The impact of early pregnancy metabolic disorders on pregnancy outcome and the specific mechanism

In early gestation, the embryo depends on glycolysis to produce adenosine triphosphate (ATP), and the most important source of the required substrates is glycogen stored in the endometrium [15]. To meet the increased metabolic needs, glucose homeostasis during pregnancy is altered, especially by the transient state of insulin resistance, which is compensated by the proliferation of pancreatic beta cells and the increasing insulin secretion capacity stimulated by glucose [16‐18]. On one hand, prolactin, as a known regulator of beta-cell growth and function, acts at multiple scales to prepare the mother for the new demands associated with the offspring. As pituitary prolactin secretion ceases, serum concentrations of leptin, oestradiol, progesterone, and other placental hormones increase with gestation, and these hormones also work together to maintain insulin resistance during pregnancy [19, 20]. On the other hand, due to the high substrate demand and the inefficiency of foetal gluconeogenesis, the fetoplacental unit forms a system for the rapid transfer of glucose from the maternal blood to the placenta. This transfer system is mainly associated with the expression of glucose transporter proteins (GLUTs), which promote insulin responsiveness [21]. There are six GLUTs confirmed in the placenta to date: GLUT1, 3, 4, 8, 9, and 12, with GLUT1, 3, and 4 assuming a significant role [22, 23]. The characteristics and regulatory factors of selected GLUTs in human placenta have been described in detail in the review by Stanirowski PJ et al. [24]. The expression of GLUT1 increased with the progression of pregnancy [25], whereas the expression of GLUT3 was significantly higher in hypoxic conditions in early gestation and diminished in late gestation [26, 27]. In general, the maternal body is regulated by several hormones to maintain the state of insulin resistance during pregnancy, while increased insulin secretion promotes the expression of GLUT in the foetal placental unit to provide adequate glucose and energy for the foetus.

Changes in maternal lipid metabolism during human pregnancy may be classified according to the anabolic and catabolic periods. Of these, the anabolic period occurs in the initial six months of human pregnancy and is dominated by increasing lipid deposition in maternal tissues [10, 12, 28]. At the macroscopic scale of lipid synthesis, during this period, the mother usually increases her food intake and the accumulation of body fat. At the molecular scale, research has shown that in rats, the conversion of glucose to fatty acids and glycerin gradually increases by the 20th day of gestation [29]. In addition, the activity of lipoprotein lipase (LPL) was enhanced in plasma; in contrast, its activity in adipose tissue did not change significantly. The functions of LPL differ between the two locations. In plasma, LPL can promote the absorption of lipids by hydrolysing triglyceride-rich chylomicrons and very low-density lipoproteins [14, 30], while LPL in adipose tissue contributes to the deposition of fat [31]. In late pregnancy, multiple hormones, including insulin, progesterone, cortisol, prolactin, oestrogen, and leptin [32, 33], mediate increased mRNA expression and the activity of hormone-sensitive LPL in white adipose tissue, thereby promoting lipolysis [34]. Thus, maternal lipid synthesis and accumulation increase in early gestation (increased intake, increased gluconeogenesis, increased plasma LPL activity), with little change in catabolism (little change in LPL activity in adipose tissue), while catabolism increases in late gestation (promoting LPL activity in adipose tissue). In contrast, the mother promotes LPL activity in adipose tissue in late pregnancy through the combination of multiple hormones, which increases lipolysis.

The maternal demand for amino acids increases during pregnancy for the establishment of the metabolic microenvironment of the endometrium in preparation for implantation and early pregnancy. Increased amino acid metabolism provides not only required proteins for exponential foetal growth, but also intermediate metabolites that promote multiple biosynthetic pathways [35]. Most amino acids are found at higher concentrations in foetal plasma than in maternal plasma, suggesting their active accumulation in the syncytial trophectoderm [36, 37]. The active uptake of neutral amino acids by syncytiotrophoblast cells is mainly mediated by the combination of sodium-coupled neutral amino acid transporters (SNATs) and non-sodium-dependent L-type amino transporters (LATs). The expression and activity of SNAT, the A system, increases with gestational age and foetal size [38, 39]. Meanwhile, the LAT, or L system, is present in only a few normal tissues and is primarily responsible for the transport of essential amino acids [40]. As a consequence, circulating amino acid concentrations increase during pregnancy and are actively transported into foetal tissues through special placental amino acid transporters (the A and L systems), which provide amino acids for foetal growth and development.

In the female reproductive system, physiological levels of reactive oxygen species (ROS) serve an important regulatory role through various signal transduction pathways in folliculogenesis, oocyte maturation, the endometrial cycle, embryogenesis, and pregnancy [41]. Under normal physiological conditions of the placenta, the balance between the elements of the intrinsic redox reaction is maintained mainly through antioxidant reactions mediated by the Keap1-nuclear factor-erythroid 2 related factor 2 (Nrf2) pathway [42‐44]. Furthermore, in the hypoxic environment of the placenta during gestation, the hydroxylation of hypoxia-inducible factor-1α (HIF-1α) (as its common degradation pathway) decreases, leading to HIF-1α accumulation and nuclear translocation [45]. HIF-1α promotes the transcription of the forkhead box protein P3 (FoxP3) gene and the production of regulatory T cells under hypoxic conditions, which promotes immune tolerance and reduces oxidative stress [46, 47]. In summary, the balance between ROS and antioxidants during pregnancy is essential and is mainly maintained by Keap1/Nrf2 and FoxP3 activated by HIF-1α in the specific hypoxic environment of the placenta.

A previous study found that women with recurrent miscarriages are more likely to have abnormal glucose metabolism [48]. In the following, we detail the studies related to abnormal glucose metabolism and abortion through three aspects: (i) glycogen synthesis; (ii) glycolysis and HIF; and (iii) CD39- and CD73-mediated ATP metabolism.

In terms of glycogen synthesis, data show that glycogen accumulates tenfold to meet the energy metabolic substrate requirements during pregnancy, and this process primarily relies on GLUT or sodium-glucose cotransporter (SGLT) [49, 50]. Studies have proven that SGLT1 gene and protein expression is significantly reduced in the endometrium of RSA patients during the implantation window [51, 52]. This was corroborated in the SGLT1-deficient mouse model, in which endometrial glycogen, litter size, and pup birth weight were lower than those of wild-type mice. This leads us to conclude that SGLT1 deficiency in the human endometrium at implantation can lead to miscarriage and intrauterine growth restriction through decreased glycogen synthesis.

It has been relatively well demonstrated that HIF-1α is involved in the regulation of glucose metabolism homeostasis under hypoxic conditions. It acts as an oxygen-sensitive transcriptional activator and can induce the transcription of a variety of genes related to gluconeogenesis [53]. Lactate (LA) is an important metabolite in hypoxia-inducible factor (HIF)-mediated glycolysis and is synthesized by lactate dehydrogenase A (LDHA) upon activation by highly expressed HIF-1α under hypoxic conditions [54]. Lactate can act as an active metabolite in physiological, immunological, and cell-biological regulation through the mediation of the monocarboxylate transporter protein (MCT) [55‐57]. It has been shown that LA content is significantly elevated in the decidua of RSA patients. LA enhances inducible nitric oxide synthase (INOS) expression in a HIF-1α-dependent manner, which in turn promotes M1 polarization of decidual macrophages, leading to the disruption of immune tolerance to trigger miscarriages [58]. The studies of mouse miscarriage models in this manuscript have also shown that blocking LA uptake with AZD3965 (MCT-1 inhibitor) could improve pregnancy outcomes, suggesting that MCT-1 could be a potential therapeutic target for RSA. Since the HIF family plays an important role in regulating glucose metabolism homeostasis, its degradation pathways are of equal interest [59, 60]. It was found that a decrease in chorionic succinate, an intermediate product of the tricarboxylic acid cycle during pregnancy, could promote HIF-1α degradation via the PHD-VHL pathway by promoting the hydroxylation of HIF-1α [61]. This leads to a decrease in HIF-1α, which inhibits angiogenesis, invasive migration of trophoblast cells and glycolysis and ultimately causes RSA. It is evident that while HIF regulates glucose metabolism homeostasis, the glycolytic products LA and succinate can also influence pregnancy outcomes by affecting HIF levels.

Abnormalities in the metabolism of ATP, the product of glucose metabolism, which is the direct provider of energy during pregnancy, are also closely associated with miscarriage. The ATP adenosine metabolic pathway modulated by CD39/CD73 has recently been suggested to play a significant role in immunosuppression [62]. CD39 hydrolyses ATP and ADP to produce AMP, and the membrane-bound 5’-nucleotidase CD73 further hydrolyses AMP into adenosine [63]. Therefore, CD39 and CD73 can convert proinflammatory immune cells driven by ATP to anti-inflammatory immune cells evoked by adenosine, thus causing immunosuppression [64, 65]. It was shown that downregulation of the TGF-β/mTOR/HIF-1α pathway leads to the inhibition of ATP-adenosine metabolism and causes a decrease in the number of CD39⁺ and CD73⁺ cells at the maternal–foetal interface [66]. This depresses the proliferation and invasion of trophoblast cells, reduces apoptosis and increases the cytotoxicity of decidual natural killer (dNK) cells, which contributes to RSA.

In conclusion, reduced GLUT protein expression during pregnancy affects glycogen transport synthesis in the endometrium, which leads to insufficient substrates for gluconeogenesis and induces RSA. After reviewing points ii and iii, we found that HIF, as an important factor regulating glucose metabolism homeostasis in the special hypoxic environment of pregnancy, can also be regulated by glucose metabolites, thus leading to miscarriages (details in Fig. 1). HIF activation by LA can promote macrophage M1 polarization to disrupt immune tolerance, leading to miscarriages (details in Fig. 1A); decreased succinate expression promotes the degradation of HIF-1α, which in turn inhibits angiogenesis, trophoblast invasion and migration and glycolysis (details in Fig. 1B); the downregulation of the TGF-β/mTOR pathway reduces HIF-1α expression, which inhibits ATP-adenosine metabolism and increases dNK cell toxicity (details in Fig. 1C).

Several studies have shown that abnormal lipid metabolism is associated with spontaneous abortion and pregnancy complications such as endothelial injury, preeclampsia, and gestational hypertension [67‐69]. In this section, we detail the studies related to abnormal lipid metabolism and abortion through three aspects: (i) peroxisome proliferator-activated receptors (PPARs); (ii) total polyunsaturated fatty acids (PUFAs), TG and inflammation; and (iii) the arachidonic acid metabolic pathway, leptin and myometrial contractions.

PPARs (PPAR-α, PPAR-β/δ, and PPAR-γ) are members of the nuclear receptor superfamily, acting as ligand-inducible transcription factors and playing crucial roles in glucose and lipid metabolism [70]. Studies have shown that PPAR deficiency inhibits fatty acid uptake and expression of fatty acid transporter proteins and promotes the production and secretion of proinflammatory cytokines, leading to impaired placental development and functional impairment [71‐73]. This might account for the high abortion rate in PPARα knockout mice in animal experiments and the fact that mutations in PPARγ and PPARδ cause infertility [74‐76]. In addition, a study has demonstrated that PPAR-agonists have antidiabetogenic, anti-inflammatory, and antioxidant effects, which are all potentially beneficial in the treatment of gestational diabetes mellitus (GDM) [77]. This suggests another possible mechanism of action for improvement of pregnancy outcomes. However, determining whether PPAR gene polymorphism is relevant to the development of GDM still requires further evidence [78].

PUFAs, especially n-6 fatty acids, contribute to lipid peroxidation fragility and the proinflammatory effects of the corresponding peroxidation products, which in turn increase oxidative stress, alter lipid metabolism, and disrupt hormones [79‐81]. This would lead to lower probability of pregnancy and live birth and increased risk of miscarriage [82]. In addition, it has been proposed that patients with insulin resistance (IR) have significantly higher triglyceride (TG) levels (which brings about an increased ratio of CD3⁺CD4⁺) and numbers of CD3⁺CD8⁺ lymphocytes, reduced insulin sensitivity, and induction of metabolic inflammation, resulting in RSA [83].

Metabolomic analysis revealed that the expression of the arachidonic acid metabolic pathway-related genes cyclooxygenase-1 (COX-1), cyclooxygenase-2 (COX-2), prostaglandin F2α receptor (PTGFR), and thromboxane A2 receptor (TBXA2R) was significantly increased in RSA patients. Related animal experiments have shown that abnormal expression of COX genes and TBXA2R can cause uterine contraction by regulating the cytoplasmic phospholipase A2α (PLA2α)/COX-2 pathway in endometrial stromal cells and inducing increased prostaglandin synthesis, ultimately leading to RSA [84‐86]. All of these findings suggest that modulation of the arachidonic acid metabolic pathway may be a prospective therapeutic strategy to alleviate symptoms in women with RSA. It has also been reported that the adipokine leptin can inhibit spontaneous and oxytocin-induced myometrial contractions by increasing NO and cGMP through stimulation of short-type leptin receptors and activation of the NO pathway in a JAK/STAT-dependent manner [87]. However, this trial demonstrated the inhibitory effect of leptin on uterine contractions only in late pregnancy, and it is not yet known whether it can be used in early pregnancy to reduce the incidence of spontaneous miscarriage.

In conclusion, the abnormal lipid metabolism provoked by the decrease in PPAR and the increases in PUFA and TG can promote inflammation and oxidative stress, which would contribute to miscarriage. In contrast, the arachidonic acid pathway increases prostaglandin synthesis to cause myometrial contractions, leading to miscarriage, which might be ameliorated by pathway modulation and leptin.

The regulation of amino acid metabolism in the endometrium is one of the most important metabolic processes to meet the increased nutritional demands of early pregnancy. It provides not only protein components but also intermediate metabolites that promote multiple biosynthetic pathways, which help to establish the metabolic microenvironment of the endometrium in preparation for implantation and early pregnancy [35, 88]. Therefore, we introduce the effects and mechanisms of amino acid metabolism on different tissues during pregnancy from three aspects: (i) autophagy of the endometrium; (ii) apoptosis of trophoblast cells; and (iii) inflammation of the maternal–foetal interface (Fig. 2).

Metabolomic analysis identified 19 decreased metabolites and 22 increased metabolites in the endometrium of women with recurrent miscarriages, including decreases in pyruvate, glutamate and succinate as well as an increase in glutamine [89]. The reduction of various amino acids accounts for the accumulation of uncharged transfer RNA (tRNA), which binds a protein kinase called eukaryotic translation initiation factor 2α kinase 4 (eIF2αK4) and activates the general control nonderepressible 2 (GCN2)/eIF2α/transcription activation factor 4 (ATF4) pathway, thereby inhibiting protein synthesis and inducing autophagy [90]. In addition, glutamine metabolism can induce autophagy through the release of NH3 [91]. This suggests that abnormal collusion between amino acid metabolism and autophagy may contribute to an impaired endometrial microenvironment, which in turn induces RSA (Fig. 2A).

It has been reported that nitric oxide (NO) can inhibit trophoblast apoptosis, thereby reducing adverse pregnancy outcomes [92]. In the placenta, NO is promoted by L-arginine via inducible NOS (iNOS) and endothelial NOS (eNOS) [93]. This process can be selectively inhibited by asymmetric dimethylarginine (ADMA), which originates from type I protein L-arginine methyltransferase (PRMT)-mediated protein degradation [94, 95]. A mechanistic study on the effect of decidual macrophages (DMs) on the regulation of trophoblast apoptosis showed that DM can promote trophoblast apoptosis by activating the PRMT3/ADMA/NO pathway, which reduces the concentration of NO in metaphase and leads to RSA [96]. Furthermore, animal experiments have also revealed that the PRMT3 inhibitor SGC707 could significantly reduce the embryo uptake rate in a mouse model prone to miscarriage, which demonstrates that the PRMT3/ADMA/NO pathway could be a potential target for the treatment of miscarriage. However, clinical and safety trials still need to be completed (Fig. 2B).

Homocysteine, a thiol-containing amino acid, is involved in sulfation and methylation metabolic pathways. Studies have shown that elevated homocysteine levels can activate proinflammatory pathways through endothelial dysfunction and lead to leukocyte-endothelial cell interactions and leukocyte recruitment, causing vascular inflammatory changes that contribute to microembolism at the maternal–foetal interface, which ultimately results in RSA [97‐99]. In addition, the trans-sulfuration pathway of homocysteine metabolism requires cystathionine beta-synthase (CBS) and the cofactor vitamin B6. Since the CBS gene is a target of the vitamin D receptor, vitamin D deficiency can increase homocysteine levels by decreasing CBS, inducing increased NK cell cytotoxicity, which in turn further promotes the inflammatory immune response at the maternal–foetal interface and leads to RSA [100‐102]. This demonstrated the relationship between amino acid metabolism and inflammation in pregnancy and suggested the necessity of vitamin D and B6 supplementation during pregnancy in RSA patients (Fig. 2C).

In conclusion, abnormal amino acid metabolism can act in several pregnancy-related tissues, leading to adverse pregnancy outcomes. Abnormal amino acid metabolism in the endometrium can induce autophagy leading to an impaired microenvironment; at the maternal–foetal interface, it can be regulated by DM to activate the PRMT3/ADMA/NO pathway and promote trophoblast apoptosis; and increased homocysteine levels at the maternal–foetal interface can induce the inflammatory immune response, leading to abortion, which can be alleviated by vitamin D supplementation.

Research on hyperandrogenemia and insulin resistance-related miscarriage in women showed that the important factors contributing to embryonic damage in polycystic ovary syndrome (PCOS)-like conditions include excessive production of ROS, mitochondrial dysfunction, and the inhibition of superoxide dismutase 1 (SOD1) and Keap1/Nrf2 antioxidant responses in the placenta [103]. This suggests that the disturbance in the balance between oxidative stress (ROS production) and antioxidants is responsible for the initiation and development of pathological processes affecting female reproduction [104, 105]. Therefore, we elaborate on the mechanisms by which dysregulation of redox reaction balance contributes to abortion in terms of abnormalities in both oxidative stress and antioxidants.

It has been reported that increased levels of malondialdehyde and lipid peroxides in placental tissue increase ROS, which can lead to sudden and premature formation of maternal placental perfusion while damaging the trophectoderm, resulting in RSA [106, 107]. In addition, activation of the Fas/FasL signalling pathway in villi tissue can promote oxidative stress-induced apoptosis of trophoblast cells, contributing to miscarriage. The molecular mechanism is associated with inhibition of the Notch1 signalling pathway and upregulation of epithelial cadherin (E-cadherin), soluble vascular endothelial growth factor receptor 1 (sFlt-1), and vascular endothelial growth factor (VEGF) expression [108]. It follows that excessive activation of oxidative stress can lead to premature placental perfusion, induction of apoptosis, and destruction of the trophectoderm, which would result in miscarriage.

Changes in the consumption of antioxidants can also lead to disturbances in the balance of pro-oxidant and antioxidant factors, which could lead to miscarriage [109, 110]. Glutathione and glutathione peroxidase are antioxidants that neutralize free radicals and lipid peroxides to maintain intracellular homeostasis and redox balance. In a large case‒control study on genetic polymorphisms of the glutathione family enzyme glutathione S-transferase (GST), an elevated risk of RSA was found to be associated with increased oxidative stress due to null polymorphisms of the GSTM1 and GSTT1 genotypes in RSA patients [111‐113]. Nonetheless, epidemiological studies and related experiments have shown that sulfur dioxide (SO2) and its derivatives can inhibit trophoblast cell viability and the ROS/IL-6/STAT3 pathway, interfere with cell proliferation by blocking the cell cycle, induce apoptosis, disrupt the secretion of inflammation-related cytokines, and inhibit cell invasion and migration, leading to miscarriage and pregnancy complications [114]. This reflects the fact that both reduced and inappropriate use of antioxidants can lead to adverse pregnancy outcomes.

Although it is controversial whether antioxidant supplementation could change pregnancy outcomes [115‐117], some new findings on antioxidant drugs have been achieved. A study in 2020 revealed that astaxanthin significantly alleviated poor glucose tolerance and beta-cell insufficiency and improved pregnancy outcomes by restoring the Nrf2/heme oxygenase-1 (HO-1) antioxidant pathway in the livers of gestational diabetic mice, inhibiting oxidative stress in vivo, and enhancing the activity of antioxidant enzymes [118]. It has also been reported that alpha lipoic acid (ALA) and its reduced form dihydrolipoic acid (DHLA) may improve pregnancy outcomes through specific stimulatory activity on Nrf2-dependent gene transcription and by the inhibition of NF-kB activity [119, 120], but more patient samples and further studies on safety in pregnancy and the pharmacokinetics of the vaginal pathway are still needed.