Field of the Invention

The invention relates to biomarkers of early miscarriage. The biomarkers are useful for identifying women with an increased likelihood of early miscarriage.

Background to the Invention

Placental-related disorders of pregnancy are almost unique to the human species. These disorders, which affect around a third of human pregnancies, primarily include miscarriage and pre-eclampsia ¹ . In other mammalian species, the incidence of both disorders is extremely low. Although epidemiological data on animals living in the wild, such as monkeys, are limited, laboratory rodents are known to have post-implantation pregnancy loss rates of less than 10% 1. By contrast in humans, it has been estimated that approximately 50 to 70 percent of spontaneous conceptions are lost prior to completion of the first trimester, most of them during the first month after the last menstrual period ² . Unlike other species, human placentation is also associated with a 30% incidence of bleeding during the first trimester. This is due to the complex interaction between placental tissue and maternal uterine tissue immediately after implantation. Human placentation is characterized by the highly invasive nature of the conceptus which embeds itself completely within the maternal uterine endometrium and superficial myometrium and by a remodelling of the tip of the maternal uterine circulation ¹ . In normal pregnancies, the earliest stages of development take place in a low oxygen (O ₂ ) environment ³ ' ⁴ . This physiological hypoxia of the early gestational sac protects the developing fetus against the deleterious and teratogenic effects of oxygen free radicals. A stable O ₂ . gradient between the maternal uterine decidua and the feto-placental tissue is also an important factor in trophoblast differentiation and migration and normal villous development and angiogenesis ⁵ ' ⁶ . At the end of the first trimester, there is a burst of oxidative stress in the periphery of the early placenta ³ . This leads to higher local O ₂ concentrations at a stage of pregnancy when the trophoblast possesses low concentrations and activities of the main antioxidant enzymes and triggers trophoblastic apoptosis and necrosis and progressive villous degeneration essential to the formation of the fetal membranes ⁷ . The oxidative stress and rise in oxygenation may also stimulate the synthesis of various trophoblastic proteins such as human chorionic gonadotropin (hCG) and oestrogens. Maternal serum concentrations of hCG peak towards the end of the first trimester, and oxidising conditions promote assembly of the sub-units in vitro ⁸ . Our recent data showing a direct relationship between intrauterine PaO ₂ in vivo and inhibin A and sFLT-1 concentrations in early pregnancy confirm our hypothesis that specific placental proteins can be regulated by intrauterine O ₂ tension ⁹ .

In miscarriage, the development of the placento-decidual interface is severely impaired leading to early and widespread onset of maternal blood flow and major O ₂ -induced degeneration ⁷ ' ¹⁰ . The excessive entry of maternal blood inside the placenta in the early stage of most miscarriages is unrelated to conceptus karyotype ¹¹ . In more than two-third of the cases of missed miscarriage, there is anatomical evidence of defective placentation with reduced cytotrophoblast invasion of the endometrium, reduced transformation and incomplete plugging of the spiral arteries. We have also previously shown that placental bleeding in the first trimester is associated with a 10% incidence of subsequent obstetric complications including complete miscarriage and premature delivery ' ' .

Angiogenesis is characterised by increased vascular permeability and endothelial cell proliferation and migration. It is regulated by various pro- and anti-angiogenic factors, angiopoietins and matrix metalloproteinases. Anti-angiogenic and pro-angiogenic factors are known to play an important role in the pathophysiology of pre-eclampsia ¹² . Studies of maternal serum levels of these factors have shown that soluble endoglin (sEng) and soluble fms-like tyrosine kinase 1 (sFlt-1) are elevated in women presenting with PE whereas vascular endothelial growth factor (VEGF) and placental growth factor (P1GF) are decreased. Some of these changes can be detected several weeks before the appearance of clinical symptoms of PE ¹³ ' ¹⁴ .

Angiogenic factors and inflammatory have not been previously evaluated in early pregnancy. The inventors aim to determine whether threatened miscarriage which is associated with a focal oxidative stress in the definitive placenta is also associated with changes in angiogenic factors and to investigate if these levels can predict a subsequent full miscarriage or other major perinatal complications such as preterm rupture of the placental membranes and preterm labour Summary of the Invention

According to the invention, there is provided a method for identifying whether a pregnant subject has an increased risk of miscarriage comprising measuring the concentration of at least one of soluble vascular endothelial growth factor receptor (sFlt-1), Placental growth factor (PIGF) and C reactive proteins in a biological sample obtained from the subject.

The inventors have identified changes in sFLT-1, PIGF or c reactive protein concentration which are indicative of an increased risk of miscarriage and perinatal complications such as preterm rupture of the placental membranes and preterm labour in women presenting with bleeding in the first trimester.

The subject is preferable a mammal, especially a primate. In one embodiment, the animal is a human. The sample may be a biological fluid such as blood, saliva or urine obtainable from the subject, but is preferably a serum or plasma sample.

The concentration of sFLT-1, PIGF or C reactive protein in the sample may be measured by any known means. Preferable the concentration is measured by an Enzyme immunoassay (ELISA) or an automated immunoassay on a robotic platform.

The term "change in concentration" preferably means that the concentration in the sample is increased or reduced, particularly reduced in comparison with a comparable sample obtained from the same subject earlier in pregnancy or in comparison with average concentrations in comparable samples obtained from other subjects at the same stage of gestation. In the method of the invention, a change, preferably a reduction in sFLT-1, PIGF or C reactive protein is identified as being indicative of an increased likelihood of miscarriage. It is preferred that the method includes the step of assaying the concentration of at least two, more preferably all of sFLT-1, PIGF and C reactive protein and that a reduction in both is particularly indicative of an increased likelihood of miscarriage.

Comparable samples are the same type of sample as each other. For example, comparable samples are taken from animals of the same species. Equally, saliva samples should be compared with other saliva samples, urine samples with other urine samples etc. To be considered statistically relevant, the reduction in concentration is preferably a reduction of at least 10%, more preferably at least 15%, more preferably at least 20%, even more preferably at least 30%.

The method of diagnosis may comprise comparing the sample with a sample from another animal or human at the same gestational age. More preferably, though, it comprises comparing the sample with samples previously obtained from the subject. Samples may be obtained daily, every two or three days or weekly, for example. Where a subject is considered to be at risk of miscarriage, due to, for example, miscarriage in a previous pregnancy, increased blood pressure etc., samples may be obtained every two or three days or weekly and compared with normal range of gestation matched controls. It is preferred that samples are obtained early in pregnancy, for example before 10 weeks gestation, more preferably before 8 weeks gestation, even more preferably at between 5 and 8 weeks gestation.

The method of identifying subjects at risk may be used in combination with other methods that identifying subjects at risk. Such methods may include checking other bio markers especially those which have previously been linked with miscarriage, such as inhibin A, activin A, progesterone, oestradiol, and hCG or indicators of oxidative stress such as heat shock proteins and markers of inflammation such as TNF alpha, TNF alpha receptors, Interferon gamma using immunoassays (ELISAs or automated platforms) in biological fluids including blood, saliva and urine. The method of the invention may also be used to identify subjects that may benefit from treatment with an appropriate medicament, such as progesterone.

Also provided are C reactive protein, sFLT-1 and/or P1GF for use as a biomarker of early miscarriage.

Detailed description of the Invention

The invention will now be described in detail, by way of example only, with reference to the following drawings. Figure 1 : Circulating levels of serum human chorionic gonadotrophs (hCG) in the different groups of women. Group 1 (non pregnant women, n=18), group 2 (normal pregnant women, n=21), group 3 (threatened miscarriage patients with a live birth outcome, n=26), group 4 (threatened miscarriage patients with a subsequent miscarriage, n=19) and group 5 (confirmed miscarriage, n=21). General linear analysis of variance was carried out to study the statistical significance between the groups with posthoc tests. P<0.001=***, P<0.01=**, P<0.05=*

Figure 2: Circulating levels of serum soluble VEGF receptor 1 (sFLT-1) in the different groups of women. Group 1 (non pregnant women, n=18), group 2 (normal pregnant women, n=21), group 3 (threatened miscarriage patients with a live birth outcome, n=26), group 4 (threatened miscarriage patients with a subsequent miscarriage, n=19) and group 5 (confirmed miscarriage, n=21). General linear analysis of variance was carried out to study the statistical significance between the groups with posthoc tests. P<0.001=***, P<0.01=**, P<0.05=*

Figure 3: Circulating levels of serum placental growth factor (PIGF) in the different groups of women. Group 1 (non pregnant women, n=18), group 2 (normal pregnant women, n=21), group 3 (threatened miscarriage patients with a live birth outcome, n=26), group 4 (threatened miscarriage patients with a subsequent miscarriage, n=19) and group 5 (confirmed miscarriage, n=21). General linear analysis of variance was carried out to study the statistical significance between the groups with posthoc tests. P<0.001=***, P<0.01=**, P<0.05=*

Figure 4: Circulating levels of serum soluble Endoglin (sEndog) in the different groups of women. Group 1 (non pregnant women, n=18), group 2 (normal pregnant women, n=21), group 3 (threatened miscarriage patients with a live birth outcome, n=26), group 4 (threatened miscarriage patients with a subsequent miscarriage, n=19) and group 5 (confirmed miscarriage, n=21). General linear analysis of variance was carried out to study the statistical significance between the groups with posthoc tests. P<0.001=***, P<0.01=**, P<0.05=*

Figure 5: Circulating levels of serum heat shock protein 70 (Hsp 70) in the different groups of women. Group 1 (non pregnant women, n=18), group 2 (normal pregnant women, n=21), group 3 (threatened miscarriage patients with a live birth outcome, n=26), group 4 (threatened miscarriage patients with a subsequent miscarriage, n=19) and group 5 (confirmed miscarriage, n=21). General linear analysis of variance was carried out to study the statistical significance between the groups with posthoc tests. P<0.001=***, P<0.01=**, P<0.05=*

Figure 6: Levels of c-reactive protein in the different groups of women.

Figures 7a to d show the results of example 3, which further analysed the use of soluble FLT- 1 and PIGF as biomarkers of early miscarriage.

Methodology

Study groups: Blood samples were collected prospectively from normal pregnant controls (n= 21), women presenting with threatened miscarriage (TM) (n= 45) and women presenting with a confirmed-miscarriage (n= 21). In addition, non pregnant controls (20-35 years) were recruited from the same population during the same period of time. Demographic data of the different groups are summarized in table 1. Detailed pregnancy outcome was obtained for the TM group and normal antenatal controls. Multiple pregnancies and women who were not contactable to follow their outcome were excluded from the investigation.

All pregnant women were recruited from the University College London hospital (UCLH) and the non pregnant control samples were obtained from UCL staff and students. This study was approved by UCLH ethics committee. All women gave informed written consent to participate in the study. Demographic data including ethnicity, age and BMI from all women were collected. Age and BMI were similar in all groups. All blood samples were taken between 6-14 weeks in gestation. Serum samples were frozen at -80°C until analysis.

Assays: All samples were assayed for hCG, Soluble FLT-1 , PIGF, sEndoglin and heat shock protein 70 (HSP70). Maternal Serum (MS) hCG was measured using Roche Cobas automated system. Detection limit of the assay was 2mIU/ml.

Quantakine ELISA kits for PIGF and soluble FLT-1 and reagents from a duoset for Soluble Endoglin was used from R&D systems (UK) to measure the respective proteins in serum samples according to the manufacturer's protocol. All samples were assayed in duplicates and the intra and inter assay variations were <12% for all assays. The minimum detection limit for the respective assays was PIGF: 3.9pg/ml, sEng: 62.5pg/ml, sFlt-1 31.3pg/ml. Heat shock protein 70 was measured in serum using a highly sensitive assay kit from assays design (Cambridge bioscience, UK). The detection limit of HSP 70 was 0.2ng/ml.

Statistical Analysis: Data were not normally distributed. We carried out non parametric general linear model analysis of variance weighted for gestation at sampling using Dunnett's T3 posthoc tests were carried out to investigate differences between groups. Partial correlation was carried out to study the relationship between the parameters controlling for gestation. Data are presented as mean and standard deviation (SD) and P values were considered significant at P<0.05. Results

The TM group included 26 women who subsequently had a live birth and 19 who subsequently had a miscarriage.

The mean HCG concentration was significantly higher (P<0.001) in the pregnant group compared to the non pregnant controls (<detection) as expected (Figure 1). Miscarriage patients had significantly lower MS hCG levels compared to control pregnancy (P<0.001) and TM patients with a live birth (P<0.05).

The mean MS sFlt-1 level was significantly (P<0.001) higher (~20 fold) in the normal pregnancy compared to the non-pregnant group whereas mean MS sFLT-1 levels were significantly (P<0.001) lower (~6 fold) in the TM subgroup with a subsequent miscarriage and in the confirmed miscarriage group (~5 fold, P<0.001) compared to controls. TM subgroup with a subsequent live birth (figure 2) also had significantly (P<0.001) higher (~4 fold) mean MS sFLT-1 compared to TM subgroup with a subsequent miscarriage, and to the miscarriage group (-2.5 fold, P<0.05).

The mean MS PIGF levels were significantly (P<0.001) higher in pregnancy than in nonpregnant controls and significantly (P<0.001) higher in TM patients with a subsequent live birth compared to TM patients who subsequently miscarried (<2 fold) Control pregnant women had significantly (P<0.001) higher levels of MS PIGF compared to TM patients with subsequent miscarriage (>4 fold) or miscarriage (~5 fold) (figure 3).

There was no significant difference in the mean soluble Endoglin levels between the different groups and subgroups (figure 4). Mean MS HSP70 levels were also not significantly different between all groups of pregnant women and non-pregnant controls (figure 5).

Bivariate correlation analysis showed a positive correlation between gestational age and MS sFLT-1 (r=0.645, P<0.001), MS PIGF (r=0.6, P<0.001) and MS hCG (r=0.27, P=0.02) in the first trimester. Maternal age was found to be inversely proportionate to MS PIGF (r=-0.231, P<0.05) and MS hSP70 ((r=-324, P<0.02). Partial correlation controlling for gestation, BMI and age shows that MS sFLT- 1 and MS hCG (r=0.434, P=0.001) and MS PIGF (r=0.27, P=0.05) are positively correlated.

Discussion

This data indicate for the first time differences in the MS levels of sFLT-1 and PIGF in confirmed miscarriage and TM complicated by subsequent miscarriage compared to normal pregnant controls and TM with subsequent live birth. Our findings suggest that these molecules could be used as predictive markers of miscarriage in early pregnancy and of subsequent perinatal complications. The development of algorithms to predict miscarriage and premature labour could help triaging for high-risk couples with a history of recurrent miscarriages and for women presenting with bleeding in early pregnancy.

Independent of the cause of an early pregnancy failure, the premature and excessive entry of maternal blood inside the placenta has two effects on the villous tissue. A direct mechanical effect with most of the villi becoming progressively enmeshed inside large intervillous blood thrombi and an indirect widespread O ₂ -mediated effect with oxidative damage leading to major apoptosis and necrosis of the villous trophoblast ⁷ ' ¹⁰ . Overall, the consequences are placental degeneration with complete loss of syncytiotrophoblast function and detachment of the placenta from the uterine wall. This mechanism is common to all miscarriages, the time at which it occurs in the first trimester depending on the aetiology ¹ . Oxidative stress is also involved in the development of other placental-related disorders such as pre-eclampsia and premature rupture of the placental membranes.

We have recently shown that O ₂ concentration in the placental bed blood is inversely related to inhibin A and sFlt-1 in normal early pregnancy ⁹ . We have also previously shown that MS inhibin A levels are lower in pregnant women presenting with a missed-miscarriage, recurrent miscarriage and TM complicated by a subsequent miscarriage ¹⁵ ' ¹⁶ . These data suggest that an impaired placentation may be associated with placental metabolic changes before the appearance of clinical symptoms and these changes are modulated by an abnormal increase in O ₂ concentration inside the placenta soon after implantation. In missed miscarriages we also found that MS βhCG levels are significantly higher in cases presenting with diffuse intervillous blood flow on ultrasound examination ¹¹ . No difference in immunostaining for βhCG was found between placental tissues from normal pregnancies and missed miscarriages, but significantly higher villous βhCG content was found on Western blotting in miscarriage when the fetus had died recently ¹¹ . The synthesis of hCG molecules is directly linked with trophoblast differentiation suggesting that hCG changes immediately after fetal demise may reflect a temporary attempt by the trophoblast to stabilise after the initial oxidative insult.

HSP70 is a sensitive marker of oxidative stress in tissues and we have previously found a peak in the expression of HSP70 in the placental tissue from normal pregnancies at 8-10 when the membranes start to form3. We also found that the immunoreactivity for HSP70 is greater in samples from peripheral than from central regions of normal placentas immunostaining and in the whole placental tissue in missed-miscarriage ³ . Early flow is restricted to the peripheral regions of most normal placentas whereas in missed miscarriages it was most common in central regions or throughout the placenta ³ ' ¹¹ . In the present study, we found no difference in the MS HSP70 between the non-pregnant, the normal pregnancy and miscarriage group and subgroups suggesting that peripheral circulating HSP70 does not reflect focal changes in intrauterine tissues. HSP70 is therefore unlikely to be useful in the screening of oxidative-strees related placental-disorders of pregnancy.

Soluble fms-like tyrosine kinase 1 (sFLT-1) and Endoglin (sEng) are members of the TGF β family. Soluble Endoglin is a placenta derived soluble TGF-beta co-receptor. sFLT-1 is soluble vascular endothelial growth factor (VEGF) receptor 1 (sFLT-1) and antagonises the effect of VEGF. Soluble Endoglin inhibits the formation of blood vessels and induces vascular permeability and hypertensionl7. Angiogenic growth factors VEGF- A and placental growth factor (P1GF) have been investigated extensively in normal and abnormal placental vascular development ^18-20 . Recent studies suggest that its major phenotypes, hypertension and proteinuria, may be due to an excess of circulating anti-angiogenic growth factors, most notably sFltl and sEng. sFltl is an endogenous protein that is produced by the placenta. sFltl is able to bind to the angiogenic growth factors vascular endothelial growth factor and placental growth factor, thereby neutralizing their functions. In this study, we found that the levels of sFLT-1 rose by 20 fold in pregnancy state confirming that the feto-placental unit is a major source for this molecule in early pregnancy. P1GF also increased by around 5 fold in early pregnancy. By contrast, no difference was found levels in the concentrations of sEng between samples from non-pregnant women, from early pregnancy and from miscarriage suggesting that the source of this protein is extra placental.

High MS levels of sFltl and low levels of free VEGF and P1GF have been previously reported before and during to clinical manifestation of pre eclampsia (Reference). More recently, MS levels of sEng were also shown to be elevated in pre eclamptic women and levels of sEng correlated strongly with disease severity ¹³ ' ¹⁴ . There is also evidence that the circulating cytokine levels and the cytokine profile in the decidua are different in women who experience recurrent miscarriages ^21-24 but the exact interaction of each of these cytokines with the invading trophoblast is not well defined. Decreased expression of angiogenic factor genes for basic FGF, intergrin aV and VEGF has also been observed in placental samples from recurrent miscarriage patients25. There are no reliable assays available to measure 'total' VEGF in circulation and circulating levels of VEGF are almost below detection in early pregnancy9. In the present study we found significantly (P<0.001) lower levels of sFLT-1 and P1GF in confirmed miscarriages than in controls. Whereas the lower of P1GF can be related to a decreased syncytiotrophoblast synthesis, lower levels of sFLT-1 may be compensatory as the placenta may be producing more VEGF and the soluble VEGF-R1 that is unbound to VEGF in circulation could be reduced. On the other hand, both VEGF and the receptor production may be lower in patients who subsequently have a miscarriage thus reflecting lower levels of sFLT-1 in maternal circulation in these cases.

Some diseases such as maternal diabetes can generate O ₂ free radicals (OFR) in larger quantities than the early placental limited antioxidant defences can cope with, leading to DNA damage, oxidation of proteins and lipid and resulting in secondary trophoblast dysfunction ²⁶ .Our previous study in early pregnancy showed that sFLT-1 is inversely proportional to the O ₂ concentration in the placental bed blood between 6-12 weeks gestation9. In TM there is a focal bleed mainly in the membranes in the periphery of the developing placenta. This common pregnancy complication occurs mainly at the time of the formation of the membranes at 8-12 weeks and can lead to a complete miscarriage if the hematoma extends into the definitive placenta ²⁷ . Earlier bleeding i.e before 8 weeks is in general not associated with subsequent pregnancy complications whereas the formation of an intrauterine hematoma at the end of the first trimester increases the risk of pre eclampsia ²⁸ . Overall women presenting with threatened miscarriage are also more likely to deliver prematurely and/or present with preterm pre labor rupture of membranes ²⁹ . The presence of a haematoma may also be associated with a chronic inflammatory reaction in the decidua resulting in persistent myometrial activity and expulsion of the pregnancy or progressive cellular dysfunction and/or damage to cellular layers of the membranes leading to pre-term rupture and delivery ³⁰ . We previously found that serum hCG concentrations are significantly higher in cases of threatened miscarriage compared with normal controls but lower in those who subsequently miscarried ¹⁶ . The lower sFLT-1 found in the serum of TM that will subsequently miscarry suggest a compensatory mechanism where higher concentrations of pro-angiogenic factors such as VEGF are available to promote angiogenesis in a failing pregnancy. Despite the small number of miscarriage outcome in the TM patients the results are very significant indicative of these molecules as sensitive predictors of a subsequent miscarriage. Further larger studies are required to confirm this results and the predictive efficiency of these molecules of miscarriage in early pregnancy.

EXAMPLE 2

Introduction

C-reactive protein (CRP) is an acute-phase protein secreted by the liver in response to inflammation. It is not specific for infection but is a marker used for the diagnosis of many inflammatory, infective and malignant conditions (Wiwanitkit, 2005). High-sensitivity C- reactive protein (hsCRP) has been gaining recognition as an independent risk factor for cardiovascular disease. In particular, increased level of hsCRP, have been associated with increased stroke risk as well as an increased rate of atherosclerosis progression in the carotid vessels (Ridker and Silvertown, 2008). Evidence supporting a link between periodontal disease and cardiovascular disease is accumulating in the literature. Recent studies have also shown that pre-eclamptic women present a high prevalence of periodontitis, suggesting that active periodontal disease may play a role in the pathogenesis of pre-eclampsia (Vergnes, 2008).

Methodology

Study groups: Blood samples were collected prospectively from normal pregnant controls (n= 21), women presenting with threatened miscarriage (TM) (n= 45) and women presenting with a missed-miscarriage (n= 21). In addition, non pregnant controls (20-35 years) were recruited from the same population during the same period of time. Demographic data of the different groups are summarized in table 1. Detailed pregnancy outcome was obtained for the TM group and normal antenatal controls. Multiple pregnancies and women who were not contactable to follow their outcome were excluded from the investigation.

All pregnant women were recruited from the University College London hospital (UCLH) and the non pregnant controls samples were obtained from UCL staff and students. This study was approved by UCLH ethics committee. All women gave informed written consent to participate in the study. Demographic data including ethnicity, age and BMI from all women were collected. Age and BMI were similar in all groups. All blood samples were taken between 6-14 weeks in gestation.

hsCRP were measured in serum using a highly sensitive assay kit from assays design (Cambridge bioscience, UK). The detection limit of hsCRP was 600mg/ml respectively.

Statistical Analysis: Data were not normally distributed. We carried out non parametric general linear model analysis of variance weighted for gestation at sampling using Dunnett's T3 posthoc tests were carried out to investigate differences between groups. Partial correlation was carried out to study the relationship between the parameters controlling for gestation. Data are presented as mean and standard deviation (SD) and P values were considered significant at P<0.05.

Results The TM group included 26 women who subsequently had a live birth and 19 who subsequently had a miscarriage.

The MShSCRP levels were significantly (P<0.05) lower in TM subgroup with a subsequent miscarriage compared to normal control pregnant women (~4 fold) and TM patients with a subsequent live birth (~3 fold). Significantly (P<0.05) lower mean level of MShsCRP were found in the missed-miscarriage group compared to pregnant normal control women (~4 fold) (figure 6).

EXAMPLE 3

The assays of example 1 were repeated for soluble FLT-1 and PIGF, with greater subject numbers. The results are shown in figures 7a to d.

References

(1) Jauniaux E, Poston L, Burton GJ. Placental-related diseases of pregnancy: Involvement of oxidative stress and implications in human evolution. Hum Reprod Update. 2006;12:747-55.

(2) Johns G, Jauniaux E. Early pregnancy loss. In: Fetal Medicine: Basic science and clinical Practice 2nd Edt. Edts: Rodeck C, Whittle MJ. Churchill Livingstone, London, 602-619, 2009.

(3) Jauniaux E, Watson AL, Hempstock J, Bao Y-P, Skepper JN and Burton GJ (2000) Onset of maternal arterial blood flow and placental oxidative stress; a possible factor in human early pregnancy failure. Am J Pathol 157, 2111-2122.

(4) Jauniaux E, Gulbis B and Burton GJ (2003) The human first trimester gestational sac limits rather than facilities oxygen transfer to the foetus: A review. Placenta-Trophoblast Research 24, S86-S93.

(5) Red-Horse K, Zhou Y, Genbacev O, Prakobphol A, Foulk R, McMaster M, Fisher SJ. Trophoblast differentiation during embryo implantation and formation of the maternal- fetal interface. J Clin Invest. 2004 Sep;l 14(6):744-54.

(6) Burton GJ. Oxygen, the Janus gas; its effects on human placental development and function. J Anat. 2009 Jul;215(l):27-35. Epub 2008 Oct 13.

(7) Jauniaux E, Hempstock J, Greenwold N and Burton GJ (2003) Trophoblastic oxidative stress in relation to temporal and regional differences in maternal placental blood flow in normal and abnormal early pregnancies. Am J Pathol 162, 115-125. (8) Xing Y, Williams C, Campbell RK, Cook S, Knoppers M, Addona T, Altarocca V and Moyle WR (2001) Threading of a glycosylated protein loop through a protein hole: implications for combination of human chorionic gonadotropin subunits. Protein Sci 10, 226- 235.

(9) Muttukrishna S, Suri, S, Groome NP and Jauniaux E (2008) , Relationships between TGFP proteins and oxygen concentrations inside the first trimester human gestational sac, Plos one, 3: e2302 1-7.

(10) Hempstock J, Jauniaux E, Greenwold N and Burton GJ (2003) The contribution of placental oxidative stress to early pregnancy failure. Hum Pathol 34, 1265-1275.

(11) Greenwold N, Jauniaux E, Gulbis B, Hempstock J, Gervy C, Burton G. Relationships between maternal serum, endocrinology, placental karyotype and intervillous circulation in early pregnancy failure. Fertil Steril 2003; 79: 1373-1379.

(12) Khalil, A., Muttukrishna, S., Harrington, K and Jauniaux, E (2008) Effect of Antihypertensive Therapy on Levels of Angiogenic Factors in Pregnancies with Hypertensive Disorders. Plos One, e2766.

(13) Levine RJ, Thadhani R, Qian C, Lam C, Lim KH, et al. (2005) Urinary placental growth factor and risk of preeclampsia. JAMA 293(1): 77-85.

(14) Rana S, Karumanchi SA, Levine RJ, Venkatesha S, Rauh-Hain JA, et al. (2007) Sequential changes in antiangiogenic factors in early pregnancy and risk of developing preeclampsia. Hypertension 50(1): 137-142.

(15) Muttukrishna, S, Jauniaux, E, Greenwold, N, McGarrigle, H, Jivraj, S, Carter, S, Elgaddal, S, Groome, N, Regan, L. (2002) Levels of inhibin A, activin A and follistatin in missed and recurrent miscarriages. Hum. Reprod., 17:3072-3078.

(16) Johns J, Muttukrishna S, Lygnos M, Groome N, Jauniaux E. Maternal serum hormone concentrations for prediction of adverse outcome in threatened miscarriage. Reprod Biomed

Online. 2007;15:413-21.

(17) Venkatesha, S, Toporsian M, Lam c, Hanai J, Mammoto T, Kim YM, Bdolah Y, Lim KH, Yuan HT, Libermann TA, Stillman IE, Roberts D, D'Amore PA, Epstein FH, Sellke FW, Romero R, Sukhatme VP, Letarte M, Karumanchi SA (2006) Souluble Endoglin contributes to the pathogenesis of pre eclampsia. Nature Medicine 12:862.

(18) Plaisier M, Rodrigues S, Willems F, Koolwijk P, van Hinsbergh VW, Helmerhorst FM (2007) Different degrees of vascularization and their relationship to the expression of vascular endothelial growth factor, placental growth factor, angiopoietins, and their receptors in first-trimester decidual tissues. Fertil Steril. 2007 88:176-187 (19) Smith GC, Crossley JA, Aitken DA, Jenkins N, Lyall F, Cameron AD, Connor JM, Dobbie R. (2007) Circulating angiogenic factors in early pregnancy and the risk of preeclampsia, intrauterine growth restriction, spontaneous preterm birth, and stillbirth. Obstet Gynecol. 109: 1316-1324

(20) Vuorela P, Hatva E, Lymboussaki A, Kaipainen A, Joukov V, et al (1997) Expression of vascular endothelial growth factor and placenta growth factor in human placenta. Biol Reprod 56(2): 489-94.

(21) Jenkins C, Wilson R, Roberts J, Miller H, McKillop JH and Walker JJ (2000) Antioxidants: Their role in pregnancy and miscarriage. Antiox Redox Signal 2: 623-628.

(22) Makhseed M, Raghupathy R, Azizieh F, Farhat R, Hassan N and Bandar A (2000) Circulating cytokines and CD30 in normal human pregnancy and recurrent spontaneous abortions. Hum Reprod 15, 2011-2017.

(23) Baxter N, Sumiya M, Cheng S, Erlich H, Regan L, Simons A and Summerfield JA (2001) Recurrent miscarriage and variant alleles of mannose binding lectin, tumour necrosis factor and lymphotoxin a genes. Clin Exp Immunol 126, 529-534.

(24) von Wolff M, Thaler CJ, Strowitzki T, Broome J, Stolz W and Tabibzadeh S (2000) Regulated expression of cytokines in human endometrium throughout the menstrual cycle: dysregulation in habitual abortion. Molec Hum Reprod 6, 627-634.

(25) Choi HK, Choi BC, Lee SH, Kim JW Cha KY Baek KH Expression of angiogenesis and apoptosis related genes in chorionic villi derived from recurrent pregnancy loss patients. Mol Reprod Dev 66, 24-31.

(26) Burton GJ, Hempstock J and Jauniaux E (2003) Oxygen, early embryonic metabolism and free radical-mediated embryopathies. Reprod Biomed Online 6, 84-96.

(27) Johns J, Hyett J and Jauniaux E (2003) Obstetric outcome after threatened miscarriage with and without a hematoma on ultrasound. Obstet Gynecol 102, 483-487.

(28) HarviUe EW, Wilkox AJ, Baird DD and Weinberg CR (2003) Vaginal bleeding in very early pregnancy. Hum Reprod 18, 1944-1947.

(29) Johns J, Jauniaux E. Threatened miscarriage as a predictor of obstetric outcome. Obstet Gynecol. 2006; 107: 845-50.

(30) Silver RK, Wilson D, Philip J, Thorn EA, Zachary JM and Mohide P for the NICHD EATA Trial Group (2005) Late first-trimester placental disruption and subsequent gestational hypertension/preeclampsia. Obstet Gynecol 105, 587-592.