Kisspeptin in the Prediction of Pregnancy Complications

Kisspeptin and its receptor are central to reproductive health acting as key regulators of the reproductive endocrine axis in humans. Kisspeptin is most widely recognised as a regulator of gonadotrophin releasing hormone (GnRH) neuronal function. However, recent evidence has demonstrated that kisspeptin and its receptor also play a fundamental role during pregnancy in the regulation of placentation. Kisspeptin is abundantly expressed in syncytiotrophoblasts, and its receptor in both cyto- and syncytio-trophoblasts. Circulating levels of kisspeptin rise dramatically during healthy pregnancy, which have been proposed as having potential as a biomarker of placental function. Indeed, alterations in kisspeptin levels are associated with an increased risk of adverse maternal and foetal complications. This review summarises data evaluating kisspeptin’s role as a putative biomarker of pregnancy complications including miscarriage, ectopic pregnancy (EP), preterm birth (PTB), foetal growth restriction (FGR), hypertensive disorders of pregnancy (HDP), pre-eclampsia (PE), gestational diabetes mellitus (GDM), and gestational trophoblastic disease (GTD).


INTRODUCTION
Kisspeptin is best known for its role as a hypothalamic neuropeptide that regulates gonadotrophin releasing hormone (GnRH) secretion (1). Indeed, early studies showed that inactivating variants of the kisspeptin receptor result in pubertal failure due to hypogonadotrophic hypogonadism, confirming the importance of kisspeptin signalling to reproductive health (2,3).
During pregnancy, kisspeptin is produced in large amounts by the placenta and thus there is significant interest in evaluating its potential as a novel marker of pregnancy complications (4). Kisspeptin is a peptide encoded by the KISS-1 gene that binds to a G-protein coupled kisspeptin receptor (KISS-1R, previously known as the orphan receptor GPR54) (5). Kisspeptin levels in the circulation are several hundred fold higher during healthy pregnancy compared to the nonpregnant state (6,7). This review will summarise data evaluating kisspeptin's role as a putative biomarker of pregnancy complications including miscarriage, ectopic pregnancy (EP), preterm birth (PTB), foetal growth restriction (FGR), hypertensive disorders of pregnancy (HDP), preeclampsia (PE), gestational diabetes mellitus (GDM), and gestational trophoblastic disease (GTD).

KISSPEPTIN
The gene encoding kisspeptin (KISS-1) was first identified in 1996 as a metastasis tumour-suppressor gene in malignant melanoma cell lines and its peptide product was initially termed 'metastin' (8). Subsequently, it became known as kisspeptin in homage to its discovery in Hershey, Pennsylvania, USA, the hometown of the famous chocolate Hershey's kisses (8). The KISS-1 gene, located on chromosome 1q32, encodes a 145 amino acid prepropeptide that is posttranslationally cleaved into biologically active kisspeptin peptides of different amino acid lengths indicated by their suffix: e.g. kisspeptin -54, -14, -13, and -10 (5,9,10). All of these peptides bind and activate the kisspeptin receptor through their shared Cterminal region decapeptide motif (Arg-Phe-NH 2 ) (5, 10). Kisspeptin is expressed in multiple tissues including the hypothalamus, limbic system, gonads, pancreas, and liver, but is particularly abundant in the placenta, and thus is believed to play an important role in pregnancy (10,11).

KISSPEPTIN IN HEALTHY PREGNANCY
Kisspeptin plays a key role in implantation and decidualisation. Kisspeptin promotes embryo attachment to the endometrium through interaction with cell adhesion molecules, and stimulates stromal decidualisation by up-regulating leukaemia inhibitory factor (LIF) (12) (Figure 1). Kisspeptin also attenuates the excessive migration and invasion of trophoblasts through inhibition of the matrix metalloproteinases (MMP) 2 and 9 (13)(14)(15). Kisspeptin may also impact angiogenesis and uterine spiral artery modelling (16)(17)(18). A further relevant mechanism of kisspeptin in pregnancy relates to the maternal immune tolerance needed to avoid foetal rejection. Indeed, in vitro incubation with kisspeptin at levels corresponding to those found in pregnancy, results in increased differentiation of human naive T cells into T-regulatory cells (19).
The placenta is considered the main source of kisspeptin during pregnancy and the KISS-1/Kiss-1 gene is expressed in syncytiotrophoblasts, whereas its receptor is expressed in both cytotrophoblasts and syncytiotrophoblasts (12) ( Table 1). Expression of kisspeptin and its receptor is high during early pregnancy and declines as the placenta matures, thus highlighting kisspeptin's role in placentation (14). Interestingly, circulating kisspeptin levels increase linearly with advancing gestation and kisspeptin-54 immunoreactivity dramatically rises from 1230 pmol/L during the first trimester to 9590 pmol/L during the third trimester and returns to non-pregnant levels (<100 pmol/L) soon after birth (8 pmol/L) (6,7,20).
Circulating kisspeptin levels are affected by several variables in healthy pregnancy (20). Whilst gestational and maternal age are associated with raised kisspeptin levels, Afro-Caribbean ethnicity, smoking during pregnancy, and high body mass index (BMI) are associated with reduced kisspeptin levels (20). Additionally, kisspeptin levels have been shown to be lower in serum compared to plasma samples, and are influenced by preanalytical factors such as collection tube type, processing time and time to sample storage (49).

Kisspeptin in Miscarriage
Miscarriage is the spontaneous loss of an intrauterine pregnancy before 24 weeks of gestation and affects 1 in 5 clinical pregnancies (50). Miscarriage predominantly occurs during the first trimester of pregnancy and the majority of early miscarriages are due to a genetic abnormality of the developing embryo, however other causes include endocrine, anatomical, and immunological factors (51).
Miscarriage diagnosis can be challenging as often a pregnancy is failing for a time before pregnancy loss has conclusively been confirmed. This uncertainty can exacerbate the psychological burden related to investigating possible miscarriage, with up to 6% of women suffering from moderate-severe depression, 17% from moderate-severe anxiety and 18% from post-traumatic stress disorder (52). To date, there is no clinical predictor of miscarriage, however recent data demonstrates a potential for kisspeptin as a biomarker of miscarriage.
Kisspeptin levels (adjusted for gestation) are markedly reduced by 60-79% in women with miscarriage compared to healthy pregnancy (20)(21)(22)(23)(24)(25) (Table 2.1). Above average levels, when corrected for gestational age, are reassuring with a <1% chance of miscarriage (20), whereas kisspeptin levels 95% lower than the median for that gestation are associated with up to an 85% chance of miscarriage. Concordantly, KISS-1 expression is decreased in the placentae of women with recurrent spontaneous abortion compared to those who undergo voluntary termination of pregnancy (11). Furthermore, whilst kisspeptin's high diagnostic performance for identifying miscarriage is maintained in late-first trimester pregnancies (>8 weeks of gestation), that of b-human chorionic gonadotrophin (b-hCG) worsens (20). Thus, the combination of both kisspeptin and b-hCG can be used to ensure high diagnostic accuracy at all gestations (AUCROC 0.92, 95% CI 0.89-0.95) (20,22,24). Kisspeptin has also been shown to reflect different types of miscarriage, with lower levels reported in complete (no retained products of conception) versus incomplete (retained products of conception) or missed (empty gestational sac or a foetal pole with no heartbeat) miscarriage (20). Additionally, both kisspeptin and b-hCG levels decline with closer proximity to miscarriage confirmation, and therefore repeat measurements every 1-2 weeks could enable further risk-stratification of miscarriage risk in clinical practice (20).
Studies involving women with infertility who undergo assisted reproductive techniques (in vitro fertilisation, intracytoplasmic sperm insemination (ICSI) or frozen thawed embryo transfer) have found reduced b-hCG levels in miscarriage compared to controls, but no difference in kisspeptin levels (53,55). These findings may be due to the very early gestations at which kisspeptin levels were assessed (2-3 weeks following, or even before, pregnancy confirmation) (53,55). Indeed, kisspeptin may not be expressed in the placenta at high levels prior to 6 weeks of gestation, suggesting that b-hCG levels may be more useful at these very early gestations (26).

Kisspeptin in Ectopic Pregnancy
Ectopic pregnancy (EP) affects 2% of pregnancies and occurs when a fertilised ovum implants and develops outside the uterine cavity, most commonly within the fallopian tube (56). EP can result in tubal     rupture and accounts for 9-13% of all pregnancy-related deaths in developed countries and can compromise a woman's future fertility (57). EP is currently diagnosed by serial b-hCG measurements in combination with ultrasound, although laparoscopy is often required to provide a definitive diagnosis (58). The sensitivity and specificity of these tests significantly decrease in the case of pregnancies of unknown location (PUL) as false positive or negative diagnoses may occur. This is important as an incorrect diagnosis may lead to termination of a healthy pregnancy (59). Accordingly, different biomarkers have been investigated in an attempt to improve the diagnostic accuracy of EP, including kisspeptin. Some studies have found that kisspeptin levels in EP are lower than in healthy pregnancy but higher than in miscarriage (25,26). However, another study demonstrated that kisspeptin levels are not significantly altered between women with viable intrauterine pregnancies (VIUPs) and those with either EP or failing or persistent PUL, after adjusting for confounding variables (20) ( Table 2.2). Current evidence remains limited, and larger studies are required to determine kisspeptin's performance as a diagnostic marker in EP at early gestations (<6 weeks).

Kisspeptin in Hypertensive Disorders of Pregnancy and Pre-Eclampsia
Hypertensive disorders affect 5% of all pregnancies (60) and include pre-existing chronic hypertension (CH), pregnancy induced hypertension (PIH) and pre-eclampsia (PE). PIH is defined as new onset hypertension (BP ≥140/90mmHg) occurring after 20 weeks of gestation, PE is PIH with proteinuria (urine >3g/24 hours) or significant end-organ dysfunction, and severe PE is the presence of at least one of: hypertension (BP≥160/110 mmHg), visual disturbance, chest pain, dyspnoea, pulmonary oedema, seizures, or neonatal distress (61). PE is further classified, according to the onset of clinical features, into early-onset PE (EPE <34 weeks of gestation) and late-onset PE (LPE ≥34 weeks of gestation). EPE is associated with impaired trophoblast invasion, defective spiral artery remodelling and adverse perinatal complications including IUGR (62). LPE occurs due to hypoxic stress and impaired perfusion but is less likely to compromise foetal growth (63,64). Currently, PE diagnosis is based on early pregnancy risk factor screening, uterine artery Doppler velocimetry and biomarkers such as PPAP-A or placental growth factor (PlGF) (61). Kisspeptin has been implicated in the pathogenesis of PE through reduced angiogenesis, decreased cytotrophoblast invasion and increased trophoblast apoptosis, and thus could have potential in predicting PE (16)(17)(18).
Levels of circulating kisspeptin in HDP vary in the literature, and largely differ according to HDP subtype, severity, and onset ( Table 2.3). Most of the studies report reduced circulating kisspeptin levels in PE compared to normotensive pregnant controls (29,32,(40)(41)(42)(43)65) and therefore kisspeptin is considered to reflect placental dysfunction. However, expression of KISS-1, which inhibits trophoblast invasion and results in defective transformation of the spiral arteries, is increased in the placentae of PE pregnancies, thus supporting its role in the pathophysiology of PE (32-34, 36, 66) (Table 1). Nonetheless, there are also some reports of decreased KISS-1 expression in PE placentae (37,38) ( Table 1). Furthermore, evidence suggests that circulating kisspeptin levels decline as the severity of PE increases, which could also reflect reduced placental mass in more severe disease. Indeed, both circulating kisspeptin levels and placental mass is reduced in EPE compared to LPE (28,42,67). Additionally, pregnant women with pre-existing hypertension and PE, states associated with a higher burden of disease, have reduced kisspeptin levels compared to PIH (40).
Whilst most studies demonstrate reduced kisspeptin levels in PE, a recent study found that kisspeptin levels are increased in HDP during the third trimester of pregnancy (Table 2.3). However, there was no association between circulating kisspeptin levels and severity of PET (28). It is likely that complexity in the categorisation, severity, and onset of PET, and the need for correction for possible confounders such as BMI and gestational age, could explain differences between kisspeptin levels observed in the current studies. Larger observational studies that are carefully designed to address these and look at each PET-subset throughout pregnancy would therefore be valuable in resolving these inconsistencies.

Kisspeptin in Gestational Diabetes Mellitus
During pregnancy a physiological rise in maternal insulin resistance provides glucose to the developing foetus (68,69). This insulin resistance leads to maternal pancreatic b-cell adaptation and increased insulin secretion. Failure of these changes results in gestational diabetes mellitus (GDM), which affects up to 20% of pregnancies worldwide (70).
Kisspeptin receptors are expressed in pancreatic b-cells (71) and have been implicated in b-cell adaptation during pregnancy. Exogenous kisspeptin administration has variable physiological effects on the glucose-dependent regulation of pancreatic betacells. For instance, KISS-1 peptide (KP-145) (71), KP-13 (72), KP-10 (72-74) potentiates glucose-stimulated insulin secretion (GSIS) in animal and human islets in-vitro. KP-54 increases GSIS in healthy men following an intravenous glucose tolerance test (IVGTT), which induces high glucose levels (75). On the other hand, Vikam and colleagues have found that KP-13 and KP-54 drives dose-dependent inhibitory effects on insulin secretion in mouse islets in the presence of lower glucose concentrations (2.8-11.1 mmol/l), compared to controls, which is not observed at higher glucose concentrations (76). Furthermore, chronic administration of KP-10 in non-pregnant mice enhances GSIS and improves glucose tolerance (47). Interestingly, hyperlipidaemia, impaired glucose tolerance (IGT) and weight gain develops in Kiss-1r-null female mice exclusively, thus suggesting sexual dimorphism in kisspeptin's effects on metabolism and glucose homeostasis (77).
In late gestation murine pregnancy, b-cell specific Kiss-1rknockout models and pharmacological inhibition of Kiss-1r leads to reduced GSIS and development of IGT, which is not observed in non-pregnant states or wild-type controls (47). This supports a role for b-cell kisspeptin signalling in the regulation of glucose homeostasis during pregnancy. Loss of kisspeptin signalling in the ß-cell-specific Kiss-1r-knockout models also attenuates the increased ß-cell proliferation normally seen during murine pregnancy when assessed with bromodeoxyuridine (BrdU) labelling. Nonetheless, the levels are not reduced to nonpregnant levels, suggesting contribution of other signals in pancreatic b-cell proliferation during pregnancy (47,78).

Kisspeptin in Pre-Term Birth
Pre-term birth (PTB) is defined as delivery prior to 37 weeks of gestation and affects 11% of pregnancies (79,80). Kisspeptin has been proposed to initiate labour through increased oxytocin neuronal firing rate in pregnant rats and thus may play a potential role in PTB (81). Gestation adjusted kisspeptin levels are higher in PTB-affected pregnancies than in control pregnancies during the late-first trimester, with the adjusted odds of PTB being increased by 20% (95% CI, 1-42%) for every 1 nmol/L increase in plasma kisspeptin (28) ( Table 2.

5).
Furthermore, KISS-1 mRNA expression is higher in preterm placentae than in term placentae delivered vaginally or by Caesarean section thus indicating that increased kisspeptin expression could be involved in the induction of labour (27) ( Table 1). However, no alteration in circulating kisspeptin levels have been reported to date during the third trimester between healthy pregnancy and PTB and thus more data is needed to elucidate whether there are changes in kisspeptin levels preceding and around the time of spontaneous labour (27, 28).

Kisspeptin in Foetal Growth Restriction
Foetal growth restriction (FGR) encompasses both intrauterine growth restriction (IUGR, foetal weight <10 th centile for gestational age with abnormal umbilical artery doppler results) and small for gestation age (SGA, delivery weight <10 th percentile for gestational age) (82,83). FGR is thought to arise from abnormal trophoblast invasion and spiral artery remodelling that limits oxygen supply to the placenta (84,85). The resulting ischemic injury generates reactive oxygen species which lead to apoptosis and restriction of placental and foetal growth (84,85

Kisspeptin in Gestational Trophoblastic Disease
Gestational trophoblastic disease (GTD) is characterised by an abnormal proliferation of placental tissue and comprises of choriocarcinoma, invasive mole, placental site trophoblastic tumour and epithelioid trophoblastic tumour (86). Molar pregnancy is a benign form of GTD, whereas choriocarcinomas are more aggressive, however both exhibit high b-hCG levels and respond well to chemotherapy (87). Serum b-hCG measurement aids with GTD diagnosis, staging and prognostication before and after chemotherapy (88).
KISS-1 and KISS-1R expression is significantly lower in malignant choriocarcinoma cells compared to molar and healthy pregnancies (38,48) (Table 1). Conversely, circulating kisspeptin levels are elevated in malignant GTD compared to healthy pregnancies but significantly decline following chemotherapy (7) ( Table 2.7). The increased circulating kisspeptin levels could reflect an increased malignant trophoblast mass rather than an elevation in cellular KISS-1 expression (89). Thus, kisspeptin levels can be altered in choriocarcinomas and other GTDs, which is interesting when considering the original identification of KISS-1 as an antimetastatic gene.

CONCLUSION
Kisspeptin levels are markedly reduced in miscarriage; and whilst the performance of kisspeptin levels to identify women at high risk of miscarriage is maintained throughout the first trimester, that of b-hCG falls during the latter part of the first trimester. Nevertheless, kisspeptin levels are only mildly elevated at early gestations (< 6 weeks) and therefore can be difficult to detect using current collection and assay methods. Thus, measuring kisspeptin in combination with b-hCG levels could potentially overcome this deficiency at early gestations. Due to the current difficulty in miscarriage diagnosis and the lack of available biomarkers, the high performance of plasma kisspeptin suggests that it has significant potential for further development in this context. Given that kisspeptin has been proposed as a biomarker of healthy placentation, it could potentially be used to recognise late pregnancy complications characterised by abnormal placentation during the first trimester. Regarding HDP, most studies have suggested lower circulating kisspeptin levels but increased placental kisspeptin expression. Kisspeptin levels in pregnancy complications such as PE are confounded by factors such as BMI, disease severity, time of onset, and concomitant FGR, and thus could limit the use of kisspeptin diagnostically.
Overall, current evidence suggests that circulating kisspeptin levels are consistently reduced in miscarriage, EP, FGR, GDM, and increased in PTB and GTD. Larger datasets with adequately sized control cohorts that accurately adjust for gestation, BMI, ethnicity, detailed disease severity phenotype and onset are needed to enable more precise characterisation of the utility of kisspeptin levels in these settings. In summary, circulating kisspeptin is a promising biomarker for early pregnancy loss and further research is needed to assess its potential in other pregnancy complications.

AUTHOR CONTRIBUTIONS
BP, JT wrote the manuscript, designed the figures and tables. AA, WSD, ANC reviewed and edited the manuscript and are the corresponding authors. All authors have made a substantial, direct and intellectual contribution to the work and approved the manuscript prior to its submission.