Abnormal First Trimester Serum Interleukin18 Levels are Associated with a PoorOutcome in Women with a Historyof Recurrent Miscarriage

INTRODUCTION

For a pregnancy to succeed the maternal immunesystem must adapt to the presence of foreign cells.Within the immune system T cells polarize to differ-ential functions characterized by their pattern ofcytokine expression. T-helper 1 (Th1) cells produceinterleukin (IL)-12, IL-18 and interferon (IFN)-cwhilst T-helper 2 (Th2) cells produce IL-4, 5 and 10and Th0 cells produce TNFa. Since Wegmann et al.1

first proposed that pregnancy was aTh2 phenomenonmany studies have been carried out to investigate therole of Th1/Th2 cytokines in pregnancy and recurrentmiscarriage. Makhseed et al.2 produced evidenceshowing that unexplained recurrent miscarriage wasassociated with increased Th1-type reactivity whilstTh2 dominance was associated with a successfulpregnancy. The same group3 also compared cytokineproduction in recurrent miscarriage patients whosepregnancies continued with those whose pregnanciesagain failed. Whilst the samples from those whose

pregnancies failed were not obtained until aftermiscarriage was confirmed, they found that abortionprone women whose pregnancies continued were moreTh2 biased than those who aborted. The latter groupalso produced significantly higher levels of Th1-typecytokines.

Decidual studies have also supported the Th1/Th2phenomenon. Lim et al.4 investigated the role ofcytokine production from maternal peri-implantationendometrial T-helper cells in the success or failure of apregnancy. Endometrial cells from non-pregnantwomen with a history of recurrent miscarriage pro-duced primarily Th1 cytokines whilst endometrialT cells from healthy women produced reduced Th1cytokines and increased Th2 cytokines.

These studies have tended to measure only singleparameters in relatively small patient numbers. As wehave obtained a cohort of 90 healthy pregnant womenand 115 pregnant women with a history of recurrentmiscarriage it was the aim of this study to comparelevels of IL-12, IL-18 IL-4 and IFN-c in these groups.

American Journal of Reproductive ImmunologyAJRI 2004; 51: 156–159Copyright � Blackwell Munksgaard, 2004

ISSN 1046-7408

Wilson R, Moor J, Jenkins C, Miller H, Walker JJ, McLean MA,Norman J, McInnes IB. Abnormal first trimester serum interleukin 18levels are associated with a poor outcome in women with a history ofrecurrent miscarriage. AJRI 2004; 51:156–159 � BlackwellMunksgaard, 2004

PROBLEM: How the maternal immune system adapts to tolerate thefetus is not fully understood, but a successful pregnancy is associatedwith the production of Th2-type cytokines and miscarriage isassociated with the production of Th1-type cytokines.METHOD OF STUDY: Levels of interferon (IFN)-c, interleukin (IL)-4, IL-12 and IL-18 were measured in serum from 205 pregnant womenof whom 115 pregnant women had a history of recurrent miscarriage.RESULTS: Compared with healthy pregnant women those whomiscarried had increased serum levels of the Th1-associated cytokinesIFN-c, IL-12 and IL-18.CONCLUSIONS: Increased levels of IL-18 appeared to be critical inearly pregnancy and were able to discriminate between pregnanciesthat continued and those that end in miscarriage.

Rhoda Wilson1, Judith Moor3, CarolJenkins1 Helen Miller1, James J.Walker3, Marjorie A. McLean4,J. Norman2, I. B. McInnes1

Departments of 1Medicine, and 2Obstetrics andGynaecology, Glasgow Royal Infirmary, Glasgow, UK;3Department of Obstetrics and Gynaecology, St JamesHospital, Leeds, UK; 4Early Pregnancy Unit, HairmyresHospital, East Kilbride, UK

Key words: Cytokines, recurrent miscarriage, Th1-typeimmune response

Address reprint requests to Dr R Wilson, Department ofMedicine, Glasgow Royal infirmary, 10 Alexandra Parade,Glasgow G31 2ER, UK.E-mail: gcl025@clinmed.gla.ac.uk

Submitted May 19, 2003;revised July 29, 2003;accepted August 1, 2003.

� BLACKWELL MUNKSGAARD, 2004

MATERIALS AND METHODS

PatientsThe study comprised of three groups of women:

Group 1 comprised 25 healthy non-pregnant womenof mean age 30.5 � 5.3 years.

Group 2 comprised 92 pregnant women of mean age26 � 6.9 years. Of these, one had a miscarriage in thesecond trimester and one had a termination of preg-nancy. The remaining 90 women (13.8 � 2.6 weeksgestation) had no history of recurrent miscarriage, andall subsequently delivered a healthy baby at term. Priorto blood sampling, all patients had been scanned andthe fetal heart beat.

Group 3 comprised 115 first trimester pregnantwomen of mean age 31 � 5.6 years. The women had amean gestation of 6.4 � 2.4 weeks, and had a history ofrecurrent miscarriage (mean of 4.9 � 5.1 previousmiscarriages). All these women were known to benegative for antiphospholipid antibodies, had no kary-otyping abnormalities nor were there any endocrinereasons for their miscarriage. Of these women, 74 hadpregnancies that went successfully to term (Group 3S)and 41 miscarried again later in the first trimester(Group 3M), although all were pregnant at the time ofsampling. Owing to the gestational age of these womenthey were not scanned, as it is not normal clinicalpractice to scan before the fetal heart could be seen. Nowomen were on any medication. Blood was obtainedfrom all patients and the serum stored at )70�C tillrequired. Informed consent was obtained from allpatients. The ethics committee approved the study.

MethodsFor each sample volume available, the followingcytokines were measured in serum by ELISA(Biosource Ltd): IFN-c, IL-4, IL-12 and IL-18. Results

are expressed as mean and standard deviation andwere analysed for statistical significance using Mann–Whitney test. All measurements were made in dupli-cate. While patient and control samples were notnecessarily analysed on the same day, quality controlsamples were always run in each assay. The intraand inter coefficients of variation for IL-4 were 3.6 and6.9%, for IL-12 2.6 and 5.7%, respectively; intra andinter assay precision for IL-4 were 2.4 and 6.9%, forIL-18 4.5 and 7.0%, and for IFN-c 3.1 and 7.3%,respectively.

RESULTS

Cytokine DataThe results are presented in Fig. 1. Normal pregnancywas associated with significantly increased levels ofIL-4. Compared with this group, patients who werepregnant at the time of sampling but who latermiscarried had significantly reduced levels of IL-4,and significantly higher levels of IL-12, IL-18 andIFN-c. Those with a history of recurrent miscarriage,but whose pregnancies on this occasion went to term,had lower IL-18 and higher IFN-c levels than thosefound in normal pregnancy. Significant differences inlevels of IL-12, IL-18 and IFN-c were found betweenpatients in Groups 3S and 3M. As both were of similargestations, these differences may be associated withmiscarriage. A combination of elevated IFN-c, IL-18and IL-12 was associated with miscarriage. WhenIL-18 levels were compared between Groups 3S and3M (Fig. 2), the higher levels of IL-18 were foundexclusively in those who miscarried. The lower levels,which were found in the patients with on-goingpregnancies did not differ significantly form levels inGroup 2, despite the latter being of a significantly later

Fig. 1. This figure shows cytokine levels in

the four patient groups – Group 1: non-

pregnant; Group 2: pregnant; Group 3S:

pregnant with a history of recurrent mis-

carriage and had a successful pregnancy;

and Group 3M: pregnant with a history of

recurrent miscarriage and who miscarried

again. Results are given as medians and

ranges.

RECURRENT MISCARRIAGE / 157

AMERICAN JOURNAL OF REPRODUCTIVE IMMUNOLOGY VOL. 51, 2004

gestation. Using a cut off value of 200 pg/mL, theIL-18 concentration, which was obtained when allwomen were still pregnant, had a positive predictivevalue of 93% and a negative predictive value of 100%for predicting miscarriage.

DISCUSSION

Although all patients, in this study, were pregnant atthe time of sampling, cytokine profiles differed betweenthose whose pregnancies continued and those wholater miscarried. The results obtained here confirm ourprevious findings made in smaller patient numbers5

that miscarriage was associated with significantlyhigher levels of the Th1 cytokines IL-12, IL-18 andIFN-c. Early first trimester levels of IL-18 appeared todiscriminate between those pregnancies that continuedand those that did not. Group 3S patients had partic-ularly low levels of IL-18 and this may be anadaptation to compensate for the immunologicalabnormalities that exist. A potential drawback to ourstudy is that women in Group 2 were sampled atbooking antenatal clinic (i.e. around 10–12 weeksgestation), whereas women in Groups 3S and 3Mwere sampled once pregnancy was confirmed by apregnancy test (i.e. around 6 weeks gestation). Differ-ences in cytokine levels between Groups 2 and 3 maytherefore reflect the different gestations, rather than thehistory of recurrent miscarriage. However, this con-founding variable would not affect the comparisonbetween Groups 3S and 3M, as these women weresampled at the same gestations.

Peripheral blood samples were used in this study.The results confirm those obtained from decidual cellsat the maternal fetal interface where the production ofleukaemia inhibitory factor, IL-4 and IL-10 wasdefective in decidual T cells obtained from womensuffering unexplained recurrent miscarriage comparedwith those who had a termination6. However, these

defects were not found in the peripheral blood leadingthe authors to conclude this was a microenviromen-tally orientated alteration and not an inherent featureof T cells. Why they should find no significant alter-ation in peripheral cytokine levels is not clear. The onlydifference between the two studies was that in thepresent study all patients were still pregnant at the timeof sampling, while in the study of Piccinni et al.6 thepatients were suffering a miscarriage.

In a recent study, Bates et al.7 found significantlylower levels IFN-c and significantly higher levels ofIL-10 and IL-4 in pregnant women with a history ofmiscarriage compared with healthy pregnant women.These results differ from our own findings and thismay be because they used phytohaemagglutinin(PHA)-stimulated peripheral blood lymphocytes(PBL). The serum cytokine levels used in the presentstudy reflect the circulating levels in serum at a singlepoint in time. Bates et al. have chosen to measure howPBL responded to a stimulus and it is possible that thecapacity of the PBL to respond to PHA stimulation isaltered.

How the presence of Th1-type cytokines appears tolead to miscarriage is not yet fully understood, but it ispossible that the increased production of IFN-c exertsa localized effect on the cells necessary for implanta-tion, or there may be release of IL-17, a potentfibroblast-activating factor. Alternatively, Th1 activitymay be promoted by cell contact via CD40 LIGANDand LFA1. The balance of Th1 and Th2 cells appearsto influence the outcome of a pregnancy. Therefore, itis important to understand the mechanisms leading tothe preferential induction of each of the T-cell subsets.IL-12 drives the differentiation of precurser T cells toTh1 cells. The mechanism leading to the preferentialinduction of Th1/Th2 subsets is not yet clear, althoughcytokines are thought to be one of the main factors.Whilst IL-12 drives the Th1 response IL-4 drives theTh2 response. Recently, it has been shown that IL-18in conjunction with IL-4 will induce Th0 cells tobecome Th2 cells. In the present study, IL-18 appearedto distinguish between successful and unsuccessfulpregnancies. This may be due to the elevated levels ofIL-12 found in those who miscarried, in conjunctionwith IL-18 causing a Th1 response to dominate.8

Therefore, IL-18 is occupying a crucial role in promo-ting either Th1 or Th2 responses.

Hormones have also been implicated in determiningthe role of Th0 cells. Picinni et al.9 found in vitrothat progesterone could favour the development ofT-helper cells producing Th2-type cytokines, andinduce IL-4 production and CD30 expression. Thesefindings appear to contradict those of Lim et al.4,where cytokine production from endometrial T cellsappeared not to be under any hormonal influence.

Fig. 2. This figure shows IL-18 levels in healthy pregnant women

(Group 2), women with a history of recurrent miscarriage who had a

successful pregnancy (Group 3S) and those who miscarried again

(Group 3M).

158 / WILSON ET AL.

� BLACKWELL MUNKSGAARD, 2004

Whilst this and other studies have favoured the Th1/Th2 paradigm, it may be that this approach is toosimplistic. Many of the new cytokines do not fit intothe classical Th1/Th2 dichotomy. Studies by Chaouatet al.10 implied that there were many cytokinesinvolved in the maintenance and establishment of apregnancy. These cytokines all have a precise locationwhere they act and, as well as being involved inmaternal tolerance, are also involved in immuneregulation, adhesion and vascularization. Viveset al.11 examined cytokine expression in human deci-dua and trophoblast tissue from healthy women andthose suffering first trimester abortion. Whilst thesegroups are not of comparable gestations, some of thefindings favoured a Th1/Th2 model whilst others wereagainst, for example, increased IFN-c expression indecidual and trophoblast samples from healthy womencompared with those from miscarriage patients.

In conclusion, miscarriage appears to be primarilyassociated with the production of Th1 cytokines. Wehypothesize that in pregnant women who miscarrythe increased levels of IL-12 and IL-18 drive Th0cells to become Th1 cells, thus raising IFN-cproduction and depressing IL-4 production. In thosepatients who have a successful pregnancy (Group 3S)IL-18 production is depressed, IL-12 levels do notrise, thus allowing TH0 cells to develop into Th2cells enabling the pregnancy to progress successfully.These changes occur early in pregnancy. Many of theparameters that were measured showed wide varia-tions suggesting that the abnormalities may differfrom patient to patient. However, some of theresponses observed would suggest that their immunesystems were trying to overcompensate for theabnormalities present.

AcknowledgementsWe are grateful to Tenovus-Scotland for funding thisresearch and to the early pregnancy unit at theSimpson’s Maternity Hospital Edinburgh for provi-ding us with many of our samples.

REFERENCES

1. Wegmann TG, Lin H, Guilbert L, Mossman TR: Bidi-rectional cytokine interactions in maternal foetal rela-tionships: is successful pregnancy a Th2 phenomenon?Immunol Today 1993; 14:353–356.

2. Makhseed M, Raghupathy R, Azizich F et al.: Mitogen-induced cytokine responses of maternal peripheral bloodlymphocytes indicate a differential Th1-type bias I nor-mal pregnancy and pregnancy failure. Am J ReprodImmunol 1999; 42:273–281.

3. Makhseed M, Raghupathy R, Azizich F et al.: Th1 andTh2 cytokine profiles in recurrent aborters with suc-cessful pregnancies and with subsequent abortions. HumReprod 2001; 16:2219–2226.

4. Lim KJL, Li TC, Weetman AP, Cooke ID, OdukoyaOA, Ajjan RA et al.: The role of T helper cytokines inhuman reproduction. Fertil Steril 2000; 73:136–142.

5. Jenkins C, Roberts J, Wilson R et al.: Evidence of a Th1type response being associated with recurrent miscar-riage. Fertil Steril 2000; 73:1206–1208.

6. Piccinni MP, Beloni L, Livi C et al.: Defective produc-tion of both leukemia inhibitory factor and type 2T-helper cytokines by decidual T cells in unexplainedrecurrent abortions. Nat Med 1998; 4:1020–1024.

7. Bates MD, Queby S, Takakuwa K et al.: Aberrantcytokine production by peripheral blood mononuclearcells in recurrent pregnancy loss? Hum Reprod 2002;17:2439–2444.

8. Xu D, Traijkovic V, Hunter D et al.: IL-18 induces thedifferentiation of TH1 or Th2 cells depending on cyto-kine milieu and genetic background. Eur J Immunol2000; 30:3147–3156.

9. Piccinni MP, Guidizi MG, Biagiotti R et al.: Progester-one favours the development of human T helper cellsproducing Th2 type cytokines and promotes both IL-4production and membrane CD30 expression in Th1 cellclones. J Immunol 1995; 155:129–133.

10. Chaouat G, Zourbas S, Ostojic S et al.: A brief review ofsome of the cytokines expressed at the materno-fetalinterface which might challenge the classical Th1/Th2dichotomy. J Reprod Immunol 2002; 53:241–256.

11. Vives A, Balash J, Yague J et al.: Type 1 and type-2cytokines in human decidual tissue and trophoblasts fronormal and abnormal pregnancies detected by reversetranscriptase polymerase reaction (RT-PCR). Am J Re-prod Immunol 1999; 42:361–368.

RECURRENT MISCARRIAGE / 159

AMERICAN JOURNAL OF REPRODUCTIVE IMMUNOLOGY VOL. 51, 2004