1. Introduction

Preeclampsia (PE) is a pregnancy hypertensive disorder that affects approximately 5–7% of pregnancies worldwide ^[1][2]. There is no cure for PE except for the delivery of baby. The existing literature provides compelling evidence for the role of mitochondrial oxidative stress in PE. While antioxidant therapies show a benefit in the animal models of PE, clinical trials on antioxidants (vitamins C and E) showed disappointing results ^[3][4]. The possible reason behind these unsuccessful trials were thought to be due to the inability of the antioxidants to reach the site of oxidative stress (i.e., mitochondrial matrix). It has been recently shown that targeting oxidative stress within the mitochondria using mitochondrial specific antioxidants (MitoQ and MitoTempo) attenuate hypertension and improve fetal outcomes in an animal model of PE ^[5]. The studies that followed these important findings identified a role for circulating pathogenic factors (i.e sFlt1, NK cells, CD4+T cells, TNF-α, AT1-AAs) in causing mitochondrial dysfunction in animal models of PE ^[6][7][8]. However, underlying molecular mechanisms are not well understood. Thus, further studies are vital for exploring these mechanisms to identify targets to develop new treatments for PE.

2. Mitochondrial Dysfunction in Preeclampsia

2.1. Clinical Studies

The first report that a family with mitochondrial (mt) dysfunction had a high incidence of preeclampsia/eclampsia was published three decades ago ^[9]. Since then, a huge amount of literature has centered on the role of mitochondrial function or dysfunction in PE has accumulated. The studies published shortly after this first report showed evidence of abnormal mt morphology, oxidative stress, and reduced Complex IV activity in PE placental sections ^[10][11][12]. Further, the recently published study documented term PE placental mitochondria in the presence of glutamate, malate, and ADP (Complex I mediated state 3), respire at a significantly low rate in comparison to normal pregnant placental mitochondria. Further, either preterm or term PE placental mitochondria exhibit low maximal respiration rates (induced in the presence of an uncoupler; FCCP) in comparison to normal pregnant mitochondria. Similarly, in the presence of succinate and ADP, both term and preterm placental mitochondria showed a significantly low respiration rate in comparison to normal pregnant mitochondria. Moreover, while preterm PE placental mitochondria showed significant reductions in maximal respiration rate, term PE placenta showed a trend towards reduction. Furthermore, it was found that both expression and activity of Complex IV were significantly reduced in term PE placental mitochondria. Interestingly, changes were not found in expression of other complexes, albeit reduced Complex IV could explain reduced respiration rates seen in term PE placental mitochondria. However, two independent studies by other research groups revealed that preeclamptic placental state 3 (preterm PE; GA-30 weeks) or maximal respiration (term PE; GA-37 weeks) rates were significantly increased in comparison to control normal pregnancies ^[13][14]. Furthermore, while term PE placenta exhibits increased Complexes II and III expressions, no changes were seen in any of the complex levels in preterm PE placenta.

Holland et al. reported preterm PE placental mitochondria exhibit low H₂O₂ production in comparison to control pregnancies, and further showed that SOD expression in PE placental mitochondria was significantly low ^[14]. The real-time mtROS production in isolated mitochondria was also assessed here. Interestingly, both term and preterm PE mitochondria showed significantly low mtROS production in comparison to normal pregnant mitochondria ^[15]. This is particularly an interesting finding as most often reduced respiration rates are associated with increased ROS generation. This suggests that low H₂O₂ levels seen in preeclamptic placentas may be a result of low expression of ROS metabolic systems. The excessive ROS produced within the mitochondria can cause lipid peroxidation, ETC damage, mtDNA release, ultimately apoptosis and cell death ^[16]. In fact, a few earlier studies reported increased amounts of lipid peroxides, isoprostanes, protein carbonylation, and malondialdehyde (MDA) levels in preeclamptic placenta ^[17][18][19]. It is important, however, to note that oxidative stress is not only a result of increased pro-oxidants but also can result from reduced antioxidant levels altering pro vs. anti-oxidant balance.

2.2. Animal Studies

Following early clinical observations of evidence for mitochondrial dysfunction in PE, several research groups set out to explore mitochondrial oxidative stress in animal models of PE. The research were the first to establish a functional link between hypertension and mitochondrial dysfunction in PE. In the published study utilizing a RUPP (reduced uterine perfusion pressure) rat model, it was demonstrated that preeclamptic rats exhibit reduced mitochondrial respiration in isolated mitochondria from placenta or kidney ^[5]. Interestingly, this reduced respiration was associated with elevated ROS production in both placental and renal mitochondria. Assessing electron transport chain activity further revealed that Complexes I, II, and IV were impaired in mitochondria of preeclamptic rats. Subsequently, the preclinical studies exploring the mechanistic contribution of mtROS in PE consistently reproduced these novel findings ^[7][8]. Similarly, a recent study by Yang et al. showed that RUPP mice exhibit increased placental mitochondrial oxidative stress along with mitochondrial damage ^[20]. Moreover, because the kidney is an important regulator of blood pressure and one of the affected organs in PE patients, the exploration was expanded to renal mitochondria. As this research suspected, PE rats displayed impaired mitochondrial function and increased ROS production in isolated renal mitochondria. These findings were confirmed by a recently published study that documented increased mtROS production in kidneys from RUPP preeclamptic rats ^[21]. These exciting findings prompted research groups to explore mitochondrial bioenergetics in organ systems beyond placenta or kidneys. For instance, a recent study published by Booz et al. demonstrated that isolated cardiac mitochondria from RUPP rats show reduced mitochondrial respiration and Complex IV activity ^[22]. Further, Sánchez-Aranguren et al. reported that overexpression of sFlt1 in heme oxygenase-1-deficient mice causes mitochondrial dysfunction in cardiac mitochondria ^[23]. Collectively, studies reporting mitochondrial dysfunction in extra placental organ systems suggest a role for circulating factors released in response to placental ischemia mediating maternal mitochondrial damage.  

2.3. In Vitro Cell-Based Studies

The studies exploring the mitochondrial role in PE pathology also encompasses mechanistic exploration utilizing cell-based assays. While trophoblast cell lines were used to assess placental mitochondrial function, vascular endothelial cell lines were used to understand the effect of placental ischemia-induced circulating factors on mitochondrial dysfunction in the maternal systemic vasculature. Sánchez-Aranguren et al. showed that HTR-8/SVneo cells (trophoblast cell line) treated with 2% PE patient serum exhibit reduced mitochondrial respiration and elevated mtROS production ^[6]. It was showed that HUVECs (human umbilical vein endothelial cells) treated with 10% RUPP rat serum caused increased mtROS production.^[5] In line with the finding here, McCarthy et al. reported that HUVECs treated with 3% plasma from PE patients not only increased mtROS production but also reduced cellular mitochondrial respiration ^[24]. Further, Sánchez-Aranguren et al. demonstrated that treating either HUVECs or HTR-8/SVneo cells with 2% PE patient serum reduced cellular respiration, and increased mtROS production ^[6]. Moreover, the recently published study confirmed that PE patient serum causes impaired mitochondrial function and excessive mtROS generation in HUVECs ^[25]. Taken together, these studies suggest that circulating factors released in response to placental ischemia in PE placenta leads to mitochondrial dysfunction and ROS production in both placental and vascular cells. Specifically, in regard to the findings here, although real-time placental mtROS was not elevated in the patient cohort, it is demonstrated that soluble factors from PE women stimulate mt ROS in the vascular endothelium, thus further supporting the role for peripheral mtROS in patients with PE.

3. Mediators of Mitochondrial Dysfunction and Excessive ROS Production in Preeclampsia

The studies exploring pathological mediators and molecular mechanisms that underlie mitochondrial dysfunction and oxidative stress are still at their inception. The work over the past few years has identified inflammatory cells or mediators; activated NK cells, CD4+ T cells, AT1-AAs, and TNF-α as some of the crucial players in mediating mitochondrial oxidative stress in PE. It has been shown that administration of AT1-AAs (GD13–19) to normal pregnant rats causes elevated mtROS production in placenta and kidney ^[26]. Further, these findings were confirmed in the subsequent study, in which inhibition of AT1-AAs by coadministration with 7 amino acid peptide (n7AAc) was reported to reduce mtROS in preeclamptic rats ^[7]. Another downstream pathway that has been identified was NK cell activation. Granzymes A and B, serine proteases majorly secreted by activated NK cells, have been shown to cause mitochondrial dysfunction and mtROS in in vitro studies as well as in isolated mitochondria ^[27][28]. Further, the mechanisms underlying granzyme-mediated mitochondrial oxidative stress are shown to be mediated via cleavage of Complex I subunits ^[29] and caspase substrates such as BID. In support of these findings, it was shown that depletion of NK cells in PE rats improves mitochondrial function ^[8].  It has been previously demonstrated that AT1-AAs activate NK cells in pregnant rats ^[26]. Hence, it is possible that NK cell-induced mitochondrial oxidative stress is downstream of AT1-AA signaling via the AT1 receptor. Moreover, based on the existing literature on the ability of TNF-α to cause mitochondrial oxidative stress or hypertension, and reported moderately high levels of TNF-α in PE, we hypothesized that TNF-α causes mitochondrial oxidative stress in pregnant rats. When we infused TNF-α into normal pregnant rats, the increased mitochondrial oxidative stress along with a preeclamptic phenotype was noted ^[30]. Further, in a separate study, it was shown that administering etanercept (TNF-α inhibitor) to PE rats (single dose of 0.4 mg/kg, s.c, on GD18) reduces mtROS in both placental and renal mitochondria ^[31]. Further, it has been previously shown that TNF-α infusion in the pregnant rats increases circulating AT1-AA levels ^[32]. Thus, it is possible that TNF-α-mediated mitochondrial oxidative stress is both upstream and downstream of AT1-AA signaling through the AT1R; however, future studies blocking the activity of the AT1-AA in response to TNF-α are needed to answer this question. Soluble fms-like tyrosine kinase 1 (sFlt1), an antiangiogenic factor shown to mediate preeclamptic pathology, has also been shown to play a role in causing mt ROS in PE ^[6]. It was demonstrated that sFlt-1 stimulates placental and endothelial mitochondrial ROS and a role of progesterone- induced blocking factor (PIBF) to play a defensive role against hypertension and placental and endothelial mitochondrial function during pregnancy  ^[33].

In summary, the existing literature suggests that manifestation of mitochondrial oxidative stress in PE involves a complex interplay of AT1-AAs, activated NK cells, TNF-α, and sFlt1.

4. Mitochondrial Targeted Antioxidants in Preeclampsia

Although it has been four decades since the early findings on the pathological role of mitochondrial dysfunction in PE patients emerged, it was not until recently the investigators started embracing the potential of mitochondrial antioxidants to treat PE. The impetus to explore mitochondrial targeted antioxidants in PE in part stems from the literature on mitochondrial targeted antioxidants to treat hypertension in preclinical models. Dikalova et al. showed infusion of malate or mitoebselene (mitochondrial targeted H₂O₂ scavenger)-attenuated Ang-II induced hypertension in mice. Further, administration of mitochondria-targeted SOD mimetic (MitoTempo) significantly attenuated Ang-II-induced hypertension in mice. Furthermore, Graham et al. reported that 8-week treatment of spontaneously hypertensive rats with mitochondria-targeted antioxidant (MitoQ) significantly reduced systolic blood pressure by 25 mmHg ^[34]. Based on this literature, a few research groups in recent years, have begun to explore mitochondrial-targeted antioxidants in in vitro and in vivo preclinical models of PE.

4.1. MitoQ

It has been recently demonstrated that MitoQ reduced hypertension and improved fetal outcomes in preeclamptic rats ^[5]. In line with the findings, Yang et al. have recently showed administering MitoQ during late gestation to preeclamptic mice attenuated hypertension and kidney damage ^[20]. Further, It has been shown that endothelial cells treated with serum obtained from MitoQ treated preeclamptic rats completely reversed mtROS when compared to cells exposed to sera from untreated PE rats, supporting the theory that circulating factors play an important role in causing endothelial mitochondrial oxidative stress and dysfunction ^[5]. Additionally, MitoQ-treated preeclamptic rats showed significant improvement in respiration and ETC complex activities (Complex I and IV).

While these preclinical studies show promising data, furhter studies focusing on MitoQ disposition in the feto–placental unit using in vitro/ex vivo models would be entailed before considering mitoQ for clinical evaluation.

4.2. MitoTempo

RUPP rats treated with MitoTempo (at a dose of 1 mg/kg per day, GD 14–19) have been reported to show significant attenuation in elevated blood pressure ^[5]. Further, as seen with MitoQ, endothelial cells treated with sera from MitoTempo-treated PE rats showed significantly low levels of mtROS production, indicating the potential of MitoTempo in protecting endothelial cells from circulating factors. Commensurate with the findings, McCarthy et al. showed that endothelial cells (HUVECs) pretreated with MitoTempo (5 μM) for 2 hours prior to incubation with 3% plasma from PE patients resulted in significant reduction in mtROS production ^[24]. As seen in the models of hypertension, non-targeted Tempol showed benefits at a dose of 30 mg/kg/day in PE models; however, a therapeutic benefit was showed at a dose of 1 mg/kg/day, which was 30-fold lower than the non-targeted Tempol dose, indicating the importance of using mitochondrial-targeted Tempol in PE.

4.3. Ergothioneine

Williamson et al. showed that oral administration of ERG (at a dose of 25 mg/kg per day, GD11–19) attenuated hypertension and rescued fetal growth restriction in the RUPP rats. In addition, ERG treatment reduced circulating sFlt1 levels and mitochondrial H₂O₂ in kidneys of RUPP rats. Further, in a separate study published by the same group, ERG-treated RUPP rats showed higher levels of glutamylcysteine (metabolite associated with oxidative stress in mitochondria) and lower levels of metabolites associated with inflammation ^[35]. Collectively, findings from these two studies suggest that the antihypertensive and anti-inflammatory properties of ERG are mediated by preventing mitochondrial oxidative damage by its direct effects in the mitochondria. However, further studies exploring ERG beneficial effects on mitochondrial function in placenta and endothelial cells need to be conducted before ERG can be fully considered as a potential mitochondrial antioxidant for clinical use during pregnancy.

5.The Potential of Mitochondrial Antioxidants in Treating Preeclampsia

Although there have not been any clinical studies evaluating mitochondrial antioxidants in preeclamptic patients yet, the compelling preclinical data from multiple independent research groups are promising and certainly form a foundation for translational efforts of these molecules. Before considering any novel drug molecule to be a candidate for clinical evaluation in pregnancy, the impact of pregnancy-induced physiological changes and feto–placental transfer on the safety, pharmacokinetics, and efficacy should be demonstrated in animal models. In terms of an extensive characterization in animals and humans, MitoQ dominates all the other antioxidants discussed here (readers are encouraged to refer to the review on MitoQ by Smith et al ^[36]). In recent years, MitoQ has been shown to be safe in at least in four independent studies involving pregnant rodents. Further, both MitoTempo and ERG have been reported to be safe in pregnant animals, and this research was the only study that evaluated MitoTempo in pregnant animals. Importantly, no signs of maternal or fetal toxicity were observed at the dose. While pharmacokinetics of ERG have been studied in one study, pharmacokinetics of MitoTempo are unavailable at this time. Thus, further studies are warranted for a thorough understanding of the safety and disposition of these molecules in pregnant animals before considering these molecules for clinical evaluation in pregnant women.

6. Conclusions

The disappointing results from vitamin E and vitamin C clinical trials shifted the focus of PE research away from targeted antioxidant therapy in PE. However, it is widely known that delivering antioxidants directly to the source of oxidative stress could be an effective strategy to combat oxidative stress in several disease states. In fact, targeting mitochondrial oxidative stress using mitochondrial-specific antioxidants such as MitoQ and MitoTempo would attenuate hypertension and improve fetal outcomes in animal models of PE have been shown. Other antioxidants such as ERG that are not required to be tagged with chemical carriers (TPP) for mitochondrial delivery have also shown promise in PE animal models. Further, the safety and efficacy findings of these antioxidant molecules are encouraging and certainly form a strong foundation for further characterization and evaluation in PE clinical trials. However, until people have a better understanding of extenuating factors (obesity and race) and how they affect placental mt function, every one must proceed with caution and continue the preclinical studies before proceeding to utmost important clinical studies in the search for new therapeutics for an old disease.