Introduction

Preeclampsia (PE) is a pregnancy-related disorder that is associated with the unprecedented onset of hypertension (systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg). It occurs after the 20th week of gestation, and frequently near term. It is estimated that PE occurs in 3–8% of pregnant women globally [1]. Although PE is usually identified by new episodes of hypertension and proteinuria after 20 weeks of gestation, pregnant women without proteinuria may be diagnosed with the disorder if they present with one of the following: thrombocytopenia (a platelet count of less than 100,000 per μL); impaired liver function, such as an abnormal rise in the blood concentration of transaminase (to twice the normal concentration), or renal insufficiency (a serum creatinine concentration greater than 1.1 mg/dL, or in the absence of other kidney disease, doubling of the serum creatinine concentration); pulmonary edema; and the unprecedented onset of headaches that are unresponsive to medication and cannot be accounted for by an alternative diagnoses or visual symptoms [1]. Currently, a strategy for the timely detection and diagnosis of PE is urgently needed. This would avoid emergencies or existing complications with target organs. There are several related theories about the causes of PE, including chronic uterine or placental ischemia, immune disorders, genetic imprinting [2], trophoblast apoptosis and necrosis [3], and excessive trophoblast tolerance to inflammatory reactions [4]. Moreover, previous observations have indicated that an imbalance of angiogenesis factors may also play an important role in the pathogenesis of PE [5]. All these pathogenesis processes may be affected by genetic, epigenetic, environmental, and physiological factors, and there is growing evidence that epigenetics play a role in PE [6]. With regard to epigenetics, modifications to both DNA and histones are intermediately involved in the regulation of gene activity. Furthermore, the regulation of functional noncoding RNAs (ncRNAs) can alter gene activity, which modulates gene expression and transcription, chromatin structure, epigenetic memory, selective RNA splicing, and protein translation [7].

ncRNA is a type of functional RNA molecule that is not usually translated into protein. It accounts for 98% of the human genome, and includes housekeeping ncRNA (transfer RNA (tRNA), ribosomal RNA (rRNA), and small nuclear RNA (snRNA)) and regulatory ncRNA (small interfering RNA (siRNA), microRNA (miRNA), piwi-interacting RNA (piRNA), long noncoding RNA (lncRNA), and circular RNA (circRNA)) [8]. The regulatory ncRNAs can be divided into three types—lncRNA (> 200 nucleotides (nt)), miRNA (< 200 nt), and circRNA (circular structure)—which can regulate cell processes through direct interaction with each other [9]. For example, miRNAs can regulate gene expression by targeting mRNA [10] or by adjusting its stability by targeting circRNAs and lncRNAs [11,12,13]. Alternatively, circRNAs and lncRNAs can serve as “sponges” to adjust the availability of miRNAs [14, 15]. ncRNAs can also adjust the physiological function of cells by interacting with DNA or proteins [9, 16] The aberrant expression of ncRNAs or their abnormal interactions can lead to the development of various diseases, including PE and cardiovascular disease. There is increasing evidence that miRNAs, lncRNAs, and circRNAs are widely involved in the pathogenesis of PE. In the present review, we summarize the role of these transcripts in the pathogenesis of PE, and highlight the possible use of ncRNA as a noninvasive tool for diagnosing the condition.

miRNAs and preeclampsia

miRNAs are the best studied of the ncRNAs. They are small single-stranded structures of approximately 22–25 nucleotides that can act as regulators at the post-transcriptional level. When miRNA molecule binds to the 3′-untranslated region of mRNA molecule, it induces the degradation of the mRNA or prevents its translation. It is estimated that miRNAs are able to regulate the translation of more than 60% of the protein-coding genes involved in many physiological processes, such as proliferation, differentiation, apoptosis, and development [17]. Numerous studies have revealed abnormal miRNA expression in the placentas or peripheral blood of patients with PE. Aberrant miRNAs can target downstream genes and reduce the migration and invasion of trophoblasts, or increase cell apoptosis, ultimately resulting in PE.

Expression pattern of miRNAs in patients with PE

To date, approximately 70 miRNAs have been reported to be differentially expressed in PE tissues. Table 1 summarizes the differentially expressed miRNAs involved in the pathogenesis of PE. For the first time, Pineles et al. reported seven differentially expressed miRNAs in the placentas of patients with PE who gave birth to either normal babies or babies that were small for their gestational age. Among these miRNAs, only the levels of miR-182 and miR-210 increased significantly relative to the corresponding levels experienced during normal pregnancy [118]. This finding presented new targets for the pathophysiology of PE. In view of the role of miR-210 in PE, related studies with larger sample sizes further confirmed that the placental expression of miR-210 in patients with PE did indeed increase significantly compared to the corresponding levels experienced during normal pregnancy [18,19,20]. miR-210 can inhibit the proliferation, invasion, and angiogenesis of trophoblasts by acting on the downstream target genes that encode KCMF1 [20], NOTCH1 [21] and MAPK [18]. Upregulated levels of miR-210 have also been detected in peripheral blood serum [18]. It has also been reported that the disturbed expression of miR-182-5p can inhibit trophoblast proliferation, invasion, and migration by acting on the 3′-untranslated region of RND3 [24]. Zhang et al. first reported that miR-155, an inflammation-related miRNA, is overexpressed in the placentas of PE patients, and is involved in the pathogenesis of PE because it downregulates CYR61 [25]. miR-155 can bind the target genes that encode cyclin D1 [16] and eNOS [26] to affect the migration and proliferation of trophoblasts. Differentially expressed miRNAs have been found in exosomes [50, 64, 103, 112] and Mesenchymal stem cells (MSCs) [28, 51, 71, 76] as well as in placental tissues and peripheral serum or plasma. For example, miR-16 upregulation was first confirmed in the placentas of patients with PE [29]. Studies have shown that miR-16 is differentially expressed in the decidual MSCs of patients with severe PE and normal patients, and can inhibit the proliferation and migration of decidual-derived MSCs by targeting cyclin E1 and inducing cell cycle arrest [28]. More interestingly, miR-16 overexpression in decidual-derived MSCs can also reduce the ability of human umbilical vein endothelial cells to form blood vessels [28].

Table 1 Dysregulation of miRNAs in PE

Diagnostic value of miRNAs in PE

Numerous studies have confirmed that miRNAs are involved in the pathogenesis of PE and are differentially expressed in patients with this disease. Some researchers have assessed the diagnostic value of miRNA with regard to PE by drawing receiver operating characteristic (ROC) curves. For example, Zhang et al. showed that the levels of miR-942 decreased significantly in the plasma of patients with PE compared to the corresponding levels in normal patients, and had 65.4% sensitivity and 69.2% specificity with regard to PE diagnosis [81]. Table 2 summarizes the results of research into the value of miRNAs with regard to the diagnosis of PE.

Table 2 Diagnostic value of miRNAs in PE

Association between miRNA variants and the risk of PE

Given that genetic factors play important roles in the occurrence of PE, several studies have focused on the relationship between single nucleotide polymorphisms (SNPs) in miRNAs and the risk of PE. It has also been reported that the miR-146a rs2910164 polymorphism may not be associated with PE susceptibility, cytokines, or related characteristics in black women from South Africa, whereas the GC/CC genotype may reduce susceptibility to severe PE [125]. Interestingly, Salimi et al. discovered that a maternal/placental miR-146a polymorphism (rs2910164) was associated with PE or risk of severe PE, after they genotyped it in the blood samples and placentas from the Asian mainland, using polymerase chain reaction (PCR)–fragment length polymorphism [126]. Table 3 summarizes the reported results.

Table 3 Association between polymorphisms with SNPs and risk of PE

Demethylation of the miRNA in PE

In addition to the previously reported abnormal expression of miRNA in PE patients, some studies also found that the methylation level of some abnormal expression of miRNA is associated with the risk of PE. Rezaei et al. found that hypomethylation of the miR-34a promoter was associated with the occurrence and severity of PE, when they applied methylation-specific PCR to investigate samples from 104 pregnant women with PE and 119 normotensive pregnant women [133]. Moreover, studies have shown that the abnormal regulation of the miR-let-7 family is related to PE. The methylation status of miR-let-7a in PE was evaluated by methylation-specific PCR and bisulfite sequencing PCR analyses. The results suggested that hypomethylated miR-let-7a promotes PE by downregulating Bcl-xl and YAP1 [134]. The miR-510 promoter region in bisulfite-treated PE DNA samples was also found to be unmethylated, compared to the corresponding region in the control samples [135].

lncRNAs and PE

lncRNAs are composed of more than 200 nucleotides. They promote the development of human diseases by participating in various biological processes, including genomic imprinting, chromatin modification, regulation of transcription and post-transcriptional gene expression, nuclear transport, and other regulatory processes [136]. There is abundant evidence to suggest that lncRNA expression in the placenta and peripheral blood differs between healthy pregnant women and patients with PE. This indicates that abnormal lncRNA expression is involved in the pathogenesis of PE.

Expression pattern of lncRNAs in patients with PE

Table 4 summarizes the differentially expressed lncRNAs that participate in the pathogenesis of PE. For example, Liu et al. discovered that the levels of GASAL1 lncRNA were downregulated in the placental tissues of 30 patients with PE, compared to the corresponding levels in 30 normal controls. They further demonstrated that GASAL1 lncRNA can directly bind to functional RNA-binding protein SRSF1 to promote trophoblastic proliferation and progression from the G1 to the S phase through the mTOR and EBP1 signaling pathways. It can also inhibit trophoblastic apoptosis by downregulating cleaved caspase-3 and upregulating Bcl-2 [174]. Another well-known lncRNA, that of MEG3, is expressed in a variety of normal tissues, but is absent in many tumors and tumor cell lines [189]. Yu et al. found that the expression of lncRNA MEG3 in PE placental tissue decreased significantly by 72% compared with that in the normal controls, and MEG3 interruption induced the expression of E-cadherin but reduced that of N-cadherin. This confirmed that MEG3 inhibits trophoblastic migration and invasion. They also found that MEG3 downregulation affects the TGF-β/Smad pathway by inhibiting Smad7 expression, thereby suppressing epithelial–mesenchymal transition [147]. Wang and Zou also showed that by regulating the expression of NOTCH1, MEG3 can promote the apoptosis of trophoblasts, and inhibit their migration and invasion, thereby inducing PE [148].

Table 4 Dysregulation of lncRNAs in PE

Although there is no report that the methylation level of lncRNA itself is related to PE, abnormally expressed lncRNAs can regulate the proliferation, invasion, and apoptosis of trophoblasts by regulating the methylation of downstream molecules. Zhao et al. found that when lncRNA HOTAIR is expressed at high levels, it targets miR-106 by binding to EZH2 [139], which in turn inhibits the transcription of the target gene by inducing H3K27 methylation in the promoter region, ultimately suppressing the proliferation, migration, and invasion of trophoblasts [190]. Xu et al. further confirmed that EZH2 can directly interact with the promoter region of RND3 by methylating H3K27, the 27th amino acid of histone H3, thereby reducing the expression of RND3 in PE. However, the downregulation of lncRNA TUG1 in placental tissues inhibits the proliferation, invasion, and migration of trophoblasts and promotes their apoptosis and it also obstructs spiral artery remodeling by reducing the transcriptional regulation of RND3, which is mediated by recruited EZH2 proteins [143]. Li et al. reported, for the first time, that TUG1 acts as a molecular sponge for miR-29b, thereby regulating the expression of MCL1, VEGFA, and MMP2; it is therefore involved in the development of PE [14]. Many studies have shown that H19 mutations are closely related to PE. Moreover, lncRNA H19 is upregulated in PE. This activates the PI3K/AKT/mTOR pathway, which reduces trophoblast activity and increases invasion and autophagy [153]. Zuckerwise et al. also proposed that the downregulation of H19 inhibits TGF-β signaling transduction by reducing the Par6/Smurf1/RhoA pathway activated by TβR3 [152], Gao et al. found that this impairs the migration and invasion of extravillous trophoblasts in vitro by upregulating the expression of miR-let-7 and downregulating the expression of miRNA-675 [150]. TβR3 is considered to be a downstream mediator of H19–let-7 interaction. Zhou et al. demonstrated that H19 alters genome-wide methylation levels by regulating the activity of S-adenosyl-l-homocysteine hydrolase, and H19 knockout in the intron region causes the undermethylation of TβR3 [191]. Dey et al. showed that hypermethylation of the H19 promoter is associated with reduced H19 expression [192]. Furthermore, the absence of H19 imprinting in PE placental tissues reduces the invasive ability of trophoblasts, and may be associated with severe hypertension, which exacerbates PE [193].

Diagnostic value of LncRNAs in PE

Table 5 reveals that lncRNAs may serve as potential diagnostic biomarkers for PE through ROC curve analysis. In placenta, Wu et al. have found that ROC of lnc TCL6 could reach to 0.8625 [138]. The ROC values of lnc BC030099 in the whole blood cells were 0.713 [104] and that of serum lnc AF085938, and that of lnc G36948 and lnc AK002210 were 0.7673, 0.7956 and 0.7575 respectively [178].

Table 5 Diagnostic value of lncRNAs in PE

Association between lncRNA variants and the risk of PE

Few studies have been conducted on the relationship between lncRNA gene polymorphisms and PE susceptibility. There is a significant correlation between the HOTAIR rs4759314AG genotype and higher PE risk, and the HOTAIR rs10783618 polymorphism is associated with increased PE risk in recessive and allele models. However, HOTAIR gene polymorphisms rs12826786, rs920778, and rs1899663 are not associated with PE susceptibility. The CTGAC haplotype is associated with decreased risk of PE, whereas the CTGAT haplotype is associated with increased risk [194].

CircRNAs and preeclampsia

CircRNAs are covalently closed circular ncRNA molecules. They are resistant to degradation by nucleic acid exonucleases because they lack a 5′ terminal with a cap and a 3′ terminal with a poly(A) tail. These characteristics enable circRNAs to fulfill many biological functions, such as acting as molecular sponges for miRNA, regulating gene transcription and translation, and binding to RNA-binding proteins [195].

Expression pattern of circRNAs in patients with PE

In recent decades, researchers have confirmed that circRNAs are involved in a variety of diseases [196]. However, there have been few studies on the role of circRNAs in the pathogenesis of PE. Ou et al. discovered 49 differentially expressed circRNAs in the placental tissues of patients with severe PE, using RNA sequencing, and further verified the upregulated expression of hsa_circ_0001438, hsa_circ_0001326, and hsa_circ_32340 by quantitative PCR analysis. To determine the interaction between circRNAs and miRNAs, they conducted an analysis of the Kyoto Encyclopedia of Genes and Genomics database. They found that the MAPK signaling pathway was the most enriched pathway in terms of circRNAs, and that the circRNA–miRNA–mRNA interaction network generated by hsa_circ_0001438, hsa_circ_0001326, and hsa_circ_32340 might be involved in the pathogenesis of PE. They also found that miR-145-5p was closely associated with circRNAs and mRNAs [197]. Table 6 generalizes the aberrant expression of circRNA.

Table 6 Dysregulation of circRNAs in PE

Diagnostic value of circRNAs in PE

In recent years, numerous studies have confirmed the value of plasma circRNAs as potential early biomarkers of PE. Using quantitative reverse transcription PCR, Hu et al. demonstrated that the levels of circ-0036877 in blood samples taken from patients with PE were significantly higher than those in the control group. Furthermore, ROC curve analysis suggested that plasma circ-0036877 is a potential early biomarker of PE: the area under the curve value was 0.846, the sensitivity was 85.3%, and the specificity was 72.7% [205]. Zhang et al. first published reports on the analysis of circRNA expression in blood cells. Analysis of red blood cell samples taken from 32 patients with PE and 32 healthy pregnant women revealed significantly higher circ-101,222 levels in patients with PE than in the healthy women: the area under the ROC curve, the sensitivity, and the specificity were 0.706, 65.61, and 68.54%, respectively [198]. To further verify the value of protein-bound circRNAs in the early diagnosis of PE, Bai et al. combined the plasma protein ENG with circRNAs. They found that the resulting area under the curve value increased to 0.876 (95% confidence interval (CI): 0.816–0.922), sensitivity increased to 70.73%, and specificity increased to 80.49%, compared to the corresponding values for the unbound circRNAs [201]. Table 7 summarizes the reported results.

Table 7 Diagnostic values of circRNAs in PE

Relationship between lncRNAs and miRNAs

As discussed above, both lncRNAs and miRNAs have regulatory effects on the pathogenesis of PE. Moreover, they are interrelated and interactive. lncRNAs can act as competing endogenous RNAs to influence the bioavailability of miRNAs [9]. Gao et al. first reported that H19 promoted the expression of miR-let-7 and downregulated miRNA-675, which resulted in the migration and invasion of extravillous trophoblasts in vitro [150]. Zuckerwise et al. also showed that the downregulation of H19 suppressed the Par6/Smurf1/RhoA pathway activated by TβR3 to reduce TGF-β signaling. TβR3 is considered a downstream mediator of the interaction between H19 and let-7 [152]. There have been numerous investigations into the role of the lncRNA–miRNA–mRNA axis in the pathogenesis of PE. For example, Tan et al. reported that DLX6-AS1 lncRNA may contribute to PE by suppressing the proliferation, migration, and invasion of trophoblasts via the miR-376c–GADD45A axis [155]. Li et al. were the first to report that lncRNA TUG1 causes the development of PE by acting as a molecular sponge for miRNA-29b, thereby regulating the expression of MCL1, VEGFA, and MMP2 [14]. Yu et al. discovered that lncRNA TUG1 can also act as a molecular sponge for miR-204-5p, and downregulated lncRNA TUG1 suppresses trophoblast migration and invasion, partly by sponging miR-204-5p [144]. Figure 1 depicts the molecular mechanism by which ncRNA affects the pathogenesis of PE.

Fig. 1
figure 1

lncRNAs and circRNAs can act as a competing endogenous RNA(ceRNA), binding with miRNAs and regulating the effects of miRNAs on target genes, to influence trophoblast cell proliferation, invasion, migration and apoptosis and might play a role in angiogenesis and cell population at the G 0 /G 1 phase

Relationship between circRNAs and miRNAs

CircRNAs that contain miRNA response elements can serve as competing endogenous RNAs by binding with miRNAs. They act as miRNA sponges in cells, thereby regulating the effects of miRNAs on target genes and altering their expression levels [9]. In 2013, Hansen et al. found that ciRS-7, which is a circRNA sponge for miR-7 derived from the CDR1 gene, can bind and adsorb miR-7, thereby reducing its activity and indirectly upregulating the expression of miR-7-related target genes [206]. circRNAs have a stronger potential to adsorb miRNAs in the body than linear mRNAs or lncRNAs, because they are more stable. A few researchers have reported the occurrence of PE involving the interaction between circRNAs and miRNAs. circ-PAPPA, which is downregulated in both the placentas and plasma of patients with PE, can directly target miR-384 and act as a sponge for it. Finally, miR-384 overexpression inhibits the proliferation and invasion of trophoblasts by targeting STAT3 [13]. Shen et al. reported that upregulated circ-TNRC18 in the placental tissues of patients with PE also combined with miR-762 to target GRHL2 protein to regulate trophoblast epithelial–mesenchymal transition and invasion [15]. Although the role of circRNAs in the pathogenesis of PE is not fully understood, there is new evidence that they act as molecular sponges for miRNAs. Figure 2 illustrates the relationships between the various ncRNAs.

Fig. 2
figure 2

Mechanisms of interaction of ncRNA in PE. Green: protein coding, Yellow: lncRNA, Red: miRNA, Blue: CircRNA

Discussion

The heterogeneity and complexity of PE make its diagnosis, prediction, and treatment difficult. As it is currently not possible to detect the molecular signature of the main affected organ, i.e., the placenta, until the termination of pregnancy, it is difficult to monitor the progression of PE in a timely manner. Therefore, markers that circulate in the peripheral blood have great potential for noninvasive monitoring. It is currently possible to detect numerous biochemical markers in the serum, such as placental growth factor, soluble FMS-like tyrosine kinase receptor 1, placental protein 13, and placental protein A; however, their sensitivity and specificity are low [207]. Molecular biomarkers could provide a more reliable platform for the screening and diagnosis of PE than biochemical markers. In particular, the ncRNAs in the maternal peripheral blood are expected to be useful noninvasive biomarkers.

The differential expression of ncRNAs has been investigated to confirm their involvement in the pathogenesis of PE. Following numerous studies on the abnormal expression of ncRNAs in placental tissues, studies have been carried out to investigate ncRNAs in the peripheral blood of patients with PE. For example, Li et al. separated exosomes from maternal plasma by continuous density gradient hypercentrifugation, and found that seven miRNAs were differentially expressed in the exosomes from women with PE and those from a control group; however, the source of these exosomes was not determined [208]. It has subsequently been reported that exosomes derived from human umbilical cord mesenchymal stem cells that overexpress miR-139-5P, can promote trophoblast migration, invasion, and proliferation, and prevent apoptosis [209]. Studies on placenta-specific miRNA clusters in plasma samples revealed that the overexpression of miR-517-5p, miR-518b, and miR-520 h was associated with the late development of PE, and the screening of plasma miR-517-5p in early pregnancy also identified a proportion of women with subsequent PE [124]. Furthermore, Sun et al. performed univariate and multivariate analyses on the upregulation of lncRNA BC030099 in the whole blood of patients with PE, and determined that lncRNA BC030099 was an independent predictor of PE [137].

LncRNAs can regulate miRNA activity, and post-transcriptional regulation will affect the expression and function of their target mRNAs. LncRNAs have been shown to have miRNA binding sites --miRNA responsible elements, and they may potentially sponge the miRNAs, Thus, miRNA-mediated post-transcriptional regulation of target mRNAs was impaired. Dong et al. have demonstrated that LINC00511 regulates the proliferation, apoptosis, invasion and autophagy of trophoblast cells to mediate PE progression through modulating the miR-31-5p/homeobox protein A7 axis through dual luciferase reporter gene analysis [176]. When circRNAs interacted with miRNAs, they formed miRNA molecular sponges that further inhibited the transcript and lead to gene silencing. Due to the complementarity between bases, miRNAs bound to target mRNAs and performed transcriptional silencing to regulate gene expression. However, Li et al. confirmed circ_0063517 acts as ceRNA,targeting the miR-31-5p-ETBR axis to regulate angiogenesis of vascular endothelial cells in PE by dual luciferase reporting system and RNA immunoprecipitation (RIP) analysis [210]. Because lnc00511 and circ0063517 played an important role in the occurrence and development of PE through the bridge relationship of miR-31-5p, we therefore darw the conclusion that circRNA was associated with lncRNA through miRNA. In addition to that, lnc00511 functioned as a molecular spong for miR-29b-3p,antagonizing its ability to repress Cyr61 protein translation, and meanwhile overexpression of lnc00511 promoted trophoblast cell proliferation, migration and invasion [177].It is through this network that miRNA, lncRNA and circRNA are inseparable and jointly promote the occurrence and development of PE.

Although numerous studies have confirmed the differential expression of ncRNAs in placental tissues, and their pathogenic mechanism in PE, studies on ncRNAs in the peripheral blood, especially circRNAs and lncRNAs, remain scarce. More research is required to elucidate the key role of ncRNAs in PE, because they are potential stable biomarkers for the diagnosis of the disorder.

Conclusions

The present review summarizes the expression patterns of ncRNAs, i.e., microRNAs (miRNAs), long noncoding RNAs (lncRNAs), and circular RNAs (circRNAs), and the mechanisms by which they affect PE. We examine the clinical significance of ncRNAs as biomarkers for the diagnosis of PE, and discuss the contributions made to PE by genetic polymorphisms and epigenetic ncRNA regulation. We believe that our study makes a significant contribution to the literature because it highlights the clinical value of ncRNAs as noninvasive biomarkers of PE.