynecol, 2004. 191(4): p. 1362-7. Moore, S., M. Ide, M. Randhawa, J .J . Walker, J .G. Reid, and NA. Simpson, An investigation into the association among preterm birth, cytokine gene polymorphisms and periodontal disease. BJOG, 2004. 111(2): p. 125-32. Engel, S.A., HC. Erichsen, D.A. Savitz, J. Thorp, S.J. Chanock, and AF. Olshan, Risk of spontaneous preterm birth is associated with common proinflammatory cytokine polymorphisms. Epidemiology, 2005. 16(4): p. 469-77. Valdez, L.L., A. Quintero, E. Garcia, N. Olivares, A. Celis, F. Rivas, Jr., and F. Rivas, Thrombophilic polymorphisms in preterm delivery. Blood Cells Mo] Dis, 2004. 33(1): p. 51-6. 80 81. 82. 83. 84. 85. 86. 87. 88. 89. Macones, G.A., S. Parry, M. Elkousy, B. Clothier, S.H. Ural, and J .F. Strauss, 3rd, A polymorphism in the promoter region of INF and bacterial vaginosis: preliminary evidence of gene-environment interaction in the etiology of spontaneous preterm birth. Am J Obstet Gynecol, 2004. 190(6): p. 1504-8. Roberts, A.K., F. Monzon-Bordonaba, P.G. Van Deerlin, J. Holder, G.A. Macones, M.A. Morgan, J .F. Strauss, 3rd, and S. Parry, 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(5): p. 1297-302. Dizon Townson, D.S., L. Meline, L.M. Nelson, M. Vamer, and K. Ward, Fetal carriers of the factor VLeiden mutation are prone to miscarriage and placental infarction. Am J Obstet Gynecol, 1997. 177(2): p. 402-5. Aidoo, M., P.D. McElroy, M.S. Kolczak, D.J. Terlouw, F.O. ter Kuile, B. Nahlen, A.A. La], and V. Udhayakumar, Tumor necrosis factor-alpha promoter variant 2 (YNFZ) is associated with pre-term delivery, infant mortality, and malaria morbidity in western Kenya: Asembo Bay Cohort Project IX. Genet Epidemiol, 2001. 21(3): p. 201-11. Kazzi, S.N., S.M. Jacques, F. Qureshi, M.W. Quasney, U.O. Kim, and LA. Buhimschi, Tumor necrosis factor-alpha allele lymphotoxin-alpha+25 0 is associated with the presence and severity of placental inflammation among preterm births. Pediatr Res, 2004. 56(1): p. 94-8. Bazzoni, F. and B. Beutler, The tumor necrosis factor ligand and receptor families. N Engl J Med, 1996. 334(26): p. 1717-25. Komata, T., N. Tsuchiya, M. Matsushita, K. Hagiwara, and K. Tokunaga, Association of tumor necrosis factor receptor 2 (TNFR2) polymorphism with susceptibility to systemic lupus erythematosus. Tissue Antigens, 1999. 53(6): p. 527-33. Barton, A., S. John, W.E. Ollier, A. Silman, and J. Worthington, Association between rheumatoid arthritis and polymorphism of tumor necrosis factor receptor 11, but not tumor necrosis factor receptor 1, in Caucasians. Arthritis Rheum, 2001. 44(1): p. 61-5. Romero, R., S. Durum, C.A. Dinarello, E. Oyarzun, J .C. Hobbins, and MD. Mitchell, Interleukin-I stimulates prostaglandin biosynthesis by human amnion. Prostaglandins, 1989. 37(1): p. 13-22. 81 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. Mitchell, M.D., S. Edwin, and R.J. Romero, Prostaglandin biosynthesis by human decidual cells: eflects of inflammatory mediators. Prostaglandins Leukot Essent Fatty Acids, 1990. 41(1): p. 35-8. Edwards, R.K., R.J. Ferguson, and P. Duff, The interleukin-1 beta +3953 single nucleotide polymorphism: cervical protein concentration and preterm delivery risk. Am J Reprod Immunol, 2006. 55(4): p. 259-64. Witkin, S.S., S. Gerber, and W.J. Ledger, Influence of interleukin-1 receptor antagonist gene polymorphism on disease. Clin Infect Dis, 2002. 34(2): p. 204-9. Santtila, S., K. Savinainen, and M. Hurrne, Presence of the IL-IRA allele 2 (ILIRN *2) is associated with enhanced IL—1 beta production in vitro. Scand J Immunol, 1998. 47(3): p. 195-8. Witkin, S.S., S. Vardhana, M. Yih, K. Doh, A.M. Bongiovanni, and S. Gerber, Polymorphism in intron 2 of the fetal interleukin-1 receptor antagonist genotype influences midtrimester amniotic fluid concentrations of interleukin-1 beta and interleukin-1 receptor antagonist and pregnancy outcome. Am J Obstet Gynecol, 2003. 189(5): p. 1413-7. Hao, K., X. Wang, T. Niu, X. Xu, A. Li, W. Chang, L. Wang, G. Li, N. Laird, and X. Xu, A candidate gene association study on preterm delivery: application of high-throughput genotyping technology and advanced statistical methods. Hum Mol Genet, 2004. 13(7): p. 683-9]. Maymon, E., R. Romero, P. Pacora, M.T. Gervasi, R. Gomez, S.S. Edwin, and EH. Yoon, Evidence of in vivo difirerential bioavailability of the active forms of matrix metalloproteinases 9 and 2 in parturition, spontaneous rupture of membranes, and intra-amniotic infection. Am J Obstet Gynecol, 2000. 183(4): p. 887-94. Xu, P., N. Alfaidy, and J .R. Challis, Expression of matrix metalloproteinase (MMP)-2 and MMP-9 in human placenta and fetal membranes in relation to preterm and term labor. J Clin Endocrinol Metab, 2002. 87(3): p. 1353-61. Watari, M., H. Watari, M.E. DiSanto, S. Chacko, G.P. Shi, and J .F . Strauss, 3rd, Pro-inflammatory cytokines induce expression of matrix-metabolizing enzymes in human cervical smooth muscle cells. Am J Pathol, 1999. 154(6): p. 1755-62. Li, W., E. Unlugedik, A.D. Bocking, and J .R. Challis, The role of prostaglandins in the mechanism of lipopolysaccharide-induced proMMP9 secretion from human placenta and fetal membrane cells. Biol Reprod, 2007. 76(4): p. 654-9. 82 100. 10]. 102. 103. 104. 105. 106. 107. 108. Blankenberg, S., H.J. Rupprecht, O. Poirier, C. Bicke], M. Smieja, G. Hafner, J. Meyer, F. Cambien, and L. Tiret, Plasma concentrations and genetic variation of matrix metalloproteinase 9 and prognosis of patients with cardiovascular disease. Circulation, 2003. 107(12): p. 1579-85. Ferrand, P.E., S. Parry, M. Samme], G.A. Macones, H. Kuivaniemi, R. Romero, and J .F. Strauss, 3rd, A polymorphism in the matrix metalloproteinase-9 promoter is associated with increased risk of preterm premature rupture of membranes in African Americans. Mol Hum Reprod, 2002. 8(5): p. 494-501. Aschkenazi, S., S. Straszewski, K.M. Verwer, H. Foellmer, T. Rutherford, and G. Mor, Differential regulation and function of the F as/Fas ligand system in human trophoblast cells. Biol Reprod, 2002. 66(6): p. 1853-61. Fortunato, S.J., R. Menon, and SJ. Lombardi, Support for an infection-induced apoptotic pathway in human fetal membranes. Am J Obstet Gynecol, 2001. 184(7): p. 1392-7. Kanemitsu, S., K. Ihara, A. Saifddin, T. Otsuka, T. Takeuchi, J. Nagayarna, M. Kuwano, and T. Hara, A functional polymorphism in fas (CD95/APO-1) gene promoter associated with systemic lupus erythematosus. J Rheumatol, 2002. 29(6): p. 1183-8. Fuks, A., LA. Parton, S. Polavarapu, D. Netta, S. Strassberg, I. Godi, and CD. Hsu, Polymorphism of F as and Fas ligand in preterm premature rupture of membranes in singleton pregnancies. Am J Obstet Gynecol, 2005. 193(3 Pt 2): p. 1 1 32-6. Crosdale, D.J., W.E. Ollier, W. Thomson, P.A. Dyer, J. Jensenious, R.W. Johnson, and K.V. Poulton, Mannose binding lectin (MBL) genotype distributions with relation to serum levels in UK Caucasoids. Eur J Irnmunogenet, 2000. 27(3): p. 1 1 1-7. Bodamer, G.A., G. Mitterer, W. Maurer, A. Pollak, M.W. Mueller, and W.M. Schmidt, Evidence for an association between mannose-binding lectin 2 (MBLZ) gene polymorphisms and pre-term birth. Genet Med, 2006. 8(8): p. 518-24. van de Geijn, F.E., R.J. Dolhain, W. van Rijs, S.P. Willemsen, J .M. Hazes, and OJ. de Groot, Mannose-binding lectin genotypes are associated with shorter gestational age An evolutionary advantage of low MBL production genotypes? Mo] Immunol, 2008. 45(5): p. 1514-1518. 83 109. 110. 11]. 112. 113. 114. 115. 116. 117. Goepfert, A.R., M. Vamer, K. Ward, C. Macpherson, M. Klebanoff, R.L. Goldenberg, B. Mercer, P. Meis, J. Iams, A. Moawad, J .C. Carey, K. Leveno, R. Wapner, S.N. Caritis, M. Miodovnik, Y. Sorokin, M.J. O'Sullivan, J .P. Van Dorsten, and O. Langer, Differences in inflammatory cytokine and T all-like receptor genes and bacterial vaginosis in pregnancy. Am J Obstet Gynecol, 2005. 193(4): p. 1478-85. Leitich, H., B. Bodner-Adler, M. Brunbauer, A. Kaider, C. Egarter, and P. Husslein, Bacterial vaginosis as a risk factor for preterm delivery: a meta- analysis. Am J Obstet Gynecol, 2003. 189(1): p. 139-47. Leitich, H., M. Brunbauer, B. Bodner-Adler, A. Kaider, C. Egarter, and P. Husslein, Antibiotic treatment of bacterial vaginosis in pregnancy: a meta- analysis. Am J Obstet Gynecol, 2003. 188(3): p. 752-8. Le, J. and J. Vilcek, Tumor necrosis factor and interleukin 1: cytokines with multiple overlapping biological activities. Lab Invest, 1987. 56(3): p. 234-48. So, T., A. Ito, T. Sato, Y. Mori, and S. Hirakawa, Tumor necrosis factor-alpha stimulates the biosynthesis of matrix metalloproteinases and plasminogen activator in cultured human chorionic cells. Biol Reprod, 1992. 46(5): p. 772-8. Rank, RN. and J .P. Chiao, Interleukin-1 stimulates human uterine prostaglandin production through induction of cyclooxygenase-Z expression. Am J Reprod Immunol, 2000. 43(3): p. 152-9. Arechavaleta-Velasco, F., D. Ogando, S. Parry, and F. Vadillo-Ortega, Production of matrix metalloproteinase-9 in lipopolysaccharide-stimulated human amnion occurs through an autocrine and paracrine proinflammatory cytokine-dependent system. Biol Reprod, 2002. 67(6): p. 1952-8. Moore, J .H., Computational analysis of gene-gene interactions using multifactor . dimensionality reduction. Expert Rev Mo] Diagn, 2004. 4(6): p. 795-803. Suh, Y.J., Y.J. Kim, H. Park, E.A. Park, and EH. Ha, Oxidative stress-related gene interactions with preterm delivery in Korean women. Am J Obstet Gynecol, 2008. 198(5): p. 541 e1-7. 84 HHHHHH ”711113111121lllilfljlljljlll RIES H