Emerging roles of APLN and APELA in the physiology and pathology of the female reproductive system.

Human APLN gene is located on chromosome Xq25-26.1 which encodes a pre-propeptide of 77 amino acids. After cleavage of the 22-amino acid secretory sequence at N terminus by endopeptidases, the propeptide is subsequently processed into three active fragments at several dibasic residues (Arg-Lys and Arg-Arg), including APLN-36, APLN-17 and APLN-13. APLN-13 undergoes post-transcriptional cyclization at the N-terminal glutamine, generating pyroglutamate-APLN-13 (Pyr1-APLN-13) ( Tatemoto et al., 1998 ). The potency and efficacy of APLN differ from different isoforms. For instance, APLN-36, APLN-13 and Pyr1-APLN-13 are preponderantly contributed in human cardiovascular regulation ( Maguire et al., 2009 ), whereas APLN-17 plays crucial role in APLNR internalization ( El Messari et al., 2004 ). To date, APLN is abundantly distributed in female reproductive system such as ovary, oviduct, uterus and placenta. Emphatically, APLN is identified as one type of adipokines secreted by white adipose tissue, which plays a role with other adipokines in regulating the secretion of gonadotropin releasing hormone (GnRH), gonadotropins and steroids through hypothalamo-pituitary-gonadal (HPG) axis ( Bertrand, Valet & Castan-Laurell, 2015 ; Yang et al., 2019 ). APLN/APLNR activates different types of G protein and further stimulates three important signaling pathways, which are phosphorylation of phosphoinositide 3-kinase/protein kinase B (PI3K/Akt), reduction of cyclic adenosine monophosphate (cAMP) and activation of phospholipase C- β (PLC- β ), respectively ( Fig. 2A ) ( Chapman, Dupré & Rainey, 2014 ). There are two types of pertussis toxin-sensitive G α protein (G α i/o, G α q/11) at the downstream of APLNR, mediating different signaling transduction ( Masri et al., 2002 ). G α i/o activates PI3K/Akt dependent manner which is crucial for cell survival and nitric oxide (NO) induced vasodilation ( Liu et al., 2010 ). Akt phosphorylates Bcl-2-associated death promoter (Bad, a BH3-only protein) and shifts it to an inert form, which inhibits the binding of Bad and Bcl-2. Bcl-2 plays an anti-apoptotic role by disturbing the aggregation of Bak and Bax (BH123 proteins) in the mitochondrial outer membrane, and thereby attenuating the release of cytochrome c and activation of caspase-3 ( Liu et al., 2019 ). Moreover, endothelial nitric oxide synthase (eNOS) can also be activated by Akt through phosphorylation, triggering the release of NO for vasodilation ( Fig. 2A ) ( Yang et al., 2014 ). Additionally, G α -i/o inhibits adenylate cyclase (AC), following with the reduction of 3′, 5′- cAMP and protein kinase A (PKA), which could potentially regulate the glucose homeostasis ( Fig. 2A ) ( Masri et al., 2002 ). G α q/11 activates PLC- β hydrolyze phosphatidylinositol 4, 5-bisphosphate into second messengers diacylglycerol and inositol trisphosphate, which increases the release of calcium (Ca 2+ ) from intracellular store and activates protein kinase C (PKC) ( Carpéné et al., 2007 ). Amplified intracellular Ca 2+ not only mediates positive inotropic effect in cardiac smooth muscle, but also stimulates NO release in periphery via activating eNOS by calmodulin ( Dai, Ramirez-Correa & Gao, 2006 ). PKC in turn activates Ras/MAPK system, which plays a crucial role in cell proliferation ( Szokodi et al., 2002 ). Furthermore, MAPK halts the expression of pro-oxidant enzymes and subsequently attenuates the release of reactive oxygen species (ROS), which suppresses lipid metabolism and inflammatory reaction ( Fig. 2A ) ( Than et al., 2014 ). (A) Both APLN (in blue) and APELA (in orange) can classically activate G α i/o and G α q/11 mediated intracellular transduction via binding to their common receptor APLNR. (B) Non-coding APELA binds to hnRNPL and promotes p53-mediated cell apoptosis. (C) APELA also stimulates PI3K-independent NODAL/TGF β signal through alternative receptors in hESCs. AC, adenylate cyclase; eNOS, endothelial nitric oxide synthase; IncRNA, long non-coding RNA; hnRNPL, heterogeneous nuclear ribonucleoprotein L. Apela gene, located on chromosome 4 of Homo sapiens ( Fagerberg et al., 2014 ) (chromosome 8 in Mus musculus ( Yue et al., 2014 ) and chromosome 1 in Danio rerio ( Ulitsky et al., 2011 ), was originally annotated to be transcribed exclusively into a non-coding RNA in zebrafish embryo ( Chng et al., 2013 ), while recently it was confirmed to encode a 54-amino acid precursor which further undergoes proteolysis and generates four mature isoforms: APELA-32, APELA-22, APELA-21 and APELA-11. The shortest isoform is conserved across vertebra species ( Huang et al., 2017 ). Compared with APLN, APELA as the second discovered endogenous ligand of APLNR is also ubiquitously detected in placenta, heart, kidney, prostate and mammalian plasma but not as widely as APLN ( Wang et al., 2015 ). In addition, studies have reported that APELA is highly expressed in human embryonic stem cells (hESCs) where the APLNR is absent, indicating the existence of an alternative APLNR-independent transduction ( Ho et al., 2015 ). A recent report has suggested that an orphan G protein-coupled receptor 25 (GPR25), associated with blood pressure regulation and autoimmune disease, could be activated by both APLN and APELA in non-vertebrates, which is similar as APLN in decreasing the intracellular cAMP level. However, the accurate role of this putative receptor in vertebrate remains to be determined ( Zhang et al., 2018 ). Similar to APLN, APELA binds to APLNR, subsequently activating G α i/o and G α -q/11 mediated signaling cascades, including PI3K/Akt, PKC and PKC-independent Ras/MAPK pathways ( Fig. 2A ) ( Perjés et al., 2016 ; Zhang et al., 2018 ). Nevertheless, it also plays non-redundant role for its specific properties during embryo development. In mouse ESCs, Apela acts as a long non-coding RNA that binds to heterogeneous nuclear ribonucleoprotein L (hnRNPL) at the 3′ UTR, which negatively regulates the interaction between p53 and hnRNPL, and promotes p53-mediated DNA damage induced apoptosis ( Fig. 2B ) ( Li et al., 2015 ). In hESCs, APELA acts as a paracrine secreted hormone that binds to an alternative unknown receptor (non-APLNR) and activates PI3K/AKT signaling for cell survival and self-renewal. This process resembles other fibroblast growth factor like exogenous insulin and endogenous insulin-like growth factors (IGFs) in PI3K-mediated cell proliferation. However, APELA-pulsed hESCs is non-redundant as it also implicates in mesendodermal linage commitment through a PI3K-independent manner ( Ho et al., 2015 ). During zebrafish gastrulation, a proper level of APELA acting as a mitogen, indirectly mediates the internalization of ventrolateral mesendodermal cells. This process is presumably achieved via activating NODAL/TGF β signaling pathway ( Fig. 2C ), whereas its specific mechanism remains unknown ( Pauli et al., 2014 ).

Recently, it has been witnessed that APLN and its receptor also display potential behaviors in uterus among species such as rat, mouse, ewe and human. The expression of APLNR mRNA in uterus was firstly detected through a nonspecific rat tissue RT-PCR screen ( Hosoya et al., 2000 ), then its ligand APLN was described to be elevated during the secretory phase in the glandular cells of endometrial layer whereas it remained at a low level in the stromal cells ( Kawamata et al., 2001 ; Ozkan et al., 2013 ; Mercati et al., 2019 ). It is evidently deduced that the apelinergic system is stimulated by elevated steroid hormones during the uterine secretory phase also known as the luteal phase of ovarian cycle. APLN subsequently plays a spatio-temporal role in spiral arterioles maturation and interstitial edema in endometrium where angiogenesis is taking place. An in vitro study showed that APLN played a vasodilation role in suppressing both spontaneous and oxytocin-induced contraction in human myometrial fibers ( Hehir & Morrison, 2012 ). However, serum APLN was also reported to exert a positive inotropic effect in rat myometrial layer via PKC-mediated intracellular Ca 2+ amplication ( Kacar et al., 2018 ). These opposite results may be explained by the intracellular balance between vascular dilation and smooth muscle contraction mechanisms of apelinergic system, as well as the impacts of species diversity and reagent concentrations. Endometriosis is defined as an estrogen-dependent invasion of endometrial tissue from uterus to uterine adnexa ( Bulun et al., 2019 ). It is a very common cause for chronic pain in the pelvis and could even lead to infertility in moderate and significant patients (Chaljub, Medlock & Services, 2018). Current explanations of endometriosis pathogenesis are endometrial implantation, coelomic metaplasia and induction theories which are all in agreement with the impacts of steroid hormone dysregulation and inflammatory response. Similar expression pattern of APLN was seen in both eutopic and ectopic endometrium during the menstrual cycle indicated that the ectopic endometrial lesion could share some characteristics with eutopic cellular processes in endometrium regeneration ( Mercati et al., 2019 ). Additionally, the angiogenesis and vasodilation effects of APLN could potentially be one of the causes in triggering the symptoms of endometriosis, whereas more studies are expected to confirm this point.

Apelin receptor (APLNR, also known as APJ, APJR, AGTRL1 and HG11) was firstly identified as a class A G protein-coupled receptor in 1993. It consists of 380 amino acids, which has a sequence sharing 31% homology with that of the angiotensin type 1 receptor ( O’Dowd et al., 1993 ). Nevertheless, APLNR cannot actually bind to angiotensin II and remains as an “orphan receptor” until its endogenous ligand apelin (APLN, also named APEL and XNPEP2) was later extracted from bovine stomach ( Tatemoto et al., 1998 ). APLN is generally existed in functional isoforms which are cleaved and modified from the C-terminus of a 77-amino acid pre-pro-peptide encoded by APLN gene, with different affinities for APLNR and prevalent distribution ( Chapman, Dupré & Rainey, 2014 ). Both APLN and its receptor APLNR levels are universally high at multiple organs like brain, retina, heart, stomach, liver, kidney and blood vessels in many species ( Kawamata et al., 2001 ; Zeng et al., 2007 ; Kasai et al., 2008 ; Qian et al., 2011 ; Krist et al., 2013 ; Lv et al., 2017 ). Recent years, apelin receptor early endogenous ligand (APELA, also named ELABELA, Toddler and Ende) was identified as a new endogenous ligand for APLNR in both Chng and Pauli’s labs independently ( Chng et al., 2013 ; Pauli et al., 2014 ). Similar to APLN, this 54-amino acid polypeptide is also processed into several isoforms. APELA is highly enriched in the early stage of embryo and confirmed to play a vital role in embryogenesis and angiogenesis ( Norris et al., 2017 ). APLNR and its two ligands compose the apelinergic axis, which is well delineated in systemic physiological processes like cardiogenesis, angiogenesis, fluid homeostasis, vasodilation and energy metabolism. More recently, several studies have been investigating the possible intervention of apelinergic axis in female reproductive system based on its precise regulation of steroidogenesis, angiogenesis and vasodilation, before moving onto the dysregulation of this system which hypothetically causes fertility disorders and pregnancy complications like polycystic ovary syndrome (PCOS), ovarian cancer, gestational diabetes mellitus (GDM) and preeclampsia (PE) (summarized in Table 1 ). This review summarizes and evaluates the current role of apelinergic system in female reproductive system at both physiological and pathological profiles ( Fig. 1 ), as well as providing the direction for future research. Notes. For each study, the authors, year, disease type, species, samples, molecule, sample size (No. of cases and controls), analyzed expression and significance were listed. There would be a significance when p  < 0.05, and p values were listed in the table (if given). The table was ordered by diseases, molecules and year of publication. PCOS polycystic ovary syndrome; OvCa, ovarian cancer PE preeclampsia GDM gestational diabetes mellitus NA not available polycystic ovary syndrome; OvCa, ovarian cancer preeclampsia gestational diabetes mellitus not available (A) In physiological conditions, APLN (in blue textbox) and APELA (in green textbox) play diverse roles at the different parts of the ovary, uterus and placenta. (B) In pathological conditions, aberrant expression of APLN (in orange textbox) and APELA (in purple textbox) lead to female reproductive disorders such as polycystic ovary syndrome (PCOS), ovarian cancer, preeclampsia (PE), gestational diabetes mellitus (GDM) and endometriosis. * indicating potential apelinergic stimulating factors. ? indicating still unclear or controversy about the functions or contributions of apelinergic molecules in these diseases. ↑ and ↓ in the textbox means increase and decrease respectively, both indicate statistically significant changes. → means resulting.

APLNR was reported to be expressed in the angioblast of frog embryo, which would contribute to the formation of aortic arch vessels and posterior cardinal veins. APLN was detected either within or adjacent to the endothelial cells expressed by APLNR, functioning as an angiogenic agent for nascent blood vessels, especially the intersegmental vessels formation. It also showed the chemotactic ability of APLN to induce the migration of endothelial cells ( Cox et al., 2006 ). Moreover, APLN was proved possessing an anti-apoptotic role in osteoblastic cell line of humans and mice ( Tang et al., 2007 ; Xie et al., 2007 ). It releases Bcl-2 molecule from Bad via activating PI3K/Akt pathway, which subsequently attenuates the activation of downstream apoptotic factors, such as cytochrome c and caspase-3, resulting in the inhibition of osteoblastic cell apoptosis. Additionally, APELA has also been revealed to hold a key role in cardiogenesis, angiogenesis and bone formation during the embryonic development. In APELA knockout mice, the hearts are developed poorly or not developed at all, suggesting the essential role in heart morphogenesis ( Chng et al., 2013 ). It triggers the endothelial precursor (angioblasts) to migrate towards midline and coalesce underneath the notochord, and form the first axial vessels ( Pauli et al., 2014 ; Helker et al., 2015 ). Consistently, APELA-APLNR axis is involved in early placental development and angiogenesis ( Ho et al., 2015 ; Ho et al., 2017 ). In mouse placenta, APELA is robustly expressed in syncytiotrophoblasts from early-to-mid gestation, which favors the sprout of new formed blood vessels ( Ho et al., 2017 ). It is also associated with skeletal formation through inhibiting the expression of Sox32, which can bind to Pou5f3 and Nanog molecules as a transcription factor in dorsal endoderm during gastrulation, and inhibit the formation of Pou5f3-Nanog complex. APELA-APLNR pathway can reduce Sox32 expression and allow Pou5f3-Nanog complexes formation, subsequently activating bone morphogenetic protein signaling for sclerotome fate determination ( Perez-Camps et al., 2016 ). Apelinergic system has been hypothesized as a key factor in placental angiogenesis. APLN was strongly expressed in the cytoplasm of human cytotrophoblasts during the first two trimester of pregnancy, and then decreased at the third trimester. Subtle signals were also detected in the syncytiotrophoblasts during the first trimester, but it disappeared completely in the third trimester ( Cobellis et al., 2007 ). The expression of APLNR in the placenta was later than that of APLN. In the first trimester, it was relatively low and exclusively in the cytotrophoblasts. However, in the third trimester, APLNR was expressed intensely not only in cytotrophoblasts but also in syncytiotrophoblasts, smooth muscle cells and endothelial cells inside of the placental villi ( Cobellis et al., 2007 ). This change suggests a potential chemoarractant and vasculogenic role of APLN in the invasion process of interstitial and endovascular extravillous trophoblasts. In mouse, APELA was detected initially in the trophoblasts and then increased robustly after the allantoic fusion. At the mid-gestation, it was expressed restrictedly in syncytiotrophoblasts, where APLNR was wildly existed in adjacent endothelial cells, indicating a paracrine function of this system to favor the placental angiogenic sprouting ( Ho et al., 2017 ). However, in human placenta, APELA was expressed in both cytotrophoblasts and syncytiotrophoblasts synchronously during the whole pregnancy ( Ho et al., 2017 ) and its speculated role remains to be illuminated. The basic pathological changes of hypertensive disorders in pregnancy are currently recognized as insufficient spiral arteries recasting and inflammation mediated endothelial damage triggered by the intricate network of signaling cascades. APELA as mentioned above plays a crucial role in placental angiogenesis via activating PI3K/AKT/mTOR pathway ( Ho et al., 2017 ), and whether its reduction could lead to PE is now being widely studied. It was firstly discovered that APELA knockout pregnant mice exhibited a hypertensive symptom accompanied with proteinuria and glomerular endotheliosis, which were manifested as preeclampsia-like symptoms ( Ho et al., 2017 ). Scientists therefore started investigating the change of APELA in PE patients and wanted to know whether the APELA could act as a biomarker ( Zhou et al., 2019 ). In the late-onset PE (LOPE), two studies measured a significant increased concentration of APELA in the placenta and serum ( Panaitescu et al., 2020 ; Para et al., 2020 ), while one study measured significant decrease ( Zhou et al., 2019 ). And for early-onset PE (EOPE), only one study observed decrease in both APELA mRNA and protein ( Wang et al., 2019 ), while other studies report no significant change of APELA level on either protein or mRNA ( Pritchard et al., 2018 ; Villie et al., 2019 ; Zhou et al., 2019 ; Panaitescu et al., 2020 ; Para et al., 2020 ). Furthermore, it was found that hypoxia significantly decreased the expression of LIN28B, LIN28A and APELA, and the downregulation of LIN28B and APELA may play a role in PE by reducing trophoblast invasion and syncytialization ( Canfield et al., 2019 ). There are also contradictions about the expression level of APLN in PE patients. Initially, a clinical study found an increased APLN protein level in the placental samples of PE patients, indicating a speculated correlation between APLN and PE ( Cobellis et al., 2007 ). This study was further proved by an experiment which showed intravenous injection of APLN in male mice could lead to the downregulation of blood pressure, suggesting that APLN might act as a vasodilator in PE ( Lee et al., 2000 ). However, case studies also found either decreased ( Inuzuka et al., 2013 ; Yamaleyeva et al., 2015 ; Sattar Taha, Zahraei & Al-Hakeim, 2020 ) or no significantly changed ( Van Mieghem et al., 2016 ) APLN level in PE patients compared with normotensive pregnancies. Not only its ligands, the expression of APLNR is also rather conflicting. It has long been thought that APLNR level increases when the patient suffer from PE ( Cobellis et al., 2007 ), but two other studies suggest APLNR level remains unchanged when PE occurs ( Yamaleyeva et al., 2015 ; Pritchard et al., 2018 ). However, one recent study found a significant decrease both in APLNR mRNA expression and in situ expression between PE patients and normal control, and this significance can be found when controls compared to both EOPE and LOPE groups ( Zhou et al., 2019 ). Altogether, the different expression of apelinergic system could be explained by confounding factors like BMI and mean maternal age mismatches between the cases and controls. Moreover, the balance between vasorelaxant and myocardial contractile effects of apelinergic system, as well as the crosslink of apelinergic axis with intricate inflammatory and endothelial factors in PE should also be taken into consideration. Further investigations should focus on the specific molecular mechanisms of APLN and APELA in the hypertensive disorders of pregnancy. APLN as one of the adipose tissue-derived hormones has been identified to play a role in blood glucose metabolism ( Antushevich & Wójcik, 2018 ). It has been described that insulin may up-regulate the expression of APLN through PKC and PI3K signaling pathways in both murine and human adipocytes ( Boucher et al., 2005 ). Raised apelin levels were found in both insulin-resistant mice and type 2 diabetes mellitus patients ( Xu, Tsao & Yue, 2011 ), which supported the speculation that insulin can stimulate APLN secretion. Nevertheless, the correlation of APLN levels with GDM has not yet reached an agreement in clinical researches. Three studies reported a decrease of serum APLN level in GDM patients ( Aydin, 2010 ; Boyadzhieva et al., 2013 ; Akinci et al., 2014 ) while two other groups revealed an increase ( Aslan et al., 2012 ; Kiyak Caglayan et al., 2016 ). In contrast, there were also no significant association reports between normal control and GDM patients ( Telejko et al., 2010 ; Oncul et al., 2013 ). Thus, the correlation of APLN with the pathophysiology of GDM remains to be elucidated. In addition, factors like BMI, HOMA-IR and birth weight have been shown not correlated with serum APLN level according to these studies ( Aslan et al., 2012 ; Oncul et al., 2013 ), but these confounding factors varied a lot among different groups during pregnancy, which was probably one of the cases in the controversy. APELA had a positive correlation with fasting plasma glucose levels in healthy pregnant women during the second trimester, while decreased APELA circulating level was observed in GDM patients at the same time. In the third trimester, circulating APELA level decreased significantly in both GDM and healthy groups. This study suggested that APELA could be a physiological demand in glucose metabolism, and further contributions should focus on dynamic levels monitoring and mechanism analysis ( Guo et al., 2020 ).

Recently published articles and reports (within 15 years) were conducted from PubMed, Google Scholar and Queen Mary Library databases. Based on the keywords ‘APLN’, ‘APELA’ and ‘female reproduction’, articles extracted were summarized to identify the physiological and pathological roles of apelinergic axis in female reproductive system. This study was approved by Jiangxi Provincial Key Laboratory of Reproductive Physiology and Pathology, Medical Experimental Teaching Center of Nanchang University.

This review presents a landscape of the novel APLN/APELA-APLNR system in the female reproductive field ( Table 1 and Fig. 1 ). Intricate signaling pathways and crosslinks of APLN and APELA imply their multifunctional roles in different organs like ovary, uterus and placenta, during specific developmental stages. APLN as an adipokine appears to have specific effects in steriodogenesis and metabolic regulation in GCs and CL of the ovary. Insulin and IGF1-induced APLN secretion possibly plays a role in glucose regulation in GDM patients. In addition, APLN may sustain a balance between the vasodilative and myocontractile effects in the uterus which could be correlated with hypertensive disorders during the pregnancy. Similarly, APELA as a novel ligand of APLNR also has a potential role in PE, based on the angiogenic effect of spiral arterioles. APELA is essential for fetal and placental development through stimulating the invasion of extravillous trophoblasts. This process is potentially achieved through a chemo-attractant mechanism in placental angiogenic sprouting. Moreover, there is a hyperplasia effect of APELA which could be one of the causes in ovarian tumorigenesis. All the data suggest that there should be additional studies to further investigate the precise roles of this axis in female reproductive system especially at the pathological profile. In the future, it will be important to clarify the crosslink and interaction between APLN and other adipokines in sex hormone regulation and energy metabolism. Specific expression and biological effects of APELA in ovary and uterus are also needed in prospect. It may also be crucial to identify the balance of smooth muscle contraction and vasodilation in apelinergic system at a molecular hierarchy. Collectively, the apelinergic axis is still a novel project for further investigation in both physiological and pathological aspects, and probably brings better therapeutic or prophylactic intervention towards female reproductive disorders.