To our knowledge, this is the first study to investigate the longitudinal changes of oxidative stress and PON1 lactonase activity and status during gestation in women undergoing ART. We found that the lactonase activity and level of PON1 gradually decreased with pregnancy progression. However, TAC levels significantly increased during the 2nd and 3rd trimesters. NLA of PON1 and oxidative stress parameters, including TOS, OSI, MDA, and HCY, significantly increased in the ART and control groups before delivery. Moreover, we found that the lactonase activity of PON1 in the 2nd and 3rd trimesters and before delivery, the NLA of PON1 in the 1st, 2nd, and 3rd trimesters and before delivery, and TAC in the 1st, 2nd, and 3rd trimesters were significantly higher in the ART women than in the control women. Besides, we also showed that the women undergoing ART had an adverse lipid metabolic profile in the 1st trimester and a higher FGlu, FIns, HOMA-IR, and TG levels before delivery compared with the control women. We consider that an unfavorable glycolipid metabolism and a compensatory increase in PON1 activities and TAC during pregnancy in the women undergoing ART may be the partial reasons that hindered a natural conception and increased the risk of metabolic and oxidative stress-related diseases including pregnancy complications in these women.

Pregnancy is a stress state related to an atherogenic lipid profile in women [11, 24]. Pregnancy produces transient insulin resistance manifested as elevated postprandial glucose and fasting lipid levels, increased inflammatory response, and enlarged circulating blood volume [25]. Maternal metabolic changes during early pregnancy are mostly increasing lipid synthesis and body fat accumulation from hyperphagia arising in response to a hormonal stimulus [26, 27]. On the other hand, again due to hormones, there can be an initial weight loss due to hyperemesis. As pregnancy progresses, this accumulation stops or even declines during the 3rd trimester as a consequence of both enhanced lipolysis and decreased fatty acid intake in adipose tissue, which in turn causes the increases of fatty acids into the liver, the enhancement of very low-density lipoprotein synthesis and secretion, and the elevation of lipoprotein levels in circulation [27]. Meanwhile, placental hormones and other mediators facilitate peripheral insulin resistance [26]. It was previously reported that a 40–50% reduction in insulin sensitivity and a compensation of 200–250% increase in insulin secretion by pancreatic β cells to maintain maternal euglycemia [28]. Consistent with previous reports [29, 30], in this study, we showed that TG, TC, and LDL-C levels gradually increased with pregnancy progression and peaked during the 3rd trimester or before delivery in the ART and control groups. We further demonstrated that the longitudinal changes of HDL-C, AI, apoA1, apoB, FIns, apoB/apoA1 ratio, and HOMA-IR during gestation also represented similar trends. However, unlike the tendency of physiological changes in the control women, the FGlu, FIns, and HOMA-IR during gestation reached the highest levels before delivery in the ART group. The TG, TC, LDL-C, AI, apoB, and apoB/apoA1 ratio in the 1st trimester and the FGlu, FIns, HOMA-IR, and TG before delivery were significantly higher in the ART women than in the control women. Our findings suggest that ART women have an adverse lipid metabolic profile and a fragile balance in glucose metabolism, corroborating previous reports that women undergoing ART have a higher risk of GDM [6, 31].

Another characteristic of pregnancy stress is increased oxidative stress [11, 24, 32]. Implantation and placentation in early pregnancy resemble “an open wound in the uterus”, and the following inflammatory environment is essential for repairing the uterine epithelium and removing cellular debris resulting from blastocysts and trophoblasts [33]. The systematic inflammation results in increased oxidative stress in the 1st trimester and maintained a redox balance in the 2nd and 3rd trimesters for rapid fetal growth and developed again until delivery, which needs a strong inflammatory response to promote uterine contractions and delivery of baby and placenta [33, 34]. Pereira et al. [34] reported increasing oxidative stress levels in the placenta caused by high mitochondrial activity and increased ROS production. Measurement of TOS and TAC can assess the total amount of oxidant and antioxidant molecules present in serum, respectively [9, 11]. OSI, the ratio of TOS to TAC, can estimate the redox state of the body [9]. MDA is an end-product of lipid peroxidation and the hallmark of ROS-induced injury [35]. A study has reported an increased TAC and TOS during the 3rd trimester and before delivery in normal pregnancy [32]. This study showed that TOS, OSI, and MDA levels significantly increased before delivery. However, TAC increased and peaked at 2nd or 3rd trimester and decreased before delivery in the ART and control groups.

Oxidative stress is involved in the pathophysiological process of female infertility. Diseases causing infertility, such as PCOS, obesity, advanced maternal age, and reproductive system inflammation, are associated with oxidative stress [34]. In addition, operations and medications in ART could lead to oxidation-reduction imbalance. A study reported lower superoxide dismutase and higher MDA and sulfhydryl groups after ovarian stimulation, indicating increased oxidative stress after ovarian stimulation [36]. Besides, applying gonadotropins in ART may generate large amounts of ROS and lead to oxidation-reduction imbalance, directly affecting maternal serum’s oxidative stress indicators, such as antioxidant capacity, susceptibility to oxidation in vitro, and antioxidants in vivo [37]. Our findings showed that the ART women had significantly higher TAC during pregnancy than the control women, suggesting a compensatory increase in antioxidant capacity exists in these women.

PON1 plays antioxidant and anti-atherogenic roles by hydrolyzing lipid peroxides and homocysteine thiolactone (HTL) depending on its lactonase and HTase activities [15, 16, 19, 38]. HCY, along with HTL, is an essential precursor of oxidative stress, activating autoimmune, enhancing thrombosis response, and interacting with LDL to form foam cells [16, 38]. Studies found a significant decrease in serum POase activity of PON1 during the 3rd trimester and before delivery in normal pregnancy [30, 32]. In this study, we found that the lactonase activity and level of PON1 gradually decreased with pregnancy progression. However, the NLA of PON1 and HCY levels significantly increased before delivery in the ART and control groups. Furthermore, we found that the ART women had significantly higher lactonase activity and NLA of PON1 during pregnancy compared with the control women, suggesting that a compensatory increase in PON1 activities in these women.

PON1 activities and status are influenced by various factors, such as genetic polymorphisms of PON1, oxidative stress, lipoproteins, and apolipoproteins [11, 21, 39, 40]. Genetic variants of PON1 play vital roles in regulating the expression and/or activities of PON1 and may explain more than 60% of the individual differences in enzyme protein levels and activities [12, 21, 41]. Among PON1 SNPs known, the SNP C-108T, a binding site for the transcription factor Sp1 in the promoter region, has the greatest impact on the gene expression of PON1 and can account for 22.8% of the observed variability [21, 42, 43]. The SNP Q192R, a common variation in the exon region, mainly affects the enzyme activities of PON1 in a substrate-dependent manner [11, 15, 44]. The Q isoform hydrolyzes soman, sarin, diazoxon, and lipid peroxides more efficiently [15, 44], whereas the R isoform hydrolyzes paraoxon (POase activity) more rapidly in vitro [44, 45]. Oxidative stress has a reversible effect on the levels and activities of PON1. Previous studies found that serum TAC and MDA levels correlated positively with the lactonase activity and levels of PON1, respectively, suggesting that the elevated TAC, PON1 lactonase activity, and PON1 levels might compensate for increased oxidative stress in women with GDM and their neonates [11, 12]. However, hyperglycemia and excessive oxidant molecules may cause glycosylation and oxidative damage to PON1, reducing enzyme activity [46,47,48]. PON1 activities are also affected by apolipoproteins. Both apoA1 and apoE combined with PON1 in HDL could help stimulate the PON1 lactonase activity and strengthen the enzyme stability [39, 41]. In this study, the lactonase activity and NLA of PON1 were significantly higher at different stages of pregnancy in the ART group than in the control group. Consequently, the genotypic frequencies of the PON1 C-108T and Q192R polymorphisms were not significant differences between the ART and control groups, suggesting that the compensatory increase in PON1 activities might be mainly related to an unfavorable state of glycolipid metabolism and oxidative stress.

Some limitations need to be acknowledged. First, the participants in this study are limited to women with AMA; however, age might affect oxidative stress and metabolic indicators and was an independent risk factor for many pregnancy complications. A comparison between young and AMA women would help figure out the compact of age. Second, this was a single-center study, and we could not include more participants. Limited by the relatively small sample size, we failed to perform subgroup analyses, for example, reasons for undergoing ART and different technologies such as in vitro fertilization and embryo transfer, intracytoplasmic sperm injection, and preimplantation genetic diagnosis. Third, we did not genotype PON1 polymorphisms in some participants (ART = 11; control = 5) due to a missing DNA sample, which might have affected the power of these parameters.