Cell-free DNA as a potential biomarker for preeclampsia

KEYWORDS:

• Biomarker
• hypertension
• hypertensive disorders of pregnancy
• cell-free DNA
• preeclampsia
• pregnancy
• therapeutic target

1. Introduction

Preeclampsia (PE) is defined as hypertension accompanied by hematologic, liver, renal, cerebral, and/or visual disturbances after the 20th week of gestation [1]. PE constitutes a serious pregnancy complication that adversely influences perinatal outcomes. Additionally, PE has been consistently associated with increased incidence of long-term cardiovascular events in both mother and offspring. PE remains an obstetric and public health burden with recent epidemiologic studies showing an increase in its incidence (5–8% of all pregnancies worldwide) due to a higher prevalence of risk factors, such as advanced maternal age, obesity, diabetes, and chronic hypertension [2]. Currently, there is no effective therapy for PE, with the syndrome only remitting after delivery. The difficulty in identifying and testing such therapies is partially explained by the fact that PE is a multifactorial disorder involving pathways related to impaired angiogenesis, inflammation, and oxidative stress, among others that lead to generalized endothelial dysfunction. The etiology of PE seems even more complex when considered the timing of manifestation of the clinical symptoms. Early-onset PE, defined as those diagnosed before 34 weeks of gestation (5–20% of cases), has been associated with inadequate remodeling of the decidual vasculature and placental malperfusion; whereas late-onset PE, defined as those diagnosed after 34 weeks of gestation (~80% of cases), has been associated with maternal morbidities including impaired glucose tolerance, dyslipidemia, obesity, chronic hypertension, and metabolic syndrome. These different pathways seem to converge on the pathophysiology of both types of PE causing placental senescence and stress [3]. Together, these findings justify the ongoing search for useful diagnostic and prognostic biomarkers that may be eventually targeted by novel therapies in PE.

Providing new insights into potential biomarkers for PE, exaggerated levels of both maternal and fetal-derived cell-free DNA (cfDNA) fragments have been detected in the blood of PE patients. Moreover, increased circulating cfDNA levels have been associated with disease severity and unfavorable maternal-fetal outcomes in PE [4–9]. While these studies suggest that cfDNA may be useful as a diagnostic and prognostic biomarker for PE, the exact mechanism leading to the release of cfDNA into the maternal circulation and its range of consequences are not fully known. In this commentary, recent studies measuring cfDNA in early-onset, late-onset, or any type of PE are briefly summarized. The overall goal is to point shortcomings of current literature in the field, suggest alternatives to overcome the challenge of assessing circulating cfDNA, and propose required studies to determine the role of cfDNA in mediating the pathogenesis of PE.

2. Measuring increased levels of cfDNA in preeclampsia

The fetal fraction of cfDNA is the amount of fetal cfDNA divided by the amount of total cfDNA in a given sample. In normal pregnancies, it has a bell-shaped distribution peaking between 10% and 20% at 10–21 weeks of gestation. The remaining 80–90% of total cfDNA corresponds to the maternal fraction, being primarily of hematopoietic cell origin. In PE, while the amount of both fetal and total cfDNA is increased, the fetal fraction is actually reduced [10], highlighting the greater contributory role of maternal cfDNA to the exaggerated circulating total cfDNA levels in PE. However, published data evaluating circulating cfDNA as a biomarker in PE have largely focused on fetal cfDNA. These studies relied mainly on the quantification of Y-chromosome specific gene sequences by Real Time-Polymerase Chain Reaction (qPCR) and, as such, only detected fetal cfDNA in pregnant women bearing male fetuses [4]. To overcome this limitation, recent studies have assessed genes that undergo epigenetic regulation according to the tissue origin. For instance, because the promoter region of the Ras association domain family protein 1 isoform A (RASSF1A) gene is hypermethylated in placenta and unmethylated in maternal blood cells, the fetal RASSF1A copies can be quantified by qPCR after selective removal of the maternal RASSF1A copies via methylation-sensitive restriction enzyme digestion. By employing this methodology, it has been found that circulating placental/fetal cfDNA levels are elevated in women presenting any type of PE [6–8]. In addition, two distinct prospective studies showed that placental/fetal cfDNA levels were significantly increased in plasma as early as 8 weeks through 32 weeks of gestation in women who subsequently developed PE compared to those who had uncomplicated pregnancies [8,11].

In order to assess total cfDNA, research teams have used gender non-dependent genes that are present in both fetal and maternal cells. For example, by applying qPCR to measure β-globin copies, it has been demonstrated that total cfDNA levels are elevated in the circulation of women presenting any type of PE [4,7,9] as well as in the blood collected at 20 weeks of gestation of those destined to develop PE [12]. Alternatively, few laboratories have utilized fluorescence spectrometry to quantify total cfDNA [5,13]. In addition to the low cost and easy performance, the measurement of total cfDNA by fluorescence spectrometry has the advantage of mitigating the heterogeneity inherent to the selection of gene sequences for qPCR, making this assay more appealing for routine use in the clinical setting. Such studies have confirmed that circulating total cfDNA is elevated in women presenting any type of PE [5,13]. However, there is evidence showing that these levels are increased in pregnancies delivered preterm versus term as well as in pregnancies complicated by fetal growth restriction versus normal growth [5]. Thus, future studies need to determine circulating total cfDNA levels via fluorescence spectrometry in PE and normal pregnant women matched by gestational age and fetal growth. This could be accomplished with longitudinal studies performed across all trimesters of pregnancy to rigorously evaluate the predictive value of fluorescence spectrometry-based total cfDNA levels in PE.

The cfDNA originating from placental/fetal and maternal cells may be composed by nuclear and mitochondrial DNA. Indeed, studies have described elevated circulating mitochondrial cfDNA levels in women presenting any type of PE [14,15]. Conversely, it has been found that mitochondrial cfDNA levels are decreased in blood collected at 11 to 12 + 6 weeks of gestation in women who subsequently developed PE [16]. Curiously, circulating mitochondrial cfDNA levels in first trimester of pregnancy were lower in women presenting early-onset compared to those presenting late-onset PE [16]. Therefore, further prospective studies including asymptomatic pregnant women who later turn symptomatic before and after 34 weeks of gestation are warranted to determine the use of mitochondrial cfDNA as a predictive biomarker in PE.

3. Mechanisms mediating increased cfDNA and its possible consequences in preeclampsia

As mentioned, both placental/fetal and total cfDNA have been associated with adverse maternal-fetal outcomes in PE, being levels exacerbated in pregnant women who develop severe disease and in those who deliver preterm and growth-restricted newborns. Moreover, increased circulating cfDNA has been positively correlated with increased blood pressure and enzymes linked to tissue damage as well as negatively correlated with decreased blood platelet count, gestational age at delivery, and newborn weight [4–9]. Although these reports suggest that cfDNA is implicated in the pathophysiology of PE, it is unknown whether elevated levels of cfDNA are the cause or consequence of unfavorable clinical features.

It has been suggested that placental malperfusion along with exaggerated inflammation and oxidative stress cause apoptosis and/or necrosis of the syncytiotrophoblast – a huge cell covering the entire folded microvillous surface of the placenta – resulting in shedding of placental particles into the maternal circulation [3]. Indeed, in vitro studies have found that hypoxia-reoxygenation elicits cfDNA release from human placental explants, as determined by measurement of β-globin copies via qPCR, which was associated with release of oxidative stress and apoptosis/necrosis markers in cultured media [17]. Moreover, increased levels and enhanced hypomethylation degree of placental mitochondrial DNA have been associated with low oxygen partial pressure in umbilical vein blood [18,19]. These studies corroborate with the hypothesis that elevated cfDNA in the circulation of PE patients results from spillover of placental debris due to ischemia-reperfusion injury of trophoblastic cells. However, there are also reports suggesting that elevated circulating levels of fetal cfDNA in PE is due to its reduced clearance from maternal plasma [20]. Furthermore, cfDNA and other placental molecules in the maternal circulation may induce damage of endothelial cells and activation of immune cells – processes that instigate passive (rupture of cell membranes) and active (secretion of microvesicles) release of cfDNA, contributing to the pool of circulating total cfDNA in PE [21].

Few studies in experimental animals have assessed circulating cfDNA during pregnancy and especially in models of PE. Nonetheless, studies have begun to inject cfDNA and evaluate maternal and fetal outcomes. While intraperitoneal injection of human fetal cfDNA at 300 µg/day from gestational day (GD) 10 to 14 in mice caused fetal reabsorption, preterm birth, and exacerbated inflammation in uterine and placental tissues [22], administration of human adult cfDNA during GD 10–14 [22] as well as human/mouse fetal cfDNA or mouse maternal (kidneys) cfDNA during GD 14–18 using the same route and dosage in mice did not induce PE-like features [23]. In agreement with these later findings, a study in rats has demonstrated that intraperitoneal injection of rat fetal cfDNA at 400 µg/day during GD 14–18 did not alter blood pressure, proteinuria, placental, and fetal weights [24]. Possible explanations for different results among studies are the variable timing (mid versus late gestation) and source (human versus the same species) of cfDNA administration.

Another reason for distinct effects of cfDNA on maternal-fetal parameters in animal models could be based on the type of cfDNA. Adult cfDNA is mostly hypermethylated, whereas placental/fetal cfDNA is predominantly unmethylated allowing it to bind to specific receptors in immune cells [21]. Likewise, mitochondrial DNA has a circular structure and abundant unmethylated CpG islands that enable it to activate toll-like receptors (TLR) [25]. Upon activation, neutrophils and platelets secrete cfDNA in the form of neutrophil extracellular traps (NETs) and extracellular vesicles. This feed-forward cycle of cfDNA release may contribute to exacerbated humoral and cellular immune response, systemic endothelial function, and coagulopathies in PE [21]. In support of the importance of TLR signaling in mediating the effects of cfDNA on pregnancy, TLR-9 knockout mice were protected against human fetal DNA-induced adverse placental-fetal outcomes [22]. In addition, activation of TLR-9 with intraperitoneal injections of unmethylated CpG DNA at 100 µg/day on GD 14, 17, and 18 in rats induced maternal hypertension, enhanced vasoconstriction, oxidative stress, and inflammation [26]. Therefore, future animal studies need to examine not only the species and tissue sources of injected cfDNA but also distinguish its composition between nuclear and mitochondrial DNA. Furthermore, studies should verify whether circulating levels of cfDNA are chronically elevated in pregnant animals or reach levels to mimic those observed in PE women.

4. Conclusion

In summary, human studies indicate that cfDNA is elevated in PE and circulating levels correlate with disease severity and unfavorable maternal-fetal outcomes; however, there are data in the literature contradicting this postulate. Indeed, systematic reviews on this topic concluded that the clinical relevance of circulating cfDNA as an independent biomarker for the prediction and/or progression monitoring of PE is currently uncertain due to the heterogeneity of enrolled populations, lack of standardized quantification protocols, and questionable statistical robustness [27–30]. In this commentary, it is stressed that additional prospective studies on cfDNA in PE are warranted to validate its diagnostic and prognostic value. Importantly, studies applying sensitive gene sequences on qPCR or fluorescence spectrophotometry are able to capture cfDNA in both male- and female-bearing pregnancies as well as more likely to generate reproducible results. Clinical studies should also distinguish between mitochondrial and nuclear cfDNA. Moreover, future mechanistic studies using animal models and human cells/tissues are needed to determine whether different sources of circulating cfDNA mediate pathways implicated in exaggerated inflammatory response, vascular dysfunction, and hypertension in PE.

5. Expert Opinion

Due to the large clinical, social, and public health impact of preeclampsia, intense research has been directed towards the identification of non- or minimally invasive biomarkers that may predict the development and/or worse outcomes for this maternal hypertensive syndrome. Several studies have found increased cell-free DNA levels in plasma, serum and/or urine of preeclamptic women; however, there is still no consensus regarding the diagnostic/prognostic value of cell-free DNA in preeclampsia. This may be explained by the heterogeneity of enrolled populations, lack of standardized quantification protocols, and uncertain statistical robustness among published studies. In addition to the challenges of assessing cell-free DNA in preeclampsia, little is known about the mechanisms triggering the release of cell-free DNA in the circulation or the effect of cell-free DNA on mediating adverse maternal-fetal outcomes in preeclampsia. Therefore, additional studies are necessary to fully elucidate the role of cell-free DNA in preeclampsia.

Declaration of Interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer Disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This work has been funded by the National Institutes of Health [Grant number: R01HL148191] (Jackson, MS) and the American Physiological Society – John F. Perkins, Jr Research Career Enhancement Award.