Plasma F₂-isoprostane class VI isomers at 12–18 weeks of pregnancy are associated with later occurrence of preeclampsia

Abstract

Preeclampsia (PE) has long been associated with early oxidative stress whereas the symptoms occur later in pregnancy. We have hypothesized that the oxidative stress in PE, as characterized by the presence of F₂-isoprostanes (F₂-isoPs) isomers in late pregnancy, should already be present in plasma at the first regular visit of the obstetrical follow-up. There are 64 possible isomers of F₂-isoPs derived from the oxidation of arachidonic acid (AA) but only one of these isomers has been investigated so far in PE, the classical 8-iso-PGF_2α. Here, we have investigated two regioisomers of class III (8-iso-15(R)-PGF_2α and 8-iso-PGF_2α) and a mix of two isomers of class VI (±)5-iPF_2α-VI in plasma samples collected prospectively at 12–18 weeks from normotensive controls (n=60) and pregnant mothers who developed PE later in pregnancy (n=33). The plasma samples were subjected to alkaline hydrolysis followed by liquid-liquid extraction to extract total F₂-isoPs for later quantification by HPLC-MS/MS. The F₂-isoPs were normalized to either plasma volume or polyunsaturated fatty acids (PUFAs) levels measured by GC-FID in plasma phospholipids. Early in pregnancy, only the class VI F₂-isoPs isomers were found in concentrations significantly higher in women developing PE later in pregnancy (+ 13%; p=0.0074). Normalisation of F₂-isoPs to their substrate, AA, or the omega-3 to omega-6 ratio improved the predictability of PE as determined by receiver operating characteristic (from area under the curve of 0.67 to 0.68 and 0.70 respectively). Interestingly, omega-3 fatty acids were 25% higher in the control group than the PE group (P = 0.0225). Total omega-6 PUFAs correlated to F₂-Isop isomers only in cases of PE (r > 0.377; P >0.03, Spearman correlation). In sum, this study indicates that specific isomers of class VI are significant predictors of PE. This work also suggests that F₂-isoPs isomers are not all generated and eliminated at the same extent and are influence by the PUFA composition of plasma phospholipids.

Introduction

The classical 8-iso PGF_2α, an F₂-isoprostanes (F₂-isoPs), is currently the most stable and reliable indicators of oxidative stress and lipid peroxidation [1–3]. F₂-isoPs are prostaglandin (PG) F₂-like isomers, generated by the spontaneous oxidation of arachidonic acid (AA), an omega-6 polyunsaturated fatty acids (PUFAs), contained in various phospholipid membranes [4]. F₂-isoPs are carried by lipoproteins, mainly HDL, in the blood circulation [5]. There are 64 possible isomers of F₂-isoPs classified in 4 classes of regioisomers (III, IV, V, VI) that comprises 16 members each [6, 7]. The most studied isomer is the 8-iso PGF_2α (or 8-epi-PGF_2α, iPF_2α-III). The 8-iso-PGF_2α is a vasoconstrictor similar in action to thromboxane (TX) A₂, when released from membranes by the action of phospholipases A₂ [8, 9]. The other 63 F₂-IsoP isomers are much less (or for most, not at all) investigated in pregnancy. Indeed, we pioneered the detection of the abundant class VI isomers and reported the first basal values in uncomplicated pregnancies by HPLC-MS/MS [2, 10].

Uncomplicated pregnancy is associated with a mild physiological increase in oxidative stress to cope with placental and fetal development throughout pregnancy [1]. Indeed, markers of oxidative stress, such as F₂-isoPs, reach their highest level between 37 and 41, just before birth [11]. Pathological pregnancies, such as preeclampsia (PE), expressed an even higher oxidative stress response especially in severe cases, as demonstrated using free 8-iso-PGF_2α as plasma biomarker [12].

Only few studies have looked prospectively at the potential link between PE and oxidative stress markers, such as F₂-isoPs. One report showed a urinary increase of the 8-iso-PGF_2α as early as the 15^th week of gestation in women who will later develop PE (5-fold higher risk) compared to controls [11]. Accordingly, another study reported 8-iso-PGF_2α levels at 1.7-fold higher in the amniotic fluid of PE than normotensive pregnancies at an average of 17.5 weeks of gestation [13]. Notably, only one isomer was investigated in these studies. We believe that F₂-isoP isomers are not all generated and eliminated to the same extent. We hypothesized that certain F₂-isoPs isomers could be more specific to PE than the classical 8-iso-PGF_2α. Therefore, we propose that the concomitant assessment of several F₂-isoPs by HLC-MS/MS should provide more knowledge on the specific oxidative stress appearing in PE. We also theorized that the plasma fatty acid composition should also affect the generation of specific F₂-isoPs or compete for the generation of other oxidation products. The objective of this study was thus to quantify two isomers of class III (8-iso-15(R)-PGF_2α and 8-iso-PGF_2α) and a mix of two isomers of class VI, (±5)-iPF_2α-VI (or iPF_2α -VI + 5-iPF_2α-VI), in plasma samples collected prospectively at 12–18 weeks from normotensive controls (n=60) and women that will later develop PE (n=33). Concurrently, we quantified omega-3 and omega-6 fatty acids in order to correlate the levels of PUFAs with those of F₂-isoPs.

Materials and methods

Materials

The F₂-isoPs isomers, 8-iso-15(R)-PGF_2α, 8-iso-PGF_2α, (±)5-iPF_2α -VI (or iPF_2α -VI + 5-iPF_2α-VI), and their corresponding deuterated standards; 8-iso-PGF_2α -d4 and (±)5-iPF_2α -VI-d11 were purchased from Cayman Chemical (Ann Arbor, MI, USA). Butylated hydroxytoluene (BHT) was bought from Sigma-Aldrich (Oakville, ON, Canada). The standards for fatty acid profile analysis, FAME 37 mix, 18:2 cis/trans mix 18:3 cis/trans mix, 22:5, 9t-14:1, 9t-C16:1 and several 18:1 fatty acids (6c-, 11c-, 12c-, 13c-, 6t-, 11t-18:1) were all purchased from Supelco Inc (Bellefonte, PA, USA). A mixture of 31 fatty acids (GLC-411) was obtained from NuChek Prep Inc (Elysian, MN, USA). All other reagents and solvents were HPLC grade and were purchased from VWR International (Ville Mont-Royal, QC, Canada).

Patient’s characteristics and sample processing

Patients’ clinical data and biological samples were obtained from a biobank, the Maternal and Infant Research on Oxidative Stress (MIROS cohort) described previously [14]. Plasma from sixty women with uncomplicated pregnancies were recruited prospectively early between the 12^th to the 18^th weeks of pregnancy from a subset of the 665 possible uncomplicated pregnancies recruited by the MIROS study group in 17 Canadian centers over a 26-month period. Similarly 33 PE were obtained from the same biobank. The respective institutions approved the protocol and informed consents were obtained from all pregnant women. PE was defined as gestational hypertension with proteinuria. Proteinuria was defined as a urine protein dipstick test ≥2+, or the urinary excretion of ≥ 0.3 g in a 24-hour urine collection[15]. Severe PE was defined according to the criteria of the Canadian Hypertension Society Consensus Conference Report [15]. Early on-set PE was defined as delivery before 34 weeks of gestational age [16]. Venous blood was collected into EDTA tubes and plasma samples were immediately separated by centrifugation at 500 × g for 10 minutes at 4°C. Centrifuged plasma samples were rapidly frozen and stored at −80°C [14]. Butylated hydroxytoluene (BHT) was added (0.1%) before aliquoting [10].

F₂-isoPs analysis by HPLC-MS/MS

Total F₂-isoPs (free + esterified) were extracted from plasma after alkaline hydrolysis using a liquid-liquid extraction as described in details previously [10]. Briefly, the chromatography was carried out with a Shimadzu Prominence system (Columbia, MD, USA) linked to a 3200 QTRAP^® LC/MS/MS system from AB Sciex (Concord, ON, Canada) operated in the negative mode. A Shim-Pack XR-ODS column (3.0 × 100 mm, 2.2 μm) from Shimadzu was used. The column temperature was controlled at 30°C and the injection volume was 40 μL. The chromatographic separation was done using a gradient of three solvents at 0.4 mL/min as described earlier (method II [2]). The mobile phase A was 0.01% acetic acid in water, B was 0.01% acetic acid in acetonitrile and C was 0.01% acetic acid in methanol. The F₂-isoPs and their corresponding internal standards were monitored in multiple-reaction monitoring (MRM) mode at the transitions 353.3 → 193.2 and 357.3 → 197.2 m/z respectively for class III, and at the transitions 353.0 → 115.0 and 364.6 → 115.0 m/z respectively for class VI. Quantification of F₂-isoPs was performed using the Analyst® 1.4.2 Software (AB Sciex) [10].

Determination of plasma fatty acid profile by gas chromatography

Plasma fatty acids were isolated according to a method previously described [17]. Briefly, a solution of chloroform:methanol (2:1, by volume) was used to extract lipids from plasma. Then, phospholipids were separated by thin layer chromatography using a mix of isopropyl ether:acetic acid (96:4) as eluent and phospholipid fatty acids were methylated following a transesterification reaction using a mix of methanol:benzene (4:1) and acetyl chloride at 95°C for 1.5 hour. Methylated fatty acids were finally analysed by gas chromatography coupled with a flame ionisation detector (GC-FID) as explained elsewhere [2, 17]. The peroxidation index (PI) and the unsaturation index (UI) were calculated as described elsewhere [18].

Statistical analyses

Statistical analyses were performed with GraphPad Prism 6.0f (GraphPad Software Inc., 2013, San Diego, USA). Normality was verified using the D’Agostino & Pearson omnibus test. Students t-test was used to compare the groups. Since normality was not achieved for most of the data, the Mann-Whitney test was used to compare levels of both, F₂-isoPs and fatty acids. The Spearman’s rank correlation coefficient was used to study relationship between variables. The predictive nature of F₂-isoPs was determined using the receiver operating characteristic curve analysis. A p-value lower than 0.05 (two-tailed) was considered statistically significant.

Results

Clinical features of the cohort used (biobank)

The clinical characteristics of the pregnant’s cohort used in this study was published in details previously [14]. Briefly, the PE group was constituted of 33 pregnancies in which, 21 pregnancies were obtained from a predetermined high-risk group according to one or many of the following risk factors; chronic hypertension, prepregnancy diabetes, multiple pregnancy, or a history of PE. The mean gestational age at delivery was 36± 3 weeks (mean ± SD). Of note, 22 pregnancies qualified as severe and 11 pregnancies are classified as early on-set PE (< 34 weeks at delivery) [16]. Attempt to stratify according to the two latter factors did not yield significance in our statistical analysis (results not shown). The PE group was matched with a control group of 60 pregnancies in order to avoid any statistical differences for maternal age, education, ethnic origin, marital status, prenatal vitamin supplementation, smoking or season when blood was drawn. The pre-pregnancy body mass index (BMI) and weight gain were not statistically different between the control and the PE groups (results not shown). Only gestational age at delivery (term pregnancy) was statistically different (P < 0.05).

F₂-isoPs in the plasma of women that will later develop or not PE

Table 1 shows the F₂-isoPs levels of class III and VI isomers measured by HPLC-MS/MS using maternal plasma at 12–18 weeks of gestation obtained from normotensive controls and mothers who later developed PE. The data were normalized to plasma volume or AA content, the substrate generating F₂-isoPs. Only class VI isomers, iPF_2α-VI + 5-iPF_2α-VI or (±)5-iPF_2α-VI were 13% and 19% higher in the PE group compared to controls when normalized with plasma volume or AA content (P = 0.0074 and P = 0.0040 respectively). These class VI isomers are a mix of two isomers and are sold as one standard for the purpose of quantification: (±)5-iPF_2α-VI. Both class III F₂-isoP isomers, 8-iso-15(R)-PGF_2α and 8-iso-PGF_2α, were not significantly different between groups. The sums of the class III isomers are 222 ± 18 and 250 ± 27 pg/ml for normotensives and PE respectively (mean ± SEM; P = 0.36). The sums of all F₂-isoP isomers investigated here are 371 ± 20 and 428 ± 33 pg/ml respectively for the normotensive and PE groups (P = 0.12)

TABLE 1.

F₂-isoprostanes plasma levels at the first visit (12–18 weeks) of pregnancy between normotensive controls and pregnant women that will later develop preeclampsia (PE).

	F2-Isoprostanes levels1
	pg/ml of plasma		pg/μg of arachidonic acid
	Normotensives (n = 60)	toward PE (n = 33)	Normotensives (n = 60)	toward PE (n = 33)
Class III
8-iso-15(R)-PGF2α	105 [69, 161]	95 [67, 185]	0.83 [0.42, 1.13]	0.70 [0.48, 1.26]
8-iso-PGF2α	118 [0, 199]	147 [0, 222]	0.85 [0, 1.36]	0.94 [0, 1.51]
Class VI
iPF2α-VI + 5-iPF2α-VI	152 [121, 168]	172 [137, 209]*	1.02 [0.81, 1.27]	1.21 [1.03, 1.49]**

¹Values are medians and quartiles [Q1, Q3].

^*Significantly different from normotensives pregnancies (Mann-Whitney test, P = 0.0074).

^**Significantly different from normotensives pregnancies (Mann-Whitney test, P = 0.0040).

Plasma PUFAs in pregnancy

Table 2 is reporting the levels of PUFAs in plasma phospholipids at 12–18 weeks of gestation in normotensive subjects and mothers who will later develop PE. The levels of total omega-6 PUFAs, comprising AA, the substrate of F₂-isoPs, are not different between groups (P > 0.68). However, the level of omega-3 fatty acids was significantly lower in the group that will develop PE compared to normotensive controls (81 ± 20 vs 71 ± 19 μg/ml; P=0.0225). This last difference affects indirectly the omega-3 to omega-6 ratio and the calculated PI and UI indexes (P < 0.041).

TABLE 2.

Total omega-6, total omega-3 fatty acid content and ratio, unsaturation (UI) and peroxidation (PI) indexes at the first obstetrical visit (12–18 weeks) between normotensive controls and pregnancies that will later develop preeclampsia (PE).

	Fatty acid levels (μg/mL of plasma)
	Normotensives (n = 60)	toward PE (n = 33)	P
Total ω-61	486 [449, 544]	501 [444, 556]	0.8688
20:4ω-6 (AA)	142 [128, 163]	138 [118, 172]	0.6777
Total ω-32	83.2 [64.5, 92.5] (81 ± 20)	66 [59, 84] (71 ± 19)	0.0301* (0.0225*)
Ratio and indexes
ω-3/ω-6	0.16 [0.13, 0.18]	0.14 [0.12, 0.17]	0.0382*
UI	144 [139, 149] (144 ± 6)	141 [137, 144] (141 ± 7)	0.0559§ (0.0320*)
PI	115 [107, 124] (116 ± 10)	112 [103, 118] (111 ± 11)	0.0569§ (0.0406*)

Values are medians and quartiles [Q1, Q3] and (mean ± SD) if the normality postulate is achieved.

^*Significantly different from normotensives (P < 0.05; Mann-Whitney test or (student t-test)).

^§Tendency, (P < 0.1; Mann-Whitney test)

¹Total omega-6 fatty acids = 18:2 + 18:3 + 20:2 + 20:3 + arachidonic acid (AA; 20:4) + 22:2 + 22:4 + 22:5.

²Total omega-3 fatty acids = 18:3 + 18:4 + 20:3 + 20:4 + 20:5 + 22:5 + 22:6.

Unsaturation index (UI) = (%Monoenoic x 1) + (%Dienoic x 2) + (%Trienoic x 3) + (%Tetraenoic x 4) + (%Pentaenoic x 5) + (%Hexaenoic x 6).

Peroxidation index (PI) = (%Monoenoic x 0.025) + (%Dienoic x 1) + (%Trienoic x 2) + (%Tetraenoic x 4) + (%Pentaenoic x 6) + (%Hexaenoic x 8).

Correlation between F₂-isoPs and PUFAs

Spearman’s rank correlation coefficients were used to study at 12–18 weeks of pregnancies, the associations between F₂-isoPs and plasma phospholipid PUFAs in normotensive mothers and the group that will develop PE group (Table 3). Interestingly, all F₂-isoPs isomers correlated with either, AA or total omega-6 PUFAs in PE (r > 0.377, P > 0.0031), but this phenomenon was not observed in the normotensive controls. The class VI isomers were better correlated than class III isomers in the group that will later develop PE, though (r >0.531, P > 0.001). In contrast, the control group was not correlated to omega-6 PUFAs (Table 3). In normotensive pregnancies, total omega-3 PUFAs were negatively correlated with the class VI isomers whereas the omega-3 to omega-6 ratio correlated with both the classical 8-iso-PGF_2α and the class VI isomers. Of note, F₂-isoPs were not correlated to UI and PI indexes for all study groups.

TABLE 3.

Correlations between F₂-isoprostanes and arachidonic acid (AA), total omega-3, total omega-6 fatty acids content or ratio in the plasma at the first visit (12–18 weeks) of pregnancy between normotensive controls and pregnant women that will later develop preeclampsia (PE).

	20:4ω-6 (AA)	Total ω-61	Total ω-32	ω-3/ω-6
Normotensives
8-iso-15(R)-PGF2α	ns	ns	ns	ns
8-iso-PGF2α	ns	ns	ns	r = −0.286 (p = 0.028*)
iPF2α-VI + 5-iPF2α-VI	ns	ns	r = −0.275 (p = 0.034*)	r = −0.336 (p = 0.009*)
Toward PE
8-iso-15(R)-PGF2α	r = 0.412 (p = 0.017*)	r = 0.447 (p = 0.009*)	ns	ns
8-iso-PGF2α	r = 0.377 (p = 0.031*)	r = 0.480 (p = 0.005*)	ns	ns
iPF2α-VI + 5-iPF2α-VI	r = 0.531 (p = 0.001*)	r = 0.551 (p = 0.001*)	ns	ns

Values (r) are Spearman correlation coefficients (n=60 for normotensives and n=33 for PE).

^*Significantly different (P < 0.05).

ns = non-significant (P > 0.1).

¹Total omega-6 fatty acids = 18:2 + 18:3 + 20:2 + 20:3 + 20:4 + 22:2 + 22:4 + 22:5

²Total omega-3 fatty acids = 18:3 + 18:4 + 20:3 + 20:4 + 20:5 + 22:5 + 22:6

F₂-isoPs and PE prediction

A receiver operating characteristic (ROC) curve analysis was performed in order to determine how predictive were class VI F₂-isoPs for PE (Figure 1). The iPF_2α-VI + 5-iPF_2α-VI used alone yielded to an area under the curve (AUC) of 0.67 ± 0.06 (mean ± SEM; P = 0.0079). The normalization to AA and the omega-3 to omega-6 ratio improved slightly the test toward a fair prediction, the AUC being 0.68 ± 0.06 (P = 0.0044) and 0.70 ± 0.06 respectively.

Figure 1.

Receiver operating characteristic (ROC) curves for ±5 iPF_2α-VI (iPF_2α-VI + 5-iPF_2α-VI) used for PE prediction normalized to plasma volume ( ; area = 0.67) or arachidonic acid (AA) ( ; area = 0.68) or Total ω-3/total ω-6 ratio (^····; area = 0.70). Total omega-3 fatty acids (ω-3) = 18:3 + 18:4 + 20:3 + 20:4 + 20:5 + C22:5 + C22:6. Total omega-6 fatty acids (ω-6) = 18:1 + 18:2 + 18:3 + 20:2 + 20:3 + 20:4 + 22:2 + 22:4 + 22:5. n=60 controls, n=33 PE.

Discussion

We reported relatively low levels of total F₂-isoPs (free + esterified) between 105 μg/ml for class III isomers to 152 μg/ml for class VI at 12–18 weeks in normotensive pregnancies. However, these levels are in the upper range of reported levels for the classical 8-iso-PGF_2α from healthy non-pregnant individuals, which were between 40–170 pg/mL in plasma [19, 20]. Indeed, an uneventful pregnancy is considered a mild oxidative stress [21], which may explained these slightly higher levels. Compared to levels determined early in pregnancy (12–18 weeks) as shown in the current study, our previous data obtained from two independent cohorts suggest a 1.4-fold higher level for 8-iso-PGF_2α levels and a 2.4-fold higher level for iPF₂α-VI + 5-iPF₂α-VI at the end of pregnancy [2]. These findings are in accordance with study showing that plasma 8-iso-PGF_2α levels are relatively low early in pregnancy with a significant increase of approximately two-fold from 15–20 weeks to 37–41 weeks of pregnancy [52].

Only class VI F₂-isoP isomers were significantly higher in the group developing PE compared to normotensive controls. To our knowledge, this is the first prospective report of plasma isoPs in pregnant mothers developing PE later in pregnancy. These data reveal a statistically significant increase, even though modest (13%), between the control normotensive group and mothers developing PE later in pregnancy, suggesting a slight maternal systemic increase of oxidative stress already detectable in the early stage of undiagnosed PE. The normalization of the data using substrate concentration (AA) in phospholipids did not alter this conclusion, but shows instead a 19% increase. The source of this increased class VI F₂-isoPs early in maternal blood is still undetermined, but might originate from the abnormally developed placenta rather than the maternal endothelial dysfunction at this stage of the pathology [1, 22]. This current discovery would not have been possible without the simultaneous analysis of several isomers of F₂-isoPs by HPLC-MS/MS. The current study confirms that specific F₂-isoP isomers may respond differently in various pathologies. Indeed, more than eight F₂-isoP isomers were shown to be increased in urine of hypercholesterolemic patients, whereas only three isomers, including the classical 8-iso-PGF_2α and the 8-iso-15(R)-PGF_2α, were found to be increased in cases of congestive heart failure [23].

According to the present data, the prominence of class III F₂-isoP isomers, like 8-iso-PGF_2α, in the early stage of PE appears to be much less striking than anticipated based on numerous previous studies reporting class III F₂-isoP isomers as accurate indicators of oxidative stress in many pathologies [1–3]. Furthermore, class VI isomers are rarely studied and we are pioneering their measurements in pregnancy [2, 10]. Other prospective studies on PE have looked at concentrations of urinary and amniotic fluid 8-iso-PGF_2α using enzyme immunoassays. Interestingly, elevated urinary excretion of 8-iso-PGF_2α at 15 weeks was associated with a 5-fold higher risk of PE in a cohort of 307 pregnant women. However, 24-hour urinary excretion of 8-iso-PGF_2α cannot be directly compared to plasma values of total F₂-isoPs (free + esterified) measured at a given time point. Indeed, the difference with our study could be related to the increased elimination of F₂-isoPs in urine rather than their total concentrations at any given time in blood plasma [1].

In another prospective study, levels of amniotic fluid 8-iso-PGF_2α were 1.74-fold higher in women known to later develop PE compared to controls. The PE and control groups were both constituted of 85 women. The area under curve (AUC) of the ROC curve indicated 0.81 for the prediction of PE [13]. However, there are at least three major differences in this report compared to our study: biological samples were, in the average, obtained 3 weeks later, secondly biological specimen were collected from a very different ethnic population (Caucasian vs Asian), and thirdly the collection of amniotic fluid at the very proximity of the source of oxidative stress, the placenta. These may explain the more pronounced difference between PE and the control groups using a less specific marker for PE such as the 8-iso-PGF_2α.

F₂-ioPs originates from the oxidation of AA, an omega-6 PUFAs. However, we did not observe any difference for either omega-6 or AA concentrations between the study groups. This could suggest that substrate availability was not an issue for F₂-isoPs formation. Interestingly, omega-3 level was significantly lower in the group later developing PE compared to controls. It is believed that omega-3 fatty acids are more susceptible to the oxidation than omega-6 fatty acids. Of note, oxidation of omega-3 PUFAs leads, among others, to the formation F₃-isoPs and F₄-isoPs (neuroprostanes) derived from eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) respectively [1]. This suggests that higher omega-3 level in the control group could somehow protect against the oxidation of omega-6 fatty acids, explaining the lower concentrations of AA derived F₂-isoPs in the control group. Indirect evidence of the latter came from nutritional studies. In kidney transplant patients, supplementation with fish oil rich omega-3 fatty acids for periods of 3 and 6 months caused a significant decrease in the level of plasma 8-iso-PGF_2α when compared to the untreated (placebo) group [24]. In a 6-week placebo-controlled trial, 4 g daily of either EPA or DHA in overweight dyslipidemic men caused 24% and 14% decreases in plasma 8-iso-PGF_2α [25]. The same omega-3 treatment when used in treated-hypertensive type 2 diabetic patients caused reductions in the plasma 8-iso-PGF_2α levels of 19% and 23% using EPA or DHA in comparison to the olive oil control group. Interestingly, correction for AA content did not change the conclusion of these two last independent trials [25].

Since the study group developing PE showed lower level of omega-3 fatty acids, higher level of omega-6 fatty acids and, likely, higher oxidative stress than normotensive controls, we believe that correlation between F₂-isoPs and its substrate AA is achieved under PE conditions. Further, AA constitutes a large proportion (27%) of the omega-6 fatty acids that are also correlated with the AA-derived isoPs only in the group known to develop PE later in pregnancy. It is worth noting that oxidation of AA may also yield to the formation of PGD₂/E₂-like (D₂/E₂-IsoPs) and thromboxane-like (isothromboxanes) end-products in competition with the F₂-isoPs metabolites [26]. The protective effect of omega-3, as emphasized by their negative correlation with class VI isoPs, is only observed in uneventful pregnancy with mild oxidative stress. This is in accordance with the reported absence of correlation between AA and F₂-isoPs in healthy individuals followed over a 4-month period [2727].

Conclusions

In summary, this study reveals that isomers of class VI, (±)5-iPF_2α-VI, are better predictors of PE at 12–18 weeks of gestation than class III isomers such as the classical 8-iso-PGF_2α. This also suggests that F₂-isoPs isomers are not all generated and/or eliminated at the same rate during the pathogenesis of PE. This work also underlines the possible antioxidative role of omega-3 PUFAs on the production of AA-derived F₂-IsoPs and the pathogenesis of PE.

Abbreviations

AA

arachidonic acid

AUC

area under the curve

BHT

butylated hydroxytoluene

EPA

eicosapentaenoic acid

DHA

docosahexaenoic acid

F₂-isoP

F₂-isoprostane

GC-FID

gas chromatography coupled to a flame ionisation detector

HPLC-MS/MS

high performance liquid chromatography coupled to tandem mass spectrometry

IS

internal standard

PI

peroxidation index

PUFAs

polyunsaturated fatty acids

PG

prostaglandins

PLA₂

phospholipase A₂

ROC

receiver operating characteristic

ROS

Reactive oxygen species

TX

Thromboxane

UI

unsaturation index